Discussion Board This week we are going to see how eco-farming in Germany was adopted. We can see that this has many elements of public policy and gives us more knowledge on how new innovations are ad

Article   in  Agricult ure and Human V alues · July 2007 DOI: 10.1007/s10460-007-9072-2 CITATIONS 133 READS 73,918 3 author s:

Some of the author s of this public ation are also w orking on these r elated pr ojects: Collabor ative rese arch and multi-st akeholder appr oaches in f ood and f arming syst ems Vie w pr oject Upland Pr ogram Vie w pr oject Volk er Hoffmann Univ ersity of Hohenheim 81 PUBLICA TIONS    694 CITATIONS     SEE PROFILE Kir sten Pr obst 8 PUBLICA TIONS    341 CITATIONS     SEE PROFILE Anja Christinck German Instit ute f or T ropic al and Subtr opical Agricult ure (DIT SL); seed4chang e Rese … 36 PUBLICA TIONS    375 CITATIONS     SEE PROFILE All c ontent f ollowing this p age w as uplo aded by Anja Christinck on 24 June 2014. The user has r equested enhanc ement of the do wnloaded file. Fachgebiet Landwirtschaftliche Kommunikations- und Beratungslehre Prof. Dr. Volker Hoffmann Module Reader M 4301-410 Knowledge and Innovation Management (KIM ) WS 2011/12 A 7020 1 Modul: Knowledge and Innovation Management (4301- 410) Modulverantwortliche/r: Prof. Dr. Volker Hoffmann Studiengang: Agrarwissenschaften - Agricultural Economics (Master, 2004 - 03 - 22), Sem., semi - elecitve Organic Food Chain Management (Master, 2005 - 07 - 01), 1. Sem., elective Bezug zu anderen Modulen: Good completion t o Rural Communication and Extension (M5121) or Beratungslehre (B0031) and/or Fachkommunikation (B0030). Teilnahmevoraussetzungen: none Sprache: englisch ECTS gesamt: 6 credits Angebotshäufigkeit: each WS Dauer des Moduls: 3,5 weeks (B04) Studienleist ung: Modulprüfung: oral Arbeitsaufwand: 140 - 180 h Lern - und Qualifikationsziele: To understand the dynamics of continuity and change, and how to manage them it in different fields like firms, administrations, universities, unions or farms. Understand t he problem solving, knowledge generation and knowledge systems and the adoption and diffusion of innovations. Ability to apply innovation and diffusion concepts to complex cases. Lehrveranstaltung: Knowledge and Innovation Management (4301- 411) LV - Verantw ortliche/r: Prof. Dr. Volker Hoffmann Lehrform: Lecture with exercises Verbindlichkeit: compulsary SWS: 4 ECTS: - credits Prüfung: Inhalt: Information and knowledge systems. Science and everyday life. Le arning, problem solving, researching and exploring. Theories of innovation, adoption and diffusion. Economics of research and development. Protection of knowledge properties. Strategies of innovation and change. Change management. Examples of technological revolutions worldwide . Literatur: Module reader – available at ASTA and under ILIAS. Veranstaltungsort: Hohenheim Anmerkungen: Lecture with exercises, homework assignments, data - projector. Open to external participants. H OFFMANN , Volker (ed.) 2011:K nowledge and Innovation Management. Module Reader. Hohenheim University. 2 Universität Hohenheim Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre (430A) KIM-00 August 2011Hoffmann 4301-410: Knowledge and Innovation Management Indications on exam volume, content and literature The Module Knowledge and Innovation Management consists of 56 contact hours (4 SWS) and is offered for students in the M.Sc. Studies of Agricultural Economics and as an elective for all other Master Courses. The Module aims to provide students with a basic knowledge about the development and diffu- sion of innovations and related questions of information and knowledge systems, in order to gain more insight into the social aspects of agriculture. The Module is more than just the script and the presentations, which present basic concepts, se- lected examples, and cases. The Module starts with epistemology, looking at the relationship be- tween humans, reality, and the de velopments in the theories of knowledge, and the implications for modern science and philosophy. The Renai ssance was a tremendous period of change in Europe. It shows interaction between the parts (personalities) and the whole (political, cultural and technological change). It shows that history changes, and that much of today is owed to the past. Students get personalities from the renai ssance period to do homework with and to see changes from individual, to social and technological changes. Students should gain an apprecia- tion for the fact that the individu al can never make an impact without being a part of a network, which depends a lot on the context of time and space. The chapter on knowledge perspectives presents the emergence and development of a new field of study that links theory and practice in a sy stems approach to conceptualize the entire knowl- edge process. Based on this , the Module proceeds to deal with its main parts: knowledge man- agement and innovation processes. Case studies ar e presented to illustrate the purpose and prac- tices of knowledge management in an organizati on or a system. Through the case study exercises, students will understand the intera ctions between individuals, organizations and social networks and processes. The diffusion of innovations as a so cial process is covered at length with examples and case studies. The Module provides a summary of the classical approach to the diffusion of innovations (R OGERS 2003) and to the criticisms of diffus ion research. Book reviews are used to illustrate developments and gaps in diffusion re search. The Module presents the Hohenheim con- cept of the diffusion of innovations and its practical implications for advisory work or bringing about change. The consequences of innovations are a prevalent theme throughout the Module.

The Module ends with a final exam. Students ar e given 2 hours (120 minutes) to answer 6 ques- tions. Students must answer all six questions, but only the top five answ ers will count toward a student’s final exam score. The highest score is a 4.0. A score of 1.0 is required in order to pass. The reader together with the powerpoint-files cont ains all of the information and material neces- sary to be able to answer the questions. The reader is available as printout at the ASTA or can be downloaded from ILIAS as well as all powerpoint-presentations. 3 Objectives:

Students should • understand the dynamics and continuity of change • understand how to manage cha nge at different levels • gain insight into the knowledge creation and utilization processes and system dynamics • understand the development of the adopti on and diffusion of innovations research Exam volume • General definitions, concepts and appr oaches of knowledge management • Knowledge management in firms and organizations • The generation of knowledge thr ough science and the history of the European University • Basic definitions and concepts of R OGERS ’ classical innovation theory • The main critics on E. R OGERS • The Hohenheim concept of adoption and diffusion • The comparison of farmers’ and re searchers’ knowledge systems • The interrelationship between techni cal innovation and social change The various parts of the reader Number page KIM-01 The Allegory of the Cave 6 KIM-02 Creativity in Science and Technology 11 KIM-03 Knowledge Management 17 KIM-04 Models of knowledge transfe r: critical perspectives 23 KIM-05 Diffusion of Innovations-Summary 37 KIM-06 Diffusion of Hybrid Corn in Iowa 51 KIM-07 Acceptance of the Salk Polio Vacci ne – an example of the situational approach to the diffusion of innovations 57 KIM-08 Book Review: Rogers & Shoemaker, 1971 59 KIM-09 Book Review: Five editions (1962-2003) of Everett R OGERS : Diffusion of Innovations 64 KIM-10 Basic concepts for understanding adoption and diffusion 75 KIM-11 The diffusion of innovat ions - the Hohenheim concept 87 KIM-12 Nondiffusion of the Dvorak Keyboard 97 KIM-13 Learning selection 99 KIM-14 Farmers and researchers 120 KIM-15 The diffusion of eco-farming in Germany 138 KIM-16 Farmer innovation in Africa 154 KIM-17 As Hegel once examined students 163 4 3 PowerPoint presentations Title KIM-P01 The allegory of the cave KIM-P02 Knowledge management: The case of the Tea Company, Part 1 KIM-P03 How knowledge management works: The case of the Tea Company, Part 2 KIM-P04 Knowledge management: Basic understanding and definitions KIM-P05 Intellectual property rights KIM-P06 Renaissance KIM-P07 Carl Hirnbein KIM-P08 The history of the University KIM-P09 Adoption and diffusion theory KIM-P10 Case: Bull fattening with corn silage KIM-P11 Mapping knowledge systems KIM-P12 Farmers and researchers compared Further reading material / literature 1 The following literature can be helpful in gaining additional insights and orientation for the exam.

These books will be help deepen the vari ous topics covered during the module. BEAL , George, M., D ISSANAYAKE , Wimal, K ONOSHIMA , Sumiye (Eds.) 1986: Knowledge Generation, Exchange and Utilization. Westview Press, Boulder, Colorado. FGB: 3631 H OFFMANN , VOLKER . et al. (2009): Rural Extension. Volume 1: Basic Issues and Concepts (ed.) FGB:

Hdb 362,3.

L EEUWIS , C., 2003: Communication for Rura l Innovations. Rethinking Agricultural Extension. Third Edition. Blackwell. FGB: 5304.

N OTEBOOM , Bart, 2000: Learning and innovation in organi zations and economies. Oxford University Press, Oxford, New York. FGB: 5288 R OGERS , Everett, 2003: The Diffusion of Innovations. Fifth Edition. The Free Press, New York. FGB:

4437,4 R EIJ , Chris, W ATERS -B AYER , Ann, 2001: Farmer Innovation in Africa. Earthscan, London. 362 S. FGB 5066 1 For easy finding we indicate the library signat ures: FGB = Department library, BB =Economic and Social Sciences Library, UB = C entral Library of the University.

5 Universität Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-01 January 2007Hoffmann The Allegory of the Cave In short after P LATON 1. Plato realizes that the ge neral run of humankind can think, and speak, etc., without (so far as they acknowledge) any awareness of his realm of forms.

2. The allegory of the cave is supposed to explain this.

3. In the allegory, Plato compares people untutored in the Theory of Forms to prisoners chained in a cave, unable to turn their heads. All they can see is the wa ll of the cave. Behind them burns a fire. Between the fire and the pr isoners there is a parapet, along which puppeteers can walk. The puppeteers, who are behind the prisoners, hold up puppets that cast shadows on the wall of the cave. The prisoners are unable to see these puppets, the real objects, that pass behind them. What the prisoners see and hear are shadows and ec hoes cast by objects that they do not see. 4. Such prisoners would mistake appearance for reality. They would think the things they see on the wall (the shadows) were real; they would know nothing of the real causes of the shadows.

5. So when the prisoners talk, what are they talking about? If an object (a book, let us say) is carried past behind them, and it casts a shadow on the wall, and a prisoner says "I see a book," what is he talking about? He thinks he is talk ing about a book, but he is really talking about a shadow. But he uses the word "book." What does that refer to?

6. Plato gives his answer at line (515b2). The text here has puzzled many editors, and it has been frequently emended. The translation in G RUBE /R EEVE states the point correctly: " And if they could talk to one another, don't you think they'd suppose that the names they used applied to the things they see passing before them? " 7. Plato's point is that the prisoners are mist aken. they are using the terms in their language to refer to the shadows that pass before their eyes, rather than (as is correct, in Plato's view) to the real things that cast the shadow s. If a prisoner says "That's a book" he thinks that the word "book" refers to the very thing he is looking at. But he would be wr ong. He's only looking at a shadow.

He cannot see the real referent of the word "book" . To see it, he would have to turn his head around.

8. Plato's point: the general terms of our language are not "n ames" of the physical objects that we can see. They are actually name s of things that we cannot see, things that we can only grasp with the mind.

9. When the prisoners are released, they can turn their heads and see the real objects. Then they realize their error. What can we do that is analogo us to turning our heads and seeing the causes of the shadows? We can come to grasp the Forms with our minds.

10. Plato's aim in the Republic is to describe what is necessary for us to achieve this reflective understanding. But even without it, it remains true that our very ability to think and to speak de- pends on the Forms. For the term s of the language we usually get their meaning by "naming" the Forms that the objects we perceive participate in.

6 11. The prisoners may learn what a book is by their experience with shadows of books. But they would be mistaken if they thought that the word "book" refers to something that any of them has ever seen. Likewise, we may acquire concepts by our perceptual experience of physical objects.

But we would be mistaken if we thought that the concepts that we grasp were on the same level as the things we perceive.

And here the “original text”: Plato: The Allegory of the Cave, from The Republic Plato, the most creative and influential of Socrat es' disciples, wrote dialogues, in which he fre- quently used the figure of Socrates to espouse hi s own (Plato's) full-fledged philosophy. In "The Republic," Plato sums up his views in an image of ignorant humanity, trapped in the depths and not even aware of its own limited perspective. The rare individual escapes the limitations of that cave and, through a long, tortuous intellectual journe y, discovers a higher realm, a true reality, with a final, almost mystical aw areness of Goodness as the origin of everything that exists. Such a person is then the best equippe d to govern in society, having a knowledge of what is ultimately most worthwhile in life and not just a knowledge of techniques; but that person will frequently be misunderstood by those ordinary folks back in the cave who haven't shared in the intellectual in- sight. If he were living today, Plato might re place his rather awkward cave metaphor with a movie theatre, with the projector replacing the fire, the film replacing the objects which cast shadows, the shadows on the cave wall with the pr ojected movie on the screen, and the echo with the loudspeakers behind the screen. The essential point is that the prisoners in the cave are not seeing reality, but only a shadowy representation of it. The importance of the allegory lies in Plato's belief that there are invisible truths ly ing under the apparent surface of things which only the most enlightened can grasp. Used to the world of illusion in the cave, the prisoners at first re- sist enlightenment, as students re sist education. But those who can achieve enlightenment deserve to be the leaders and rulers of a ll the rest. At the end of the passa ge, Plato expresses another of his favorite ideas: that education is not a process of putting knowledge into empty minds, but of mak- ing people realize that which they already know. This notion that truth is somehow embedded in our minds was also powerfully influential for many centuries.

Judging by this passage, why do you think many peopl e in the democracy of Athens might have been antagonistic to Plato's ideas? What does the sun symbolize in the allegory?

Is a resident of the cave (a prisoner, as it were) likely to want to make the ascent to the outer world? Why or why not? What does the sun symbo lize in the allegory? And now, I said, let me show in a figure how far our natu re is enlightened or unenlightened:--Behold! human beings liv- ing in an underground den, which has a mouth ope n towards the light and reaching all along the den; here they have been from their childhood, and have their legs and necks chained so that they cannot move, and can only see before them, being prevented by the chains from turning round their heads. Above and behind them a fire is blazing at a distance, and between the fire and the prisoners there is a raised way; and you will se e, if you look, a low wall built along the way, like the screen which marionette players have in fr ont of them, over which they show the puppets.

I see.

And do you see, I said, men passing along the wall car rying all sorts of vessels, and statues and figures of animals made of wood and stone and various materials, which appear over the wall?

Some of them are ta lking, others silent.

You have shown me a strange image, and they are strange prisoners.

7 Like ourselves, I replied; and they see only their own shadows, or the shadows of one another, which the fire throws on the opposite wall of the cave?

True, he said; how could they see anything but the shadows if they were never allowed to move their heads?

And of the objects which are being carried in like manner they would only see the shadows?

Yes, he said.

And if they were able to converse with one another, would th ey not suppose that they were nam- ing what was actually before them?

Very true.

And suppose further that the prison had an echo which came from the other side, would they not be sure to fancy when one of the passers-by spok e that the voice which they heard came from the passing shadow?

No question, he replied.

To them, I said, the truth would be litera lly nothing but the shadows of the images.

That is certain.

And now look again, and see what will naturally follow if the prisoners are released and dis- abused of their error. At first, when any of them is libera ted and compelled suddenly to stand up and turn his neck round and walk and look towards the light, he will suffer sharp pains; the glare will distress him, and he will be unable to see the realities of which in his former state he had seen the shadows; and then conceive some one sayi ng to him, that what he saw before was an il- lusion, but that now, when he is approaching nearer to being and his eye is turned towards more real existence, he has a clearer vision,--what wi ll be his reply? And you may further imagine that his instructor is pointing to the objects as they pass and requiring him to name them,--will he not be perplexed? Will he not fancy that the shadow s which he formerly saw are truer than the ob- jects which are now shown to him?

Far truer.

And if he is compelled to look stra ight at the light, will he not have a pain in his eyes which will make him turn away to take refuge in the obje cts of vision which he can see, and which he will conceive to be in reality cl earer than the things which are now being shown to him?

True, he said.

And suppose once more, that he is reluctantly dragged up a steep and rugged ascent, and held fast until he is forced into the presence of the sun himself, is he not likely to be pained and irritated?

When he approaches the light his eyes will be d azzled, and he will not be able to see anything at all of what are now called realities.

Not all in a moment, he said.

He will require to grow accustomed to the sight of the upper world. And first he will see the shadows best, next the reflections of men and other objects in th e water, and then the objects themselves; then he will gaze upon the light of the moon and the stars and the spangled heaven; and he will see the sky and the st ars by night better than the sun or the light of the sun by day?

8 Certainly.

Last of all he will be able to see the sun, and not mere reflections of him in the water, but he will see him in his own proper place, and not in another; and he will contemplate him as he is.

Certainly.

He will then proceed to argue that this is he w ho gives the season and the years, and is the guard- ian of all that is in the visible world, and in a cer tain way the cause of all things which he and his fellows have been accustomed to behold?

Clearly, he said, he would first see the sun and then reason about him.

And when he remembered his old habitation, and the wisdom of the den and his fellow-prisoners, do you not suppose that he would felicitate himself on the change, and pity them?

Certainly, he would.

And if they were in the hab it of conferring honors among themse lves on those who were quickest to observe the passing shadows and to remark which of them went before, and which followed af- ter, and which were together; and who were therefor e best able to draw conclusions as to the fu- ture, do you think that he would care for such honors and glories, or envy the possessors of them?

Would he not say with Homer, Better to be the poor servant of a poor master, and to endure anything, rather than think as they do and live after their manner? 1 Yes, he said, I think that he would rather suffer anything than entertain these false notions and live in this miserable manner.

Imagine once more, I said, such a one coming suddenly out of the sun to be replaced in his old situation; would he not be certain to have his eyes full of darkness?

To be sure, he said.

And if there were a contest, and he had to compete in measuring the shadows with the prisoners who had never moved out of the den, while his sight was still weak, and before his eyes had be- come steady (and the time which would be needed to acquire this new habit of sight might be very considerable), would he not be ridiculous? Men would say of him that up he went and down he came without his eyes; 2and that it was better not even to think of ascending; and if any one tried to loose another and lead him up to the light, let them only catch the offender, and they would put him to death. 3 No question, he said.

1 This refers to a famous passage in Homer's Odyssey in which the gh ost of the great hero Achilles, when asked if he is not proud of the fame his deeds have spread throughout the world, answers that he would rather be a slave on a worn-out farm than king over all of the famous dead. Interestingly, Plato quotes the same passage else where as disapprovingly as depicting life after death in such a negative manner that it may und ermine the willingness of soldiers to die in war. 2 The comic playwright Aristophanes had mocked Socr ates by portraying Plato's master, Socrates, as a foolish intellectual with his head in the clouds.

3 Plato undoubtedly had in mind the fact that the Athenians had condemned to death his master Soc- rates, who Plato consider ed supremely enlightened.

9 5 This entire allegory, I said, you may now appe nd, dear Glaucon, to the previous argument; the prison-house is the world of sight, the light of the fire is the sun, and you will not misapprehend me if you interpret the journey upw ards to be the ascent of the soul into the intellectual world ac- cording to my poor belief, which, at your desire , I have expressed--whether rightly or wrongly God knows. But whether true or false, my opinion is that in the world of knowledge the idea of good appears last of all, and is seen only with an effort; and, when seen, is also inferred to be the universal author of all things beautiful and right, pa rent of light and of the lord of light in this visible world, Here Plato describes his notion of God in a way that was influence profoundly Christian theologians. and the immediate source of reason and truth in the intellectual; and that this is the power upon which he woul d act rationally either in public or private life must have his eye fixed.

I agree, he said, as far as I am able to understand you.

Moreover, I said, you must not wonder that those who attain to this beatific vision are unwilling to descend to human affairs; for their souls are ever hastening into the upper world where they desire to dwell; which desire of theirs is very natural, if our allegory may be trusted.

Yes, very natural.

And is there anything surprising in one who passes fr om divine contemplations to the evil state of man, misbehaving himself in a ridiculous manner; if, while his eyes are blinking and before he has become accustomed to the surrounding darkness, he is compelled to fight in courts of law, or in other places, about the images or the shadows of images of justice, and is endeavoring to meet the conception of those who have ne ver yet seen absolute justice?

Anything but surprising, he replied.

Any one who has common sense will remember that the bewilderments of the eyes are of two kinds, and arise from two causes, either from comi ng out of the light or from going into the light, which is true of the mind's eye; and he who reme mbers this when he sees any one whose vision is perplexed and weak, will not be too ready to laugh; he will first ask whether that soul of man has come out of the brighter life, and is unable to see because unaccustomed to the dark, or having turned from darkness to the day is dazzled by excess of light. And he will count the one happy in his condition and state of being, and he will pity the other; or, if he have a mind to laugh at the soul which comes from below into the light, there will be more reason in this than in the laugh which greets him who returns from ab ove out of the light into the den.

That, he said, is a very just distinction.

But then, if I am right, certain pr ofessors of education must be wrong when they say that they can put a knowledge into the soul which was not there before, like sight into blind eyes.

They undoubtedly say this, he replied.

Whereas our argument shows that the power and capac ity of learning exists in the soul already; and that just as the eye was unable to turn fr om darkness to light without the whole body, so too the instrument of knowledge can only by the movement of the w hole soul be turned from the world of becoming into that of being, and learn by degrees to endure the sight of being and of the brightest and best of being, or in other words, of the good.

Source: Translated by Benjamin Jowett. Available at http://www.wsu.edu:8080/~wldciv/world_ci v_reader/world_civ_reader_1/plato.html , accessed on 4. 1. 2006 10 Universität Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-02 January 2007Hoffmann Creativity in Science and Technology 1 Steven H. K IM The advance of science is not comparable to the changes of a city, where old edifices are pitilessly torn down to give place to ne w, but to the continuous evolution of zo- ologic types which develop ceaselessly and end by becoming unrecognizable to the common sight, but where an expert eye finds always traces of the prior work of the cen- turies past. One must not think then that the old-fashioned theories have been sterile and vain . Jules Henri Poincaré 2 The scientific enterprise is one of generating new knowledge for its own sake. The technological process, in contrast, refers to the application of knowledge to satisfy human needs aside from cu- riosity. The spectrum of technologi cal activities, from science to engineering and marketing, is depicted in Figure A.l. Types of Discovery If the fields of science and tec hnology are to progress in an orderly fashion, it is important to de- velop a systematic theory for the nature of resu lts attained in these domains. To this end, a framework is presented for the types of discove ries as well as their methods of derivation.

The process of scientific discovery is depicted in Figure A.2. An investigator, working individu- ally, or in concert with collea gues, envisions a result based on her knowledge of the universe.

This knowledge may result from direct personal observation or indirect knowledge through the work of others.

The result is generated through some methodology. The specific method may rely on some cogni- tive mechanism that we do not yet comprehend, such as the realization that sunrise and sunset are due to the rotation of the Earth, rather than the motion of the sun. We call these yet-unknown mechanisms "intuition" or "inspiration." On the other hand, certain methods are more strai ghtforward, as exemplified by the technique of proof by contradiction or refutation. To illustrate this, we may show that a specific statement in predicate logic must be derivable from an initia l set of hypotheses, by showing that the negation of the statement would imply some inconsistency in the overall set of statements. In fact, this refutation procedure is routinely used as th e basis for the programming language of Prolog.

The results of scientific research may be classifi ed into the following four categories: alignment, possibility, impossibility, and tr ade-off. These groups of results may be obtained by methods of construction or contradiction.

1 From: K IM, Steven H., 1990: Essence of Creativity. A Guide to Tackling Difficult Problems. Oxford University Press, Oxford, 88-92 2 POINCARE 1904: Valeur de la Science, as given in P OINCARE 1946, 208.

11 Table A. l shows examples of results by category and method of proof. These classifications are discussed further below.

Alignment Alignment refers to the fit between our models and the world around us, or among the models themselves. In attaining such harmony, our perception of the universe takes a simpler form. The issue of alignment may be furt her classified into two types: paradigm and unification.

12 A paradigm is defined by our perception of the universe. Hence a paradigmatic result refers to a shift in our views 3. For example, the replacement of the geo centric view of the solar system by the heliocentric paradigm represented a major adva nce, and was instrumental for further advances in astronomy.

On the other hand, unification refers to the alignment among our models or views of the world. A unifying structure provides a general framework for organizing results that previously seemed un- related. The structure may take the form of a fr amework, model, theory, or some combination of the three.

An example of a unifying structure is the developm ent of the periodic table, and the classification of elements into related groups based on their electron configur ations. Another example is found in the theory of electromagnetism, which unifies the seemingly unrelated phenomena of electric- ity and magnetism. This integrative model al so accounts for many types of radiation, from gamma rays at one end of the sp ectrum to radio waves at the othe r. We now recognize ultraviolet emissions, as well as light and heat, as variations on the single theme of electromagnetic radia- tion.

Another unifying structure is found in the laws of thermodynamics, relating the conservation of energy and the tendency of systems toward incr easing disorder. These two laws encapsulate ob- servations that arise in all realms of natural science and engineering.

A subcategory of unification relate s to the laws of invariance. Invariance refers to the constancy among objects that appear to be diffe rent at first sight. Such principles assert, for example, the immutable nature of certain obj ects despite transformations.

The conservation laws of physics ty pify the category of invariance principles. For example, the principle of matter-energy conser vation states that matter and energy may change from one form into another, but the total ener gy of a system remains the same.

The single most important result in the field of statistics is the Central Limit Theorem. This theo- rem states that a particular probability law call ed the Gaussian distribution is pervasive. The 3 KUHN 1962, Ch. 1 13 Gaussian function, familiar to many people as the "bell-shaped" curve, serves as a good model in many practical and theoretical applications 4. A common example is intelligence scores, whose bell curve peaks at 100 and trails off toward either end to reflect the fact that decreasing numbers of individuals possess either very high or very low scores. Engineers often rely on dimensional analysis to show the validity of thei r reasoning. This approach is ba sed on the idea that the physi- cal properties of systems may depe nd only on the combinations of certain characteristics, rather than their individual values. The field of fluid mechanics, for example, uses the Reynolds num- ber: R = pvd l µ where p is the density of a fluid having viscosity µ. and flowing with velocity v in a conduit di- ameter d. The units of these parameters are such that R is a dimensionless number. The fluid travels in orderly, laminar flow for low values of R, and becomes turbulent for high values.

Crude oil flowing in a transcontinental pipeline has a different character from water running through a garden hose. Their densities differ as well as their viscosity or internal resistance to flow; the diameters of the conduits will be dissimilar, and the speeds of flow may also vary. In other words, the values of the density p, the velocity v, the diameter d, and the viscosity µ. are distinct for each fluid. But as long as the compound factor pvd l µ is much less than the threshold of about 1000, the flow will be laminar. On the othe r hand, if the number is much higher than this threshold, the flow will be turbulent. The identity of the fluid itself is of little consequence in this determination. Only the composite parameter in the form of the Reynolds number is a reliable in- dicator of turbulence, short of actually observing the fluid under the stated conditions of flow.

The field of automata theory uses a model of computation called the Turing Machine. This model depicts computational procedures as a set of s imple operations. Much of the work in automata theory deals with the invariance of computationa l power among different versions of the Turing Machine. One such result is the equivalence of all existing computers to the Turing Machine - and therefore to each other - in the range of problems they can resolve.

Einstein's Theory of Relativity may also be rega rded as an invariance result. In particular, the laws of physics are unchanged by the choice of a particular frame of reference. Possibility Much of the work in the sciences and in engineer ing deals with showing what is possible. One of the most convincing ways is proof by construction. The most cogent means of showing that hu- mans can attain powered flight is to build a flying machine. The seminal experiment by the American chemist Stanley Lloyd Miller in the mid-1950s showed that amino acids - the basic components of life - can be formed from a broth of lifeless chemicals when exposed to a flux of energy. 5 A good deal of the theoretical work in the sciences is also one of construction. This relates to the development of general models, frameworks, or theories that can accommodate diverse empirical observations.

4 The mathematically inclined reader may be aware that according to the Central Limit Theorem, the Gaussian distribution is pervasive in the following sense. Given any set of random variables that are mutually independent and have an finite variance, their sum tends toward a Gaussian distribution when the set is large. This resu lt is unchanged by the fact that the independent variables may have similar or dissimilar probabilit y distributions, any of which (or none) might be Gaussian.

5 M ILLER , 1957.

14 Analytic studies in engineering are often dedicated to the determination of what is possible. For example, the hypothesis that manned spacecraft can visit Mars and return to Earth can be con- firmed from our knowledge of the chemical prope rties of propellants, the mechanical properties of materials, and the physics of interplanetary flight. Impossibility A negative result, if proven, is as useful as a positive result. In fact, most of the major scientific advances in the 20th century are of the negative kind: Einstein's Theory of Relativity says that there are no absolutes; Heisenberg's Uncertainty Principle states that position and momentum cannot both be determined simultaneously with arbitrary precision; Gödel's Incompleteness Theorem says that there is no decisive algorithm to prove invalidity in predicate logic; according to Arrow's Impossibility Theorem, a particular set of reasonable assumptions will admit no con- sistent economic welfare function. These discoveries of the im possible actually serve to define the limits of the possible.

Trade-offs An important class of results relates to interdimensional trade-offs 6. These may relate to the rela- tionships between performance and effici ency, or time versus space, etc.. The student of economics quickly learns that life involves the pursuit of happiness under resource constraints. A simplified economy may have sufficient resources each year to produce exactly one of the following baskets of goods: 5 million mu ffins; or 800 videos; or 2 million muffins and 400 videos. These three alternatives define the "production possibi lity frontier" for the economy.

The actual choice among the three alternatives will depend on the collective disposition of the consumers and producers. The nature of the choi ce is largely a subjective matter; but the trade- offs between muffins and videos is an objec tive phenomenon whose understanding facilitates the subjective decision. Happiness may be a subjective subject, but its pursuit can be supported by ra- tional decision making.

In the realm of computer science, the area know n as complexity theory deals with the conse- quences of differing algorithms for computational efficiency, and the trade-offs between require- ments for memory versus computational time in solving a specific problem. These results support the design of computer systems, just as production possibility frontiers assist in formulating eco- nomic policy. Nature of the Categories The classification of scientific results into a set of categories provides a convenient framework for exposition and discussion. However, the various categories are not intended to be independent or mutually exclusive.

For example, the Theory of Relativity stipulates the lack of absolute frames of reference. This re- sult may be classified as an example of the impossibility of determining an absolute reference as well as one of invariance of physical laws across reference frames.

Moreover, a result that falls into one category may engender results in other categories. As dis- cussed previously, the laws of thermodynamics represent a unifying structure. The first law, how- 6 Kim, 1990, Ch. 2.

15 6 ever, stipulates the invariance of the total amount of energy in any insulated system. In addition, the second law can be used to deduce the impossibility of a perpetual motion machine.

The four categories of results hi ghlight the different types of contributions. This framework can help to promote creativity by providing a platform for the orientation of research efforts and out- lining the nature of result s that may be achieved.

Bibliography:

KIM, S. H. 1990: Designing Intelligence: A Framework for Smart Systems. Oxford University Press, New York.

K UHN , T. S., 1962: The Structure of Scientific Revolutions. 2nd ed. University of Chicago Press, Chi- cago, 1970.

M ILLER , Stanley L., 1957: The Mechanisms of Synthesis of Amino Acids by Electrical Discharges. In:

Biochimica and Biophysica Acta , v.23(3), 480-489.

P OINCARÉ , Henri, 1946: The foundations of Science: Science and Hypothesis; The Value of Science; Science and Method: transl. By G.B. Ha lsted. Science Press, Lancaster PA.

16 Universitat Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-03 November 2008Hoffmann Knowledge management 1 A combination of the two disciplines, struct uring learning processes and organization and management, has recently become popular in the field of organizational development: know- ledge management. We start with clarifying some basic terms and concepts and then go to the implications for learning of individuals and organizations.

Figure 1: The knowledge pyramid (following A AMODT & NYGARD , 1955) management knowledge Symbol competence (reading and writing) data management information management knowledge management signs data information action Ö Figure 1 shows that knowledge links informatio n with action. Information is built on data, which is composed of signs. Si gns stand for something else. Their meaning is fixed by con- vention and must be learned. Data are signs to be analyzed, and this is also possible when they are electronically stored as bits and bytes. The handling of the elements of the pyramid can be called management. Going up and down the knowl edge pyramid requires certain competences, which are normally acquired in literate societies.

Can knowledge be stored? No, at least not direc tly, not outside of living human brains. But it can be stored indirectly as information. Know ledge can be acquired through learning and in- ternalizing information, and it can be passed on by teaching and externalizing it into informa- tion. That means information can be regarded as externalized knowledge and also as processed data serving a purpose or reducing uncertainty. Computers and other machines can process da- ta, but not information. Semantic interpretation, i.e. the assessment of meaning, is a capacity requiring human intelligence, often paired with creativity and no machine can be taught this, even if the term artificial in telligence suggests it. This exte nds our knowledge pyramid as shown in Ö Figure 2. 1 1 Parts of this chapter are based on H OFFMANN , 2007.

17 Figure 2: The extended knowledge pyramid Knowledge Signs Data Information Without meaning, no context Meaning, related to facts Reduces uncertainty, purposeful Conceptualized information, interpreted, evaluated, interlinked To learn, internalize To teach, externalize Semantics Syntax Pragmatics Action Semiotics The knowledge pyramid only depicts a part of all knowledge processes, the explicit part. This can be used to pass information based on signs and data. However, there is also implicit or ta- cit knowledge. P OLANYI (1985) said that " we know more than we can tell ".

Implicit knowledge is knowledge fr om experience, is the part of knowledge that is difficult to describe, is used subconsciously and intuitivel y, is a special feeling, a talent, e.g. "green fin- gers" when dealing with plants, and cannot be co pied or programmed for robots, and is a basis for sustainable competitive advantage of enterprises.

The special trick is to find ways of rendering implicit knowledge explicit, of converting it into information that can be handed on to others. This is where the distinction ari\ ses between knowledge and capability. To acquire skills, readi ng is not very helpful. Progress only comes through repeated practice, by doi ng something again and again, and by training. Skills are im- plicit. Training that is successful in guiding a nd motivating is another teaching skill. Not every expert is also a good trainer or teacher.

Ö Figure 3 goes more into details of individual learning. The bigger circle represents the learner explaining learning as information pro cessing. Signals are perceived and interpreted under the influence of arising emotions and o ccurring cognitions, which are then stored and lead to action. Observing the own action transforms this into data and information that again enters perception. That is what we call learning by doing or experiential learning (K OLB , 1984) and what improves skills by manifold repetition called training. Instead of observing the own actions, we can also learn from the behavior of others called model learning or imitative learn- ing. Apart from behavior, we i nherit the products of previous learning processes, like cultural landscapes, tools and domestic plants and animals, which we go on using. The same happens, when we just buy products or services; there we also profit from the learning results 2 18 Figure 3: The knowledge cycle: How individuals learn Information-Processing Codes 3 of former times and other persons. The most fre quent source of learning nowadays is symbolic learning, accessing information by symbolic comm unication, mainly in school like settings, but also in individual work in libraries or distance learning. Th e lion share of such information is carried by text, spoken or written (linguistic) , often illustrated by different forms of pictures (iconic), sometimes combined with numbers or formulas (logical), and also body language or music can help to ease the information upt ake und understanding or memorizing, e.g. by ac- companying spoken words, and by crea ting additional emotional effects. N ANOKA and T AKEUCHI (1997, 85) were the first internat ionally recognized academics to point out the importance of converting knowledge between imp licit and explicit forms, which they depicted with their famous spiral, shown in Ö Figure 4.

If knowledge is available in expl icit form, then acquiring it means internalizing it. Once inter- nalized, it is implicit, an d to acquire it one must go through so cialization in this field of know- ledge and action. You grow into it and acquire it by helping and assisting, by imitation and training, and often to a great degr ee by trial and error. When you wa nt to pass it on to others without time-consuming socialization, you should ex ternalize it and make it explicit. If differ- ent contents of exp licit knowledge arise, new knowledge may emerge through a process of combination that has to prove its usefulness in actions. Therefore, it has to be internalized first.

ordering shaping Objets Environment Tools Plants Animals... Feeling Learning Behaviour Own Behaviour Beh. of Others Symbols linguistic gestural, mimical iconic logical musical Memory Thinking Reco gnizin g Action Individual Knowledge Individuum Information Uncertainty Reduction „between the ears“ Selective Projektive Data Bits and Bytes Learning Acqaintance Capacities Potential Sources of Infor- mation Perce ption Ordering Shaping 19 Figure 4: Transforming knowledge (N ANOKA & TAKEUCHI , 1997, 85) Socialization Externalization Internalization Combination explicit implicit implicit explicit Because this process between humans is never stat ic, the spiral is an appropriate symbol for representing it. The authors describe how to ex ternalize knowledge, even against all resistance, in their story of the development of the bread-baking machine. When the leader of the development department was near to despair because the bread made in the new automatic machine did not taste good, she became an apprentice to a famous baker in a five-star hotel in Tokyo. Observing what he was doing did not he lp, so she started to pound dough herself, and practice makes her perfect. After some weeks, her bread was as good as his. But in addition, she could explain in technical terms (alt hough he could not) what she was doing. Pounding dough means pulling, pushing and twisting it. The machine was only twisting the dough. To pull and push, some side bands had to be added. The rest was routine work. So by systematic variation her engineers were ab le to optimize the machine. In the first six months 500,000 small automatic household units were sold.

With this example they point out that the implic it knowledge of the co-workers is the decisive comparative advantage in international competition. When it is possible to externalize it, it can be patented, computerized, automatized, used in robots but also sold for license fees.

Organizations, as well valid for extension organizations must constantly adapt to changing conditions and co-evolve w ith their environments. This again is learning. But we stated that knowledge only exists in liv ing brains. If “knowledge is between the ears” (R ÖLING , 1994), how can organizations learn, not having ears? Ö Figure 5 gives a first answer. In an organi- zation, files and people interact. The persons can learn and can improve the information in the files, insofar their learning results can be made explicit. And in creating the outputs, the com- pany performance, the members interact by using their implicit as well as their explicit know- ledge. By interlinking files and people, imp licit and explicit knowledge, the organizations learns and creates synergies. But it would never be sufficient to withstand international competition, if an organization would depend solely on own information sources. Information is nowadays a commodity, which is partly free of charge in public domai n and partly commercialized as indicated on the right side of Figure 6, under external acquisiti on. To “make or buy” is the general question dealt with in the figure, and often buying is cheaper.

Figure 5: How organizations learn I 4 20 5 File Person Firm Ex plicit Explicit Im plicit Explicit Implicit Interlinked Information Knowing Mastering Wanting Learning Or ganization The whole is more than the sum of it‘s parts ! To make the picture more complete, Ö Figure 6 stresses on two additional points: learning by feedback through monitoring results and when exchanging information, services, goods and assets with outside partners or customers. In striving for cons tantly improved reach of objec- tives, not only the members learn and the files imp rove, but also the structures of the organiza- tion are constantly adapted and the standard procedures are improved Ö Figure 7.

In this perspective, we can understand organiza tional development also as a process of im- proving communication and knowledge management.

Figure 6: Knowledge acquisition Externally Internally Try, train Investigate, look up Buy information/ license Buy product/service Consult expert/advisor Train/educate co-workers Hire expert Figure 7: How organisations learn II 21 Services Goods 6 Members Files Structures, Procedures Information Assets, Rights By information processing Bibliography AAMODT , A., NYGARD , M., 1955: Different roles and mutual dependencies of data, Information and Knowledge. In: Data & Knowledge Engineering , 16, 191-222.

H OFFMANN , Volker, 2007: Knowledge management: what are we talking about? In: GTZ Services for Rural Development , No 15, 6-8.

K OLB , D.A., 1984: Experiential learning. Learning experience as a source of learning and development.

Prentice Hall, New York.

N ONAKA , I., TAKEUCHI , H., 1997: Die Organisation des Wissens. Campus, Frankfurt am Main, 299 pp.

P OLANYI , M. 1985: Implizites Wissen. Suhrkamp, Frankfurt am Main. RÖLING , Niels, 1994: Agricultur al Knowledge and Information Systems. In: B LACKBURN (ed.): Extension Handbook, 2 nd edition. Thompson Educational Publishing, Toronto. 57-68. 22 Universität Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-04 January 2007Hoffmann Models of knowledge transfer: critical perspectives 1 Everett M. R OGERS The field of knowledge transfer today has much reason to view its past accomplishments with considerable pride. In the last couple of decades this field of scholarly activity has attracted a growing number of dedicated rese archers and theorists, several outstanding books have appeared that synthesize work on this topic, and a numb er of university-level courses and programs of graduate-level have been launched. On the prag matic side, most government agencies (both in developing and in industriali zed nations) recognize their re sponsibility for conducting know- ledge-transfer activities. In fact, many agenci es see knowledge transfer as one of their main ac- tivities.

Critical perspectives Recently, I met with representatives of a dozen U.S. government agencies (in health, mental health, education, public transpor tation, etc.) to review their knowledge-transfer strategies. Each was allocating a portion (albeit sm all) of its total budget for knowledge transfer, and each had an office or division established to carry out knowledge-transfer functions. Significantly, each of these U.S. government agencies was, in its act ivities, questioning certain aspects of the conven- tional wisdom about knowledge transfer.

I regard this as a healthy sign. Scholarship and practice on knowledge transfer have advanced to the point where we should be questioning our pa st models, and searching for improved alterna- tives, rather than just "doing more of the same." It is in the light of such critical perspectives that the present chapter is written. A theme of this chapter is that a fundamental shift may have oc- curred in recent years as we have realized grad ually that centralized knowledge-transfer systems are not "the only wheel in tow n." While such centralized approaches have advantages under many conditions, in certain cases a more decentral ized model of knowledge transfer may be more appropriate.

Every field of scholarly activity makes certain simplifying assump tions about the complex reality that it studies. Such assumptions are built into th e intellectual paradigm that guides every field.

Often these assumptions are not very fully recognized, even as they affect such important matters as what is studied and what is ignored and which research methods are favored and which are re- jected. So, when a scholar follows a theoretical paradigm, he or she puts on a set of intellectual blinders that help the researcher avoid seeing much of reality. "The prejudice of training is al- ways a certain 'trained incapacity': The more we know about how to do something, the harder it is to learn to do it differently" (K APLAN 1964,31). Such trained incapacity is, to a certain extent, ne- cessary; without it, a scholar could not cope with th e vast uncertainties of the research process in his field. Every research worker, and every field of science, has many blind spots.

1 Source: Everett R OGERS , 1986: Models of Knowledge Transfer : Critical Perspectives. In: B EAL , George, M., D ISSANAYAKE , Wimal, K ONOSHIMA , Sumiye (Eds.): Knowledge Generation, Exchange and Utilization. Westview Press, Boulder, Colorado, 37-60 23 The growth and development of a research field is a gradual puzzle-solving process by which im- portant research questions are identified and ev entually answered. The progress of a scholarly field is helped by realization of its assumptions, biases, and wea knesses. Such self-realization is greatly assisted by intellectual criticism. Unfort unately, the field of knowledge transfer has not been subjected to much critical review, a defici ency that we hope to remedy in this chapter.

What is knowledge transfer?

Past scholarship on issues of knowledge generation, exchange, and utilization grew out of several different disciplines, each of wh ich favored certain theoretical viewpoints, research approaches, and terminologies. While the general trend is toward integration of this intellectual diversity, such academic unity is yet far from being accomplished. Perhaps a certain degree of difference in ap- proach is a good thing, at least up to a certain point, but one area in which diversity still causes troublesome problems is in terminology. What I loos ely refer to in this chapter as "knowledge transfer" is also known as knowledge, utilization, technology, transfer, and the diffusion of inno- vations (although these concepts are not exact synonyms).

We often use "innovation," "technology," and "know ledge" as synonyms, but in fact they are not the same. An innovation is an idea perceived as new. A technology is a design for instrumental action that reduces the uncertainty in the cause-e ffect relationships involved in achieving a de- sired outcome (R OGERS 1983,12). A technology usually has two components: (1) a hardware as- pect, consisting of the tool that embodies the technology as a mate rial or physical object, and (2) a software aspect, consisting of the information ba se for the tool. Both the hardware and software dimensions of a technology encompass knowledge, but of course there are many other kinds of knowledge besides the new knowledge that is involved in an i nnovation or a technology. Never- theless, most of the past studies of knowledge transfer have actually been re searches focusing on innovation or technology transfer.

The scope of knowledge transfer th at has been studied in the past has been the process through which technological information resulting from an R & D system is transferred by a linking sys- tem (e.g., an agricultural extension system) to a user system (e.g., farmers). This conception of knowledge transfer implies that it is mainly a one-way process: in actuality (or at least ideally), the R & D may have been initiated at the request of the user system, or at least in order to meet certain of their needs, further, once the users have received the knowledge and put it into use, feedback (as to how well the knowledge meets th e preexisting needs) may be conveyed back to the R & D system. So it is an ove rsimplification to think of knowle dge transfer as a one-way, top- down process. 2 The agricultural extension model Any discussion of models of know ledge transfer must begin with agriculture extension, both for historical reasons and because the agriculture exte nsion model has so influenced all of our think- ing about this topic. While ou r intellectual dependence upon this model was mainly functional in the past, we have also been unf ortunately limited in the scope of our conceptualizations about 2 This discussion of the c oncept of knowledge transfer raises t he issues of what should be the main dependent variable(s) (1) in research on knowledge transfer, and (2) in practice, as an indicator of performance. In past studies of the diffusion of innovations, the usual dep endent variable has been adoption versus rejection of a technological innovation. But there are many other possible dependent variables in knowledge transfer research and/or practice: Awareness-knowledge of a technological innovation or of another idea, de velopment of a favorable or unfavorable attitude toward the innova- tion or another idea, or beneficia l consequences of adoption or rejecti on of the idea (to meet the orig- inal needs).

24 knowledge transfer. The first step in breaking outside the bounds of our prior thinking is to real- ize that certain alternative models may be possi ble. Of course, it may be advantageous to com- bine certain elements of a rela tively centralized model like agricu ltural extension with parts of a decentralized model to formulate a knowledge transfer system that is especially suited to a set of particular conditions. This contingency approach to knowledge transfer is more academically sound than the numerous descriptions of a knowledge transfer system in the past, which stated or implied that that system was the best alternative for a wide range of conditions. For example, it has been claimed that the agricultural extensi on model could be effectively applied to solve knowledge transfer problems in education, fam ily planning, vocational rehabilitation, and so forth. An examination of these "extensions" of the agricultural extension model, however, has shown them to be relativel y unsuccessful unless major m odifications were made (R OGERS , EVEL- AND , and B EAN 1984). The agricultural extension model is a set of assumptions, principles, and organizational struc- tures for diffusing the results of agricultural research to farm audiences in the United States. This "model" is based directly on the experience of the U.S. government agency responsible for diffus- ing agricultural innovations; it closely parallels the conventional conceptions of a research and development/diffusion/u tilization process.

Eight main elements constitute the agricultural extension model:

1. A critical mass of new technology , so that the diffusion system has a body of innovations with potential usefulness to practitioners.

2. A research subsystem oriented to utilization , as a result of incentives and rewards for re- searchers, research funding policie s, and the personal ideologies of the agricultural researchers.

3. A high degree of user control over the know ledge transfer/research utilization process, as evidenced through client participation in policy determination, attention to user needs in guiding research and diffusion decisions, and the importance accorded feedback from clients on the sys- tem's effectiveness. 3 4. Structural linkages among the research utilization system's components , as provided by a shared conception of the system, use of a comm on "language" by members of the system, and by a common sense of mission.

5 A high degree of client cont act by the linking subsystem, which is facilitated by reasonable agent/client ratios and by a rela tively homogenous client audience. 6. A spannable social distance across each in terface between components in the system (where social dist ance might occur in levels of professi onalism, formal education, technical ex- pertise, and specializatio n). Generally, these variables decrease as one moves from the research subsystem (where Ph.D. 's are usually employed) , through linkers, to the client subsystem.

7. Evolution as a complete system , rather than the knowledge transfer system having been grafted on as an additional component to an existing research system.

8. A high degree of control by the system over its environment, thus enabling the system to shape the environment rather than passively reacti ng to changes. Such a system is less likely to face unexpected crises or competit ors, and is able to obtain adequate resources. The degree of 3 While much rhetoric is given to this feedback about needed research from farmers through the ex- tension service to agricultura l scientists, it is actually a fairly rare occurrence.

25 control is expressed through the system's power base, its perceived legitimacy, and its policial- legal influence.

The following generalizations are offered about the agricultural extension model:

1. In response to alterations in the environment, the agricultur al extension model has changed considerably since its origin in the United States in 1911. To a large extent, these adjustments are a reason for its relative success.

2. The agricultural extension model is based on cl ient participation in identifying local needs, program planning, and evaluation and feedback.

3. Agricultural research activities are oriented toward potential utili zation of research results. This pro-utilization policy facil itates the linking function of the extension workers.

4. State-level extension specialists are in close social and spatial contact with agricultural re- searchers in their specialty, whic h facilitates their performance in linking research-based know- ledge to farmer problems.

5. The agricultural extension model was more eff ective in diffusing agricultural production tech- nology to farmers (such as in crop and livestock production) than in its latter-day extensions to farmers on other subjects and to non-farm audiences.

6. The agricultural extension model recognizes the importance of communication as a basic process-skill for extension change agents an d provides communication training on an in-service basis.

7. The agricultural extension model includes not onl y a systematic procedure for the transfer of innovations from researchers to farmers but also an institutionalized means for orienting research activities toward users' needs. Thus, the land-grant college/agricu ltural experiment station/extension service complex is a total knowledge utilization system, which includes innova tion-diffusion as only one of its components.

The federal investment in agricultural extensi on represents a heavy commitment, compared to that in agricultural research. Fe derally funded extension activitie s represent about 40 to 60 per- cent of the annual federal investment in agricu ltural R & D. For example, the USDA recently al- lotted $423 million for R & D. This figure would be considerably higher (over $600 million) if state funding were also included. 4 The annual federal budget for extension was $200 million:

with state and county government contributions, the total annual budget for the extension services was about $500 million. Thus the total extension budget almost approaches the total public agri- cultural R & D budget. Even if only the federal investment is consider ed, extension receives about half the funding of R & D. Comparable figur es for federal extension-type activities as a proportion of federally supported R & D are much, much smaller in other fields:

1. Law Enforcement Assistance Administration 14 % 2. National Institute of Education 10 % 3. U. S. Department of Labor 3 % 4. National Institute of Mental Health 2 % 5. National Aeronautics and Space Ad ministration (NASA) 0.17% 4 The activities of the extensi on services over the years have focused somewhat narrowly on imme- diate technical problems in agriculture, rather th an on the longer range social, political, economic, and ecological consequences of technological change in U.S. agriculture (H IGHTOWER 1972).

26 Undoubtedly one of the reasons for the success of the agricultural extension services is their rela- tively high, stable budget. The financial success is, in turn, aided by the support given to the agri- cultural extension services by the powerf ul American Farm Bureau Federation.

Extending the agricultural extension model What factors drawn from the agricultural exte nsion model can be applied to other knowledge transfer systems, and which are unique to the agricultural extension services? In other words, can the agricultural extension model be extended to other situations? R OGERS , EVELAND , and B EAN (1984) compared seven selected attempts to ex tend the agricultural extension model on the eight main elements of the model (that we stated pr eviously). The seven "extensions" occurred during the 1955 to 1975 period and represent cases with which the analysts were personally acquainted.

These seven "extensions" have (in most cases) been extensively evaluated , and so rather definite conclusions are possible.

The general pattern of extension system development in the agricultural case, and the relative successes and failures evidenced in the other seve n cases, suggest some broad conclusions about knowledge transfer. The historical development of the agricultural extension system stretches over about 100 years. Comparatively speaking, know ledge transfer efforts in education, vocation- al rehabilitation, and other fields appear woefully underfunded and to have been treated like un- wanted children of over-expectan t parents. Two experiences (agricultural extension and family planning) in the developing countries of Latin America, Africa, and Asia show a lack of under- standing of the importance of cultural adaptation of elements of the agricultural extension model (even when the model is applied to agricultural problems).

The county extension agent in th e United States was a product of commercial agriculture, not subsistence farming. Until American agriculture be gan to modernize, there was not much need for an extension service. Subsistence (precomme rcial) farming in developing countries has not embraced the agricultural extensi on model with much success, a fact that suggests that the suc- cessful introduction of a knowledge transfer system must be carefully timed so that a feeling of need for its services exists or can soon be developed.

Attempts to introduce one or two elements of the agricultural extension model to non-agricultural settings should not be undertaken without adeq uate appreciation of the difficulties involved. The time and resources required to permit these knowledge transfer elements to prove their utility and to become assimilated into the culture of the host system can be easily underestimated. The fail- ure of modestly funded efforts to transplant spec ific elements of the agricultural extension model into other sectors sugge sts that an extension system approach needs to be taken. When only cer- tain elements of the agricultural extension model were introduced without support from the other elements, they usually failed.

Knowledge cannot be transferred effectively unless the goals of such transfer are very clear. The goals of the agricultural extension services were fairly di rect and unambiguous: to produce more food and to raise farm incomes. In education and in rehabilitation, for example, the goal situation is much more complicated, with multiple, conflicting goals for knowledge transfer.

The agricultural extension services begin with users' needs and problems, and the system operates to find useful information to meet these needs, while many other, less effective knowledge trans- fer systems take an opposite approach of conducti ng research largely in answer to researchers' needs, and then attempting to find some use for th e results. Naturally, the research topics usually do not match with users' needs. An effective know ledge transfer system must begin with users' needs.

27 Decentralized diffusion systems We have already implied that there is considerable flexibility in the way the eight elements of the agricultural extension model can be adapted in se lected knowledge transfer systems. During very recent years, a diffusion system in marked contrast to the centralized diffusion system of the agri- cultural extension model has been id entified: decentralized diffusion.

Centralized Diffusion System In 1971, Professor Donald S CHON of MIT wrote that " theories of diffusion have characteristically lagged behind the reality of emerging systems. " S CHON particularly singled out classical diffu- sion theory for criticism: he te rmed classical diffusion a "cente r-periphery model." This model, S CHON (1971, 81) said, rests on the basic assumption that " An innovation to be diffused exists ful- ly realized in its essentials, prior to its diffusion " and that the diffusion process can be centrally managed. The best-known example of a centralized diffusion system is the agricultural extension services.

In this classical diffusi on model, an innovation originated from some expert source (often an R & D organization). This source then diffuses the innovation as a uniform package to potential adop- ters, who accept or reject the i nnovation. The role of the adopter of the innovation is that of a rel- atively passive accepter. This classical model owes much of its popularity to the success of the agricultural extension services and to the fact th at the basic paradigm for diffusion research grew out of the R YAN and G ROSS (1943) hybrid corn study. Much agri cultural diffusion in the United States is relatively centralized, in that key decisions about which innova tions to diffuse, how to diffuse them, and to whom are made by a small num ber of technically expert officials near the top of a diffusion system. While S CHON noted that it fails to capture the complexity of relatively decentralized diffusion systems in which innovati ons originate in numerous sources and evolve as they diffuse via horizontal networks.

During the late 1970s, I gradually became aware of diffusion system s that did not operate at all like the relatively centralized diffusion system that I had described in previous publications. In- stead of coming out of formal R & D systems, innovations ofte n bubbled up from the operational levels of the system, with the inventing done by users. Then the new ideas spread horizontally via peer networks, with a high degree of re-inventing occurring as the innovations were modified by users to fit their particular c onditions. Such decentralized diffusi on systems usually are not run by a small set of technical experts. Instead, decision making in the diffusion system is widely shared with adopters making many decisions. In many cas es, adopters served as their own change agents.

Gradually, I begun to realize that the centralized diffusion model was not the only wheel in town.

Comparing centralized versus decentralized diffusion systems How does a decentralized diffusion system differ fr om its centralized counterpart? Table 1 shows six of the main differences betw een centralized and decentralized diffusion systems. This distinc- tion is somewhat oversimplified because it suggests a dichotomy, rather than a continuum, of centralized/decentralized diffusion systems. In real ity, an actual diffusion system is usually some combination of the elements of a centralized an d a decentralized diffusion system. For example, the agricultural extension services in the United St ates are nearer the more centralized end of the centralized/decentralized continuum, although they ha ve certain characteristics of a decentralized system.

In general, centralized diffusion systems are based on a linear, one-way model of communication.

Decentralized diffusion systems more closely follow a convergence model of communication, in 28 Table 1: Characteristics of centralized a nd decentralized diffusion systems Characteristics of diffusion Centralized diffusion systems Decentralized diffusion systems The degree of centralization in decision-making and power. Overall control of decisions by national government adminis- trators and technical subject- matter experts. Wide sharing of power and con- trol among the members of the diffusion system; client control by local officials/leaders.

Direction of diffusion.

Top-down diffusion from ex- perts to local users of innova- tions. Peer diffusion of innovations in- novations through horizontal networks.

Sources of innovations.

Innovations come from formal R & D conducted by technical ex- perts. Innovations come from local ex- perimentation by nonexperts, who often are users.

Who decides which innova- tions to diffuse?

Decisions about which innova- tions should be diffused are made by top administrators and technical subject-matter spe- cialists. Local units decide innovations should the basis of their evalua- tions of the innovations.

How important are clients' needs in driving the diffu- sion process?

An innovation-centered ap- proach; technology-push, em- phasizing needs created by the availability of the innovation. A problem-centered approach; technology-pull, created by local- ly perceived needs and prob- lems.

Amount of re-invention ?

A low degree of local adapta- tion and re-invention of the in- novations as they diffuse among adopters. A high degree of local adaptation and re-invention of the innova- tions as they diffuse among adopters.

Source: R OGERS 1983,335 which participants create and share information with one another in order to reach a mutual un- derstanding (R OGERS and K INCAID 1981). A fundamental assumption of decentralized diffusion systems is that members of the user system have the ability to make sound decisions about how the diffusion process is managed. This capacity of the users to run their own diffusion system makes the most sense (1) when the users are highl y educated and technically competent practitio- ners (for example, cardiovascular surgeons), so that all the users are experts, or (2) when the in- novations being diffused are not at a high level of technology (for example, home energy conser- vation or organic gardening versus building a nucle ar power plant), so that intelligent laymen have sufficient technical expertise.

The fact that relatively decentralized diffusion syst ems exist in a wide variety of fields and loca- tions suggests that in the past we may have se verely underestimated the degree to which the user system was capable of managing its own knowledge transfer process. Our understanding of de- centralized diffusion systems is still limited, owing to the general lack of investigations of such user-dominated diffusion. However, it seems appare nt that certain elements of decentralized dif- fusion systems might be combined with certain asp ects of the centralized model to fit a particular situation uniquely. In other words, the classica l diffusion model is being questioned in certain very important ways.

29 Advantages and Disadvantages of Decentralized Diffusion Decentralized diffusion systems have both advantages and disadvantages. Compared to centra- lized systems, the innovations that decentralized systems diffuse are likely to fit with users' needs and problems more closely. Users feel a sense of control over a decentralized diffusion system, as they participate in making many of the key decisi ons, such as which of their perceived problems need most attention, which innovations best meet these needs, how to seek information about each innovation and from what source, and how much to modify an innovation as they adopt and implement it to their particular setting. The high degree of user control over these key decisions means that a decentralized diffusion system is gear ed closely to local needs. Problems of change agent/client heterophily are minimized. It is main ly user motivations to seek innovations that drive a decentralized diffusion pro cess, and this may be more cost-efficient than situations in which professional change agents manage the diffu sion process. User self-reliance is encouraged in a decentralized system, finally, decentralized diffusion is publicly popular: users generally like such systems. Several disadvantages, however, often characterize decentralized diffusion sys- tems:

1. Technical expertise is sometimes difficult to bring to bear on decisions about which innova- tions to diffuse and to adopt, and it is possible for "bad innovations" to diffuse through a decen- tralized system because of this lack of "quality control." So wh en a diffusion system is dissemi- nating innovations that involve a high level of technical expertis e, a decentralized diffusion sys- tem may be less appropriate than a mo re centralized diffusion system. 2. Furthermore, extremely decentralized diffusion systems lack a coordinating role (that is, the "big picture" of the system, where problems exis t and which innovations might be used to solve them). For example, a local user may not know wh ich other users he or she could visit to learn about an innovation. Thus, completely decentralized diffusion systems suffer from the fact that local users, who control the system, lack certain aspects of the big picture about users' problems and about available innovations to meet these problems.

3. A highly decentralized system will not be appr opriate for innovation for which potential users do not feel a need. An example is family pl anning in developing nations, which a government may regard as a high priority but which people may not want. There are very few decentralized diffusion systems for contraception in Latin America, Africa, and Asia. Thus, our present discussion suggests that: 1. Decentralized diffusion systems are most appropria te for certain conditions, such as for diffus- ing innovations that do not involve a high level of technical expertise, among a set of users with relatively heterogeneous conditions. When these conditions are homogeneous, a relatively more centralized diffusion system may be most appropriate.

2. Certain elements of centralized and decentralized diffusion systems can be combined to form a diffusion system that uniquely fits a particular situation. For example, a diffusion system may combine a central-type coordina ting role, with decentralized decisions being made about which innovations should be diffused and which users others should site-visit. Technical evaluations of promising innovations can be made in an otherwise decentralized diffusion system.

Biases in knowledge transfer The constructive criticisms that have been made of knowledge transfer models in very recent years help us identify several biases in such work, and they also suggest ways of overcoming such biases.

30 The pro-innovation bias The pro-innovation bias is the implication that an innovatio n should be diffused and adopted by all members of a social system, that it should be diffused more rapidly, and that the innovation should be neither re-invented nor rejected. Seld om is the pro-innovation bias straightforwardly stated in scholarly publications. Rather, the bias is assumed and implied. This lack of recognition of the pro-innovation bias makes it especially troublesome and potentially dangerous in an intel- lectual sense. The bias leads researchers to ig nore the study of ignorance about innovations, to underemphasize the rejection or discontinuance of innovations, to overlook re-invention, and to fail to study anti-diffusion programs designed to prevent the diffusion of "bad" innovations (like marijuana or drugs or cigarettes, for example). Th e net result of the pro-innovation bias is that we have failed to learn about certain very important aspects of the diffusion of innovations. What we do know about diffusion (and other aspects of knowledge transfer) is unnecessarily rather limited.

But it need not be so.

How did the pro-innovation bias originally occur? Part of the r eason is historical. Undoubtedly, hybrid corn was profitable for each of the Iowa farmers in the early R YAN and G ROSS (1943) dif- fusion study, but most other innovati ons that have been investigated do not have this extremely high degree of relative advantage. Many individuals, for their own good, should not adopt them.

Perhaps if the field of diffusion research had not begun with highly profitable agricultural innova- tions in the 1940s and the 1950s, th e pro-innovation bias would have been avoided, or at least recognized and dealt with properly.

During the 1970s, several critics of diffusion res earch recognized the pro-innovation bias. For ex- ample, D OWNS and M OHR (1976,700) stated; " The act of innovating is still heavily laden with positive value. Innovativeness, like efficiency, is a characteristic we want social organisms to possess. Unlike the ideas of progress and growth, which have long since been casualties of a new consciousness, innovation, especially when seen as more than purely technological change, is still associated with improvement . " What causes the pro-innovation bias in diffusion research?

1. Much diffusion research is funded by change agencies; they have a pro-innovation bias (un- derstandably so, since th ey are in the business of promoting innovations), and this viewpoint has often been accepted by many of the diffusion researchers whose work they sponsor, whom they call upon for consultation about their diffusion pr oblems, and whose students they may hire.

2. "Successful" diffusions leave a rate of adoption that can be retrospectively investigated by dif- fusion researchers, while an unsuccessful diffusion does not leave visible traces that can be very easily studied. For instance, a reje cted and/or a discontinued innovati on is not so easily identified and investigated by a researcher by interroga ting the rejectors and/or discontinuers.

As a general result of the pro- innovation bias, we know much more (1) about the diffusion of ra- pidly diffusing innovations, (2) a bout adoption than about rejection, and (3) about continued use than about discontinuance. The pro-innovation bias in diffusion research is understandable from the viewpoint of financial, logistical, methodol ogical, and practical policy considerations. The problem is that the pro-innovation bias is limiti ng in an intellectual sense: we know too much about innovation successes and not enough about i nnovation failures. While we have largely dis- cussed the pro-innovation bias here in terms of the diffusion of innovations, it also permeates all other aspects of the knowle dge transfer process. How might the pro-innova tion bias be overcome? 31 1. Alternative research approaches to post hoc data-gathe ring about how an innovation has dif- fused should be explored in knowledge-transfer re search. Diffusion research does not necessarily have to be conducted after an innovation has di ffused completely to the members of a system.

Such a rearward orientation to most diffusion st udies helps lead them to a concentration on suc- cessful innovations. It is also possible to investigate the diffusion of an innovation while the dif- fusion process is still underway, or , in fact, before it even begins.

2. Researchers should become much more questioni ng of, and careful about, how they select their innovations of study. Even if a successful innovation is selected for investigation, a scholar might also investigate an unsuccessful innovation that failed to diffuse widely among members of the same system. Such a comparative analysis would help illuminate the seriousness of the pro- innovation bias. In general, a mu ch wider range of innovations should be studied in knowledge- transfer research.

3. Researchers should investigate the broader cont ext in which an innovation diffuses, such as how the initial decision is made that the innova tion should be diffused to members of a system, how public policies affect the rate of diffusion, how the innovatio n of study is related to other in- novations and to the existing practice (s) that it replaces, and how it was decided to conduct the R & D that led to the innovation in the first place. This wider scope to research studies would help illuminate the broader system in whic h the knowledge-transfer process occurs.

4. We should increase our understanding of the motivations for adopting an innovation. Strange- ly, such "why" questions about adopting an innovation have only seldom been probed by diffu- sion researchers; undoubtedly, motivations for adoption are a difficult issue to investigate. Some adopters may not be able to tell a researcher why they decided to use a new idea. O\ ther adopters may be unwilling to do so. Seldom are simple, direct questions in a survey interview adequate to uncover an adopter's reasons for using an innovati on. But we should not give up on trying to find out the "why" of adoption just because valuable data about adoption motivations are difficult to obtain by the usual methods of di ffusion research data-gathering.

It is often assumed that an economic motivation is the main thrust for adopting an innovation, es- pecially if the new id ea is expensive. Economic factors are undoubtedly very important for cer- tain types of innovations and thei r adopters, such as the use of agricultural innovations by U. S.

farmers, but the prestige secured from adopting an innovation before one's peers may also be an important factor. Certainly the first and most impo rtant step in shedding a pro-innovation bias in knowledge-transfer research is to recognize that it may exist.

The Individual-Blame Bias in Knowledge Transfer In addition to a pro-innovation bias in much past diffusion research, there has also been a source- bias, a tendency for diffusion resear ch to side with the change agencies th at promote innovations rather than with the audience of potential adopters. This source-bias is perhaps even suggested by the words that we employ to describe this field of research: "Diffusion" research might have been called something like "problem-solving," "innov ation-seeking," or the "evaluation of innova- tions" had the audience originally had a stronger influence on this research. One cannot help but wonder bow the diffusion research approach might have been different if the R YAN and G ROSS (1943) hybrid corn study had been sponsored by the Iowa Farm Bureau Federation (a farmer's or- ganization) rather than by an agricultural resear ch center like the Iowa Agricultural Experiment Station. And what if the Columbia University drug study (C OLEMAN , KATZ , and M ENZEL 1966) had been sponsored by the American Medical Asso ciation, rather than by the Pfizer Drug Com- pany? The source-sponsorship of early diffusion studies may have given these investigations not only a pro-innovation bias but may have also struct ured the nature of diffusion research toward individual-blame.

32 Individual-blame is the tendency to hold an individual responsible for his or her problems, ra- ther than the system of which the individual is a part (C APLAN and N ELSON 1973). In other words, an individual-blame orientation implies th at "if the shoe doesn't fit, there's something wrong with your foot." An opposite point of view would blame the system, not the individual: it might imply that the shoe manufacturer or the mark eting system could be at fault for a shoe that does not fit.

Of course it is likely that some of the factors underlying a particular social problem may indeed be individual in nature, and that any effective solution to the problem may have to deal with changing these individual factors. However in ma ny cases the causes of the social problem lie in the system of which the individual is a part. Ame liorative social policies that are limited to indi- vidual interventions will not be very effective in solving system-level problems. How a social problem is defined is an important determinant of how we go about solving it, and therefore of the effectiveness of the attempted solution. A frequent error in defining a social problem is to overstress individual-blame and to underestimate system-blame.

System-blame may be defined as the tendency to hold a system responsible for the problems of individual members of the system. How else can the person-blame bias be overcome?

1. Researchers must attempt to keep an open mind about the causes of a social problem, at least until exploratory data are gathered, and guard against accepting others' definitions of knowledge- transfer problems, which often tend to be in terms of individual-blame. 2. All the participants should be involved, including potential adopters, in the definition of a re- search problem, rather than just those indivi duals who are seeking amelioration of a problem.

3. Social and communication structural variables, as well as intra-individual variables, should be considered in knowledge-transfer research. Past diffusion studies largely consisted of audience research, while seriously neglecting source rese arch. The broader issues of who owns and con- trols (1) the R & D system that produces innovations and (2) the communication systems that dif- fuses them, and to whose benefit, also need atte ntion in future knowledge-transfer investigations.

As in the case of the pro-innovation bias in diffusion research, perhaps one of the first and most important ways to guard against the individua l-blame bias is to be aware that it exists. To what extent does knowledge-transfer re search have an individual-blame bias? It is difficult to as- sess the degree of individual-blame in past rese arches accurately, but, on careful reading, there seems to be a certain flavor of individual-blam e in many of the resulting publications. An indi- vidual-blame orientation is not, in and of itself, always inappropriate. Perhaps individual-level variables are the most appropriate to investigate in a particular study. By no means do we advo- cate the complete discarding of all individua l-level, psychological variables in knowledge- transfer research, but in almost all cases, such a psychological approach centering on individual- level variables is not a complete explanat ion of the behavior being investigated. The generation of innovations Knowledge transfer consists of much more than just diffusion. Past investigations have over- looked the fact that a great deal of relevant activities and d ecisions usually occurred long before the diffusion process began: A perceived proble m, funding decisions about R & D activities that led to research work, inventi on of the innovation and then its development and commercializa- tion, a decision that it should be diffused, transfer of the innovation to a diffusion agency, and its communication to an audien ce of potential adopters. Then the first adoption occurs.

This entire pre-diffusion series of activities and d ecisions is certainly an important part of the in- novation-development process, of which the diffusion phase is but one component. The impor- 33 tance of what happens prior to the beginning of an innovation's diffusion (especially those events that affect the nature of diffusion later on) has been almost entirely ignored in past research.

The innovation-development process consists of all of the deci sions, activities, and their im- pacts that occur from recognition of a need or problem, through research, development, and commercialization of an innovation, through diffusi on and adoption of the innovation by users, to its consequences. Here we take up each of the main steps in the innovation-development process, which corresponds roughly to the process of knowledge transfer.

1. Recognizing a Problem or Need. One of the ways in which the innovation-development process begins is by recognition of a problem or need, which stimulates research and develop- ment activities designed to create an innovation to solve the problem/need. In certain cases, a scientist may perceive a forthcoming problem and launch research to find a solution. An example is the agricultural scientist at the University of California at Davis who foresaw a severe labor shortage for California tomato farmers when the bracero program ended and initiated an R & D program to breed hard tomato variet ies that could be machine-picked.

In other cases, a problem/ need may rise to high priority on a system's agenda of social problems through a political process. Research and development to develop safer cars and highways had been conducted and accumulated for several years, but the results were not put into practice until the mid-1960s when a series of highly public ized legislative hearings and Ralph N ADER 's (1965) book, Unsafe at Any Speed , called national attention to the high rate of traffic fatalities. The so- cial problem of auto safety rose to a high national priority owing to higher fatality rates in the ear- ly 1960s, when the annual death rate reached 50,000. But the interpretation of this dangerous trend was in large part a political activity.

2. Basic and Applied Research . Most innovations that have b een investigated in diffusion re- searches have been technological innovations. Most such innovations are created by scientific re- search activities, alt hough they often result from an interp lay of scientific method and practical operations. The knowledge base for a technology usually derives from basic research, defined as original investigations for the advancement of sc ientific knowledge that do not have the specific objective of applying this knowledge to practical problems. In contrast, applied research con- sists of scientific investigati ons that are intended to solve practical problems. Scientific know- ledge is put into practice in order to design an innovation that will solve a perceived need or prob- lem. Applied researchers are the main users of basic research. Thus, an invention may result from a sequence of (1) basic research followed by (2) applied research leading to (3) development.

3. Development . The abbreviation R & D corresponds closel y to the concept that it represents:

"R" always appears together with "D" and, more over, always precedes "D"; development is al- ways based on research. In fact, it is usually difficult or impossible to separate research and de- velopment, which is why the term "R & D" is so often used.

Development of an innovation is the proc ess of putting a new idea in a form that is expected to meet the needs of an audience of potential adopters. This phase nor mally occurs after research but prior to the innovation that stems from research.

4. Diffusion and Adoption . Perhaps the most crucial decision in the entire innovation-develop- ment process is the decision to begin diffusion of an innovation to potential adopters. On the one hand, there is usually pressure to approve an i nnovation for diffusion as soon as possible, as the social problem/need that it seeks to solve may ha ve been given a high priority. Public funds may have been used to sponsor the research, and such financial support is an unrealized public in- vestment until the innovation is adopted by users. On the other hand, the change agency's reputa- tion and credibility in the eyes of its clients rests on only recommendi ng innovations that will 34 have beneficial consequences for their adopters. Scientists tend to be cautious when it comes time to translate their scientific findings into practice.

A novel approach to gatekeeping medical innovati ons is followed by the National Institutes of Health through the cond uct of "consensus development conferences." Consensus development is a process that brings together biomedical res earch scientists, practicing physicians, consumers, and others in an effort to reach general agr eement on whether a given medical technology is safe and effective (L OWE 1980). The technology may be a device, a drug, or a medical or surgical procedure. A consensus conference differs from th e usual state-of-the-art scientific meeting in that a broadly based panel is cons tituted to address a set of predetermined questions regarding the particular medical innovation unde r review. A three-day consensus conference typically begins with a series of research synthesis papers that are discussed by the expert investigators, users of the technology, and their consumers, A consensus statement is prepared by the panel and read on the final day of the conference to the audience, who then react to it. The final consensus state- ment is then published by the U.S. Government Printing Office and widely disseminated to phy- sicians, the mass media, medi cal journals, and the public.

Consensus conferences were begun in 1978 in recogniti on of the fact that the medical field lacked a formal process to assure that medical research discoveries were identified and scientifically evaluated to determine if they were ready to be used by doctors and other health-care workers. It was feared that some new technologies might have been disseminated without an adequate scien- tific test, while other well-validated medical t echnologies might be diffusing too slowly. The con- sensus panels have, in fact, occasionally reco mmended against using a given medical or surgical procedure, device, or drug under certain conditions. So, they serve an important function in gate- keeping the flow of medical innovati ons from research into practice.

Some other fields also utilize a formal procedure for deciding when an innovation should be dif- fused. Most knowledge transfer systems however, do not evaluate innovations for diffusion in such a rigorous way. Here, perhaps, we see an example of how one knowledge-transfer system can learn and adapt useful lessons from another such system. Such transfer of knowledge-transfer methodologies can be greatly facilitated by th e world of scholars of the knowledge-transfer process, as they engage in comparative analyses and evaluations of knowledge transfer systems. The entrepreneurial transfer of knowledge Previously in this chapter we argued that past research on the knowledge transfer process has been unduly limited in scope. There are many types of knowledge transfer that have been ignored by scholars. One of these is techno logy transfer that occurs between private firms and that is dri- ven by market forces, rather than by public pol icies enacted through activities of a government agency. We should not forget that most of wh at we now understand about the nature of know- ledge transfer is based, rath er narrowly, upon the transfer of innovations from a national govern- ment agency to individuals; here again we see the considerable influence of the agricultural ex- tension model upon our thinking about knowledge tr ansfer. Yet a great deal of knowledge trans- fer obviously must take place in th e context of for-profit firms that are competitively seeking to market innovative products to consumers. One sp ectacular illustration is provided by Silicon Val- ley, the high-technology complex in Northern California that is the world center of the microelec- tronics industry. Silicon Valley produces the se miconductor chips, microcomputers, video games, and lasers that are transforming industriali zed nations into information societies.

At the heart of Silicon Valley is severe compe tition in continuous technological innovation; each company tries to gain an advantage over its competitors by being the first to market (perhaps by only a month or two) with a new product. In this setting, gaining technical information about a new idea translates directly into profits. In the early days of Silicon Valley (in the 1960s), much 35 14 of the information came from research laboratories at Stanford University. Today much of the new information comes from R & D that is co nducted by the firms themselves. One reason for the high rate of job mobility (estimated to be a bout 30 percent per year) in Silicon Valley is that each firm tries to hire key employees away from its competitors in order to learn company se- crets. Further, one or several ke y employees frequently spin-off of an established firm in order to start their own company, which usually is in dire ct competition with its parent. The new firm is usually organized around a new product, which, if successful, can make the founders of the start- up millionaires within a few years.

Here we see an entirely different type of knowle dge transfer from that represented by the agricul- tural extension model. The U.S. government has not played a direct role in the rise of Silicon Val- ley, nor in the dozen or so other "Silicon Vall eys" that are springing up around the United States.

Significantly, each of these new high-technology co mplexes has a research university at its center (an example is HIT in the Route 128 complex around Boston, the University of North Carolina in the Research Triangle complex, etc.).

So each of these high-technology ce nters represents an example of a special kind of technology transfer. Perhaps the research university repres ents today the central institution in the emerging information society (R OGERS and L ARSEN 1983).

Much more needs to be learned about this entrepreneurial model of know ledge transfer. Perhaps it illustrates an important lesson for scholars and practitioners of knowledge transfer: that they need to broaden their definitions and con ceptions of the knowledge-transfer process. References CAPLAN , N., and S. N ELSON . 1973. "On Being Useful. The Natu re and Consequences of Psychological Research on Social Problems." In: American Psychologist 28, no. 3: 199-211.

C OLEMAN , J. S., E. KATZ , and H. M ENZEL . 1966. Medical Innovation; A Diffusion Study. New York: Free Press. D OWNS , G. W., and L. B. M OHR . 1976. "Conceptual lssues in t he Study of Innovations." In: Adminis- trative Science Quarterly 21, no. 4; 700-714.

H IGHTOWER , J. 1972. Hard Tomatoes, Hard Times; T he Failure of America's Land Grant College Complex. Washington, D.C.: Ag ribusiness Accountability Project; and Cambridge: Schenkman.

K APLAN , A. 1964. The Conduct of Inquiry. San Francisco; Chandler.

L OWE , C. U. 1980. "The Consensus Development Programme: Technology Assessment at the Na- tional Institute of Health." In: British Medical Journal 280, no. 6231: 1583-1584.

R OGERS , E. M. 1983. Diffusion of Innovations. New York: Free Press.

R OGERS , E. M., and D. L. K INCAID . 1981. Communication Networks; A New Paradigm for Research.

New York; Free Press.

R OGERS , E. M., and J. K. L ARSEN . 1983. Silicon Valley Fever; The Rise of High-Technology Culture.

New York: Basic Books.

R OGERS , E. M., J. D. EVELAND , and A. S. BEAN . 1984. Extending the Agricultural Extension Model. Washington, D.C.; University Press of America.

R YAN , B., and C. G ROSS . 1943. "The Diffusion of Hybrid Seed Corn in Two Iowa Communities." In:

Rural Sociology 8, no. 1; 15-24.

S CHON , D. 1971. Beyond the Steady State. London: Temple Smith. 36 Universität Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-05 January 2007Hoffmann Diffusion of Innovations 1 1 Everett M. R OGERS Summary Chapter 1 – Elements of diffusion Diffusion is the process by which an innovation is communi- cated through certain channels over time among the members of a social system. Diffusion is a special type of communica- tion concerned with the spread of messages that are perceived as new ideal. Communication is a process in which partici- pants create and share information with one another in order to reach a mutual understanding. Diffusion has a special cha- racter because of the newness of the idea in the message con- tent. Thus some degree of uncer tainty and perceived risk is involved in the diffusion process. An individual can reduce this degree of uncertainty by obtaining information. Informa- tion is a difference in matter ener gy that affects uncertainty in a situation where a choice exists among a set of alternatives.

The main elements in the diffu sion of new ideas are: (1) an innovation (2) that is communicated through certain channels (3) over time (4) among the members of a social system. 1 Elements of diffusion 2 A history of diffusion re- search 3 Contributions and criticisms of diffusion research 4 The generation of innovation 5 The innovation-decision process 6 Attributes of innovations and their rate of adoption 7 Innovativeness and adopter categories 8 Diffusion networks 9 The change agent 10 Innovation in organizations 11 Consequences of innova- tions 1.1 Innovation An innovation is an idea, practice, or object perceived as new by an individual or other unit of adoption. Most of the new ideas discussed in this book are technological innovations. A tech- nology is a design for instrumental action that reduces the uncertainty in the cause-effect rela- tionships involved in achieving a desired outcom e. Most technologies have two components:

(1) hardware , consisting of the tool that embodies the technology as a material or physical object, and (2) software, consisting of the knowledge base for the tool.

The characteristics of an innovation, as percei ved by the members of a social system, deter- mine its rate of adoption. Five attributes of innovations are: (1) relative advantage, (2) compa- tibility, (3) complexity, (4) triala bility, and (5) observability.

Re-Invention is the degree to which an innovation is changed or modified by a user in the process of its adopti on and implementation.

1.2. Communication Channels A communication channel is the means by which messages get from one individual to another.

Mass media channels are more effective in creating knowledge of innovations, whereas inter- personal channels are more effective in forming and changing attitudes toward a new idea, and thus in influencing the decision to adopt or reject a new idea. Most individuals 1 Source: Everett R OGERS , 2003: The Diffusion of Innovations. Fifth Edition. The Free Press, New York.

37 evaluate an innovation not on the basis of scien tific research by experts but through the subjec- tive evaluations of near peers who have adopted the innovati on. These near peers thus serve as role model, whose innovation behavior tends to be imitated by others in their system.

A distinctive aspect of diffusion is that at least some degree of heterophily is usually present in communication about innovations. Heterophily is the degree to which two or more individuals who interact are different in certain attributes, such as beliefs, education, social status, and the like. The opposite of heterophily is homophily, the degree to which two or more individuals who interact are similar in certain attributes . Most human communication takes place between individuals who are homophilous, a situation that leads to more effective communication.

Therefore, the heterophily that is often present in the diffusion of innovations leads to special problems in achieving effective communication. 1.3. Time Time is involved in diffusion in (1) the i nnovation-diffusion process, (2) innovativeness, and (3) an innovation's ra te of adoption. The innovation decision process is the process through which an individual (or other decision-making unit) passes fr om first knowledge of an innova- tion to forming an attitude towa rd the innovation, to a decision to adopt or reject, to implemen- tation of the new idea, and to confirmation of this decision. We conceptualize five steps in this process: (1) knowledge, (2) pers uasion, (3) decision, (4) implemen tation, and (5) confirmation.

An individual seeks information at various stages in the innovation-decision process in order to decrease uncertainty about an innovation's ex pected consequences. The decision stage leads (1) to adoption, a decision to make full use of an innova tion as the best course of action avail- able, or (2) to rejection, a decision not to adopt an innovation. 2 38 3 1.4. Social System A social System is a set of interrelated units that are engaged in joint problem solving to ac- complish a common goal. A system has structure, defined as the patterned arrangements of the units in a system, which gives stability and regu larity to individual behavior in a system. The social and communication structur e of a system facilitates or impedes the diffusion of innova- tions in the system. One aspect of social structure is norms, the established behavior patterns for the members of a social system.

Opinion leadership is the degree to which an in dividual is able to influence informally other individuals' attitudes or overt behavior in a desired way with relative frequency. A change agent is an individual who attempts to influence clients innovation-decisions in a direction that is deemed desirable by a change agency. An aide is a less than fully prof essional change agent who intensively contacts clients to influence their innovation-decisions.

We distinguish among three main types of innovation-decisions: (1) optional innovation- decisions, choices to adopt or reject an innovation th at are made by an individual independent of the decisions of other members of the system, (2) collective innovation-decisions, choices to adopt or reject an innovation that are made by consensus among the members of a system, and (3) authority innovation-decisions, choices to adopt or reject an innovation that are made by relatively few individuals in a system who possess power, status, or technical expertise. A fourth category consists of a sequential combination of two or more of these three types of innovation decisions: Contingent innovation-decisions are choices to adopt or reject that are made only after a prio r innovation-decision.

A final way in which a social syst em influences diffusion concerns consequences, the changes that occur to an individual or a social system as a result of the adoption or rejection of an in- novation.

Summary Chapter 2 – A history of diffusion research This chapter showed that although diffusion research began as a series of scientific enclaves, it has emerged as a single, integrated body of concepts and generalizations, given though the investigations are conducte d by researchers in different scientific disciplines. A research tra- dition is a series of investigations on a similar topic in which successive studies are influenced by preceding inquiries. The major diffusion trad itions described are anthropology, early soci- ology, rural sociology, education, public health/medical sociology, communication, marketing, geography, and general sociology.

Eight main types of diffusion research were identified:

1. Earliness of knowing about innovations. 2. Rate of adoption of different innovations in a social system.

3. Innovativeness. 4. Opinion leadership. 5. Diffusion networks. 6. Rate of adoption in diffe rent social systems.

7. Communication channel usage.

8. Consequences of innovation.

When scholars follow an intellectual paradigm in a research field, it enables them to pursue a coherent set of research directions. The paradigm also imposes and standardizes a set of as- sumptions and conceptual biases that, once begun, are difficult to recognize and overcome.

That is the challenge for the next generation of diffusion scholars. In my first book on diffu- 39 4 sion (R OGERS 1962,x), I stated, "This book suggests that students of diffusion have been work- ing where the ground was soft . . . The challenge for future research is to expand the area of digging and to search for different objectives than those of the past. Perhaps there is a need to dig deeper, in directions that theory suggests." Summary Chapter 3 – Contributions and criticisms of diffusion research We reviewed four major shortcomings of diffusi on research in this chapter. We conclude that the beginnings of diffusion research left an indelible stamp on the approaches, concepts, me- thods, and assumptions of the field. The biases that we inherite d from our research ancestors have been inappropriate for certain important diff usion research tasks of today. It is ironic that the study of innovation has itsel f been so traditional.

The four major criticisms of diffusion re search, discussed in this chapter are:

1. The pro-innovation bias, the implication of most diffusi on research that an innovation should be diffused to and adopted by all members of a social system, that it should be diffused rapidly, and that the inno vation should be neither re-invented nor rejected.

z. The individual-blame bias, the tendency to hold an indivi dual responsible for his or her problems, rather than the system of which the individual is a part.

3. The recall problem in diffusion research, which may lead to inaccuracies when respondents are asked to remember the time at which they adopted a new idea.

4. The issue of equality in the diffusion of innovations, as socioeconomic gaps among the members of a social system are often widene d as a result of the spread of new ideas.

Alternatives to the usual diffusion research approaches were proposed for overcoming each of these four criticisms of diffusion research.

Summary Chapter 4 – The generation of innovations Past diffusion researches usually began with the first adopter of an innovation, that is, with the left-hand tail of the S-shaped diffusion curve. Events and decisions occurring previous to this point have a considerable infl uence upon the diffusion process. Th e scope of future diffusion research should be broadened to include study of the entire process through which an innova- tion is generated.

The innovation-development process consists of all the decisions, activities, and their impacts that occur from recognition of a need or pr oblem, through research, development, and com- mercialization of an innovation, through diffusion and adoption of the innovation by users, to its consequences. Recognition of a Problem or need may occur when a social problem rises to a high priority an the agenda of topics which deserve research.

Many, but not all, technological in novations come out of research. Basic research is defined as original investigations for the advancement of scientific knowledge and that do not have the specific objective of applying th is knowledge to practical problem s. The results of basic re- search may be used in applied research, which consists of scientific investigations that are intended to solve practical problems. Lead users develop innovations and then convince a manufacturing company to pr oduce and sell the innovation, ofte n after the lead user has created a prototype of the innovation. The usua l next stage in the innovation development process is development, defined as the process of putting a new idea into a form that is ex- pected to meet the needs of an audience of potential adopters. Technological determinism is the belief that technology causes changes in society. An opposite viewpoint is social construc- tionism, which states that social factors shape a technology. A next stage, commercialization, is defined as the production, manufacturing, packaging, marketing, and distribution of a prod- 40 uct that embodies an innovation. Commercialization is carried out mainly by private firms.

A particularly crucial point in the innovation-de velopment process is the decision to begin diffusing an innovation to potential adopters. How are innovations evaluated for their efficacy, safety, and other factors? Finally, an innovation may diffuse, be adopted, and, eventually, cause consequences, the final stage in the innovation-development process. The six stages de- scribed here may not always o ccur in a linear sequence, the tim e order of the stages may be different, and certain stages may not occur at all.

Summary Chapter 5 – The i nnovation-decision process The innovation-decision process is the process through which an individual (or other deci- sion-making unit) passes from first knowledge of an innovation, to forming an attitude toward the innovation, to a decisi on to adopt or reject, to implemen tation of the new idea, and to con- firmation of this decision. This process consists of five stages: (1) knowledge, when the indi- vidual is exposed to th e innovation's existence and gains an understanding of how it functions; (2) persuasion, when the individual forms a favorable or unfavorable attitude toward the inno- vation; (3) decision, when the individual engages in activities that lead to a choice to adopt or reject the innovation; (4) implementation, when the individual puts an innovation into use; and (5) confirmation, when the individual seeks reinforcemen t for an innovation-decision already made but may reverse the decision if e xposed to conflicting messages about it. Earlier knowers of an innovation, when compared to later knowers, are characterized by more formal education, higher social status, greate r exposure to mass media channels of communi- cation, greater exposure to inte rpersonal channels of communi cation, greater change agent contact, greater social participation, and greater cosmopoliteness. 5 Re-invention is the degree to which an innovation is ch anged or modified by a user in the process of its adoption and implementation. Re-i nvention occurs at the implementation stage for many innovations and for many adopters . A higher degree of re-invention leads to (1) a faster rate of adopti on of an innovation and (2) a greater degree of sustainability of an innova- tion . Sustainability is the degree to which an innovation is continued over time after a diffu- sion program ends.

41 6 Discontinuance is a decision to reject an innovation afte r having previously adopted it. Dis- continuance can be of two types: (1) replacement discontinuance, in which an idea is rejected in order to adopt a better idea which superseded it, and (2) disenchantment discontinuance, in which an idea is rejected as a result of dissatis faction with its performance. Later adopters are more likely to discontinue innovati ons than are earlier adapters.

We conclude that stages exist in the innovation-decision process, although further study of this issue is needed.

A communication channel is the means by which a message gets from a source to a receiver.

We categorize communication channels (1) as eith er interpersonal or mass media in nature and (2) as originating from either localite or cosmopolite sources. Mass media channels are means of transmitting messages that involve a mass medi um such as radio, television, newspapers, and so on, that enable a source of one or a few individuals to reach an audience of many. In- terpersonal channels involve a face-to-face exchange be tween two or more individuals.

Mass media channels are relatively more important at the knowledge stage, and interpersonal channels are relatively more important at the persuasion stage in the innovation-decision process. Cosmopolite channels are relatively mo re important at the knowledge stage, and loca- lite channels are relatively more important at the persuasion stage in the innovation-decision process. Mass media channels are relatively mo re important than interpersonal channels for earlier adopters than for later adopters. Cosmopolite channels are relatively more important than localite channels for earlier adopters than for later adopters.

The innovation-decision period is the length of time required fo r an individual or organization to pass through the innovation-d ecision process. The rate of awareness-knowledge for an in- novation is more rapid than it s rate of adoption. Earlier adopters have a shorter innovation- decision period than do later adopters.

Summary Chapter 6 – Attributes of i nnovations and their rate of adoption This chapter suggested five attributes of innovations by which an innovation can be described.

Individuals' perceptions of thes e attributes predict an innovation's rate of adoption. We rec- ommend that measures of the five perceived attr ibutes should be developed in each diffusion study, rather than utilizing existing scales borrowed from previous investigations.

Rate of adoption is the relative speed with which an innovation is adopted by members of a social system. In addition to the perceived attr ibutes of an innovation, such other variables affect its rate of adoption as (1) the type of innovation-decision, (2) the nature of communica- tion channels diffusing the innovatio n at various stages in the innovation-decision process, (3) the nature of the social system ; and (4) the extent of change agents' efforts in diffusing the innovation. Most past research, ho wever, concentrated on predicting the rate of adoption by the five perceived attributes of innovations.

Relative advantage is the degree to which an innovation is perceived as better than the idea it supersedes. The relative advantage of an innova tion, as perceived by members of a social sys- tem, is positively related to its rate of adoption. Overadoption is the adoption of an innovation when experts feel that it should be rejected. Preventive innovations, defined as new ideas that an individual adopts now in order to lower the probability of some unwanted future event, diffuse more slowly than incr emental (nonpreventive) innovations.

Compatibility is the degree to which an innovation is pe rceived as consistent with the existing values, past experiences, and n eeds of potential adopters. The compatibility of an innovation, as perceived by members of a social system, is positively related to its rate of adoption. Nam- ing an innovation and positioning it relative to pr evious ideas are important means of making 42 an innovation more compatible. Change agen ts often ignore indigenous knowledge systems, which provide one means by which individua ls give meaning to an innovation.

Complexity is the degree to which an innovation is perceived as relatively difficult to under- stand and to use. The complexity of an innovation, as perceived by members of a social sys- tem, is negatively related to its rate of adoption.

Trailability is the degree to which an innovation may be experimented with on a limited basis.

The trialability of an innovation, as perceive d by members of a social system, is positively related to its ra te of adoption.

Observability is the degree to which the results of an innovation are visible to others. The ob- servability of an innovation, as perceived by memb ers of a social system, is positively related to its rate of adoption.

A basic theme of this chapter is that change agents and diffusion scholars must understand how potential adopters perceive new ideas. Such perceptions count in determining the nature of the diffusion process. 7 43 Summary Chapter 7 – Innovativeness and adopter categories Adopter categories are the classifications of the members of a social system on the basis of innovativeness, the degree to which an individual or othe r unit of adoption is relatively earlier in adopting new ideas than other members of a system. A variety of categorization systems and titles for adopters have been used in past st udies. This chapter described the standard five adopter categories that are widely followed today in diffusion research, and their applications.

Adopter distributions tend to follow an S-shaped curve over time and to approach normality .

The continuum of innovativeness can be partitio ned into five adopter categories (innovators, early adopters, early majority, late majority, a nd laggards) on the basis of two characteristics of a normal distribution, the mean and the standa rd deviation. The dominant attributes of each category are: Innovators-venturesome; early adopters-respect; earl y majority-deliberate; late majority-skeptical; and laggards- traditional. The relatively earlier adopters in a social system are no different from later adopters in age, but they have more years of formal education, are more likely to be literate, and have higher social status, a greater degree of upward social mo- bility, and larger-sized units, such as farms, companies, schools, and so on. These characteris- tics of adopter categorie s indicate that earlier adopters have generally higher socioeconomic status than do later adopters.

Earlier adopters in a system also differ from la ter adopters in personality variables. Earlier adopters have greater empathy, less dogmatism, a greater ability to deal with abstractions, greater rationality, greater intell igence, a more favorable attitude toward change, a greater abil- ity to cope with uncertainty and risk, a more favorable attitude towa rd science, less fatalism and greater self-efficacy , and higher aspirations for formal education, higher-status occupa- tions, and so on.

Finally, the adopter categories ha ve different communication behavior. Earlier adopters have more social participation, are more highly interc onnected in the interpersonal networks of their system, are more cosmopolite, have more cont act with change agents, greater exposure to mass media channels, and greater exposure to interpersonal communication channels, engage in more active information seeking, a nd have greater knowledge of innovations, and a higher degree of opinion leadership. 8 44 9 Past research thus shows many important differe nces between earlier and later adopters of innovations in (1) socioeconomic status, (2) pe rsonality variables, and (3) communication be- havior. The distinctive characteri stics of the five adopter categories mean that these adopter categories can be used for audience segmentation , a strategy in which different communi- cation channels and/or messages are used to reach each sub audience.

Summary Chapter 8 – Diffusion networks This chapter dealt with opinion leadership, co mmunication networks, and the critical mass.

Opinion leadership is the degree to which an individual is able to influence informally other individuals' attitudes or overt behavior in a desired way with relative frequency. Opinion lead- ers play an important role in diffusion networks, and are often id entified and utilized in diffu- sion programs.

Homophily is the degree to which individuals who communicate are similar. Heterophily is the degree to which individuals who in teract are different in certain attributes. Interpersonal diffu- sion networks are mostly homophilous. Homophily can act as an invisible barrier to the rapid flow of innovations within a social system, as similar people interact in socially horizontal patterns, thus preventing a new idea from tric kling down from those of higher socioeconomic status, more education, and gr eater technical expertise.

When interpersonal diffusion networks are heterophilous, followers generally seek opinion leaders of higher socioeconomic status, with mo re formal education, greater mass media expo- sure, more cosmopoliteness, greater contact with change agents, and more innovativeness.

Compared to followers, opini on leaders have greater mass media exposure, more cosmo- politeness, greater contact with change agents, greater social participation, higher social status, and more innovativeness. Opinion leaders confor m more closely to a system's norms than do their followers. When a social system's norms favor change, opinion leaders are especially innovative.

A communication network consists of interconnected individuals who are linked by patterned flows of information. An individual's network li nks are important determinants of his or her adoption of innovations. The network interconnectedne ss of an individual in a social system is positively related to the i ndividual's innovativeness. Interconnectedness is the degree to which the units in a social system ar e linked by interpersonal networks.

Networks provide a certain degree of structure and stability in th e predictability of human be- havior. Communication structure is the differentiated elements that can be recognized in the patterned communication flow s in a system. This structure cons ists of the cliques within a sys- tem and the network interconnections among them that are provided by bridges and liaisons.

Individuals are identif ied as belonging to cliques on the basis of communication proximity, the degree to which two linked individuals in a network have personal communication networks that overlap. A personal network consists of those interconnected individuals who are linked by patterned communication flows to a given individual. Personal networks that are radial (rather than interlocking) are more open to an individual's environment, and hence play a more important role in the diffusion of innovations. The information exchange potential of commu- nication network links is negatively related to their degree of (1) communication proximity and (2) homophily. This generalization expresses Mark G RANOVETTER 's theory of "the strength-of-weak-ties." Individuals tend to be linked to others who are close to them in physi- cal distance and who are relatively ho mophilous in social characteristics.

The critical mass occurs at the point at which enough individuals in a system have adopted an innovation so that the innovation's further rate of adoption becomes self-sustaining. The cri- tical mass is particularly important in the diffu sion of interactive innovations such as e-mail, where each additional adopter increases the utility of adopting the innovation for all adopters.

45 10 Interactivity is the degree to which participants in a communication process can exchange roles in, and have control over, their mutual di scourse. As more individuals in a system adopt 46 11 a noninteractive innovation, it is perc eived as increasingly beneficial to future adopters (this is a sequential interdependence effect on later adopters). However, in the case of an interactive innovation, the benefits from each additional adoption increase not only for all future adopters, but also for each previous adopter (this is reciprocal interdependence).

A threshold is the number of other individuals who must be engaged in an activity before a given individual will join that activity. An innovator has a low threshold of resistance to adopting a new idea, and so few (or no) interper sonal network influences are needed for adop- tion. In contrast, a late majority individual has a much higher th reshold that must be overcome by near-peer network influences in order to overcome resistance to the innovation. Thresholds act for individuals in a somewhat parallel way to the critical mass at the system level. An indi- vidual is more likely to adopt an innovation if mo re of the other individuals in his or her per- sonal network adopted previously. Summary Chapter 9 – The change agent Change agents operate interventions, defined as actions with a coherent objective to bring about behavior change in order to produce iden tifiable outcomes. For example, an HIV pre- vention program such as ‘stop AIDS’ in San Francisco was designed to slow the rate of HIV infection. Targeting (defined as the process of customiz ing the design and delivery of a com- munication program on the basis of the characteristics of an intended audience segment) is one means of segmenting a heterogeneous audience so that customized messages that fit each indi- vidual's situation are delivered. Currently, the Intern et is often utilized to deliver such targeted messages.

A change agent is an individual who influences clients' innovation decisions in a direction deemed desirable by a change agency. Change ag ents face two main problems: (1) their social marginality; due to their position midway between a change agency and their client system, and (2) Information overload, the state of an individual or a system in which excessive com- munication inputs cannot be processed and use d, leading to breakdown. Seven roles of the change agent are: (1) to develop a need for cha nge on the part of clients, (2) to establish an information exchange relationshi p, (3) to diagnose problems, (4) to create an intent to change in the client, (5) to translate intentions into action, (6) to st abilize adoption and prevent discon- tinuance, and (7) to achieve a term inal relationship, with clients.

A change agent's relative success in securing the adoption of innovations by clients is positive- ly, related to (1) the extent of the change agent' s effort in contacting clients, (2) a client orien- tation, rather than a change ag ency orientation, (3) the degree to which the diffusion program is compatible with clients' needs, (4) the change agent's empathy with clients, (5) his or her homophily with clients, (6) credib ility in the clients' eyes, (7) the extent to which he or she works through opinion leaders, and (8) increasin g clients' ability to evaluate innovations.

Further, we propose that contact by change agents is positively related to (1) higher socioeco- nomic status among clients, (2) greater social participation, (3 ) higher formal education, and (4) cosmopoliteness.

An aide is a less than fully professional change ag ent who intensively contacts clients in order to influence their innovation-decisions. Not only do aides provide lower-cost contacts with clients than is possible with professional change agents, but they are also able to bridge the heterophily gap between professionals and clients, especi ally lower socioeconomic status clients. Aides have less competence credibility, the degree to which a communication source or channel is perceived as knowledgeable and expert, but they have greater safety credibility, the degree to which a communication source or cha nnel is perceived as trustworthy. An aide's safety credibility is due to his or her homophily with the client system. Inauthentic professio- nalism is the process through which an aide takes on the dress, speech, or other identifying, 47 12 marks of a professional change agent. In recent decades diffusion scholars have become aware that an alternative to the classical diffusion mode l exists in the form of decentralized diffusion systems. These diffusion programs have outrun the classical model (a relatively centralized approach). In centralized diffusion systems, such as the agri cultural extension services in the United States, overall control of diffusion decisions (such as which innovations to diffuse, which diffusion channels to use, and to whom to diffuse innovations) is held by government officials and technical subject-matter experts. Diffusion in centralized systems flows from the top down, from experts to users. In contrast, decentralized diffusion systems are client- controlled with a wide sharing of power and control among the members of the diffusion sys- tem. Instead of coming out of R&D systems, innovations in decentralized systems bubble up from local experimentation by non expert users. Local units decide which innovations should diffuse through horizontal networks, allowing a high degree of re-invention. Decentralized diffusion systems are based upon convergence comm unication, in which participants create and share information with one another in or der to reach a mutual understanding. Decentra- lized diffusion systems are (1) most appropriate for certain conditions and (2) can be com- bined with elements of centralized systems to form a hybrid diffusion system.

Summary Chapter 10 – Innovation in organizations An organization is a stable system of indi viduals who work together to achieve common goals through a hierarchy of ranks and a division of labor. Individual behavior in an organization is relatively stable and predictable because organi zational structure is characterized by predeter- mined goals, prescribed roles, an authority structure, rules and regulations, and informal pat- terns. Although behavior in organizations is relatively stable, innovation is ongoing.

At first, innovation in organizations was main ly studied by correlating independent variables with organizational innovativeness in cross-sectional data analysis. A consistent finding in this organizational innovativeness resear ch was that larger organizations are more innovative. Ra- ther low correlations of characteristics variables with orga nizational innovativeness were found, perhaps because the organizational structure variables that were studied were related to innovation in one direction during the initiation subprocess of the innov ation process and in the opposite direction during the implementation subprocess. For instance, low centralization, high organizational complexity, and low formali zation facilitate innovation in the initiation subprocess, but impede implementation. Today, research an organizational innovativeness is much less likely to be conducted than is st udy of the innovation process in organizations.

The presence of an innovation champion contributes to the success of an innovation in an or- ganization. A champion is defined as a charismatic indivi dual who throws his or her support behind an innovation, thus overcom ing the indifference or resistance that the new idea may provoke. Research has shown that innovation cham pions may be powerful individuals in an organization, or they may be lower-level individuals who possess the ability to coordinate the actions of others. The degree to which champi ons are powerful seems to depend on the nature of the innovation and the organizati on in which it is gaining acceptance.

Studies of organizational innov ativeness tended to be replaced by research on the innovation process in organizations. We divide the i nnovation process into two subprocesses: (1) initia- tion, all of the information gather ing, conceptualizing, and planni ng for the adoption of an innovation, leading up to the decision to adopt and (2) implementation, all of the events, ac- tions, and decisions involved in putting an innovation into use. Th e two initiation stages are (1) agenda-setting and (2) matc hing. The three implementation st ages are (1) redefining/ re- structuring, (2) clarifyi ng, and (3) routinizing.

Agenda-setting occurs in the innovation pr ocess when a general organizational problem that 48 13 may create a perceived need fo r an innovation is defined. A performance gap, the discrepancy between an organization's expectations and its actual performance, can trigger the innovation process. Matching is the stage in the innovation process at which a problem from the organiza- tion's agenda is fit with an innovation, and this match is planned and designed.

Redefining/restructuring occurs when the innovation is re-invented so as to accommodate the organizations needs and structure more closely and when the organization's structure is mod- ified to fit with the innovation. Both the innovation and th e organization usually change during the innovation process.

Clarifying occurs as the innovation is put into more wi despread use in an organization, so that the meaning of the new idea gradually become s clearer to the organization's members. Routinization occurs when the innovation has become incorporated into th e regular activities of the organization and loses its separate identity. Sustainability, a c\ los ely related concept to routinization, is defined as the degree to whic h an innovation continues to be used after the initial effort to secure adoption is completed. Sust ainability is more likely if widespread partic- ipation has occurred in the innovation process, if re-invention took place, and if an innovation- champion was involved. This fifth stage, rou tinization, marks the end of the innovation process in an organization.

Summary Chapter 11 – Consequences of innovations Consequences are the changes that occur to an individual or to a social system as a result of the adoption or rejection of an innovation Although obviously im portant, the consequences of innovations have received inadequate attention by change agents and by diffusion researchers.

Consequences have not been studied adequately because (1) change agencies have overem- phasized adoption per se, assuming that an innovation's consequences will be positive, (2) the usual survey research methods may be inappropr iate for investigating consequences, and (3) consequences are often difficult to measure.

Consequences are classified as (1) desirable vers us undesirable, (2) direct versus indirect, and (3) anticipated versus unanticipated. Desirable consequences are the functional effects of an innovation for an individual or for a social system. Undesirable consequences are the dysfunc- tional effects of an innovation for an individua l or for a social system. Many innovations cause both positive and negative conseque nces, and it is thus erroneous to assume that the desirable impacts can be achieved without also experienci ng undesirable effects. We conclude that the effects of an innovation usually cannot be managed so as to separate the desirable from the undesirable consequences.

Direct consequences are the changes to an individual or a system that occur in immediate re- sponse to an innovation. Indirect consequences are the changes to an individual or a system that occur as a result of the di rect consequences of an innova tion. They are the consequences of the consequences of an innovation.

Anticipated consequences are changes due to an innovation that are recognized and intended by the members of a system. Unanticipated consequences are changes due to an innovation that are neither intended nor recogn ized by the members of a system.

The undesirable, indirect, and unanticipated consequences of an innovation usually go togeth- er, as do the desirable, direct, and anticipated consequences. An illustration is provided by the introduction of the steel ax among Australia n aborigines, which caused many undesirable, indirect, and unanticipated consequences, including breakdown of the family structure, the emergence of prostitution, and misuse of the i nnovation itself. The case of the steel ax illu- strates three intrinsic elements of an innovation: (1) form, the directly observable physical ap- pearance and substance of an innovation, (2) function, the contribution made by the innovation 49 14 to the way of life of individuals or to the social system, and (3) meaning, the subjective and frequently subconscious perception of the innovation by members of the social system.

Change agents more easily anticipate the form and function of an innovation for their clients than its meaning.

Stable equilibrium occurs when almost no change is occurring in the structure or functioning of a social system. Dynamic equilibrium occurs when the rate of change in a social system is commensurate with the system's ability to cope with it. Disequilibrium occurs when the rate of change is too rapid to permit the system to adjust. Change agents generally wish to achieve a rate of change that leads to dynamic e quilibrium, and to avoid disequilibrium.

One goal of diffusion programs is to raise the le vel of Good in a system. A second dimension of consequences is whether the distribution of Good among the members of a system becomes more or less equal. The consequences of th e diffusion of innovations usually widen the so- cioeconomic gap between the earlier and later adopting categories in a system. Further, the consequences of the diffusion of innovations usually widen the socioeconomic gap between the audience segments previously high and low in socioeconomic status.

A system's social structure par tly determines the equality vers us the inequality of an innova- tion's consequences. When a system's structure is already very unequal, the consequences of an innovation (especially if it is a relatively high-cost innov ation) will lead to even greater inequality in the form of wider socioeconomic gaps.

What strategies could be followed in order to narrow gaps? The answer depends on three main reasons why socioeconomic gaps ordinarily widen as a consequence of diffusion: (1) "ups" have greater access to informati on that creates awareness of innovations; (2) they have greater access to innovation-evaluation information from peers; and (3) "ups" possess greater slack resources for adopting innovations than do "downs ". When special efforts are made by a diffu- sion agency, it is possible to narrow, or at least not to widen, socioeconomi c gaps in a social system. In other words, wideni ng gaps are not inevitable. The digital divide is the gap that exists between individua ls advantaged by the Internet and those individuals relatively disadv antaged by the Internet. This inequality exists both within the United States and between North America a nd Europe versus developing nations. Efforts to bridge the digital divide, such as providing public access to computers and the Internet in cyber cafés and telecenters, are under way.

50 1 Universität Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-06 January 2007Hoffmann Diffusion of Hybrid Corn in Iowa 1* Everett M. R OGERS RYAN and G ROSS ’s (1943) study of the diffusion of hybrid seed corn in Iowa is the most in- fluential diffusion study of all time, despite the 5,200-plus diffusion investigations conducted since this pioneering study. The hybrid corn investigation include s each of the four main ele- ments of diffusion (an innovation, communication ch annels, time, and the social system) that we have just discussed and serv es to illustrate these elements.

Hybrid corn became one of the most important ne w agricultural technologies after it was re- leased to Iowa farmers in 1928. The new seed ushered in the agricultural innovations in the 1930s through the 1950s that led to an agricultu ral revolution in farm productivity. Hybrid seed was developed by agricultural scientists at Iowa State University and other state land- grant universities. The diffusion of hybrid seed was heavily promoted by the Iowa Agricultur- al Extension Service and by salesman from seed corn companies. Hybrid corn yielded an in- creased harvest of about 20 per cent per acre over the open-pollinat ed varieties that it replaced.

It was also more drought-resistant and better suite d to harvesting with mechanical corn pick- ers. The seed lost its hybrid vigor after the first generation, so farmers had to purchase hybrid seed each year. Previously, farmers had saved th eir own seed, selected from their best-looking corn plants. The adoption of hybrid corn meant that an Iowa farmer had to make important changes in his corn-growing behavi or. Hybrid seed corn ushered in a new era of farmers' de- pendence on agribusiness companies that sold chemical fertilizers, pesticides, and other farm inputs.

When Professor Bryce R YAN , fresh from his Ph.D. studies in sociology at Harvard University, arrived at Iowa State Univers ity in 1939, he chose hybrid corn as the innovation of study in his investigation of social factors in economic decisions. This interest drew him to study how Iowa farmers' social relationships with their neighbors influenced them to adopt hybrid corn.

Ryan had read anthropological work on diffusion while he was at Harvard, so he cast his Iowa study of hybrid corn in a diffusion framework. But unlike the qualitative methods used in anthropological studies of diffus ion, the Iowa investigation mainly utilized qualitative data from survey interviews with Iowa farmer s about their adoption of hybrid corn seed.

In the summer of 1941, Neal C. G ROSS , a new graduate student in rural sociology, was hired as a research assistant on the hybrid corn diffusion project. R YAN and G ROSS selected two small Iowa communities located some fifty mile s west of Ames and conducted personal inter- views with all of the farmers living in these two systems. Using a structured questionnaire, Neal G ROSS , who did most of the data gathering, interviewed each respondent as to when the farmer decided to adopt hybrid co rn (the year of adoption was to become the main dependent variable in the data analysis ), the communication channels us ed at each stage in the innova- tion-decision process, and how much of the fa rmer’s corn acreage was planted in hybrid (ra- 1 Source: Everett R OGERS , 2003: The Diffusion of Innovations. Fifth Edition. The Free Press, New York, 31-35, 273.

* This case illustration is based on R YAN and GROSS (1943), GROSS (1942), RYAN and GROSS (1950), and VALENTE and ROGERS (1995). Fig. 1 –4 and Table 1 are added from the original publi- cation, R YAN and G ROSS 1943, by Volker Hoffmann.

51 ther than open-pollinated seed) each year. In addition to these recall data about the innovation, the two rural sociologists also asked each resp ondent about his formal education, age, farm size, income, frequency of travel to Des Moines and other cities, readership of farm maga- zines, and other variables that were later correlated with innovativeness (measured, as the year in which each farmer decided to adopt hybrid corn).

Neal G ROSS was from an urban background in Milw aukee, Wisconsin, and initially felt somewhat uncomfortable interviewing Iowa farmers. Someone in Ames told G ROSS that farm people got up very early in the morning, so on his first day of data gathering, he arrived at a respondent's home at 6:00 A.M., while it was stil l half dark. By the end of this first day, G ROSS had interviewed twenty-one people, and he averaged an incredible fourteen Interviews per day for the entire study! Today, a survey in terviewer who averages four interviews per day is considered hardworking. During one personal interview, an Iowa farmer, perhaps slyly leading him on, asked G ROSS for advice about contro lling horse nettles. G ROSS had never heard of horse nettles. He told the farmer that he should call a veterinarian to look at his sick horse (horse nettles are actually a kind of noxious weed). Source: R YAN & G ROSS 1943,22 Neal G ROSS personally interviewed 345 farmers in the two Iowa communities, but twelve farmers operating less than twenty acres were discarded from the data analysis, as were 74 respondents who had started farmi ng after hybrid corn began to diffuse. Thus, the data analy- sis was based on 259 respondents.

When all of the data were gathered, R YAN and G ROSS coded the farmers' interview responses into numbers. The diffusion researchers analyzed the data by band tabulation and with a desk calculator (Computers were not available for data analysis until some years later). Within a year, G ROSS (1942) completed his masters thesis on the diffusion of hybrid corn, and shortly thereafter R YAN and G ROSS (1943) published their research fi ndings in the journal Rural So- ciology (this article is the most widely cited publication from the study, although there are several others). This paper became the founding document for the research specialty of the diffusion of innovations. Several previous studi es had been completed on the diffusion of agricultural innovations, but they did not lead to a research tradition because they did not create a research paradigm fo r the diffusion of innovations (V ALENTE and R OGERS 1995). The 2 52 3 ugh they did not use t the new idea. The overall shape of the rate of adoption looked like an "S" (see Figure. 7-1).

R YAN and G ROSS (1943) study established the customary research methodology to be used by most diffusion investigators: retrospective surv ey interviews in which adopters of an innova- tion are asked when they adopted, where or fr om whom they obtained information about the innovation, and the consequences of adoption. R YAN and G ROSS (1943) popularized the term "diffusion" (which had previously been used by Anthropologists), altho the concept of "innovation”. That term would come from later scholars.

All but 2 of the 259 farmers had adopted hybrid corn in the thirteen years between 1928 and 1941. When plotted cumulatively on a year-by-year basis, the adoption rate formed an S- shaped curve over time. After the first five years, by 1933, onl y 10 percent of the Iowa far- mers had adopted. Then the adoption curve "too k off," shooting up to 40 percent adoption in the next three years (by 1936). Then the rate of adoption leveled off as fewer and fewer far- mers remained to adop Rogers, 2003, 273 Source: ROGERS 2003 53 4 oines, Iowa's largest city, located about seventy- five miles from the two study communities.

f about one acre, before deciding to plant 100 percent of his corn acreage in hybrid varieties.

Table 1: Medi ual Years by e W h ra irs ed br eed Farmers were assigned to adopter categories on the basis of when they adopted the new seed (G ROSS 1942). Compared to later adopters, the innova tors had larger-sized farms, higher in- comes, and more years of formal education. The innovators were also more cosmopolite, as measured by their number of trip s to Des M Although hybrid corn was an i nnovation with a high degree of relative advantage over the open-pollinated seed that it replaced, the t ypical farmer moved slowly from awareness- knowledge of the innovation to adoption. The innovation-decision period from first know- ledge to the adoption decision aver aged about nine years for all respondents, an indication that the innovation-decision process invo lved considerable deliberation, even in the case of an in- novation with spectacular results. The average re spondent took three or four years after plant- ing his first hybrid seed, usually on a small trial plot o an Per Cent of Corn Acr eage in Hybrid for Individ Y ar in hicOpe tor F t us Hyid S Year first used hybrid 1933 1934 1935 1936 1937 1938 1939 1940 1941 No. of cases Before 19341 38,0 50,0 67,0 100100100100100 100 24 1934 20,0 29,0 42,067,095,0 100100 100 16 1935 18,0 44,0 75,0 100100100 100 21 1936 20,041,062,5 100100 100 36 1937 19,055,0 100100 100 61 1938 25,079,0 100 100 46 1939 30,091,5 100 36 1940 60,514 100 1941 54,0 3 Total 257 Never accepted 2 Total sample 259 1 The median hybrid planting for this group in first year of acceptance was 12 per cent of total corn acreage. Source: R YAN and G ROSS 943,19 Communication channels played different roles at various stages in the innovation-decision process. The typical farmer first heard of hybrid seed from a salesman, but neighbors were the most frequently cited channel leading to pe rsuasion. Thus salesmen were more important channels for earlier adopters, and neighbors were more impor tant for later adopters. The R YAN and G ROSS (1943) findings suggested the important role of interpersonal networks in the diffusion process in a system. The farmer-to-farmer exchanges of their personal expe- riences with hybrid seed were at the heart of diffusion. When enough such positive expe- riences were accumulated by the innovators, and especially by early adopters, and exchanged with other farmers in the community, the rate of adoption took off. This threshold for hybrid corn occurred in 1935. After that point, it would have been impossible to halt its further diffu- sion. The farm community as a social system , including the networks linking the individual farmers within it, was a crucial element in th e diffusion process. In order to understand the 54 5 43) should have asked - nity members as if they were unrelat ed respondents in a random sample" (K ATZ et al. 1963).

role of diffusion networks and opinion leadership, R YAN and G ROSS (19 sociometric questions of their respon dents, such as "From which other Source: R YAN and G ROSS 1943,17 farmers did you obtain information about hybrid co rn?" The sample design, which consisted of a complete enumeration in two communities, would have made the use of such communi- cation network questions appropriate. But " information was simply collected from all commu Source: Ryan & Gross, 1943, 20-21 55 6 ividuals who are influenced to follow their lead. Diffusion is fundamentally a social process.

m. The hy- brid corn study has left an indelible stamp on the histor y of all diffusion research.

GR e Diffusion of a Culture Trait in Two Iowa Townships. MSc Thesis, Iowa K A s of Research on the Diffusion of R Y 43: The Diffusion of Hybrid Seed Corn in Two Iowa Communities.

R Y Two Iowa V A radigm as an Exam ple of Scientific Growth. In: Science Communication 16(3):238-269 Even without sociometric data about diffusion networks, R YAN and G ROSS (1943) sensed that hybrid corn had spread in th e two Iowa communities as a kind of social snowball: "There is no doubt but that the behavior of one individual in an interacting population affects the beha- vior of his fellows. Thus, the demonstrated su ccess of hybrid seed on a few farms offers new stimulus to the remaining ones." The two rural sociologists intuitively sensed what later diffu- sion scholars were to gather more detailed evid ence to prove; that the heart of the diffusion process consists of interpers onal network exchanges and social modeling by those individuals who have already adopted an i nnovation to those ind Study of the invisible college of rural sociologists investigating diffusion as of the mid-1960s identified the researchers who first utilized a new concept and/or methodological tool in stud- ying diffusion (C RANE 1972), R YAN and G ROSS launched fifteen of the eighteen most widely used intellectual innovations in the rural sociology diffusion research tradition. So Bryce R YAN and Neal G ROSS played key roles in forming the classical diffusion paradig Bibliography: CRANE , Diana, 1972: Invisible Colleges. Univ ersity of Chicago Press, Chicago. OSS , Neal C. 1942: Th State College, Ames. TZ , Elihu, L EVIN , Martin L., H AMILTON , Herbert, 1963: Tradition Innovations. In: American Sociological Review , 25:237-253. AN , Bryce, G ROSS , Neal C. 19 In: Rural Sociology 8: 15-24. AN , Bryce, G ROSS , Neal C. 1950: Acceptance and Diffusion of Hybrid Seed Corn in Communities. Research Bulletin 372, Agricu ltural Experiment Station, Ames, Iowa. LENTE , Thomas W., R OGERS , Everett M. 1995: The Origins and Development of the Diffusion of Innovations Pa 56 1 Universität Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-07 January 2007 Albrecht, Hoffmann Acceptance of the Salk Polio Vaccine – an example of the situational approach to the diffusion of innovations An early American study by J. C. BELCHER 1 indicates that the diffusion of innovations must be seen and explained in a situ ational approach. It shows that the relations between adoption behavior and independent variab les are not always as relevant as assumed and do not always follow the expected direction. Until today the importance of this study has been overlooked in diffusion literature, or at least ignored 2, despite showing the fallacies of taking factors as fix, and always influencing in the same direction, instead of taking the special conditions of the situation into consideration, a nd understanding that a factor’s influence depends on the con- text.

B ELCHER began his investigation about the diffusion of the Salk Polio Vaccination with the hypothesis that, in this case, as in the many kno wn studies regarding the diffusion of agricul- tural innovations, factors like income, education level, expos ure to information (reading jour- nals, access to television, etc.), living standard etc. would be definitive in people’s readiness to get their children vaccinated. However, he found out in his study in Central Georgia with 701 households that there were to tally different innovators and early adopters. The first ones were Afro-American families with low soci o-economic status and education level.

B ELCHER tried to follow up this phenomenon of unexp ected behavior by searching for expla- nations. He studied the course of the public discussion about this new vaccine, interviewed some concerned persons, and combined these findi ngs with some simple facts. This enabled him to conclude some rather convincing interpretations.

The press had recently reported in great detail and on high ranks about the unsatisfying results of previous polio vaccinations. This resulted in a rather skeptical attitude towards that new vaccine from the families, reading newspapers – usually the more educated ones with better income and of European origin. The vaccination was offered free of charge by th e Public Health Service, an institution that the white population considered to be created especi ally for black and low income people. There- fore, the more affluent families did not like to bring their children to the Public Health Service for vaccinations. However, the children of low income Afro-American families were driven by school busses free of charge to go to vaccination, and this was not t\ he case for most child- ren from affluent white families. White persons in social professions, being trusted by Afro- Americans, supported the vaccination program. Teachers were in a competition with each 1 Acceptance of the Salk Polio Vaccine. In: Rural Sociology, 23, 1958,2,158-170 2 Everett R OGERS (1962) shows knowing this study, first on p. 24 , where he refers to a list of stu- dies by Katz and Levin indicating „ 21 drug studies plus additional analysis of medical innovations such as polio vaccine “, and again polio vaccine on p. 45, then in a list of „ studies about public health ideas “ on p. 37, in footnote 16 he cites Belcher 1958, and on p. 174, in a chapter about adopter categories, where he lists studies, confirming the hypothesis: Higher Social Status Than Later Adopters , he states in the related footnote 23: „ Research studies which support this state- ment are .... Belcher (1958) . ..“. This is doubtless false, may be not intentionally, because the op- posite is the finding of Be lcher, the early adopters had t he lowest social status.

57 2 other to reach the 100% vaccinations in their cl asses. They even made the promotion to the next class dependent on being vaccinated. Fina lly, the influence of Afro-American preachers, with whom the Public Health Service co-operated for years in addressing the Afro-American clientele, was effective in convinc ing most of the Afro-Americans.

Whether or not these facts and factors were influe ntial in this specific case is not really rele- vant beyond that example. But knowing at least this plausible interpretation, it cannot be de- nied that factors measured independently from the specific situation are not sufficient to ex- plain the diffusion process. Instead, for any ne w situation, the question about what the rele- vant factors are, has to be asked again. It can al so be concluded that it is not enough to take a situational approach to influentia l factors because the course of events, the specific process is as decisive as the structure of factors in adoption and diffusion.

58 1 Universität Hohenheim 430 Fachgebiet:

Landwirtschaftliche Kommunikations- und Beratungslehre KIM-08 January 2007Hoffmann Book Review: Rogers & Shoemaker, 1971 1 Hartmut A LBRECHT This book is substantially an enlarged, reedit ed and complemented new edition of the best known book of Rogers, Diffusion of Innovations from 1962. The order of presentation is es- sentially the same as in the previous edition, which gave a comprehensive overview on diffu- sion research. The following is emphasized as new:

1. The old book was written for stud ents in later semesters; the new book “is directed to social scientists with an academic interest in the microanalysis of communication and change, and to change agents whose purpose is to diffuse innovations”(p. xviii).

2. The new book shows in many parts a new orientation, generally taking examples? from communication theory. This is in line with the ne w title, the review of the research tradition of diffusion investigations, as well as with the ch oice of new terms, like the replacement of the adoption process by innovation-deci sion process and the related phases: knowledge, persua- sion, decision, confirmation.

3. The book shows two entirely new chapters, sh owing a relevant enlargement in theme com- pared to the first edition: instead of looking at innovation and diffusion exclusively from situa- tions of free choice for single individuals, now group- or collective decisions as well as deci- sions in organizations receive th eir own chapters. The look at co llective decisions is based on orientations mainly developed by George M. B EAL (1966, 233ff) and became known under the label social-action-concept, and through the inve stigations of political scientists, who tried to find out, who mainly influences decisions in la rger communities (power holders). The discus- sion of decisions in organizati ons is based on the sociological ly and social-psychologically oriented work of the gene ral organization literature.

Special importance is devoted to the appendi x. The bibliography lists about 1200 empirical and 300 non-empirical studies from all over the world, classified into research traditions of the respective authors (Cultural anthropology, ag ricultural economy, communication, education, early sociology, agricultural ex tension, geography, general econom ic science, sociology, engi- neering, journalism, marketing, medical soci ology, psychology, administration, rural sociolo- gy, statistics and rhetoric). This overview al one makes the purchase of the book worthwhile for all, who are active in this field of research and teaching.

The classification of the empirical studies according to main research methods applied and main results (done in the Diffusion Documentati on Center at Michigan State University, East Lansing) created the possibilit y, to link the generalizations about diffusion with all studies confirming the stated relationshi p, contradicting it, or showing that such a relationship could not be assessed at all; e.g. “early adopters have larger size d units (farms, and so on) than do later adopters (152 studies, or 67 % support; 75 studies do not support) ”(p. 361). I do not know any publication, offering a comparab le overview on diffusion research.

1 Source: A LBRECHT , Hartmut, 1973, In: Sociologia Ruralis, 13(3,4)294-299 Book Review of: Rog- ers, E.M. and F.F. Shoemaker (1971): Communica tion of Innovations, The Free Press, New York, 476 pp. Translation from German to English by Volker Hoffmann.

59 2 The critic, presented in the following, is intended to stimulate the rethinking of the relevant aspects of diffusion research and th e analysis of its results by R OGERS and S HOEMAKER , who come from a different orientation.

Following P OPPER , the value of a theory is proved if seri ous efforts to refute it fail; because it is not difficult, in a recurrent investigation of similar situati ons to find confirmation of estab- lished hypotheses. From this orientation, studies that contradict (‘not supporting’) the hypo- theses have to be seen as refutation. Under this perspective, - if confirmation is only seen at a limit at minimum 90 % of the studi es – only 41 out of 101 generalizations persist. For these 41 not rejected generalizations near ly without exception less than 10 studies exist, which in gen- eral have not been designed as ‘effort of refutation’ (P OPPER ). Practically this means, that the core of the derived relationships (to which the generalizations of the studies refer) in the pre- sented form cannot be confirmed. This claims evidently for a reformulation of the generaliza- tions (hypotheses). Either the c onditions of validity must be sp ecified more clearly or relation- ships in another theoretical or ientation have to be found. The necessity for this has been recognized for a long time. Striking for me is the investigation of B ELCHER , on participation in a recommended vaccination campaign against polio in Geor- gia, USA, published 1958 in the journal “R ural Sociology.” The study was designed on the basis of the generalizations that came from th e rural sociological diffusion research at that time, and checked the vaccination acceptance ag ainst characteristics like income, school edu- cation level, contact to media, cosmopoliteness, et c. It was discovered that most of the rela- tionships did not match the expectations, in fa ct they often contradicted the hypotheses. The fact, that a great deal of hypot hesized generalizations have no t been confirmed by the observa- tions, means that the same factor (age, size of en terprise, contact to media, etc.) does not have the same direction and strength of effect upon a doption behavior in different situations. There- fore effects cannot be explained by isolated factors, but only by constellations of factors , - the whole context of the specific situation. Respecting this fact, L EWIN had developed and tested the approach of field theory in 1931. R OGERS and S HOEMAKER make use of the studies that came out of this orientation at different places throughout their book (L EWIN , LIPPITT , CHIN and others).

Approaches to this perspective are also visible in the first edition (diffusion of innovations) in the final chapter, written together with A. Eugene H AVENS (pointing to C OTTERELL ’s ‘situa- tional field’, 1942). However. theses approaches have not been used so far for the further de- velopment of a theoretical foundation for diffusion research.

As a consequence, the new book also does not offer a satisfying interpre\ tation of the expe- rience, that – successful – diffusi on processes frequently show a wave like curve, that means, that the curve of adoption per unit of time at first inclines only slowly, than more quickly, turns, and fades out again slowly. R OGERS and S HOEMAKER make analogy to explain the curves of the individual l earning processes, (p. 178: “If a social system is substituted for the individual in the learning curve, it seems reasonable to expect that experience with the innova- tion is gained as each successive me mber in the social system adopts it.” ) and the course of epidemic diseases. The first explanation equa tes learning (recall) performance with decision and action, and equates individual and social syst em – two risky theoretical jumps. The second explanation equates decision and action with passive suffering (being infected). Another ex- planation is also given, (in the old edition ‘interaction effect’) which says, that many informed persons, as compared to only a few informed persons, can inform (influence) more non- informed people per unit of time. The slowed c ourse of diffusion in the second half of the process then can only be based on the reason, (an assumption far from reality) that towards the end of the process it becomes more and mo re difficult to find non-knowers to whom the knowers could tell about the innovation (p. 179).

60 3 The course of successful and failed diffusion processes, and the critical consideration of inno- vators in the early stage of th e diffusion process can be interpre ted on the basis of well proofed theory, if the following points are considered:

1. In the initial situation the potential readiness for action of all members of the social system are already differentiated (accor ding to attitude, relative attachment to norms, knowledge and skills, subjectively felt urgency, relative advant age of an innovation, objective conditions and options for action, etc.). In many cases for these factors norma l distribution can be assumed.

(only a few cases with exclusively driving fact ors, and only a few with exclusively inhibiting factors).

2. The situations of action are to be seen as force fields, in which the factors in different strength act as driving or restraining forces (H RUSCHKA 1964, 117ff).

3. The change of the constellation of forces cau ses behavior change, and is decisive: the sub- jective appreciation of the assumed consequences of an innovation (in case of succeeding) as well as the likelihood of success and the fear of the possible negative consequences of the adoption of an innovation (in case of failu re) as well as the likelihood of failure.

4. In small scale social system s (with which advisory work a nd more general support normally deal) the perception fields of its members overlap. The behavior of ot hers (relevant members of the social system) influences a change of the action situation in the remaining ones (reci- procal dynamic action-reaction relations).

5. Firstly, Innovations cause uncertain ties (defense attitude of the others). With the proof that desirable results can be realized, they then intr oduce driving forces into the action situation of others and reduce restraining forces (risk). This effect is multiplied, if influential persons (comparable relevant others) adopt and rende r the innovation into the new norm. These dy- namic effects are situated – in successful diffusi on processes – timely in the first half of the whole process. Afterwards, in substance, only the differentiation of the action situations inhi- bits action. This explains the slow fading out of the process (for the whole process compare e.g. A LBRECHT 1969, 2566ff.; B OESCH 1966; E MERY 1962).

Such an interpretation, without additional assumptions, or anothe r theoretical approach, makes it possible, to also explain processes, that so far unfortunately have not gained recognition in diffusion research, namely those where promoti onal efforts create active resistance, actuate conflicts in the social system and fail partially or totally. The underlying reasons for these so- cial psychological processe s have been analyzed by S PIEGEL (1961). With the help of the se- mantic differential profile (polar ities profile) he assessed the distribution of opinions in the social field (concerning a pr oduct, an object of opinion, an innovation) and could thereby make the emergence and the crystallization of th e fronts of rejection visible. Based on this and other studies, the likely processes of the introduction of innovations opposing norms or the introduction of innovators not fit ting to the norm can be described and receive their first orien- tation in reality (A LBRECHT 1969,268ff.).

Lastly,: Rogers and Shoemaker state explicitly, th at directed change, or planned change is the main theme of their book. An adviser or change agent, seeking helpful orientation for his work, - unfortunately – will not find much help. Two facts seem to be relevant here:

1. The great majority of diffusion studies are based on ex-post interv iewing of the ‘adopting units’. Together with the surely necessary and revealing consideration of the time factor (the relative position of the adopti on in time compared with other adopters): ‘innovativeness of members of a social system,’ has led research to nearly exclusively successful innovation processes but not ones that have failed. The causes of failure woul d have been of special inter- est. Without investigating the cau ses of failure of directed change, diffusion research can only describe such change without methodologically targeted support. The effects of restrain 61 4 ing factors in specific situations are not a matter of research. Therefore, no insights can be gained about how these restraining forces can be lowered or overcome.

2. As an advisor, one can learn from diffusion re search, that individuals with better education, higher income, greater prestige, cosmopolite pers onal contacts, increased media access , etc., adopt innovations earlier than others. These are factors of the situation, scarcely influenced by an advisor. He can only infl uence his methodological appro ach. The methodological proce- dures of support organizations were researched in too few studies in connection with the adop- tion and diffusion of innovations, that this category of studies does not meas ure (see p.72f.), as even a possible category of studies. Field experi ments – that would give answers to questions about purposeful methods of promotion – are exp licitly cited as necessary by the authors. They state (p. 65), that recently there is a trend in th is direction. Therefore, it would be worthwhile to direct the theoretical orientat ion, that up to now nearly exclusively concerns the target sys- tem, with a same emphasis on th e support system. If support fails, then causes have to be searched for in the whole field of interaction (that is in the support system, the target system and between the two systems). C HIN 1962, and R ILEY and R ILEY 1959, among others, give valuable theoretical contri butions on this issue. For the actual state of the art diffusion resear ch, referred by Rogers and Shoemaker therefore is valid, what B ENNIS (1965,339) criticized correctly, but in another context: “ They are theo- ries of change, and not of changing”. That is the reason to point here to work of B ENNIS , BENNE AND CHIN (1962) and of H AVELOCK (1971), in which the possibilities and problems of planned change are dealt with.

Alltogether: the book of R OGERS and S HOEMAKER is extremely informative, stimulating, and highly readable – it needs readers who can critically check it.

Bibliography ALBRECHT , H. (1969): Innovationsprozesse in der Land wirtschaft. Eine kritische Analyse der ag- rarsoziologischen 'adoption'- u nd 'diffusion'- Forschung in bezug auf Probleme der landwirt- schaftlichen Beratung. D. Br eitenbach Verlag der SSIP Schriften, Saarbrücken.

B EAL , G. M. (1966): Decision making in social change. In: Schweitzer, H. J. (Eds.): Rural sociology in a changing urbanized society. Dept. of Agricultural Economics, AES, Urbana, III.

B ELCHER , J. C. (1958): Acceptance of the Salk polio vaccine. In: Rural Sociology, 23, pp. 158-170. BENNIS , W. G., (1965): Theory and method in applying be havioral science to planned organiza- tional change. In: The Joumal of Appli ed Behavioral Science, Washington, D.C., 1 , 337-360.

B ENNIS , W. G., B ENNE , K. D., and C HIN , R., (1962): The planning of change. Readings in the ap- plied behavioral sciences. Holt, Rinehart and Winston, New York.

B OESCH , E. E. (1966): Psychologische Theorie des sozialen Wandels. In: B ESTERS , H. und B OESCH , E. E. (Eds.): Entwicklungspolitik. Handbuch und Lexikon. Matthias Grünewald, Stutt- gart, Berlin. Kreuznach u. Mainz, Col. 335-416.

C HIN , R. (1961): The utility of syst em models and developmental models for practitioners. In: B EN- NIS , W. G., BENNE , K. D. U. CHIN , R. (Eds.): The planning of cha nge. Holt, Rinehart and Winston, New York:, pp. 201-214.

C OTTRELL , L, S. (1942): The analysis of situational fields in social psychology. In: Americ. Socio- logical Review. New York, 7, pp. 370-382 .

62 5 EMERY , F. E. (1962), Group dynamics in rural extension. A case study of the diffusion and adop- tion of a new practice. International Agricultural Center, Wageningen.

H AVELOCK , R. C. (1971): Planning for innovation throug h dissemination and mobilization of know- ledge. 3. edition. Institute for Social Research, Center for Research on Utilization of Scientific Knowledge, Univ. of Michigan, Ann Arbor.

H RUSCHKA , E. (1964): Psychologische Grundlagen des Beratungsvorgangs. In: Probleme der Be- ratung. Eugen Ulmer, Stuttgart, pp. 107-136.

L EWIN , K. (1931): Der Übergang von der Aristoteli schen zur Galileischen Denkweise in Biologie und Psychologie. In: Erkenntnis, 1, pp. 421-466.

L EWIN , K. (1958): Group decision and social change. In: M ACCOBY , E. E., NEWCOMB , T. M. AND HARTLEY E. L.(eds.): Readings in social psychology. 3. edition. Holt, Rinehart and Winston, New York, pp. 197-211.

L IPPITT , R., W ATSON , J. and W ESTLEY , B. (1958):The dynamics of planned change. Harcourt, Brace & Co., New York.

R ILEY , J. W,. jr., and R ILey, M. W., (1959): Mass communication a nd the social system. Merton, Broom and Cottrell (Eds.): Sociology today . Basic Books, New York, pp. 537-578.

R OGERS , E. M., 1962: Diffusion of Innovations. T he Free Press of Glencoe, New York.

S PIEGEL , B., 1961: Die Struktur der Meinungsverte ilung im sozialen Feld. Das psychologische Marktmodell. Hans Huber, Bern und Stuttgart.

63 Universität Hohenheim Fachgebiet: Landwirtschaftliche Kommunikations- und Beratungslehre KIM-09 January 2007Hoffmann Book Review: Five editions (1962-2003) of Everett ROGERS:

Diffusion of Innovations 1 Volker Hoffmann Preface Innovation and diffusion was a main topic of study in the early professional life of Hartmut Albrecht, my academic teacher and predecessor. This interest connected him to colleagues such as Eugene Wilkening, Herbert Lionberger, Everett Rogers in the USA; Anne van den Ban in the Neth- erlands; and to many of the ea rly participants of the ESEE 2 group. I find my-self, now, one of the senior members of this group. Out of respect and appreciation, I am bringing to completion, in the footsteps of Erna Hruschka and Hartmut Albrecht, some of their unfinished work. A new MSc bloc- course on ‘Knowledge and Innovation Management ’ brought about the opportunity to study the books of Rogers again, and to add to the book re-view of Hartmut A LBRECHT of the second edition now one of all five. This review of his work is, indeed, late in coming especially as Albrecht’s re- view in 1973 was published only in German in Sociologia Ruralis, along with his “habilitation” (1969) about innovation processes in agriculture which provided his alternative theoretical concept.

As Everett R OGERS in edition two refers to 3 of Albr echt’s earlier German publications (1963, 1964, 1965), it is noticeable that A LBRECHT ’S later and even more significant writings (1969, 1973) do not appear in editions three to five. When th e three of us met on 17 June 1996, at Schloss Thur- nau in Bavaria for an afternoon symposium among economists, Rogers acknowledged Albrecht “as his most important German colleague” . Unfortunately, Albrecht and Rogers are no longer with us to read and react to this review themselves. Ho wever, life and science goes on and from this per- spective it is never too late.

Positive Assessment Judging by the number books sold and ci tations given, Rogers is the most successful scientist to ever come out of the tradition of Rural Sociolo gy. This begs an explanation. Everett Rogers is a highly gifted writer, a very tale nted communicator and has a feeling for relevant and challenging is- sues. Much of his communication success is due to his ability to make complex things simple. He gives his audience a feeling of understanding by grasping the few common features and factors un- derlying the confusing diversity of reality. He appears as the great simplifier, showing that within research on innovation and diffusion thousands of phenomena, cases, and studies from many differ- ent research domains and research traditions can be traced back to a few simple principles and a limited number of generalizations.

In each edition, his carefully selected cases and ex amples provide persuasive evidence that impress and often surprise the reader. Some of these, such as ‘Water Boiling in Peru’, ‘Steel Axes for Stone- 1 H OFFMANN , Volker, 2007, In: Journal of Agricultural Education and Extension, Vol. 13, No. 2, 147-158, with minor modifications 2 ESEE = European Seminar on Ext ension Education, bi-annual informal conference meeting of re- searchers and University teac hers, mainly from Europe.

64 Age Aborigines’ or ‘Hard Tomatoes in California’ have circled the globe and entered innumerable Social Science textbooks.

Despite the growing boom of new studies – primarily sparked by his own writings - his efforts to keep up-to-date did not diminish. Each new edition of his famous first book was revised as he strove to integrate new developm ents and findings as well as answer criticisms. This is demon- strated by the additional pages, new chapters and bib liographies listing recent titles in each edition, including the fifth edition (2003) which appeared one year before he passed away at 83 on 31 Octo- ber 2004.

Albrecht’s critique was based on R ogers’ books, especially the first tw o editions; as is this updated review version. Many of our criti ques get their inspiration from Rogers’ texts. His work is a fore- runner without which many of our alternative concepts and theoretical frame-works (e.g. A LBRECHT et al. 1989) could not have been developed. In this sense, we are proud to be his scholars and see our contributions as complementary to a unique a nd paradigmatic masterpiece of applied social sci- ence in agriculture.

Critical Considerations By highlighting Rogers’ efforts, and not just his achievements in the positive assessment above, some seed of criticism appears. Ultimately, Rogers must be measured by his self-established yard- stick and aspirations. Rigor is required, especially as his outstanding success sees him widely re- garded as the classic aut hor in this field. Where he fails, coun tless novices fail with him. Even where he does not fail but only gives rise to the s lightest misunderstanding, he shares responsibility for the resulting consequences for science and practice.

Here I will state a few general criticisms, followed by several sub- chapters which discuss specific points in greater detail, al ong with evidence from sources.

Herbert L IONBERGER (1960) summarized the findings of th e rural sociology tradition in his book “Adoption of New Ideas and Practices”, obviously writ ten for extension staff and use in the work of Land Grant Colleges. Along with the key findings a nd generalizations, most of the special termi- nology can be found there. He even included the most prominent of the early medical sociology studies (C OLEMAN & M ENZEL . 1955, COLEMAN ET AL. 1957). The famous diffusion curve with adopter categories and percentages along mean and standard deviation is presented on page 37, re- ferring back to R OGERS & BEAL (1958, 33).

R OGERS refers to L IONBERGER regularly in his first two editions, in the third edition his name only appears as the co-author of a study about Taiwan, and in the last two editions L IONBERGER is no longer mentioned. Rogers claims to go far beyond the rural sociology tradition and to summarize findings and approaches from all di fferent disciplines (six in the first edition, eight in chapters and 16 as categories in the bibliography of the fift h edition). His subtitle of edition two “A Cross- Cultural Approach” even includes the traditional so cieties in developing countries. Rogers writes for scientists but also for other educated readers like “advanced co llege students enrolled in” social science courses. He presents a more scie ntific attitude and writing style than L IONBERGER although there is not much difference in substance. While L IONBERGER reads like a summary of experiences gained from research and extens ion work, Rogers reads like a su mmary of findings only from re- search, and as the theoretical foundation of a new tradition of interdisciplinary research. We see a clear strategy of confirming importan ce by applying the “more of the same” principle.

L IONBERGER (1960) refers to exactly 100 diffusion studies. ROGERS (1962) reviews 506 diffusion studies in edition one (p. 5), continues with more than 1500 in edition two (p. 41), and cites 5200 studies in the preface of edition fi ve (p. xvii). Here the interested reader can also find a type of an 65 academic CV with the author’s evaluation of the different stations and their performances over his career. Compared with the marked increase in st udies, the promised qualitative gains are less re- markable: “The stream of diffusion scholarsh ip over the past sixty years represents both similarities and differences, continuities and discontinuities, and so does this book. By no means, however, do I seek only to synthesize the important findings from past re-search. I also strive to criticize this work (including my own) and to suggest direct ions for the future that are different to the past. I have once again titled this book diffusion of innovations to identify it with the forty- year sequential tradition of diffusion studies marked by my 1962 book of the same title. ”(V:

xviii) Over the five editions it is difficult to avoid the impr ession that his first and foremost aim is to build up, defend and maintain the author’s reputation. This desire, in moderation, fuels most outstanding performances but it should not aff ect scientific quality. Errors are always attributed to others; criti- cisms only concern other researchers, the readers, or whom ever, but never Everett Rogers himself.

At no point do we read ‘here I have failed, here I have to correct previous statements or explana- tions’. Yes, progress is seen in terminology and in sights; yes, Rogers was incomplete, therefore, more studies were included and complementary chapters added, but he himself was never wrong.

Only under the heading of “Shortcomings of the Pre- sent Approach” (II: 91ff.) does he deal with his own approach to summarizing findings, but in a defe nsive way that is without further consequences.

(See my paragraph on generalizations, later.) This unwillingness for self-criticism and self- correction leads to more and more inconsistencies and internal contradictions from edition to edi- tion.

A Critical Attitude, but Not App lied to his Own Thoughts and Writings In I: 38: “One general criticism of the rural sociology tr adition which has been voiced by rural soci- ologists themselves (examples are Lionberger 1952 and 1960), is the lack of attention to so- ciological theory. There is a noticeable tendency for many rural sociology dif-fusion studies to approach raw empiricism, with little emphasis u pon the sociological significance of findings.” And two pages earlier: “ a great number of later rural sociological studies have followed an uni- maginative ‘factors-related-to-innov ativeness’ approach. The results add very little, in many cases, to present knowledge of how new ideas diffuse except further verification of previous findings.”(I:

36f.) In Table 2.1 he points out that the only co mmon interest of five re-search traditions (except anthropology) was to create findin gs on correlates of innovativeness. At the end of the book, he lists 52 non-linked (context free) generalizations, making it easy for generations of students and scholars to select their research hypotheses and do exactly wh at he had pointed out to be useless (I311-315).

In the later editions, these “useless” studies help increase the number of studies supporting his gen- eralizations. Meanwhile this is done mainly by applying multivariate statistics, mostly Probit and Logit regression models and taking premises and ma gnitude of the regression coefficients as true, even if sometimes there is no way to find any plau sible explanation. The most glaring inconsistency is in his attempt at posing a theory first and fo llowing it with 52 “generalizations” which should provide “a skeleton summary of the major conclusions of what is known about the diffusion of inno- vations ” in his chapter I,XI, (I: 300-315. 101 generalizati ons detailing all studies supporting or not supporting are mentioned in II: 346- 385). More about this inconsistency between a holistic “situa- tionist” theory and isolated cont ext-free generalizations follows.

66 Contextual dependency of Causal Relations - In Search of the Appropriate Theory „ A science without a theory is b lind because it lacks that element which alone is able to or- ganize facts and give direction to research. Even from a practical point of view the mere gathering of facts has very limited value. It canno t give an answer to the question that is most important for practical purposes namely, what must one do to ob tain a de-sired effect in given concrete cases? To answer this question it is necessary to have a theory, but a theory which is empirical and not specula tive. This means that theory and facts must be closely re- lated to each other. L EWIN , 1936“ Rogers chose this statement of L EWIN as a preface to his chapter I, XI; certainly a good choice. He then states: “A search of the diffusion literature reveals (1) a general lack of agreement upon sociological concepts involved in adoption behavior, and (2) absence of a synthesis of these concepts into a general theory that might be tested by empirical research. The many studies that have been completed provide an excellent base for an attempt to formu- late a general theory of the diffusion and adopt ion of innovations. Ordinarily, one would ex- pect theoretical considerations to appear in the early chapters of a book. They could then serve as a framework for the entire volume. How ever, in the present case it is our belief that any theoretical statement must be so highly tentative that it is more appropriately placed at the end rather than at the beginning of the present work. The purpose of\ this chapter is to state a direction in which analysis should proc eed toward a general theory of the diffusion and adoption of new ideas. ” Rogers continues under the headli ne „Theoretical Approach“:

„ Perhaps one of the most effective means of conceptualizing adoption and diffusion behavior is first to view this behavior in its most basic and elementary form, and then to develop some of the complex variables affecti ng this behavior. At one level of conceptualization, adoption of a new idea by an individual is a type of ac tion. According to Parsons and Shils (1952, p. 56), an act consists of three basic elements: (1) an actor (2) orienting to (3) a situation. This con- ceptualization of human behavior implies:

1. Behavior is oriented toward attaining ends or goals.

2. It takes place in situations.

3. It is normatively regulated. 4. It involves an expenditure of effort or "motivation." After identifying an ultimate goal as an individual’s desire for security, he defines security as the „ subjective state of well-bei ng which minimizes tension “(I: 301). This again reminds me of Lewin’s field theory of behavior and beha vior change, where driving and inhibiting forces orient action to reach an equilibrium of forces under different levels of tension. And so do the next two citations:

„ Behavior takes place in situa tions. Individuals do not exist as a mass of disconnected units.

They are members of social systems, and these memberships in social systems have important effects upon their behavior. The situational fields3 in which behavior occurs do not necessar- ily follow community or organizational boundarie s. One may be psychologically identified with a group and take the group's perspective as his own wit hout being on the membership list. Of course, physical proxim ity, along with social status and other psychological identifica- tions, are factors influencing freq uency of interaction. 3. The term "situational field" (Cottrell, 1942) is somewhat preferable to "situation," as the former does not imply time boundaries, while the latter does. "Situation" tends to conno te a given time and place. Situational field is 67 defined as that part of the environment which is perceived by an actor as significant for him.“ (I: 302) „ Perception - The concept of perception is a key dimension in understanding the diffusion of ideas. Although a new idea may be regarded as advantageous by experts in some field, a par- ticular actor may not perceive the innovation in a similar manner. Perception is the way in which an individual responds to any sense or impression which he detects (Lindesmith and Strauss, 1956, p. 85). Perception is a function of the situational fields within which the indi- vidual operates. Knowledge of these situational fields, the manner in which the individual identifies himself, his sense of security, and the normative regularities may enable the theo- retical specification of some of the conditio ns for adoption behavior. As Cottrell (1924) 3 stated, "Items of behavior such as attitudes, traits, etc., studied apart from the context pro- vided by the actor's definition of the situation, yield meaningless results." Thus, it is essential that the present model for adopti on behavior account for the actor's perceptions of the situa- tion. “ Rogers had the right th eoretical perspective, whether based on L EWIN or other more or less contem- porary sources does not really matter, he could clearly define it. The pity is that he did not apply it. Not in the previous chapters, for which the theore tical attempt came too late, nor in his subsequent conclusions. He does not come back to this theoretical approach in the following four editions nor does he apply it. Instead, he falls back on the li st of 52 context-free and independent “generaliza- tions”, unfortunately, ex actly what people were looking for a nd what has survived until today. Gen- eralizations resistant to deeper insights like w eeds resistant to herbicides, which he carefully enlarged and confirmed th rough four more editions.

At least a bit of “situationism” appears when he e xplains why he restricts himself to those strategies which “ apply to a broad range of change agent-client relationships ”. “Such recommendations are often useless outside of a very sp ecific situation because they are rarely general in their applica- tion ” (I: 278). In his later editions two to five, he is more detailed and methodology -oriented in his respective chapters about the change agent.

B ELCHER (1958) as an Example of How Excepti ons are Used to Confirm the Rule Clear empirical evidence for the contextual dependenci es of causal relations is given in the form of the study of B ELCHER (1958) published in Rural Sociology on the acceptance of polio vaccination.

This study contradicts all common hypothesis from the rural sociology diffusion research tradition up to that point. Everett R OGERS (1962) shows knowledge of this study, first on p. 24 where he re- fers to a list of studies by Katz and Levin indicating “21 drug studies plus additional analysis of medical innovations such as polio vaccine “. Polio vaccine is also mentioned on p. 45; then again in a list of „studies about public health ideas” on p. 37. In footnote 16 he cites B ELCHER 1958 and on p. 174 in a chapter about adopter categories he lists 17 studies confirming the hypothesis: “Higher Social Status Than Later Adopters ”. He states in the related footnote 23: „ Research studies which support this stat ement are .... B ELCHER (1958) ...“. This is doubtlessly false because B ELCHER found the opposite - the early adopters had the lowest social status. By committing this error R OGERS could maintain the hypothesis more easily. His readers did not have a ny incentive to check the original study as it was part of a longer list, and one of 500 studies reviewed overall as Rogers states in his preface. In the second edit ion, the study is classified correc tly as not supporting the generali- zation, but the list of suppor ting studies totals 275 agains t 127(II: 357, 359). The name B ELCHER is not included in the author index (many others are also missing) and in editions three to five his name is no longer mentioned.

3 Typing error, should read 1942 68 Generalizations and How to Confirm Them?

In the second edition there is an extra little chapter “Purpose of This Book”. There we find: “ The primary purpose of this book is to synthesize a series of generalizations from research on the diffu- sion of innovations. Each of thes e generalizations represents the relationship found between two or more concepts.” (II: 41) But what are these generalizati ons? Obviously, they are not hypothe ses about causal relations be- tween variables or groups of variables because they do not take any context into consideration. Are they correlations? Again, no, because they do not statistically relate variables. The basis of these generalizations is a given number of studies. As more studies confirm rather than contradict the generalization, the more likel y it is to be confirmed again in future studies – provided the context is the same as in the studies review ed, or that the context does not matter. This would produce a type of a general socio-economic law, and no contradict ing studies would be found from the past nor in the future.

Rogers himself is well aware of this problem in edition two: “ Another shortcoming of our generali- zations in the follo wing chapters is the deceit of their neatness and simplicity. Or generalizations deal almost entirely with pairs of concepts, wherea s the real nature of diffusion is certainly a cob- web of interrelationships among numerous variables ”…Why not include more variables? ” Unfortunately it cannot be. Most of the empirical diffusion studies reviewed in this book fo- cus upon only two-variable hypotheses, and we c annot summarize findings that do not exist.

Further, our ability to understand three-variable, four-variabl e, and so on generalizations usually suffers in direct proportion to the numbe r of variables included. Therefore for the sake of clarity and because we lack an empirical basis to do other-wise, the generalizations in this volume, with only a few exceptions, deal with two concepts. 1 1) However, where the original research publication provided a basis for doi ng so, we coded a generalization as ‘condi- tional’, meaning that the rela tionship found between two variabl es depends upon a third vari- able.” … ”However, only 331 (about 5 percent) of the 6,811 empirical generalizations avail- able as of July 1968, were conditional. All the re st (95 percent) are two-variable generaliza- tions. In Appendix A we consider the conditional relationships as not supporting each diffu- sion generalization. ” (II: 93f. and III: 131, but without the footnote. The entire Appendix A is omitted in editions three-five, but a smaller num ber of generalizations is presented throughout the chapters and in the summaries.) At first glance the procedure, clearly made explic it, might convince quick readers. But what does it really mean? “ For the sake of clarity ” the appropriateness of the gene ralizations is purposefully sac- rificed. He counts “ conditional relationships ” as not supporting the two- variable relation. The only reason given to maintain and even increase the number of generalizations from the first edition is the scarcity of such studies with conditional re lations, and the over-whelming majority of studies using inappropriate study design.

Unfortunately, many readers and resear ch scholars did not see it like that, but took it as widely con- firmed hypotheses valid for all types of innovation-situations.

Does the Diffusion Curve Often Follow a Normal Distribution?

Early mainstream social sciences tried to follow the standards of natural sciences. That meant using quantification, statistical analysis and testing hypothesis, as general as possible, to find “social laws” analogous to natural laws. This was abandoned later on, all human action came to be regarded as context bound and only roughly pr edictable when the context remained more or less the same and qualitative research increased in importance ag ain. At the time the first edition came out, it was a sensation that a social phenom enon like the diffusion of innovati ons should follow the normal dis- 69 tribution. This, never really true but carefully established and maintained myth, contributed a large part to the public interest in diffusion research and Everett ROGERS’ books.

“ The major findings from the hybrid study are: 1. The first use of hybrid seed followed a bell shaped (but not exactly normal) distribution wh en plotted over time. (Ryan and Gross, 1943)”.

(I: 34) “ Testing Adopter Distributions for Normality. A ge neral finding of past investigations is that adopter distributions follow a bell shaped curve over time and approach normality . There are useful implications of th is generalization for a standard me thod of adopter categorization.

Eight adopter distributions test ed by Rogers (1958b) were bell shaped and all approached normality, although half of those tested were found to deviate significantly from normality (Table 6 2). Four additional studies appear in the l iterature on the normality of adopter dis- tributions. None of these analyses utilized the most precise statistical tools for determining normality, but each found that adopter distri butions approached normality. 1. Ryan and Gross (1943) found the distributi on of dates first use of hybrid corn was nearly normal.... “ Looking at Table 6-2, the correct re sult of Ryan and Gross is presented, significantly different from normal distribution at one percent level of signi ficance (Ryan and Gross underline that by figure 4, which shows the curve they observe d next to the one normally expected). In the commentary to his table 6-2, Rogers suggests possible reasons for deviations come from how the time of adoption (trial versus full adoption) was determined and by how farmers were considered, for instance, did they start farming during the course of diffusion. At the end of the four cases he cites, none of which really support normal distribution but only a certain similarity to it, he sums up:

„ Most adopter distributions closely appr oach normality and many are normal. 8 Further re- search is needed to determin e specifically why some adopter curves are normal and some are not. 8.) No claim is made, however, that adopter distributions for all innovations are neces- sarily normal. Sorokin (1959, p. 684) has attacked such a claim: "The convincing logical con- siderations as well as the factual tests do not give any basis for a belief in the existence of any 'normal' or even typical curve of diffusion or di ffusion rate for all cultural values in all cir- cumstances. Such a 'normal' curve is but a myt h." I prefer to disagree with Sorokin, as do most diffusion researchers. The normal adopter dist ribution is useful if viewed as an "ideal type" that provides a standard from which statistical goodness-of-fit can be computed.“ (I:

158f.) In the second edition this reads: “Research has generally shown t hat the adoption of an innova-tion follows a normal, bell shaped curve when plotted over time on a frequency basis.” (II: 176f.) Later, he again gives details to support this although the facts presented do not confirm the generalization so clearly (II: 179f., c ontinued up to V: 275).

Finally, he concludes by introducing his famous diffusion curve with the adopter categories accord- ing to mean and standard deviations of the normal di stribution in Figure 6-1 in I: 162. (Fig. 5-2: II, 182; Fig 7-3, V: 281). The Problem with Innovativeness “The criterion for adopter categorization is innova tiveness, which is the degree to which an individual is relatively earlier to adopt new ideas than other me mbers of his social system.

Thus, it is plain that innovativeness is a "relative" concept. One has either more or less in- innovativeness than others in a social system. It is essential to specify the social system whose members one is classifying on th e basis of their innovativeness. “ (I: 159f.) 70 This is right when only one innovation is considered. Often, Rogers and his followers computed in- novativeness across several innovations over a number of years. In this case, it is not only the type of social system but also the type of innovations considered that in fluences the measured construct.

Those readers who are not trained in the social sciences will regard innovativeness as an absolute measure of a personal characte r trait. Even though Rogers knows better, he enhances research around innovativeness and devotes much space in his books to this question of correlates of innova- tiveness. That he knows better is shown in his little chapter on „Consistency of Innovativeness.“ (I: 186f.) „ There is no clear-cut evidence, as to whether or not i nnovating behavior is completely con- sistent.“...“There is less evidence, however, that a farm innovator is also an innovator in po- litical ideology, consumer behavio r, or other areas of life. In any event, it is doubtful whether an individual who is an innovator for one idea is a laggard for another idea. “ (I: 187) The last sentence is a statement of opinion, not ba sed on facts. Life experience shows that the oppo- site is also likely. People do not follow the latest fashions in all fields of life even when, for exam- ple, they are very innovative in th eir job. Or they may be risk averse and reluctant to innovate in their job, but very innovative in their hobby. Rogers admits this indirectly, in the subchapter on „ Changes in Adopter Categories over Time “, (I: 189), and he states that it is a fact that also comes out in several panel studies. But then Rogers bri ngs in a pile of potatoes analogy that attempts to demonstrate that innovativene ss is more consistent. “ In fact, the shifting of individuals among adopter categories over time may be likened to a bell-shaped pile of potatoes. The potato pile rest s precipitously near the edge of a table. As the potatoes rearrange their relative positions within the stack over time, an occasional potato is shoved over the edge of the table and out of the pile and distribute themselves throughout the stack. While the pile reta ins its bell shape over a time period, individuals within may be changing positions.” (I: 191). Changes occur because it is a rela tive construct. and so it is clearly not a stable character trait.

Most of his „generalizations“ must be seen as an artifact of method because they are based on a ma- jority of studies reviewed. While a minority of studies do not support them, this majority results from the type of innovations selected, better even to say innovation-situations selected. Most of studies reviewed come from Rural Sociology and deal with farm modernization. Evidence from other research fields is then ecl ectically added, most of the cases again deal with modernization is- sues, to give an impression of the ove rall validity of the generalizations.

By establishing „ideal types“ of diffusion curves and of adopter categories and their characteriza- tion, he mixes up normative and empirical considerati ons leaving the reader with an impression that this all is based on empirical evidence.

Explanation of the Second Part of the Diffusion Curve The most difficult task for Rogers – and he never solved the problem – was to explain why the dif- fusion curve, shaped like a normal distribution, went down in its second part: “The writings of early sociologis ts, learning psychologists, and students of the interaction ef- fect provide theoretical reasons for expecting adopter distributions to be normal. The interac- tion effect is the process through which indivi duals in a social system who have adopted an innovation influence those who have not yet adopted. Adopter distributions follow a bell- shaped curve over time and approach normality. ” (I: 191f.) 71 The writings of early sociologists only indicate a similarity to a bell-shaped curved without offering any explanation. The analogy to learning curves – wh ich in themselves are not widely accepted in learning psychology – is pure speculation. Rogers unconvi ncingly offers the inter-action effect as an explanation: “If the first adopter of the innovat ion discusses it with two other members of the social system, and these two adopters pass the new idea along to tw o peers, the resulting distribution follows a binomial expansion. This mathematical func tion follows a normal shape, when plotted.” (I:

154) “ The interaction effect begins to level off after the second half of the individuals in a social system have adopted because each new adopter finds it increasingly diffi\ cult to tell the new idea to a peer who has not yet adopted.” (I: 155, V: 274) The first statement is wrong because the y = x 2 function gives a J-curve. The second statement may be right, but is irrelevant because while informat ion is a precondition for adoption it is not the only factor of influence. The standard case of hybrid corn is sufficient to prove this empirically: “ Nonadopters are often aware of an innovation but are not motivated to try out and adopt it.

Ryan and Gross (1943) reported that almost all of the Iowa farmers in their study heard about hybrid seed corn before more than a handful were planting it.”(I: 108) “Knowing about an innovation is often quite a diffe rent matter from using the idea. Most individuals know about many innovations which they have not adopted. ” (II: 108) Nevertheless, this “ information trickle down ” explanation is kept alive up to the fifth edition (V:

274).

Pro-innovation Bias In editions two-five Rogers criticizes diffusion research as suffering from a “ pro-innovation bias”.

In later stages he also admits that rejection or discontinuance is sometimes more rational and better for the actor than “unwise” adoption, because i nnovations are not equally good for all potential adopters. “There may be both rational and irrational disc ontinuances just as there are both rational and irrational adoption decisions. ”(I: 91) ” Most past research on the diffusion of innovat ions investigated either rational adopters, or irrational under-adopters, or else compared the two types of individuals. Few studies are available on irrational over- adopters or rational rejectors.”(I: 142) “ Our discussion should not be interpreted to me an that traditional norms are necessarily un- desirable. In many cases, tradi tion may lend stability to a soci al system where it is undergoing rapid change and the danger of disorganization. ”(I: 62) But the following text in the same chapter is so clearly in favor of “ modern” and so disparaging of “ traditional ” that the cited statement is discredited. (V: 282-285). This is also implied in calling the latest adopter category “ laggards”, a terminology maintained through all five editions, instead of using a more neutral term like “ latest adopte