This is not a paper but a question and answer type format.That's all I need is 500-600 words. If possible.In this activity, you will review the Rainey article to evaluate the strengths and weaknesses

JPART21:i321–i345 PUBLICNESS AND ORGANIZATIONAL PERFORMANCE: A SPECIAL ISSUE Sampling Designs for Analyzing Publicness:

Alternatives and Their Strengths and Weaknesses Hal G. Rainey University of Georgia ABSTRACT Research comparing public and private organizations and otherwise analyzing ‘‘publicness’’ involves complex challenges. These include the challenge of designing and attaining adequate samples to represent the two complex categories of ‘‘public’’ and ‘‘private,’’ as well as dimensions of publicness, and subcategories and control variables needed for valid comparisons. This review of sampling alternatives begins with discussion of the strengths and weaknesses of the presumably optimal design, a national probability sample of organizations. Due to the expensive nature of such a design—the discussion concentrates on the one such study ever conducted—the discussion then considers the strengths and weaknesses of the purposive samples and samples of opportunity that most research have used. In spite of limitations in representativeness and in accounting for all variables needed to eliminate alternative interpretations, studies using such samples can be aggregated to support conclusions about differences between public and private organizations that are by now well-founded. Researchers should continue to seek opportunities for the optimal large representative samples. Lacking such opportunities, researchers can contribute usefully to analysis of publicness by carefully designing their studies to make them consistent with previous studies and to support aggregation with previous studies.

Do public organizations and the people in them have characteristics distinct from those in other types of organizations, such as business ﬁrms and private nonproﬁt organizations?

Many organizations mix characteristics traditionally associated with public and private organizations, such as high levels of government funding of a privately owned business ﬁrm and many other mixed or hybridized situations. How does such hybridization affect the organizations? Such matters are fundamental to the existence of a ﬁeld such as public administration and play an important role in many public policy decisions. If there is noth- ing distinctive about a public category of administration and organizations, we can simplify matters by abolishing all academic programs in public administration. If it makes no dif- ference whether an organization operates under more public than private auspices, then we can dispense with decisions about privatization of government services since it makes no difference whether or not we privatize them.

Address correspondence to the author at [email protected].

doi:10.1093/jopart/mur029 ªThe Author 2011. Published by Oxford University Press on behalf of the Journal of Public Administration Research and Theory, Inc. All rights reserved. For permissions, please e-mail: [email protected] These points imply that the need for analysis of publicness is obvious. Many analysts of organizations, however, object to distinctions among categories such as public versus private. The categories lump together in the public, private, and nonproﬁt categories, a vastly diverse array of organizational types within each category. Many of these types within these broad ‘‘sector’’ categories resemble similar types in the other sectors (e.g., public and private schools, public and private electric utilities and transportation activities, and for-proﬁt and nonproﬁt hospitals). Many of the activities of people and organizations in the different sectors are very similar—making decisions, leading and motivating, setting strategy, designing organizations, and work processes. The most prominent of organization theorists have raised these sorts of objections and have emphasized the commonalities among public, private, and nonproﬁt organizations (e.g., Simon 1995; Thompson 1962).

Resolving these conﬂicting perspectives through systematic analyses of publicness involves major challenges in research design. Given the complexities of the sectors and of publicness, how can one design research that elucidates these categories and concepts?

How can we represent such large, diverse categories in research designs. Factors such as size, task or function, and industry characteristics can inﬂuence an organization more than its status as a governmental entity. Research needs to show that these alternative factors do not confuse analysis of differences between public organizations and other types. Obvi- ously, for example, if you compare large public agencies to small private ﬁrms and ﬁnd the agencies more bureaucratic, size may be the real explanation. Also, one would not compare a set of public hospitals to private utilities as a way of assessing the nature of public organizations. Among many research design challenges, this matter of how one designs a sample of organizations or people is one of the very important ones. The sections to follow examine alternative designs for sampling procedures that researchers have used, to consider their pros and cons, to identify possibilities and challenges.

First, the discussion focuses on a rare example of a large national probability sample of organizations—apparently the only such sample of organizations. Although not aimed at analyzing publicness, this study found differences between public and private organiza- tions. The study not only represents some ideal characteristics for a study of publicness but also illustrates many of the challenges that analysis of publicness must confront. These challenges have led researchers to employ various less-than-ideal designs, including those described in table 1. 1Researchers have analyzed responses to national social surveys to compare public and private employees’ perceptions about their work. Others have inter- viewed executives who have experience in the pubic and private sectors or have conducted case studies of individual public organizations and drawn conclusions about their distinc- tiveness. Researchers have studied opportunity or judgmental samples of public and private organizations, rather than random or probability samples, sometimes with small samples and sometimes with large, diverse samples. Others have studied public and private organ- izations in the same functional category, such as universities, schools, or nursing homes, or have compared public and private versions of the same service, such as refuse collection.

Some researchers have analyzed ‘‘publicness’’ as a dimension, rather than a dichotomy between public and private categories. Each of these designs has limitations, but each has also contributed to the analysis of public-private distinctions and of publicness. Each design remains a potential alternative for researchers, who need to consider their strengths 1 For a similar but alternative categorization of designs for analyzing publicness, see Bozeman (1987, 43ff). i322 Journal of Public Administration Research and Theory and weaknesses as does the following discussion. The discussion concludes with observa- tions about ways in which many of these studies have produced similar ﬁndings that support generalizations and ways in which future studies can continue to do so.

As implied already, the discussion will refer to public versus private comparisons and to analyses of publicness interchangeably. There are important distinctions between these two approaches, discussed later, but a comparison of public and private organizations can be considered an analysis of publicness. In addition, sampling designs represent only one of many research design challenges (such as measurement procedures and theories or models to guide the design), but the discussion will concentrate on sampling challenges and closely related matters.

THE IMPOSSIBLE DREAM? A LARGE, NATIONALLY REPRESENTATIVE SAMPLE Ideally, an analysis of publicness requires a convincing sample, with a good model that accounts for other variables besides the public-private dimension. Ideally, studies would also have large, well-designed samples of organizations and employees, representing many functions and controlling for many variables. Such studies require abundant resources and have been rare and for certain purposes nonexistent in organizational or managerial research.

As suggested by the ﬁrst entry in table 1, the National Organizations Study (NOS) is an exception to this generalization. For the NOS, an analysis of organizational structures and human resource practices, the researchers drew a probability sample of organizations de- signed to represent the population of organizations in the United States (Kalleberg et al.

1996, 2001). It is apparently the only study of a national probability sample of organizations ever conducted. The NOS exempliﬁes not only possibilities for a study of a probability sample but also complications that can arise in implementing such a study. NOS was ﬁrst conducted in 1991, and another, similar NOS was conducted in 2002. For purposes of the present discussion, similarities between the two studies make it appropriate to refer to both of them at various points, in spite of differences in the two studies.

Although the discussion to follow emphasizes limitations and challenges to which the NOS points, the study produced important ﬁndings about publicness. The NOS researchers did not intend the study as an examination of publicness. The results of the survey and the analysis, however, indicated that public, nonproﬁt, and private organizations differed sig- niﬁcantly on a measure of formalization and a measure of decentralization. On the measure of formalization, which was actually a measure of formalization of personnel procedures (e.g., the existence of documents on hiring/ﬁring procedures, and documents on personnel evaluation), public organizations were signiﬁcantly higher than business ﬁrms. Public or- ganizations were signiﬁcantly lower than business ﬁrms in the degree to which personnel decisions were decentralized. These results are consistent with results of other studies (e.g., Feeney and Rainey 2009), as discussed later. The NOS has also made possible additional analyses of public, private, and nonproﬁt differences (Frumkin and Galaskiewicz 2004).

Kalleberg et al. (2001) used the sample for the General Social Survey (GSS) as basis for their sample of organizations. The GSS sample is a probability sample of individual respondents in the United States that asks a large number of questions about many aspects of the respondents’ lives, including questions about their work and their attitudes about their work. The NOS researchers contacted the respondents to the GSS, who were randomly selected, to identify the names, addresses, and telephone numbers of the organization in i323 RaineySampling Designs for Analyzing Publicness i323 Table 1 Alternative Designs for Comparisons of Public and Private Organizations Example Sectors and Units Strengths Weaknesses of Analyses 1. Survey of a probability sample of organizations representing the population of organizations in the United States National Organizations Survey (Kalleberg et al. 1996, 2001)Organizations in the United States including public, private, and nonproﬁt organizations.A representative sample of organizations, including the three types.Smallnfor categories within the three main categories. Very general conclusions about the three categories (e.g., public organizations have higher levels of formalization of personnel systems than private organizations). No measures of ‘‘publicness.’’ 2. Social surveys of probability samples of the population of the United States, with analysis of questions about work and the workplace Crewson (1995a, 1995b, 1997)Respondents to the GSS, across 20 years.A representative sample of Americans. Repeated ﬁndings of differences between public and private employees in reward preferences.Smallnfor categories within the public and private categories.

Very general conclusions about public and private respondents.

Very limited information about the characteristics of their employing organizations.

3. Surveys of members of the same profession, in public and private sector jobs Crewson (1995a, 1995b, 1997) and Langbein and Lewis (1998)Respondents to survey of a random sample of members of the Institute for Electrical and Electronic Engineers.Sampling within an occupational category provides some control for occupation and task. The survey provideddata for analysis of demographic and professional characteristics of respondents.Questions about generalizability from one occupational category to others.

4. Interviews or essays by executives experienced in the public and private sectors Blumenthal (1983), Cervantes (1983) and Rumsfeld (1983)Corporate executives with experience as executives in government agencies.Direct observations by experienced individuals. Similar observationsNo clear evidence of validity of executive observations. Limited sample size raises issues of Continued i324 Table 1(continued) Alternative Designs for Comparisons of Public and Private Organizations Example Sectors and Units Strengths Weaknesses of Analyses by different executives support generalizations.representativeness and generalizability. The executive perspective is not necessarily generalizable or applicable to various important topics (e.g., employee perceptions of the incentive system).

5. Case studies of individual government agencies, leading to generalizations about government agencies Warwick (1975) Case study of a failed attempt at organizational change in the US Department of State.In-depth observation and analysis of organizational processes and behaviors.No direct evidence of applicability of ﬁndings to other government organizations. No direct comparisons to private organizations.

6. Studies of large samples of the same functional type of government agency, leading to generalizations about government agencies Meyer (1979) A national sample of state and local public ﬁnance agencies.Focus on one functional type provides some control for function and task. Large sample provides evidence of generalizability within the functional category.No direct evidence of applicability of ﬁndings to government organizations in other functional categories. No direct comparisons to private organizations.

7. Studies of large samples of public and private organizations within the same functional category Holdaway et al.

(1975)Canadian universities with varying degrees of government control.

Chubb and Moe (1990)Respondents to a survey of administrators and teachers in public and privateFocus on one unctional category provides some control for function and task. Large sample provides evidence of generalizability withinNo direct evidence of applicability of ﬁndings to public and private organizations in other functional categories. The nature and effects Continued i325 Table 1(continued) Alternative Designs for Comparisons of Public and Private Organizations Example Sectors and Units Strengths Weaknesses of Analyses schools in the United Statesthe category. Results can support policy and decisions for that categoryof public and private distinctions and variations in publicness may be distinct for different functional categories. Within functional categories, public and private organizations may operate in very distinct ways. They may serve different populations of clients, (e.g., Amirkhanyan et al., 2008).

Tolbert (1985) Universities in the United States with varying degrees of public and private inﬂuence.

Amirkhanyan, et al (2008)Very large sample of public, private, and nonproﬁt nursing home facilities (N514,423).

8. Surveys of respondents from small samples of public and private organizations Rainey (1983), Gabris and Simo (1995), Kurland and Egan (1999), Brewer and Lam (2009), Kurke and Aldrich (1983), and Andersen (2010)Respondents to surveys in small samples of public and private organizations (two to seven organizations per sector).Findings across different types of public and private organizations provide evidence about generalizability of the ﬁndings across those two categories.

Similarity of ﬁndings across these small sample studies, and with other can also support general conclusions.Small opportunity samples leave questions about representativeness and generalizability. Both positive and null ﬁndings could be due to inclusion of particular subtypes of public and private organizations in the study. Some authors are not careful to describe clearly the organizations in the study, and use simple government ownership Continued i326 Table 1(continued) Alternative Designs for Comparisons of Public and Private Organizations Example Sectors and Units Strengths Weaknesses of Analyses or auspices as the deﬁnition of public. This leaves open the possibility, for example, that the public sample includes a government corporation or user fee-charging public organization that reduces differences with the private sector sample.

9. Surveys of respondents from large, diverse samples of public and private organizations Rainey, Pandey, and Bozeman (1995)Respondents to a survey of managers in 196 public, private, and nonproﬁt organizations in the area Syracuse and Albany, New York. The public sample included a very diverse set of state and local government agencies. The private sample included a diverse set of private nonproﬁt and private for-proﬁt organizations, drawn randomly from a directory of such organizations in the area.The larger, more diverse sample enhances sample representativeness.

Similarities between ﬁndings with this sample and ﬁndings in other studies support generalizations about public and private differences. The characteristics of the sample represented well the characteristics of the population.The sample is still not conclusively probabilistic and representative.

It represents one geographic area of the nation and does not represent all levels of government.

Feeney and Rainey (2009)Respondents to a survey of managers in government and nonproﬁt organizations in Illinois and Georgia.Representation of two states, one with recent personnel reforms to increase managerial discretion over personnel and the other with heavy unionization, enhances control for geographic location and stateThe sample is still not conclusively probabilistic and representative.

It represents only two state governments and therefore represents only two geographical Continued i327 Table 1(continued) Alternative Designs for Comparisons of Public and Private Organizations Example Sectors and Units Strengths Weaknesses of Analyses government institutional. Findings of sharp differences between public and nonproﬁt organizations that closely resembled other studies of public vs. private differences support conclusions about the explanatory capacity of the ‘‘public’’ category.and institutional contexts, and only one level of government.

Hickson et al. (1986) Strategic decisions in 30 British organizations,including public and private service organizations and public and private manufacturing organizations.

10. Comparisons of public and private versions of service delivery of the same service Savas (2000) and Hodge (2000)Numerous studies comparing public and private service delivery in a wide variety of types of service.Numerous studies provide a great deal of evidence. Comparisons of the same service provide some control for service, function, and task characteristics.Mixed ﬁndings. Proponents cite numerous studies showing greater cost efﬁciency of private service delivery (e.g., Savas, 2000). Other studies, such as an international meta-analysis, conclude that private sector superiority in efﬁciency occurs only a very few service areas, such as refuse collection (Hodge, 2000). Designing well-controlled comparisons can be very challenging; public and private service delivery may have different goals, serve different clients, or have other differences. Continued i328 Table 1(continued) Alternative Designs for Comparisons of Public and Private Organizations Example Sectors and Units Strengths Weaknesses of Analyses 11. Studies of ‘‘dimensional publicness’’ Bozeman and Bretschneider (1994) and Bozeman (1987)Organizations that vary in ‘‘resource publicness’’ and and publicness of organizational control. A sample of 733 research and development laboratories.Supports conceptualization and analysis of mixed or hybridized public and private status. Supports comparison of ‘‘core’’ public approaches to dimensional approaches.Very limited number of studies. The conception and measures of publicness vary in different studies.

Rainey, Pandey, and Bozeman (1995)Respondents to a survey of 196 managers in public, nonproﬁt, and private for-proﬁt organizations. i329 which the respondent and his or her spouse worked. Since these respondents had been ran- domly sampled, this made their work organizations a probability sample (Spaeth and O’Rourke 1996, 23). 2 The NOS researchers then contacted the work organizations to identify personnel ofﬁ- cials from the organization to act as respondents for their organizations. The researchers conducted telephone interviews with these ofﬁcials, using a questionnaire that asked about the characteristics and practices of the organizations, especially in areas related to person- nel and human resources. The researchers contended that the sampling procedure provided a sample of all types, ages, and sizes of work establishments. Indeed, the sample did include a distribution of organizations by industry that matched well against the distribution of industries and functions in which respondents to the GSS and the Current Population Sur- vey said they worked (Spaeth and O’Rourke 1996, 34–5). The sample also showed diversity and dispersion by organizational size.

The NOS asked respondents to identify whether their work establishment was for- proﬁt or not-for-proﬁt. If the response indicated not-for-proﬁt, the respondent was then asked if the work establishment was public or private. This provides for a distinction be- tween public not-for-proﬁt organizations and private not-for-proﬁt organizations. 3The NOS researchers analyzed differences among public, private, and nonproﬁt organizations by using the for-proﬁt category as the base category, with dummy variables for nonproﬁt (the private not-for-proﬁts) and for public (the public not-for-proﬁts). As described earlier, they found statistically signiﬁcant differences indicating that public organizations had high- er levels of formalization of their personnel procedures and lower levels of decentralization of decision making, as compared with the private organizations. The NOS analysis of such differences leads to a number of observations and conclusions that indicate challenges and issues for studies aiming to analyze publicness. Response Rate Challenges Implementation of the survey involved complex challenges. For the 1991 survey, from an original sample of 1,127 work establishments, the ﬁnal sample of completed and usable responses included 727 or 64.5% of the original total. The attenuation resulted mainly from refusals to participate. 4The telephone interviews required numerous contacts with respond- ents, with median interview involving six contacts and the maximum number of contacts for an individual interview reaching a total of 58 contacts for one interview. The researchers obtained the total of 727 organizations in part by way of a mail survey that they sent to some respondents upon their request, as an alternative to the telephone interview. Refusal rates for this mail survey were quite high. 2 If the respondent worked in a larger organization, that made it more likely that another respondent would work in the same organization, so the probability of selection into the sample of organizations was proportionate to organizational size. The researchers contended that a sample for which probability of selection is proportionate to size of the work establishment is statistically optimal for populations such as organizations that vary widely in size. 3 The 1991 survey also asked the respondent to select the type of organization the not-for-proﬁt represented, from among categories of government agency, hospital, school, college or university, and others. It also asked whether the organization was a federal, state, or local agency. The 2002 NOS did not ask these questions. 4 Interviewees were allowed to request a mailed version of the questionnaire, and 434 such questionnaires were mailed to respondents. Only 29.3% of these respondents returned the mail survey, whereas 44.2% refused. i330 Journal of Public Administration Research and Theory Sample Size and Representativeness Challenges The sample size was ultimately rather small for certain purposes. As noted above, the sample included a diverse array of organizational types. The researchers reported the dis- tribution of establishments by industry (agriculture, durable and nondurable manufactur- ing, transportation, retail trade, professional services, and other categories). This distribution approximated that of other national probability samples such as the GSS and the Current Population Survey, which are not surveys about organizations, but ask about place of employment (Spaeth and O’Rourke 1996, 34). Whereas this provides ev- idence of the general representativeness of the sample, the sample characteristics also show the limited representation of particular organizational types, such as public organizations.

For reasons not fully explained in their text, the NOS researchers’ categorization of estab- lishments by industry includes the category of ‘‘public administration.’’ This category included only 8.1% of the sample (Spaeth and O’Rourke 1996, 34). In the United States, there are fewer public and nonproﬁt organizations and employees than in the private sector, and the public and nonproﬁt categories will make up smaller proportions of a probability sample. 5This obviously imposes limits on the analysis of public organizations. For exam- ple, one might want to analyze differences among federal, state, and local public organ- izations or among types of public organizations such as labor departments, transportation departments, and social welfare departments. One might want to analyze similarities and differences between these types of public organizations and for-proﬁt private ﬁrms. If these categories of public organizations are represented at all, however, their sample size within the larger sample would be too low for meaningful statistical analysis (using controls for size and other variables). In the 1991 NOS, for example, only 17 respondents reported that they worked for federal agencies.

Another issue in sample representativeness concerns the variations within the public and private categories mentioned earlier. In the population of organizations in the United States and other nations, some organizations fall into the categories of public agencies and private ﬁrms without much dispute (Dahl and Lindblom 1953). The US Department of State is a public or governmental agency. Microsoft may have large contracts and sales with governments, but it is a private ﬁrm. Other organizations are hybridized versions of varying levels of publicness (or ‘‘privateness’’) such as government corporations, heavily regulated private ﬁrms (electric utilities), and ﬁrms with very high percentages of their revenues from government contracts.

Sampling all these types of organizations appears challenging enough, but there are still other categories within the public and private categories that complicate the sampling process. Schools, universities, utilities, hospitals, transportation services (trains, buses, and airlines) and other functional types of organizations operate under governmental and pri- vate auspices in various nations, but represent categories distinct from typical governmen- tal agencies. The NOS analysis apparently places in the public category all these types of organizations whose representative responding to the survey says are in the not-for-proﬁt and public categories. That is, public schools go in the public category with government agencies and private schools go in the category with private manufacturing and service ﬁrms. As discussed later, this raises questions about the extent to which these different 5 In the 2002 survey, for example, 112 (21.7%) of the 515 establishments in the sample were public, as opposed to private nonproﬁts (n539, 7.6%) or private for proﬁt organizations (n5360, 70%). RaineySampling Designs for Analyzing Publicness i331 types within the two sectors should be represented in a sample, which in turn raises issues about stratifying the sample, and about increasing sample size and resource costs for the study. An opposite sampling issue concerns representation of organizations that exist in only one sector. Consider taking a random sample of US Departments of Defense, of State and of the Treasury. There is only one of each. Although one could stratify to include public and private schools, hospitals, and other types represented in both sectors, there are no appropriate private counterparts for some government agencies. 6 As noted above, the refusal rate was high, and this creates additional representative- ness issues. Other researchers have found that variation in response rates when surveying respondents in different sectors (e.g., Pandey and Kingsley 2000; Pandey and Welch 2005), across cities (e.g. DeHart-Davis 2007), and across states(e.g. DeHart-Davis and Bozeman 2001). Researchers have also found that representatives from private ﬁrms show a higher likelihood of nonresponse than those from public organizations. Surveys of nonproﬁt or- ganization representatives tend to have lower response rates than the response rates for public organization respondents. Achieving high, or even acceptable response rates, rep- resents a major challenge, as does the likelihood of unbalanced response rates for the sector types (i.e., lower response rates from private organizations than from public organizations).

Unbalanced response rates raise problems of eliminating the possibility of response bias or differential representativeness of the samples. Resource Requirements The high refusal rate corresponds with the level of difﬁculty encountered, and the resources required to generate good response rates. This is reﬂected in the numerous contact attempts by the NOS researchers, the accompanying mail survey, and their other efforts. To better represent organizational types, especially within the public sector, one needs a larger sam- ple. One needs also to consider whether the sample should involve some stratifying or ‘‘matching.’’ That is, the sample could include a sufﬁciently large subsample of organi- zations that operate in both sectors, and with varying levels of publicness, such as hospitals and schools, to analyzed difference due to levels of publicness. This will require a very large sample for the overall study. With the other complications such as refusal rates, this will make the optimal sample very expensive.

Limitations of Surveys The NOS was a survey and surveys have well-known limitations. These include threats of monomethod bias, response biases, limited ﬁdelity or richness of the responses, and other limitations. Alternative methods, however, such as close qualitative observation or in-person interviews, and even analysis of archival data, can require much heavier investments of time and resources. The reliance on survey responses also limits inclusion of various contextual variables that can support analysis of why organizations might differ on certain variables. For example, the NOS data set does not support analysis of why public organizations have more 6 The US Postal Service can also be considered unique to the public sector. Private package delivery ﬁrms make very inappropriate comparison groups for the US Postal Service, for a lot of purposes, due to differences in size, mission and task (daily mail deliveries, and others), unionization, and other characteristics. The US Postal Service and the Department of Defense employ about half of the civilian workforce of the US Government. i332 Journal of Public Administration Research and Theory formalized personnel procedures than private organizations. The researchers simply con- clude that this must be due to stronger institutional inﬂuences on public organizations (Kalleberg et al. 1996, 326). The data set does not make possible a rich analysis of such questions as whether this higher formalization in government organizations reﬂects higher formalization in organizations operating under civil service systems but does not characterize certain types of government organizations (again, such as hospitals, electric utilities, or government authorities or enterprises).

The NOS results offer very general conclusions, such as the ﬁnding mentioned above about public organizations being more highly formalized that private organizations. Such a general conclusion has theoretical and practical value. It obviously lacks many nuances, however, about variations within the ‘‘public’’ category and about such matters as whether public and private organizations similar in functional type (e.g., hospitals) differ. These nuances are not readily analyzable given the limited public sector sample.

In spite of all these considerations about limitations, the ﬁndings of this rare national probability sample of organizations have considerable signiﬁcance for the study of public and private organizations. With such a sample, organizational sociologists with no special interest or concern about public versus private distinctions found signiﬁcant differences.

They found that private, public, and nonproﬁt organizations differed in the degree to which personnel procedures are formalized and decentralized. Without necessarily looking for public versus private distinctions, the researchers found them. In addition, these ﬁndings show consistency with those of other studies, described below, with more limited samples.

These ﬁndings from the only probability sample of organizations contrast sharply with the assertions mentioned earlier that there are no signiﬁcant differences between public and private organizations (e.g., Simon 1995).

Social Surveys of Probability Samples of the Population of the United States, with Analysis of Questions about Work and the Workplace The second entry in table 1 refers to studies that compared public and private samples from census data, large-scale social surveys, or national studies (Crewson 1995a, 1995b; Light 2002; Smith and Nock 1980; US Ofﬁce of Personnel Management 2000). The GSS is sim- ilar to NOS in that it draws a national probability sample, but it differs in that it asks very few questions about the respondents’ work organizations. Others analyze a very large, pur- portedly representative sample of a large population, such as the US Ofﬁce of Personnel Management (OPM) survey of a very large sample of federal employees; OPM then com- pared some of the results to the survey responses of a large sample of private sector employ- ees. These studies produce highly aggregated ﬁndings that prove difﬁcult to relate to the characteristics of speciﬁc organizations and the people in them. Crewson’s (1995a) analysis of the GSS, for example, cannot go into depth about different organizations in the sample because the survey asks questions about work attitudes but is not a study of organizations.

The studies do provide meaningful results, of course. For example, Crewson (1995b) shows that across 20 years, respondents to the GSS, a survey of social attitudes and per- ceptions of a nationally representative sample of Americans, have provided evidence of differences between persons employed in public organizations, as compared to those em- ployed in private organizations. Public sector respondents have consistently placed a lower ranking on ‘‘high income’’ as the most important aspect of their jobs, as compared to private sector respondents who gave a higher ranking to high income. The public sector RaineySampling Designs for Analyzing Publicness i333 respondents also gave consistently higher rankings than private sector respondents to work that is useful to society, work that helps others, and work that provides a feeling of accom- plishment. The differences between the public and private respondents were not always huge, and sometimes they were not statistically signiﬁcant. They were, however, very con- sistent across time. 7These ﬁndings support very general generalizations. Obviously, there are variations among public employees in such responses, by occupation, organizational level, organizational function, and other differences. These variables are not represented in the sample in frequencies high enough to analyze such variations. Nevertheless, these ﬁnd- ings are consistent with other ﬁndings from studies with purposive samples and samples of opportunity (e.g., Rainey 1982). Taken together, then, studies with different sampling de- signs can be aggregated to support generalizations about differences between public and private organizations and the people in them.

Examining the NOS and these other large surveys suggests the characteristics needed in a sample that will support rich analysis of publicness, in many different types of organ- izations. The sample would have to be very large and the study would require very high levels of resources. To represent the populations of public and private organizations, as well as types of organizations that operate in both sectors or that vary in publicness, the sample would have to involve stratifying or matching to represent various types of organizations.

For example, the sample would ideally include large subsamples of such organizations as public and private schools, utilities, transportation organizations, and hospitals. The sample should also include, of course, organizations that may be unique to public or private aus- pices, such as the Departments of Defense and State. Refusal rates will likely be high, a factor that will contribute to higher resource requirements to carry out follow-ups and other ways of increasing participation. Given the challenges that such a large and expensive study will involve, it becomes important to consider alternative designs. The next section does so. THE ART OF THE POSSIBLE: LESS COMPREHENSIVE DESIGNS AND SAMPLES OF OPPORTUNITY The challenges raised by a survey such as NOS suggest why researchers have adopted a variety of less comprehensive approaches such as judgmental samples and samples of opportunity. A review of these alternatives shows that although they often have sharp lim- itations compared to an optimal design, they have strengths as well, and have contributed to progress in analyzing publicness. In addition, given the challenges that a large probability sample imposes, researchers need to assess more feasible alternatives.

Before turning to empirical studies, a review of the publicness literature should note that some authors have theorized about publicness on the basis of assumptions, previous literature and research, and their own experiences (e.g., Dahl and Lindblom 1953; Downs 7 One should also note that the public and private samples are highly aggregated, involving respondents from all organizational levels of many different types of public and private organizations. Other evidence indicates that respondents at higher hierarchical levels and higher salary levels in public organizations tend to be much more likely to rank high income lower, and to rank helping others and doing work that is useful to society higher, than public employees at the lower hierarchical and salary levels. The mixing of these levels in the sample therefore attenuates or lessens the public versus private distinction at the higher levels, making the consistent overall differences all the more meaningful. i334 Journal of Public Administration Research and Theory 1967; Wilson 1989). Similarly, but less systematically, some books about public bureau- cracies have simply provided a list of the differences between public and private based on the authors’ knowledge and experience (Gawthorp 1969; Mainzer 1973). These contribu- tions have provided theoretical and practical guidance for empirical researchers, but this review will concentrate on empirical studies.

Surveys of Members of the Same Profession, in Public and Private Sector Jobs As indicated by the third entry on table 1, researchers have taken the opportunity to com- pare respondents in the same professional specialization, but working in the public or pri- vate sectors. Langbein and Lewis (1998) used survey results from a large random sample of members of the Institute for Electrical and Electronic Engineers, to compare those in gov- ernment jobs to those in private sector jobs on estimates of productivity levels and how they relate to pay levels. The results suggested that the engineers in public sector jobs might be somewhat less productive than those in private sector jobs, but they were paid less than private sector counterparts, even taking into account their comparative productivity levels.

Crewson (1995a, 1995b) used the same survey to compare the public and private sector engineers on their ratings of the importance to them of various job factors or incentives. The public sector electrical engineers were signiﬁcantly higher in their ratings of work that is useful to society and work that is helpful to others. These ﬁndings are consistent with many other studies that have found that public sector respondents place higher importance on altruistic and public service rewards in their work than do private sector respondents.

Chubb and Moe (1990) analyzed responses to a large survey of teachers and admin- istrators in public and private schools in the United States. Among other ﬁndings, their results indicated that respondents from public schools perceived less control over the school’s standards, procedures, and curriculum than did private school respondents.

The public schools, then, appeared more public in the level of external institutional control and intervention. Although public organizations can vary widely in autonomy and ‘‘power’’ (Carpenter 2001: Meier and Bothe 2007), various studies from different nations with dif- ferent samples have found that public organizations tend to show less autonomy and more inﬂuence of external institutions than more private organizations (e.g., Hickson et al. 1986; Kalleberg et al. 1996, 326).

The studies of respondents within the same professional specialization provide some control for occupational, technological, and task contexts, although the controls are not nec- essarily precise (e.g., engineers and teachers might differ among themselves on professional and task dimensions). Such comparisons do provide some protection against specious results due to procedures such as comparing very different professionals or occupations in the public and private sectors (e.g., comparing public sector social workers to private sector salesper- sons). As suggested in table 1, comparisons within professional categories do leave ques- tions about whether results for a professional category generalize to other categories. As described above, however, correspondence between results for comparisons within profes- sions and ﬁndings with types of studies contribute valuably to cumulation of ﬁndings.

Interviews or Essays by Executives Experienced in the Public and Private Sectors Executives and managers who have served in both public agencies and private business ﬁrms usually emphasize important differences between the two settings, although they also RaineySampling Designs for Analyzing Publicness i335 emphasize general similarities in leadership roles in the different sectors (Blumenthal 1983; Cervantes 1983; Hunt 1999; IBM Endowment for the Business of Government 2002; Rumsfeld 1983; Weiss 1983). These testimonials offer a richness of descriptive detail and explanation, as opposed to responses to highly structured questionnaire items. These interviews or essays frequently advance very similar observations about the government setting. They emphasize administrative constraints such as civil service systems, the stron- ger role of external political institutions and processes than in the private sector, the greater difﬁculty in assessing performance due to complex goals and the absence of such indicators as proﬁts and sales, and other differences from the private business setting. These simi- larities among observations by different executives, and their consistency with evidence from other types of studies, contribute to support for general conclusions about the impli- cations of publicness. As forms of evidence in themselves, however, they have the limi- tations described in table 1. They apply primarily to the executive and managerial levels, and differences might fade at lower levels. As one gains the richness and nuance of more qualitative approaches such as these, sample size attenuates. There remain questions about whether the executive’s observations might reﬂect a particular governmental setting such as a particular type of agency—a highly politicized agency, for example, as opposed to a rel- atively autonomous agency.

Case Studies of Individual Government Agencies, Leading to Generalizations about Government Agencies Occasionally scholars have conducted research projects that measure or observe only pub- lic bureaucracies, but then draw conclusions about their differences from private organ- izations. Some of these studies have concentrated on one agency. For example, Warwick (1975) conducted a case study of a failed attempt at organizational change in the US Department of State. He drew general conclusions about why governmental bureau- cracies tend toward high levels of bureaucratization and administrative constraint and of external political inﬂuence and internal resistance to change. Where these observations coincide with the conclusions of other studies of publicness, they can add to support for generalizations across studies. As table 1 suggests, such analyses include no evidence of the characteristics of private sector organizations, however, and leave obvious questions about validation of their conclusions.

Studies of Large Samples of the Same Functional Type of Government Agency, Leading to Generalizations about Government Agencies Researchers have also studied a large sample of government organizations from the same functional category and drawn conclusions about the characteristics of government organ- izations. Meyer (1979) analyzed a sample of 240 state and local government ﬁnancial administration agencies. He found evidence that the agencies were more subject to change than conventional stereotypes might suggest. In addition, such variables as organizational leadership and effective claims to domain mediated the environmental inﬂuences on the agencies such that the inﬂuences led to change in some but not others. Meyer ultimately drew conclusions about the nature of government organizations. Among other conclusions, he argued (190–221) that government agencies usually have no market or quasi- market alternatives. This results in vague goals and limits government agencies’ choices i336 Journal of Public Administration Research and Theory of organizational forms to that of Weberian hierarchy, with emphasis on conformity to rules and procedures, leading to ‘‘rule boundedness.’’ These ‘‘conformity criteria’’ displace efﬁciency criteria with increased external ‘‘societal determination’’ by national bureau- cratic institutions. 8As table 1 suggests, studies of this type obviously provide no direct private sector comparison sample. They also leave questions as to whether their ﬁndings apply to public organizations in other functional categories. For example, might public ﬁnance agencies, due to the nature of their tasks and particularly high accountability re- quirements, tend more toward bureaucratic hierarchy and conformity criteria than most other government agencies? Studies of Large Samples of Public and Private Organizations Within the Same Functional Category As the seventh entry of table 1 indicates, researchers have compared large samples of public and private organizations within the same functional category. This avoids the problem of having no representation of private organizations within the sample. As the table indicates, these studies have provided evidence about distinctions between public and private hos- pitals (Savas 2000, 190), schools (Chubb and Moe 1990), nursing homes (Amirkhanyan et al. 2008), airlines (Backx et al. 2002), research and development (R&D) activities (Bozeman and Bretschneider 1994), and electric utilities (Atkinson and Halversen 1986). 9These studies have found differences that show that the public-private distinction appears meaningful even when the same general functional types of organizations operate under both auspices. As the table also emphasizes, however, the studies do not provide evidence about whether their conclusions generalize to public and private organizations in other functional categories. Although space limits preclude a full review of the variations among these studies, the variations indicate that the nature of the public-private distinctions and of variations in publicness can vary across functional categories, due to variations in the inﬂuences of task and technology, different market environments, or variations in public policy domain or in clients and customers. Surveys or Observations of Respondents from Small Samples of Public and Private Organizations Other researchers compare small sets of public and private organizations, usually comparing survey responses from managers or employees. There have been many of these studies, and although they focus on a diverse array of dependent variables, some of them have reported sim- ilar ﬁndings on the same variables that contribute to cumulation of evidence. The studies have compared respondents from public and private organizations on organizational commitment, work satisfaction, perceptions of organizational structure and reward systems, leadership behaviors, conﬂict resolution, and other variables (e.g., Andersen 2010; Brewer and Lam 8 These conclusions resemble those of some other analysts of government bureaucracy, but not others such as, Goodsell (2004), Kelman (2005), and Rainey and Steinbauer (1999). 9 Similarly, other studies compare a function, such as management of computers or the innovativeness of information technology in government and business organizations (Bretschneider 1990; Bretschneider and Wittmer 1993; Moon and Bretschneider 2002). Still others compare state-owned enterprises to private ﬁrms (Hickson et al. 1986; MacAvoy and McIssac 1989; Mascarenhas 1989). RaineySampling Designs for Analyzing Publicness i337 2009; Buchanan 1974, 1975; Kurke and Aldrich 1983; Kurland and Egan 1999; Mintzberg 1972; Porter and Lawler 1968; Rainey 1979, 1983). The studies obviously include different organizations from the public and private sector to try to represent in the sample the diversity of the sectors or at least some of the diversity. Especially for very small samples—some studies compare to public organizations to two private one (e.g., Gabris and Simo 1995)—questions remain about how well the small samples represent the full populations. They really cannot account well for the variations within the two sectors in organizational task, size, age, and many other variables.

Researchers interested in focusing on variations in publicness as a dimension have pointed out that these small sample comparisons have often dichotomized public and pri- vate organizations simply on the basis of ownership (Bozeman and Bretschneider 1994).

This omits attention to variations in the degree of publicness, represented by differences in the level of government funding and government control. Actually, some of these studies do not dichotomize only on the basis of public versus private ownership or auspices but also on the basis of public versus private sources of resources or funding (e.g., Rainey 1983).

This means that the study compares the organizations that Dahl and Lindblom (1953) re- ferred to as ‘‘agencies’’ and ‘‘enterprises’’ and that Wamsley and Zald (1973) classiﬁed as publicly owned and funded versus those that are privately owned and funded. These dichotomies at least avoid including hybridized organizations such as government corpo- rations into the sample. They do not offer a lot of information about variations in public- ness, but they do compare ‘‘core’’ public and private organizations that represent poles at the end of the dimensions of publicness. Still, this set of studies tends to involve dichot- omous comparisons of public and private organizations based on hypotheses drawing on theoretical arguments and previous empirical evidence that they should differ. They do not really test directly the assertions about why public and private organizations should differ and do not directly measure and assess the factors and causal dynamics that produce the differences. This raises a challenge for future research, to which the discussion turns in the conclusion section below.

Some subsests of these studies have reported similar ﬁndings about public versus pri- vate differences in managers’ perceptions of organizational structure and reward systems, about public organizational settings imposing more administrative constraints, about reward systems in the public sector making it more difﬁcult to ﬁre poor performers and reward good performers with higher pay, and about higher levels of public service motivation, altruistic and social service values, and lower emphasis on the importance of pay as a work reward. The studies vary so much in focal dependent variables, however, that they do not support high levels of cumulation of ﬁndings except in the several areas just mentioned, and do not support meta-analytic studies.

Surveys of Respondents from Large, Diverse Samples of Public and Private Organizations As the ninth entry on table 1 indicates, some researchers have drawn large, diverse samples of different types of public and private organizations. One of the subsamples in the two studies listed on the table were drawn using a random sampling method. These samples provide more assurance against possible biases due to very small samples that misrepresent the populations. These studies, however, still leave questions about representing the full populations. For example, the Rainey, Pandey, and Bozeman (1995) sample consisted of i338 Journal of Public Administration Research and Theory organizations from areas in New York State, so the sample is not veriﬁably representative of the entire nation. These studies of larger samples share many of the characteristics of the smaller samples described in the section above, such as the variables on which they focus and their tendency to cumulate on ﬁndings similar to those of the small sample studies.

Often, they have added more convincing evidence of distinctive aspects of public manage- ment. For example, since there are public and private sector manufacturing and service ﬁrms in England, Hickson et al. (1986) were able to compare strategic decisions in the private manufacturing ﬁrms to those in the public manufacturing ﬁrms and in the private service organizations to those in the public service organizations. Among other ﬁndings, their results indicated that in both types of public organizations, strategic decision processes tended toward a more ‘‘vortex sporadic’’ character than in the private organizations. That is, the public sector decision processes were more turbulent, with more external intervention, than the processes in the private organizations, and thus more akin to ‘‘garbage can’’ pro- cesses as depicted by Cohen et al. (1972). 10 Thus, these studies provide more opportunities for testing hypotheses about why pub- lic and private organizations differ or resemble each other and the nature of the differences and the inﬂuences on them. There remains the same challenge mentioned above for smaller samples, of providing more direct tests of the theoretical postulates about the inﬂuences that lead to public versus private differences. Comparisons of Public and Private Versions of Service Delivery of the Same Service Still another source of evidence about publicness in the form of public versus private com- parisons comes from analyses of public versus private delivery of a particular service.

These studies are often distinct from those described in seventh entry in table 1 because they focus on outcomes of public or private provision of a service, usually with emphasis on the cost efﬁciency of service provision, but often with attention to the quality of the service as well. With that focus, these studies often pay little or no attention to organizational char- acteristics and managerial dynamics. Most of these studies are oriented to the policy de- bates over whether a particular service should be provided by the public or private sector and over the privatization of government services. The wide array of studies have led to conﬂicting conclusions. Savas (2000), for example, reviews numerous studies of public and private provision of numerous public services and concludes that private providers usually deliver services more efﬁciently with equivalent quality. The preponderance of studies do appear to support this conclusion. Hodge (2000), on the other hand, reports a meta-analysis of research in multiple nations and concludes that higher levels of efﬁciency due to private provision are restricted to only a few service areas, such as refuse collection and building maintenance. Other experts have added expressions of skepticism about the claims that private sector service provision is superior (e.g., Sclar 2000). Studies of one functional type, however, may not apply to other functional types. The public-private distinction apparently has different implications in one industry or market environment, such as 10 As another example, Pandey and Scott (2002) used the data set that Rainey et al. (1995, see table 1), with additional data from administration of the same survey in Florida and Colorado, were able to show that red tape and formalization of rules in organizations are distinct concepts in both public and private (including private nonproﬁts) organizations.

Such a ﬁnding for a smaller, less diverse sample would have raised more doubts about its representativeness and validity. RaineySampling Designs for Analyzing Publicness i339 hospitals, compared to another industry or market, such as refuse collection (Hodge 2000).

Yet another complication is that public and private organizations within a functional cat- egory may not actually do the same thing or operate in the same way (Kelman 1985). For example, private and public hospitals may serve different patients and operate under dif- ferent rules and policies, and public and private electric utilities may have different funding patterns. For example, Kelman (1985) pointed out that organizations managing government hospitals operate under different rules for hospital construction contracts than private sector hospital management ﬁrms. The public hospitals also face uniform rules and policies about the size of hospital rooms and other details that inﬂuence costs of construction.

Studies of ‘‘Dimensional Publicness’’ Following upon Bozeman’s (1987) analysis of publicness as a continuum along two dimen- sions, some researchers have operationalized the dimensions and examined their relations to organizational characteristics. Bozeman suggested that rather than treating public versus private as a dichotomy, publicness should be conceived as varying along a dimension of ﬁnancial or resource publicness and another dimension of publicness based on political authority. When people in an organization have more authority over the ﬁnancial resources of the organization, the organization has less ﬁnancial or resource publicness. Drawing on property rights theory from economics, Bozeman argued that the control of the ﬁnancial resources of an organization has very signiﬁcant effects on managerial incentives and de- cisions in the organization that in turn inﬂuence other characteristics of the organization. In very public organizations where people in the organization have little control of this sort, these effects on managerial incentives and decisions attenuate. Organizations also vary in their level of political authority. Organizations receive varying levels of authorization to act on behalf of government or as agent of the government. Government agencies have more of this authorization than do most private business ﬁrms, but ﬁrms with high amounts of fund- ing from government contracts in a sense are acting as agents for the government, as are ﬁrms operating under high levels of government regulations.

Bozeman and Bretschneider (1994) put these ideas to work in an analysis of survey results for 733 R&D laboratories and the relations between publicness and the laboratories’ outputs and bureaucratic characteristics (forms of ‘‘red tape’’). Resource publicness was measured by questions asking about the percentage of a laboratory’s budget provided by government grants, appropriations, and equipment purchases. The researchers measured publicness of the laboratories’ goals and agendas with questions about the importance of government funding to the laboratory’s existence and research focus and measured ‘‘communications publicness’’ with questions about the percentage of phone calls and mail that were exchanged with government personnel. To be able to compare publicness results to results using the more frequently studied public versus private comparisons, they also measured ‘‘core publicness’’ by whether the laboratory was located in government, indus- try, or a university.

Concerning laboratory outputs, the resource publicness variables related signiﬁcantly to the proportion of time spent on producing journal articles (a public domain output) and on obtaining patents (a more proprietary output). Higher levels of resource publicness were associated with more concentration on producing articles, the public domain outputs. In the core publicness comparison, status as a government or university laboratory was also sig- niﬁcantly and very strongly related to concentration on journal articles. To analyze i340 Journal of Public Administration Research and Theory bureaucratic characteristics, the survey asked for estimates of the time required to hire a new person, ﬁre a person, purchase equipment, obtain permission to submit a paper for publication, and for research to be circulated outside the organization. Resource pub- licness in the form of higher levels of government appropriations related signiﬁcantly to the latter two forms of red tape—time required to submit a paper for publication and to circulate research results. Core publicness on the other hand—status as a government laboratory—related very strongly to the time required for hiring and ﬁring. Thus, the results indicated that both dimensional publicness and core publicness relate to important char- acteristics of the laboratories but to different characteristics.

Rainey et al. (1995, see ninth entry on table 1) also employed dimensional and core publicness variables in analyzing organizational red tape, using survey results for repre- sentatives of 196 public, private, and nonproﬁt organizations. The survey included ques- tions about resource publicness that asked about the percentage of the organization’s budget that came from government contracts and grants. ‘‘Inﬂuence publicness’’ questions asked about the extent to which external government organizations and ofﬁcials exerted strong control and inﬂuence on the organization. The organizations were also compared by ‘‘sector’’ or status as a government, private, or private nonproﬁt organization. Resource publicness related signiﬁcantly to higher levels of general red tape—a general question about the level of red tape in the organization. Inﬂuence publicness related signiﬁcantly to survey responses indicating that the respondent perceived personnel rules to constrain the relationship between performance and rewards. Sector related signiﬁcantly to both of these variables—to general red tape and to personnel rule constraints—but was also the only independent variable signiﬁcantly related to questions about how long it takes to hire and ﬁre, and this relationship was very strong. These results, together with those of Bozeman and Bretschneider (1994), are consistent with the conclusion that both dimen- sional publicness and sector (core publicness) can relate to organizational outputs, but in different ways. The sector comparisons appear to reﬂect such distinctions as public organizations falling under civil service systems and the legal environment of public em- ployment law. The dimensional publicness results, and especially those for resource pub- licness, appear to capture the governmental accountability requirements that increase with the level of governmental contracts and grants, and the inﬂuence of government funding on whether organizational outputs are more in the public domain as opposed to proprietary.

The sampling implications of analyses of dimensional publicness include the need for large samples of organizations that vary in levels ofgovernment funding and government inﬂuence on the organizations. The studies to date indicate that the resource or ﬁnancial publicness dimension can be measured by simply asking respondents about levels of government funding through various sources, such as grants, contracts, and appropriations. The other dimension, that is supposed to represent variations in the degree of political authority, has varied in their deﬁnition and measurement in different studies and need further conceptual development.

CONCLUSIONS The sampling designs reviewed here raise challenges for researchers, but the body of research offers as much cumulation on certain ﬁndings, using different sampling approaches, as one can ﬁnd for any topic in the social and administrative sciences. As noted, the studies show evidence that public or governmental organizations show higher levels of external institutional inﬂuences than do business organizations. Government organizations RaineySampling Designs for Analyzing Publicness i341 show higher levels of certain types of organizational structuring and rule constraints, es- pecially in personnel administration. People in government express certain differences in work-related attitudes and values than people in business ﬁrms, such as higher levels of altruistic and public service values and lower valuation of high income as an ultimate reward for work. These and other ﬁndings cumulating in the research need further conﬁr- mation and reﬁnements, of course.

In pursuit of further evidence, the most desirable sample is a large, representative one, including many different organizational types and settings. This will involve stratifying the sam- ple to represent such types, including types of organizations that operate in both the public and private sectors and that vary in publicness. Yet this review has discussed the challenges that such a design imposes, including high resource requirements, the likelihood of difﬁculties in attaining high response rates, and others. In spite of the limitations of more opportunistic designs reviewed here, researchers have continued to publish them in large part because they have produced signiﬁcant ﬁndings and meaningful results. Given the success of these less expensive and more feasible studies, researchers need to continue to consider them.

In so doing researchers need also to consider the limitations and weaknesses identiﬁed here and to work to address them. Now that a body of research has built up, researchers should concentrate on aligning additional studies with previous research, to the extent possible. Es- pecially for smaller samples, researchers need to be clearer about the nature of organizations in the sample, in terms of their functions and tasks, of government funding and control, as well as other control variables. Researchers need to more carefully review previous studies to consider how their concepts, variables, and measures, and the organizations in the sample, can resemble those of other studies in ways that provide for cumulation of ﬁndings. Many of the studies have been presented as if they are exploratory. The literature reviews tend to say, basically, that there are reasons to argue that government organizations or employees should differ from those of private sector employees, so let’s see if they do. Future research needs to develop clearer theoretical rationales for why and how publicness, in the sense of core pub- licness and dimensional publicness, makes a difference. Elements of those rationales need to be better represented in the research designs. The small set of studies of dimensional pub- licness provide examples of one approach to such elaborated analysis. They show the fea- sibility that representation of such factors as variations in government funding and government control or inﬂuence can be included in research in ways that lead to meaningful ﬁndings and elaborated explanations of those ﬁndings.

REFERENCES Amirkhanyan, Anna A., Hyun J. Kim, and Kristina T. Lambright. 2008. Does the public sector outperform the nonproﬁt and for-proﬁt sectors? Evidence from a national panel study on nursing home quality and access.Journal of Policy Analysis and Management27:326–53.

Andersen, Jon A. 2010. Public versus private managers: How public and private managers differ in leadership behavior.Public Administration Review70:131–41.

Atkinson, Scott E., and Robert Halversen. 1986. The relative efﬁciency of public and private ﬁrms in a regulated environment: The case of U.S. electric utilities.Journal of Public Economics2:281–94.

Backx, Mattijis, Michael Carney, and Eric Gedajlovic. 2002. Public, private, and mixed ownership and the performance of international airlines.Journal of Air Transport Management8:213–20.

Blumenthal, W. Michael. 1983. Candid reﬂections of a businessman in Washington. InPublic management, eds. J. L. Perry and K. L. Kraemer, 22–33. Palo Alto, CA: Mayﬁeld.

Bozeman, Barry. 1987.All organizations are public: Bridging public and private organizational theories.

San Francisco, CA: Jossey-Bass. i342 Journal of Public Administration Research and Theory Bozeman, Barry, and Stuart Bretschneider. 1994. The ‘publicness puzzle’ in organization theory: A test of alternative explanations of differences between public and private organizations.Journal of Public Administration Research and Theory4:197–223.

Bretschneider, Stuart. 1990. Management information systems in public and private organizations: An empirical test.Public Administration Review50:536–45.

Bretschneider, Stuart, and Dennis, Wittmer. 1993. Organizational adoption of microcomputer technology:

The role of sector.Information Systems Research4:88–108.

Brewer, Brian, and Gilbert K. Y. Lam. 2009. Conﬂict handling preferences: A public-private comparison.

Public Personnel Management38:1–14.

Buchanan, Bruce. 1974. Government managers, business executives, and organizational commitment.

Public Administration Review35:339–47.

———. 1975. Red tape and the service ethic: Some unexpected differences between public and private managers.Administration and Society6:423–8.

Carpenter, Daniel P. 2001.The forging of bureaucratic autonomy: Reputations, networks, and policy innovation in executive agencies, 1862–1928. Princeton, NJ: Princeton Univ. Press.

Cervantes, Alfonso J. 1983. Memoirs of a businessman-mayor. InPublic management, eds. James L. Perry and Kenneth L. Kraemer, 350–352. Palo Alto, CA: Mayﬁeld.

Chubb, John E., and Terry M. Moe. 1990.Politics, markets, and America’s schools. Washington, DC:

Brookings Institution.

Cohen, Michael D., James G. March, and Johan P. Olsen. 1972. A garbage can model of organizational choice.Administrative Science Quarterly17:1–25.

Crewson, Phillip E. 1995a. A comparative analysis of public and private sector entrant quality.American Journal of Political Science39:628–39.

———. 1995b.The public service ethic. PhD diss., American University, Washington, DC.

———. 1997. Public-service motivation: Building empirical evidence of incidence and effect.Journal of Public Administration Research and Theory7:499–518.

Dahl, Robert A., and Charles E. Lindblom. 1953.Politics, economics, and welfare. New York: Har- perCollins.

DeHart-Davis, Leisha. 2007. The unbureaucratic personality.Public Administration Review67:892–903.

DeHart-Davis, Leisha, and Barry, Bozeman. 2001. Regulatory compliance and air quality permitting: why do ﬁrms ’overcomply?Journal of Public Administration Research and Theory11:471–508.

Downs, Anthony. 1967.Inside bureaucracy. New York: Little, Brown.

Feeney, Mary K., and Hal G. Rainey. 2009. Personnel ﬂexibility and red tape in public and nonproﬁt organizations: Distinctions due to institutional and political accountability.Journal of Public Administration Research and Theory20:801–26.

Frumkin, Peter, and Joseph Galaskiewicz. 2004. Institutional isomorphism and public sector organiza- tions.Journal of Public Administration Research and Theory14:283–307.

Gabris, Gerald T., and Gloria Simo. 1995. Public sector motivation as an independent variable affecting career decisions.Public Personnel Management24:33–51.

Gawthorp, Louis C. 1969.Bureaucratic behavior in the executive branch. New York: Free Press.

Goodsell, Charles T. 2004.The case for bureaucracy. Washington, DC: CQ Press.

Hickson, David J, Richard J. Butler, David Cray, Geoffrey R. Mallory, and David C. Wilson. 1986.Top decisions: Strategic decision making in organizations. San Francisco, CA: Jossey-Bass.

Hodge, Graeme A. 2000.Privatization: An international review of performance. Boulder, CO: Westview Press.

Holdaway, Edward A., John F. Newberry, David J. Hickson, and R. Peter Heron. 1975. Dimensions of organizations in complex societies: The educational sector.Administrative Science Quarterly 20:37–58.

Hunt, Albert R. 1999. Reﬂections of a heavyweight.Wall Street Journal,July 1, 1999, A23.

IBM Endowment for the Business of Government. 2002. A conversation with Charles Rossotti, Com- missioner, Internal Revenue Service.The Business of GovernmentWinter:13–17.

Kalleberg, Arne L., David Knoke, and Peter V. Marsden. 2001.The National Organizations Survey (NOS) 1996–1997. Ann Arbor, MI: Inter-Univ. Consortium for Political and Social Research (ICPSR3190). RaineySampling Designs for Analyzing Publicness i343 Kalleberg, Arne L., David Knoke, Peter V. Marsden, and Joe L. Spaeth. (eds). 1996.Organizations in America: Analyzing their structures and human resource practices. Thousand Oaks, CA: Sage.

Kelman, Steven. 1985. The Grace Commission: How much waste in government?Public Interest 78:62–82.

———. 2005.Unleashing change: a study of organizational renewal in government. Washington, DC:

Brookings Institution Press.

Kurke, Lance B., and Howard E. Aldrich. 1983. Mintzberg was right! A replication and extension of the nature of managerial work.Management Science29:975–84.

Kurland, Nancy B., and Terri D. Egan. 1999. Public vs. private perceptions of formalization, outcomes, and justice.Journal of Public Administration Research and Theory9:437–58.

Langbein, Laura I., and Gregory B. Lewis. 1998. Pay, productivity, and the public sector: The case of electrical engineers.Journal of Public Administration Research and Theory8:391–412.

Light, Paul C. 2002. The content of their character: The state of the nonproﬁt workforce.Nonproﬁt QuarterlyFall http://www.brookings.edu/articles/2002/1002governance_light.aspx (accessed June 6, 2011).

MacAvoy, Paul W., and George S. McIsaac. 1989. The performance and management of United States federal government enterprises. InPrivatization and state-owned enterprises, eds. Paul W. Mac- Avoy, W. T. Stanbury, George Yarrow, and Richard J. Zeckhauser, 77–135. Boston, MA: Kluwer.

Mainzer, Lewis C. 1973.Political bureaucracy. Glenview, IL: Scott, Foresman.

Mascarenhas, Briance. 1989. Domains of state-owned, privately held, and publicly traded ﬁrms in in- ternational competition.Administrative Science Quarterly34:582–97.

Meier, Kenneth J., and John Bohte. 2007.Politics and the bureaucracy, 5th ed. Belmont, CA: Thomson- Wadsworth.

Meyer, Marshall W. 1979.Change in public bureaucracies. Cambridge: Cambridge Univ. Press.

Mintzberg, Henry. 1972.The nature of managerial work. New York: HarperCollins.

Moon, M Jae., and Stuart Bretschneider. 2002. Does the perception of red tape constrain IT innovativeness in organizations? Unexpected results from a simultaneous equation model and implications.Journal of Public Administration Research and Theory12:273–92.

Pandey, Sanjay K., and Gordon A. Kingsley. 2000. Examining red tape in public and private organizations:

Alternative explanations from a social psychological model.Journal of Public Administration Research and Theory10:779–800.

Pandey, Sanjay K., and Patrick G. Scott. 2002. Red Tape: A review and assessment of concepts and measures.Journal of Public Administration Research and Theory12:553–80.

Pandey, Sanjay K., and Eric W. Welch. 2005. Beyond stereotypes: A multistage model of managerial perceptions of red tape.Administration and Society37:542–75.

Porter, Lyman W., and Edward E. Lawler, III. 1968.Managerial attitudes and performance. Burr Ridge, IL: Irwin.

Rainey, Hal G. 1979. Perceptions of incentives in business and government: Implications for civil service reform.Public Administration Review39:440–8.

———. 1982. Reward preferences among public and private managers: In search of the service ethic.

American Review of Public Administration16:288–302.

———. 1983. Public agencies and private ﬁrms: Incentive structures, goals, and individual roles.

Administration and Society15:207–42.

Rainey, Hal G., Sanjay, Pandey, and Barry Bozeman. 1995. Public and private managers’ perceptions of red tape.Public Administration Review55:567–74.

Rainey, Hal G., and Paula J. Steinbauer. 1999. Galloping elephants: Developing elements of a theory of effective public organizations.Journal of Public Administration Research and Theory9:1–32.

Rumsfeld, Donald. 1983. A politician-turned-executive surveys both worlds. InPublic management, eds.

James L. Perry and Kenneth L. Kraemer, 34–40. Palo Alto, CA: Mayﬁeld.

Savas, Emmanuel S. 2000.Privatization and public-private partnerships. New York: Seven Bridges Press/ Chatham House.

Sclar, Elliott D. 2000.You don’t always get what you pay for: The economics of privatization. Ithaca, NY:

Cornell Univ. Press. i344 Journal of Public Administration Research and Theory Simon, Herbert A. 1995. Organizations and markets.Journal of Public Administration Research and Theory5:273–94.

Smith, Michael P., and Steven L. Nock. 1980. Social class and the quality of life in public and private organizations.Journal of Social Issues36:59–75.

Spaeth, Joe L., and Diane P. O’Rourke. 1996. Designing and implementing the National Organizations Study. InOrganizations in America: Analyzing their structures and human resource practices, eds.

Arne L. Kalleberg, David Knoke, Peter V. Marsden, and Joe L. Spaeth, 23–44. Thousand Oaks, CA:

Sage.

Thompson, James D. 1962. Common and uncommon elements in administration.Social Welfare Forum 89:181–201.

Tolbert, Pamela S. 1985. Resource dependence and institutional environments: Sources of administrative structure in institutions of higher education.Administrative Science Quarterly30:1–13.

US Ofﬁce of Personnel Management. 2000.Survey of federal employees. Washington, DC: US Ofﬁce of Personnel Management.

Wamsley, Gary L., and Mayer N. Zald. 1973.The political economy of public organizations. Lexington, MA: Heath.

Warwick, Donald P. 1975.A theory of public bureaucracy. Cambridge, MA: Harvard University Press.

Weiss, Herman L. 1983. Why business and government exchange executives. InPublic management, eds.

James L. Perry and Kenneth L. Kraemer. Palo Alto, CA: Mayﬁeld.

Wilson, James Q. 1989.Bureaucracy. New York: Basic Books. RaineySampling Designs for Analyzing Publicness i345 Copyright of Journal of Public Administration Research & Theory is the property of Oxford University Press / UK and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.