Masters in Forensic Science Thesis and PowerPointEyewitness Testimony: How Accurate is it?

MISTAKEN IDENTITY IN EYEWITNESS IDENTIFICATIONS A RESEARCH PROJECT SUBMITTED TO NATIONAL UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF FORENSIC SCIENCE JULY 2004 By Suzanna L. Smedley Masters Thesis Committee:

Wayne Moorehead, Chairperson Marcy Worthington, Supervisor Thomas Green, Ph.D., Supervisor MASTERS THESIS APPROVAL FORM We certify that we have read the project of Suzanna L. Smedley entitled MISTAKEN IDENTITY IN EYEWITNESS IDENTIFICATIONS and that, in our opinion; it is satisfactory in scope and quality as a thesis for the degree of Master of Forensic Sciences at National University.

Approved by:

Wayne Moorehead, Chairperson Date: _____________ Marcy Worthington, Supervisor Date: _____________ Thomas Green, Ph.D., Supervisor Date: _____________ Dr. Ismail Sebatan, Director of Forensic Sciences Program Date: _____________ ACKNOWLEDGEMENTS I would like to personally thank Dr. Ismail Sebatan, Dr. Gary Hoban, Dr. Charles Tatum, and the entire committee for assisting me in using National University graduate classes to collect my data. In addition, I would like the National University teachers and students who participated in this study, as well as the Teachers and Administrators at Wing Lane Elementary School in Valinda, California.

I would also like to thank Dr. Raj Singh, Director of Graduate Studies at Hope International University. I appreciate your valuable input on this project and for continuing to be a mentor to me, even beyond my studies at HIU. I would like to thank Patricia Nowosinski from Priceless Moments Video for filming the mock crime, as well as the “actors” in the video. Thank you to Tay Nicholai (the Suspect), Jack Nowosinski (the Witness), Kristen Weir (the Victim), Heather Nicholai, Kyler Nicholai, and Sherry Blankenship. I appreciate you all spending a Sunday afternoon in the park to create this crime video.

Finally, thank you to each of the men who participated in the suspect lineup: Jack Nowosinski (Suspect #1), Jeff Field (Suspect #2), John Cooper (Suspect #3), Ken Weir (Suspect #4), Terry Bass (Suspect #5), and Tom Cochran (Suspect #6).

ABSTRACT The intention of this project was to determine what factors play a role in mistaken identity. This project was inspired by the exoneration of 138 men by the Innocence Project, who were wrongfully convicted of crimes they did not commit. Mistaken identity played a major role in those convictions in two-thirds of those 138 cases. A mock crime, depicting a suspect stealing a woman’s purse, was videotaped and shown to five participant groups. Participants were asked to complete a questionnaire and the accuracy of the results was studied. Participants were also shown a photo lineup, two days after they witnessed the crime video, and were asked to select the suspect. The lineup results were analyzed to determine if participants would mistakenly select a witness, whose photograph was included in the lineup, and if participants would be compelled to select a suspect, even though they are not confident in their selection. Table of Contents ACKNOWLEDGEMENTS………………………………………………………….…..iii ABSTRACT..…………………………………………………………………………….iv Table of Contents…………………………………………………………….………..…v List of Tables…………………………………………………………………………..…vi Chapter 1: Introduction……………………………………………………………………1 Chapter 2: Literature Review……………………………………………………….……10 Chapter 3: Methodology…………………………………………………………….…...33 Chapter 4: Results………………………………………………………………………..36 Chapter 5: Discussion, Conclusions, and Recommendations……………………………48 Appendix A………………………………………………………………………………59 References……………………………………………………………………………….60 List of Tables Table 2-1: Means of Number of Descriptors, Length and Precision of Person Descriptions as a Function of Situational Factors for Victim and Bystanders…………..21 Table 4-1: Questionnaire Results of Crime Detail Summary………………………...…37 Table 4-2: Questionnaire Results of Suspect’s Spit mentioned…………………………38 Table 4-3: Questionnaire Results of Suspect’s Ethnicity………………………………..39 Table 4-4: Questionnaire Results of Suspect’s Age……………………………………..40 Table 4-5: Questionnaire Results of Suspect’s Height…………………………………..41 Table 4-6: Questionnaire Results of Suspect’s Weight………………………………….43 Table 4-7: Questionnaire Results of Suspect’s Clothing Description…………………...44 Table 4-8: Questionnaire Results of Suspect Lineup…………………………………....46 1 Chapter 1: Introduction Problem Statement The Innocence Project, created by Barry Scheck and Peter Nuefield at the Benjamin N. Cardozo School of Law in 1992, has used DNA testing to assist in freeing 144 men who were wrongly convicted of various crimes. According to the Innocence Project website (http://www.innocenceproject.org/causes/mistakenid.php), in two-thirds of the first 138 overturned convictions, mistaken identity played a major role in the original conviction. In another study, analysis of 40 cases in which DNA evidence established the innocence of a wrongfully imprisoned person revealed that 36 of them (90%) involved mistaken eyewitness identification (Schacter, 2001). The number of overturned convictions continues to grow as efforts, like the Innocence Project, use advanced scientific methods that were not necessarily available at the time of the convictions. Prior to the introduction and acceptance of DNA testing in the courtroom, eyewitness testimony was often all that linked a suspect to a crime. Even now, a number of cases rely only on a witness’s identification of a suspect.

In recent years, eyewitness testimony has come under more scrutiny as being unreliable. Studies are proving that people do not always recall the correct detail when questioned, however witness statements are a critical part of an investigation into a crime. In conjunction with other compelling evidence, suspect identification by an eyewitness can be a powerful tool for conviction. However, on its own, can it be trusted in court? If the jury’s decision to convict is based on an eyewitness’ identification and circumstantial evidence, should the accused be convicted? Wrightsman (1991) discusses a case in which Father Bernard T. Pagano, Assistant Pastor of a Catholic church in Bethesda, Maryland, was accused of committing a series of 2 robberies in Delaware. These robberies were unique because the assailant would apologize to his victims after robbing them, earning him the nickname “Gentleman Bandit.” A composite sketch, which closely resembled Father Pagano, was drawn and released to the public. Father Pagano was subsequently arrested and charged with the robberies, despite his claims of innocence. During the trial, at least seven witnesses testified that Father Pagano was, in fact, the robber. Although there was little other evidence against him, a conviction seemed inevitable. Before the trial concluded, a man serving time for another offense confessed to being the “Gentleman Bandit.” The robber was 15 years younger than Father Pagano, however their appearances were very similar. How could seven witnesses testify in court that an innocent man was the robber? The witnesses were confident that Father Pagano was the robber and research shows confident witnesses are more likely to be believed by a jury (Schacter, 2001). Father Pagano was fortunate that the actual robber came forward, but many others have either served years in prison before they were exonerated or never have been able to convince the courts of their innocence.

Purpose and Objectives This project serves to examine the reliability of eyewitness testimony as it relates to suspect. This project will look at (a) the biology of the brain, (b) how the mind processes information, (c) how a person recalls information from memory, (d) the difficulties of giving verbal suspect descriptions, (e) whether observing a person of a different race can cause the identification process to be more unreliable, (f) misattribution and suggestibility in witnesses, (g) interview techniques that may improve witness recall, 3 and (h) Neil v. Biggers, the court case that set guidelines for accepting eyewitness identifications in court. An experiment will be conducted in which a mock crime will be videotaped and shown to five participant groups. Each participant will be asked to complete a questionnaire (Appendix A), which requests detailed information regarding the crime and the suspect they have just witnessed. Two days after each participant completes the questionnaire, they will be asked to view a photo lineup to determine if they recognize the suspect. The suspect’s photo will not appear in the lineup to determine if people feel compelled to pick a suspect even if they are unsure. However, a witness from the video will appear in the lineup. The objectives of this experiment are to determine (a) the detail and accuracy of participant’s recollection of the crime, (b) the detail and accuracy of participant’s recollection of the suspect, and (c) if participants will misattribute the witness as the suspect because they may recognize his face, (d) if participants feel compelled to select a suspect even if they are unsure.

Rationale for the Project The primary rationale for this project is to better understand how the mind processes information when observing traumatic situations. The mind has many secrets yet to be revealed and a better understanding of how the mind works could be the first step to improving the way witnesses recall information. Secondly, a number of people have served prison sentences for crimes they did not commit based on questionable witness identifications. The project serves to explore why some suspect identifications can be unreliable. 4 Definition of Terms Acetylcholine – A white crystalline derivative of choline, C 7H17NO 3, that is released at the ends of nerve fibers in the somatic and parasympathetic nervous systems and is involved in the transmission of nerve impulses in the body.

Action Potential – A momentary change in electrical potential on the surface of a nerve or muscle cell that takes place when it is stimulated, especially by the transmission of a nerve impulse.

Amino Acid – An organic compound containing an amino group (NH 2), a carboxylic acid group (COOH), and any of various side groups, especially any of the 20 compounds that have the basic formula NH 2CHRCOOH, and that link together by peptide bonds to form proteins or that function as chemical messengers and as intermediates in metabolism.

Axon – The usually long process of a nerve fiber that generally conducts impulses away from the body of the nerve cell.

Chromosomes – A threadlike linear strand of DNA and associated proteins in the nucleus of eukaryotic cells that carries the genes and functions in the transmission of hereditary information.

Cytoplasm – The protoplasm outside the nucleus of a cell.

Dendrite – A branched protoplasmic extension of a nerve cell that conducts impulses from adjacent cells inward toward the cell body.

Deoxyribonucleic Acid (DNA) – A nucleic acid that carries the genetic information in the cell and is capable of self-replication and synthesis of RNA. DNA consists of two long chains of nucleotides twisted into a double helix and joined by hydrogen bonds 5 between the complementary bases adenine and thymine or cytosine and guanine. The sequence of nucleotides determines individual hereditary characteristics.

Dopamine – A monoamine neurotransmitter formed in the brain by the decarboxylation of dopa and essential to the normal functioning of the central nervous system.

Gamma-Aminobutyric Acid (GABA) – An amino acid, C 4H9NO 2 that occurs in the central nervous system and is associated with the transmission of nerve impulses.

Glutamate – A salt or ester of glutamic acid, especially one that functions as a neurotransmitter that excites cells of the central nervous system.

Ion – An atom or a group of atoms that has acquired a net electric charge by gaining or losing one or more electrons.

Ion Channel – Tiny pore in the cell membrane, through which ions travel.

Kilogram – The base unit of mass in the International System, equal to 1,000 grams (2.2046 pounds).

Long Term Potential (LTP) – A cellular mechanism that causes the connection of a series of synapses to strengthen.

Membrane – A thin, pliable layer of tissue covering surfaces or separating or connecting regions, structures, or organs of an animal or a plant. Millivolt – A unit of potential difference equal to one thousandth (10- 3) of a volt.

Molecule – The smallest particle of a substance that retains the chemical and physical properties of the substance and is composed of two or more atoms; a group of like or different atoms held together by chemical forces. 6 Neurons – Any of the impulse-conducting cells that constitute the brain, spinal column, and nerves, consisting of a nucleated cell body with one or more dendrites and a single axon.

Neurotransmitter – A chemical substance that transmits nerve impulses across a synapse.

Noradrenaline – a hormone secreted by the adrenal medulla and also released at synapses.

Nucleus – A large, membrane-bound, usually spherical protoplasmic structure within a living cell, containing the cell's hereditary material and controlling its metabolism, growth, and reproduction.

Potassium – A soft, silver-white, highly or explosively reactive metallic element that occurs in nature only in compounds. It is obtained by electrolysis of its common hydroxide and found in, or converted to, a wide variety of salts used especially in fertilizers and soaps. Atomic number 19; atomic weight 39.098; melting point 63.65°C; boiling point 774°C; specific gravity 0.862; valence 1.

Protein – Any of a group of complex organic macromolecules that contain carbon, hydrogen, oxygen, nitrogen, and usually sulfur and are composed of one or more chains of amino acids. Proteins are fundamental components of all living cells and include many substances, such as enzymes, hormones, and antibodies that are necessary for the proper functioning of an organism.

Ribonucleic Acid (RNA) – A polymeric constituent of all living cells and many viruses, consisting of a long, usually single-stranded chain of alternating phosphate and ribose units with the bases adenine, guanine, cytosine, and uracil bonded to the ribose. The 7 structure and base sequence of RNA are determinants of protein synthesis and the transmission of genetic information.

Sodium – A soft, light, extremely malleable silver-white metallic element that reacts explosively with water, is naturally abundant in combined forms, especially in common salt, and is used in the production of a wide variety of industrially important compounds.

Atomic number 11; atomic weight 22.99; melting point 97.8°C; boiling point 892°C; specific gravity 0.971; valence 1.

Synapse – The junction across which a nerve impulse passes from an axon terminal to a neuron, muscle cell, or gland cell.

Synaptic Cleft – the space between neurons at a nerve synapse across which a nerve impulse is transmitted by a neurotransmitter.

(http://dictionary.reference.com) Limitations of the Study The main limitation of this research is the number of participants in the experiment. There were 110 participants, each an adult over the age of 22. Most of the participants were graduate students, who hold a four-year degree. Sample size guidelines suggest if the population the study is representing is greater than 5,000 people, as in this case, the sample size should be at least 400 participants, however, the larger the sample size the better (Gay, 1996). This limits the study because the small number of participants may not be a good representation of the population.

Another limitation of the study is the inability to question each participant in person. Proxies were given written instructions on how to conduct the experiment. It was impossible to interview each participant, so the questionnaire was provided. The 8 questionnaire does adequately represent the interview process between an investigator and a witness. There was no attempt at soliciting more information from the witnesses on details of the crime or the suspect. Research Hypotheses There are four hypotheses that will be analyzed when reviewing the data collected from each participant group. The first hypothesis for this project is that all the participants will be able to accurately describe the crime detail, suspect’s gender, and suspect’s ethnicity, however it is expected that most participants will overlook a specific detail of the crime that could provide the suspect’s DNA. As the suspect walks onto the scene, he spits on the grass. It is expected that very few participants will note this crucial element of the suspect’s behavior. Research (Ellis, Deregowsky, & Shepherd, 1975) on the Other-Race effect has shown that participants are more likely to make an accurate identification when they are the same ethnicity of the suspect. Based on this information, it is expected that White participants will correctly describe the suspect as White more often than the other ethnicity groups participating in the experiment. The second hypothesis is that age, height, weight and clothing descriptions of the suspect are expected to vary by participant. The age ranges attributed to the suspect are expected to be younger than the suspect’s actual age because of the type of clothing he was wearing during the crime, supported the research of Sporer (1996). It is likely that the height and weight estimates will be below the suspect’s actual height and weight, especially in female participants, also supported by the research of Sporer (1996). The clothing description is expected to vary because a witness, who chases the suspect after the crime, is the last image the participants see before the video clip ends. Due to 9 misattribution, it is expected that many participants will attribute articles of the witness’s clothing (shorts, a sweatshirt and a baseball cap) to the suspect.

The third and fourth hypotheses relate to the photo lineup. In the photo lineup, the photo of the witness who chases the suspect will appear in the selection. In the third hypothesis, it is expected that 40-50% of the participants will select the witness as the suspect, simply because they recognize his face. For hypothesis four, three of the five participant groups will not be told they can answer “unsure” if they are not confident in their response. The expected outcome is the number of participants in these groups specifying they are unsure if the suspect is in the lineup will be very low.

Summary of Remaining Chapters Eyewitness accounts are extremely important in police investigations. They are valuable in pointing an investigator in the right direction, however it should be understood that a witness’s identification of a suspect could be unreliable. The remaining chapters of this project will include a Literature Review, which will address each of the objectives listed above, the Methodology of the research experiment, the Results of that experiment related to the hypotheses given above, and Discussion, Conclusions, and Recommendations based on the results. 10 Chapter 2: Literature Review Biology of the Brain The human brain weighs about 1.5 to 2 kilograms and contains 100 billion neurons (Bourtchouladze, 2002). The size and shape of the neurons differ and they can be shaped like rods, spirals, spheres, beans, spiders, or octopi. Although individual neurons are dedicated to a specific job, each is part of a very organized community of cells, which is called the neural network. The number of neurons and the way they are connected affect the learning capacity and memory in the different animals. The fundamental design of the neurons is the same, even though the size and shape may vary. Each neuron has an outer wall, called a membrane, and a nucleus. The membrane defines the shape of the neuron and the nucleus, which is located at the heart of the cell, acts as the central command post for the cell. Within the nucleus, chromosomes and strands of deoxyribonucleic acid (DNA) can be found. The double helix of DNA has the genetic encoding of all cell function. Between the inner lining of the membrane and the outer lining of the nucleus is cytoplasm. Cytoplasm is a semi-fluid material that contains a complex of molecules that produces and packages proteins. Proteins are made up of individual amino acids that join together into a chain of three or more. There are twenty different amino acids, which can create an astronomical number of possible combinations within a complex protein. A giant molecule, called ribonucleic acid (RNA), whose synthesis is controlled by the cell’s DNA, encodes each protein (Bourtchouladze, 2002). Neurons use two kinds of specialized structures, dendrites and axons, to send and receive messages. Each nerve cell contains one axon and as many as 100,000 dendrites (Ratey, 2001). Dendrites are the main way by which neurons pass on information to 11 other neurons. The neuron and its thousands of neighbors send out roots (axons) and branches (dendrites) in all directions, which intertwine to form an interconnected tangle with 100 trillion constantly changing connections (Ratey, 2001). There is a chemical exchange between the axon and its surroundings setting up an electrical difference across the membrane. When a cell is resting, the difference is about minus sixty-five millivolts (Bourtchouladze, 2002). The communication process begins when a dendrite or the body of a cell is stimulated by a message causing the electrical difference to become disturbed and the membrane’s voltage to swiftly swing up and down. There is chain reaction, stimulating other cells with the incoming message. If there is enough input at the same time, a wave of electrical activity, called action potential, will travel through the axon. When the ion concentration fluctuates, the action potential is the result. The fluctuation in ion concentration is caused by the two-way traffic of sodium and potassium ions traveling through tiny pores in the cell membrane, called ion channels. Along the signal’s path, the ion channels open and close sequentially, until it reaches the end of the axon. The end of the axon branches into tiny filaments, which form presynaptic terminals. Once the signal reaches the presynaptic terminal, contact is made with a dendrite of the next neuron. The junction of the presynaptic terminal and the dendrite branch is called a synapse. One neuron can have thousands of synapses with other neurons. When the signal arrives at the synapse, a major modification occurs. The electrical signal’s arrival causes the release of packets of chemical neurotransmitters, changing the signal from electricity to chemistry. The packets of neurotransmitters were discovered in the 1950s by Bernard Katz (Bourtchouladze, 2002) and typically contain glutamate, gamma-aminobutyric acid 12 (GABA), acetylcholine, noradrenaline, and dopamine. The neurotransmitters jump the synaptic cleft and bind to protein molecules on the surface of the recipient cell. This process takes a fraction of a second to complete and is repeated for most activities of the brain, including learning, memory, perception, movement, and emotion (Bourtchouladze, 2002).

When the brain receives the stimulus of a new event or experience, the message fires along a path of neurons, causing the connection between some synapses to be strengthened, while other connections will be weakened. The new connection may disappear if it is not transferred to its permanent location by a process called Long-Term Potential (LTP) (Ratey, 2001). LTP is a cellular mechanism that causes the connection of a series of synapses to strengthen. Once an event or idea is encoded with a series of synapse connections, the same event or idea will fire along the same path causing the bonds to become stronger. The more the same stimulus is repeated, the stronger the connections become. During this process, the neurons along the path will recruit neighboring neurons, eventually developing an entire network of neurons devoted to one skill, word, idea, or any number of subjects. Once the network is formed, this subject is transformed into memory. Memory This process can take many repetitions or one instant to create enough neurons to cement the subject into memory. While learning one new skill may take four or five passes to get a good understanding of the process, another can be instantly remembered. A dramatic occurrence, like the terrorist attacks of September 11, 2001, can put LTP into overdrive and create an instant neural network employing neurons from all areas of the 13 brain. For this reason, one might remember in specific detail where they were, what they were wearing, and whom they were with at the time of the occurrence. People still discuss where they were when they learned that John F. Kennedy was assassinated in 1963 or the Space Shuttle Challenger exploded in 1986, with clarity as if the events just occurred. This phenomenon is called Flashbulb Memory (Ratey, 2001).

The biology of the brain is well understood by the scientific community, yet how the information converts from short-term memory to long-term memory is debatable. There are many theories as to how this transfer occurs. Research (Ratey, 2001) has suggested that the transfer from short-term memory to long-term memory happens in a small window of time. This window is the time needed for the neurons to synthesize the necessary proteins for LTP to occur. Key proteins that are produced by the first neuron stimulated must find the synapse and bind to it for the memory to be cemented into place. The binding of the protein to the synapse changes the structure of the synapse making it more sensitive to incoming signals.

Two other models of memory are Broadbent’s Filter Theory and the Atkinson- Shiffrin Buffer Model (Ainsworth, 1998). Broadbent’s Filter Theory states that the human brain is not capable of processing all the information the sensory organs provide at any one time. There is a filtering system, which allows humans to attend to what is most important in that moment and filter out all the other stimuli. When one stimulus is being filtered out in one moment, it can easily become the focus in the next. For example, if a person is watching television and others are talking in the background the person might filter out the voices of the others until they mention his name, then his attention will shift to the live conversation. In legal circumstances, this theory could play 14 a role in understanding that witnesses cannot be expected to take in every aspect of a crime, crime scene, or suspect. Information that is filtered out does not make it to short- term memory so it cannot be transferred to long-term memory, therefore it cannot be retrieved by methods like hypnosis. Critics of this theory believe it is too simplistic and assumes humans cannot multi-process.

The Atkinson-Shiffrin Buffer Model shows memory as a three-stage system. As information enters the sensory organs, it is briefly held in a sensory storage area. If attended to, the information passes to a short-term memory store, which can only hold a limited amount of information for about half a minute. After 30 seconds the information is either replaced by new information or transferred to long-term memory. The “buffer” in the theory title relates to the short-term memory store, which protects the long-term memory from too much information at one time. Information stored in short-term memory is in chunks of five or six digits, as they are easier to remember than longer segments. Atkinson and Shiffrin (1971) state that only information that has been rehearsed can be transferred into long-term memory. In studies of free recall (Bourtchouladze, 2002), participants were given a list of items to memorize. Results indicated that they were more likely to remember the first few items and the last few items on a list. This is called the Primancy and Recency effect. The first and last items are the more likely to be rehearsed than the items in the middle of the list. This effect supports Atkinson and Shiffrin’s rehearsal theory.

Suspect Descriptors Three key stages to the memory process are, acquiring information, retaining information, and retrieving information, and there can be a failure at any one of them. 15 Experiments in London and the United States have tested a person’s ability to remember information acquired while distracted. In the 1950s, the Brown-Peterson Distracter Technique resulted in 90% of new information being forgotten after 18 seconds of complete distraction (Bourtchouladze, 2002). Forgetting in such a short time is attributed to the Trace Decay theory, which states that if an item is not repeatedly rehearsed, it will be forgotten. Retention of information, as discussed above, is not fully understood and works automatically without conscious effort. The main problem investigators have with eyewitnesses is in the retrieval of vital information. Two ways witnesses will retrieve information are recollection and recognition. Recognition is easier than recollection. If you are asked to recall all 50 states you may forget some, but you will recognize them as states if you see the names written down. With eyewitnesses, the process of taking the image of a suspect and verbalizing it into descriptors can be very difficult. One problem is that a suspect’s image is stored as a whole, but when giving key descriptions, it is done in pieces. In 1986, Yuille and Cutshall studied a shooting that had over 21 witnesses. Thirteen of the 21 witnesses volunteered to take part in their follow-up research study five months after the incident occurred. Yuille and Custshall interviewed the 13 witnesses again and found their accounts to be surprisingly detailed and accurate. This was later attributed to multiple police interviews, news media stories, and recounting the story to other people. This reiteration of the details served to cement them into long-term memory. Yuille and Cutshall went beyond follow-up interviews with the witnesses and examined the original police interviews of all 21 witnesses. From those original 16 interviews, there were 392 action, 180 person, and 77.5 object descriptions. Most of these descriptions turned out to be correct, however most errors occurred in the 180 person descriptions. Height, weight, and age were the most commonly incorrect descriptors among the witnesses with 23% in error. Second, color and style of hair and clothing had an error rate of 18.1%. These errors may seem relatively low among 21 witnesses, however when police arrive at a crime scene, they want to send a bulletin to other officers and agencies with a suspect description, in case the suspect is still in the area. Errors in the initial description of a suspect may reduce the likelihood of capturing a suspect if he remains in the area or result in an innocent party being picked up by police. Studies (Sporer, 1996) related to the accuracy of suspect descriptions, height and weight normally produce inaccurate results. The tendency was to underestimate the height and weight, although this was more common with female study participants. One reason for this discrepancy could be that the suspect or target in the study is typically male so it is more difficult for females to judge. In other studies (Sporer, 1996), it was determined that participants use their own height and weight as an anchor for comparison only when the target is the same gender as the participant. Most witnesses attempt to give an age range when relaying suspect information to investigators, however studies have shown this estimate can differ depending on the style of clothing worn by the suspect. In one study (Sporer, 1996), the target first acted aggressively while wearing a red windbreaker then changed into a nice leather jacket to play the role of a business associate. In the first scenario, the average age given by participants was 23.5, while the second scenario yielded an average age of 27.5. 17 Facial Descriptions Facial descriptions are very important in an investigation, especially if a composite sketch of the suspect will be drawn. The main difficulty in giving facial descriptions is taking a whole image of a face and transferring verbally into individual elements. Most witnesses give information related to the upper half of the face, including hair color, hairstyle, and eye color. Although this information is very useful to police, these features are easily changed. Cheekbone structure, chin shape, and forehead shape are not easily changed, but witnesses cannot easily describe them. Many factors determine how easily a face is remembered and recalled. Some studies (Shepherd & Ellis, 1973) have shown that faces that a person finds attractive is more likely to be remembered than ones they find unattractive. Others (Light, Kayra- Stuart, & Hollander, 1979) suggest that if a face stands out in some way, it is more likely to be recognized and recalled than more average faces. Distinctive or unusual faces can evoke an arousal in the witness and the image may be stored differently than normal faces (Goldstein & Chance, 1981). Even experiences and training may determine how well a person will recall certain features. A dentist might notice teeth or an eye doctor might notice eyes. Studies to determine if police officers make better witnesses due to their training have had inconsistent results (Sporer, 1996). Some studies show no difference between police and civilian participants, while others show that police officers have significantly better results.

Other-Race Effect Another major factor in mistaken identification is known as the Other-Race Effect. This effect relates to the phenomenon of witnesses having a more difficult time 18 describing or recognizing suspects of a different race. A study by Ellis, Deregowsky, and Shepherd (1975) had 12 African (Black) and 12 Scottish (White) teenagers view 16 faces and describe four of them, one Black male, one Black female, one White male, and one White female. Black participants described more features than the White participants and the Black faces were given more descriptors than white faces by both sets of participants. Although the study was not originally designed to determine if participants could better recognize people of their own race, the results yielded some interesting observations on how participants described the faces of their own race. White participants more often described hair color and texture and eye color. Black participants described ears, chin, eyebrows, size and whites of the eyes, position of the hair, and the outline of the face. One problem attributed to cross-racial identification is inappropriate encoding strategies (Ainsworth, 1998). People tend to use face-encoding strategies that have worked for them in the past. A White person may pay closer attention to hair color and style and eye color when looking at the face of another White person; however when looking at a Black person, this strategy may be inappropriate because hair and eye color tend to be uniform among Black people. If the problem were related to encoding, then training could possibly improve in the recognition of other races. Woodhead, Baddeley, and Simmonds (1979) report three studies where participants were evaluated on recognition accuracy both before and after a three-day training course. In two of the studies, there was improvement, however in the third study the number of correct identifications decreased. Two possible explanations were offered for these results (Malpass, 1981). First, for most people, learning faces is a well-established skill, which is difficult to change. Second, the type of training used in these studies may have been 19 inappropriate for improving facial recognition. Four different training courses were analyzed and assessed for effectiveness. It was determined that no one type of course was better than the other and in many cases recognition performance went down (Malpass, 1981). A theory on why these courses failed is the length of these training were too short, under 12 hours. Longer trainings seem to have an improved effect on participants (Ellis, 1984). Chance and Goldstein (1996) reviewed studies of the other-race effect and offer two possible explanations. First, ethnic attitudes may either impair or enhance a person’s ability to recognize faces of other racial groups. Second, differing levels of experiences and familiarity with other groups might modify the person’s face recognition and/or memory process. They said, “Although attitudes toward other groups may influence the degree and quality of contact with them, no straightforward relationships between attitudes and memory occur” (Chance & Goldstein, 1996, p. 170).

Another explanation is that less familiar faces can be harder to recognize and stereotyping may compound this (Brown, 1986). Humans tend to assign similar characteristics to members of their own group and when viewing other races they do not pay enough attention to features that would later assist in distinguishing one face from another. Witness Factors Shapiro and Penrod (1986) reviewed the results of 128 eyewitness identifications and facial recognition studies, examining 960 experimental conditions and 16,950 subjects. They found that there were certain characteristics that made for a better witness. 20 Female witnesses were slightly more likely to make correct identifications than males; however, females were also more likely to make false identifications. In terms of race, Black participants made more correct identifications, with 54.5%, Anglos rated second at 42.6%, and Mexican-Americans at 38.1% accuracy rates. Verbal skills also affect accuracy rates, the better verbal skills the witness has the more likely they are to select the correct suspect. Age of the witness did not show significant differences except in young children (age 4-12) and in older adults (over 60 years old). Young children and older adults are less likely to select correctly, especially when the suspect is absent from a photo lineup. Some participants were rated on their recognition skills, classified as either “good recognizers” or “poor recognizers.” As expected, good recognizers significantly outperformed poor recognizers in facial recognition. Training and job skills, as mentioned above, show significant improvement in some studies and no difference in others. External Factors In some cases it is not the witness’s powers of observation, but external factors that contribute to inaccurate descriptions. Sporer (1992) did a study to determine varying levels of exposure to different factors affects the accuracy of eyewitness accounts. The study tested low, medium, and high exposures to alcohol intoxication, illumination, duration of the event, time to observe, number of targets and disguises worn by targets. Table 2-1 summarizes the results of this study.

21 Table 2-1 Means of Number of Descriptors, Length and Precision of Person Descriptions as a Function of Situational Factors for Victim and Bystanders (N=64) Predictor Variable Description Measure None/Low Medium High Very High F Descriptors 14.07 3.86 9.86** Lines 9.53 6.29 1.55 Precision 5.45 2.14 15.93** * Alcohol Intoxication (55) (7) Descriptors 6.92 15.08 14.36 4.57* Lines 5.31 11.25 10.00 3.51* Precision 3.31 6.17 5.42 7.10** Illumination (13) (12) (36) Descriptors 14.27 12.13 12.82 13.71 <1 Lines 12.91 7.13 8.36 10.14 2.06 Precision 6.36 4.73 5.14 4.71 1.39 Duration of Event (11) (15) (22) (14) Descriptors 12.57 10.43 13.73 14.50 <1 Lines 10.48 5.00 8.45 11.17 1.81 Precision 5.48 3.57 5.55 4.38 1.59 Time too observe (21) (7) (22) (12) Descriptors 14.49 12.09 10.57 2.19 Lines 9.00 7.36 10.57 <1 Precision 5.79 4.36 3.71 5.67** Number of Targets (39) (11) (14) Descriptors 13.27 12.22 <1 Lines 10.15 7.13 3.35 Precision 5.39 4.57 1.95 Disguise (41) (23) Note: Sporer (1992) *p .05. **p .01. ***p .001.

The results were judged by the number of descriptors, the number of lines, and precision in statements. In cases of high levels of intoxication and low levels of illumination, the results were as expected, the number of descriptors and lines were much lower than more desirable conditions. Surprisingly, longer duration of an event and more time to observe an event produced a lower number of descriptors and lines of description. If there were fewer targets, there were more descriptors but fewer lines and the results were better if the target did not have a disguise. 22 Misattribution Another, yet less understood, problem is misattribution, also called unconscious transfer. Houts (1956) highlighted one case from the 1950s, where a railway booking clerk was robbed at gunpoint and his money was stolen. The clerk was asked to attend an “identity party” where he identified a suspect. The identified suspect had an iron clad alibi and was not charged with the robbery. Authorities, trying to establish how the clerk could have identified the wrong person, found that the wrongfully identified man was the only person that looked familiar to the clerk. The man was a sailor stationed in the area, who had purchased tickets from the clerk on numerous occasions. The clerk mistakenly associated him with the robber. In this case, the sailor was fortunate that he had an alibi at the time of the robbery or he could have served time for a crime he did not commit. Because it is difficult to prove, some researchers do not feel this is a large problem in wrongful convictions (Ainsworth, 1998). In one study (Ainsworth, 1995), participants were asked to read a short passage from a newspaper. Adjacent to the article were two photographs, one of the suspect and the other of a Good Samaritan who had intervened and thwarted the crime. One week after reading the article, without warning, the participants were asked to pick the suspect out from a six-person photo lineup. The participants were separated into three groups. The first group saw the suspect’s photo, along with the photos of five others. The second group saw the Good Samaritan’s photo, along with the photos of five other, excluding the suspect. The third group saw neither the photo of the suspect or the Good Samaritan, just six other photos. The first group selected the suspect in nearly 40% of the cases, while the second group selected the Good Samaritan in nearly 50% of the cases. The second group had a misattribution rate of 23 nearly 50%. Had this been an actual case, it is possible that a Good Samaritan or innocent passerby could be identified as the suspect. Suggestibility The way a witness is interviewed can greatly affect their statements, as humans can be highly suggestible. Merely answering “okay” to a witnesses statement may lead them to believe they are giving accurate information, even if they were previously uncertain. A confident witness is much more likely to be believed by a jury than an uncertain one (Schacter, 2001). One study (Schacter, 2001) had participants watch a video of a man entering a Target store. They were told that the man entered the store and murdered a security guard, but the crime was not shown on the video. The participants were then given a set of photos, in which the gunman was not included. Some participants were told, “Good, you identified the actual suspect,” others did not receive any feedback, and others received disconfirming feedback upon selecting a suspect. All participants were asked to assess how well they were able to view the suspect, evaluate their memory with certainty and clarity. Those who received positive feedback claimed higher rates of trust and confidence in their memories and their ability to select the correct suspect, even though every participant selected the wrong suspect. If an actual witness was given confirming feedback during police interviews, they would be more confident when giving testimony in court and more likely to be believed by the jury.

Cognitive Interview Technique Police agencies are now recognizing the importance of how they handle interviews with witnesses and are training their officers of new interviewing techniques. One such technique is the Cognitive Interview Technique (CIT), developed by Ronald P. 24 Fisher and R. Edward Geiselman in the 1980s, which has since been revised and refined (Ainsworth, 1998). There are two major areas of emphasis on this technique, memory and communication. The witness needs to be able to retrieve the vital information from memory, translate the images to words, and communicate it to investigators. What is special about this technique is that it puts the witness in the dominant role, while the investigators acts as a guide or assistant. Fisher and Geiselman (1992) suggest the interaction emulate more of a social interaction than a rigid interview. Below are some key guidelines to the Cognitive Interview technique.

• The first task of the interviewer is to make the witness feel at ease and to encourage his or her active participation in the process. If the witness is less likely to be led by the investigators, the more likely they are to give uncontaminated information regarding the events. • If the witness is also the victim and is having a problem communicating to officers due to embarrassment, some time should be taken to build a rapport before getting into questions about what happened. • One technique to get the witness or victim to feel more at ease with talking about it may be to encourage them to relay events as if it happened to someone else. • Two ways the interviewer can build up the trust of the witness o Make the interview personal - use the witnesses first name and allow the witness to use the interviewer’s first name o Communicate with empathy – make the witness or victim feel like this is not just another crime or interview for the investigator, but that they care about the outcome. 25 • If a sexual assault occurred, the interviewer should be the same sex as the victim, if possible.

• It is important to get an understanding of the witness’s mental state at the time of the crime, what they were doing, where they were going, if they were in a hurry, before the interview starts.

• The interviewer should ask open-ended questions, allowing the witness to give more expansive answers and try to avoid Y/N questions, especially in the beginning. • The witness should be allowed to talk at length without being interrupted and the interviewer should be an active listener.

• The interviewer should encourage the witness to relax not only at the beginning of the interview, but after each question. The witness might want to take some deep breaths or envision a pleasant scene in order to relax. • The interviewer should always avoid making judgments or acting in a confrontational manner. The interviewer should not make the witness feel as if they are not doing as well as they could be or that they are taking a test, with right and wrong answers. Each witness should be encouraged and reassured throughout the session that they are doing well – but care should be taken to avoid saying things that might bias future accounts the witness might give.

• If there are inconsistencies in the story, it may be error of misunderstanding. The interviewer should not bring up such inconsistencies right away, so as not to put the witness on the defensive. If clarification is needed, make notes and wait for the witness’s thought process to complete before asking clarifying questions. 26 • Interviewers should encourage witnesses to use relative rather than absolute judgments about height, weight, age, etc. Was the offender taller or shorter than me?, rather than How tall? • Use of props such as different knifes, or a similar weapon that was used during crime to not only help determine what the weapon looked like, but also in what manner it was wielded. • Encourage witnesses to attempt to recall details they were previously unable to retrieve – could be a “tip of the tongue” scenario where the piece of information is there but cannot be immediately retrieved. • Fisher and Geiselman (1992) believe that witnesses store both general and highly detailed information in their memory. They believe there are two different types of coding and that witnesses store information in both a concept and an image code. The interviewers task is to elicit as much information from the image code as possible. The witness should focus on the surrounding scene and pay attention to visual, auditory, tactile, and other sensory representatives rather than just recalling visual.

The CIT method has been evaluated through numerous studies and most concur that witnesses were able to give more detailed and accurate information when this technique was employed. An early British study stated they were able to elicit 55% more information from witnesses (Ainsworth, 1998). Fisher and Geiselman (1992) have also tested their method and believe it works for students and non-students, rookie and veteran investigators, and civil as well as criminal investigations. 27 There are some limitations to this method. First, it may be difficult for investigators to take a passive role in an interview when they are used to being in control. Another limitation is that this technique does not work with obstinate or reluctant witnesses. If the witness does not want to participate, other tactics should be employed. The length of time required to use this technique makes it difficult for busy investigators. Some investigators may not want to take the time to build a rapport with a witness and the time it takes will differ with each witness. Because of this limitation, investigators may only employ this technique on more serious crimes. They may rate the crimes to determine if it is worth the extra effort (Ainsworth, 1998).

Department of Justice Guidelines for Eyewitness Evidence In 1998, former Attorney General Janet Reno formed a working group consisting of psychologists, police, and attorneys to develop guidelines for collecting eyewitness evidence. This group published a set of guidelines based on rigorous scientific studies (Schacter, 2001). A copy of these guidelines can be found at website www.ncjrs.org/pdffiles1/nij/178240.pdf.

Suspect Identification Techniques As stated previously, recognition of a suspect is much easier than recalling their physical description. There are several techniques police agencies use to find their suspect. The first thing in all cases is to get a detailed description of the face, build, and clothing. Once police agencies have the description, they can make a decision on what techniques to employ. 28 Mugshot Inspections With this technique, a witness will look through a collection of mugshots of people who have previously been convicted of a similar offense. Even if a witness cannot recall many specific details about a suspect’s face, they can still recognize them when seeing the whole face. One benefit of this technique is that witnesses can rule out a large number of suspects. Another is that witnesses see the mugshots sequentially so each photo can be compared against the image of the suspect in the witness’s mind rather than selecting the most familiar suspect in a group. A third benefit is that witnesses can select multiple possible suspects so police can follow up on the leads to determine if one is the actual perpetrator.

One problem with this technique is that it only includes photos of those previously convicted of a crime and does not take into consideration first time offenders or those that have evaded the law. In many cases, if a witness does not see the suspect the first time through, they are encouraged to look again. Studies have shown that the more photographs a witness looks at, the more likely they are to select the wrong person (Davis, Shepherd, & Ellis, 1979). Wells (1988) recommends that witnesses should not be given more than 50 mugshots to view at one time. New technology is emerging where the features are rated on a number of dimensions and the data are stored in a computer database. A witness will be asked to do a similar rating of the suspect they witnessed and the results will be cross-referenced with the database, so only the most likely suspects will be shown to the witness. Even though these procedures are in early stages, the results have been encouraging so far (Shepherd 29 & Ellis, 1996). However, this rating system only works if the witness’s memory is accurate and it takes into consideration the subjective nature of descriptions. Another problem may occur if police have a suspect in mind, they might draw attention to that mugshot. Although this behavior is frowned upon, it can occur and it is difficult to prove. When a witness does make a decision, they are more likely to carry that choice through to photo or live lineups in order to be consistent in their story (Brigham & Cairns, 1988). Conversely, if one suspect is overlooked during the mugshot viewing, the same suspect will continue to be overlooked. Photo lineup Identifications Photo lineups really narrow down the number of photographs a witness has to look at. This method is especially useful when police have a suspect in mind, so they can incorporate the suspect into the photo spread. The photographs chosen to be in a photo lineup should be as similar to the witness’s description of the suspect as possible. An advantage to this method over live lineups is that the witness cannot see if a suspect is nervous or fidgety, which can sway a witness’s selection (Ainsworth & King, 1988). A concern with this method is that the witness is looking at all the photographs at the same time, so may be comparing them to each other rather than against the image in memory. The witness will be instructed to look at each picture separately and judge it individually, but may be inclined to survey all the photos and rule out each one by one, until they have one photo left, thus selecting the suspect. Another disadvantage to this method is that the photographs are static and the witness cannot see mannerism or behaviors the suspect might have had during the crime. Yet another disadvantage is that the suspect’s appearance could have changed dramatically since the photograph was taken. Many 30 studies have shown that photo lineup can produce accurate results, however this method most often leads to the misidentification of an innocent party. Buckhout (1980) showed a short film of a purse snatching on television. In the video, a man was seen walking down a hallway, with his back to the camera, he then knocked a woman over and grabbed her purse, and turned to flee, running back toward the camera so the audience could get a good look at his face. Viewers were then shown a photo lineup of six men who resembled the assailant and were asked to call a special phone number if they thought they could identify the correct person. Viewers were also told that the suspect might not be included in the photographs, but in fact he was. There were over 2,000 calls from viewers and only 14.1% selected the correct person. Over 85% made an incorrect selection, even though they were only asked to call if they were confident in their selection. Cutler, Berman, Penrod, and Fisher (1994) reviewed a large number of studies on photo lineups and determined the results were the same as those of video or live lineups. Live Lineups Live lineups have an advantage over mugshots or photo spreads because the witness can see mannerisms and behaviors of the potential suspects. The potential suspects may be directed to say a word or sentence similar to what was said during the crime so the witness can attempt to identify the suspect’s voice. Other behaviors can also be acted out during the lineup. As in a photo lineup, this technique is most useful when police have a suspect in mind. The suspect and a number of foils, or innocent participants, will go before the witness to determine if the police were correct in their hunch. The biggest requirement to assure a fair lineup is that the foils must look very 31 similar in appearance to the suspect. Five foils plus the suspect is acceptable in the United States, however in England they require at least seven. There are a few problems with live lineups. First, the witness’s attitude can present a problem, especially if the witness was also the victim. A victim would have high motivation to select an assailant so as to help police solve the case. If the actual suspect is not in the lineup, the victim might still feel compelled to select the person that most resembles the actual suspect. Second, the same problem can occur as in the photo spreads, where the witness will rule out suspects until there is one and select the most likely person rather than the actual suspect. A good way to combat this problem is to have each person in the lineup walk out one at a time, forcing the witness to compare the face against the image in memory rather than the other people in the lineup. Suspects can decide what place they will hold in the lineup and also have limited abilities to object to the composition. Once the composition of the lineup is determined, the suspect has little say. A problem for the suspect is nervousness or an uncomfortable demeanor might cause them to be selected. Witnesses can pick up on non-verbal communications and make a decision for a suspect when they were otherwise uncertain. A nervous participant would stand out among foils who have nothing to lose by participating in the lineup. Although this may not sound like a problem because the nervous suspect should be selected, if the suspect is innocent, he could be just as nervous at the prospect of being identified for a crime he didn’t commit, so nervousness does not necessarily insinuate guilt.

Neil v. Biggers In 1972, a woman was confronted by an attacker in the dimly lit kitchen of her 32 home then taken to a remote moonlit area where she was raped. Biggers was convicted of the rape in Tennessee on evidence partly consisting of the victim’s visual and voice identification during a “show-up” seven months after the attack occurred. The suspect was brought to the station and the witness came to view his mannerisms and voice without other participants in a lineup. The victim claimed to get a good look at the assailant both in the dim light of the kitchen and in the moonlit area outdoors and she was very confident in the identification. The Supreme Court held that based on reliable, despite suggestive, procedure due process was not violated (Loftus, 1996). The court found the victim’s identification reliable after examining the following five factors, which have become a standard for eyewitness identification evidence in court (Loftus, 1996).

o Did the witness have an opportunity to observe the defendant?

o What was the witness’s degree of attention?

o How accurate were the witness’s accounts at the confrontation (identification)?

o How certain was the witness at the confrontation?

o How much time has lapsed between the crime and the confrontation?

Although these factors must be considered when eyewitness identifications are entered into court, the responses to some of these inquiries can be subjective. The witness may be very certain and feel they are accurate, but in reality they have identified the wrong person. As the studies above have shown, the process of identifying a suspect can break down at any stage of the process, from the first witness’s statement to the jury’s conviction. It is an imperfect system that needs to be continually studied and improved upon. 33 Chapter 3: Methodology A mock crime was videotaped and shown to five groups of participants. The video depicts a group of four adults and a baby sitting at a picnic table in an outdoor setting. A white man in his early thirties enters from the right of the screen. He walks by the family’s table, spits on the grass, then sits at an adjacent picnic table. After observing the family and his surroundings for a few seconds, the man gets up and walks slowly past the family’s table, grabs a black purse from a woman sitting at the table then runs out of sight between two buildings. Another woman at the table yells, “Hey, he took my purse” and the man sitting next to her chases the suspect out of sight.

Each group was shown the 30-second video and participants were given a questionnaire to complete (Appendix A). The questionnaire requests the demographic information of each participant, including gender, age and ethnicity, as well as specific information related to the crime and suspect they just witnessed.

Two days after each group witnessed the crime video, the participants were shown a six-pack photo lineup and were asked to select the suspect they witnessed in the purse snatching. Two of the five groups were told that the suspect may not be in the lineup and it was acceptable to answer “unsure”. The remaining three groups were not given specific instruction, just shown the photo lineup. Included in the lineup was the photo of the witness who chased the suspect after the purse was stolen. There were 110 respondents, within the five groups that participated in the experiment. Each participant was classified into four demographic areas: Participant Group, Age, Gender, and Ethnicity. 34 Qualitative Results Participants were asked on the questionnaire to specifically describe the crime they had just witnessed. These qualitative answers were converted to quantitative results for statistical analysis. In order to convert the responses, each respondent’s summary of the crime that occurred was rated on detail and accuracy. There were six key details, or elements, of the crime, all specific to the suspect’s behavior, that helped determine how detailed their summaries were. The six elements are:

1. A man walked into an area where a family or group sat at a picnic table.

2. He spit in the grass.

3. He sat at a neighboring picnic table.

4. He looked back at the potential victims.

5. He walked back by the victim’s table and stole a purse.

6. He was chased by a man who was sitting at the victim’s table.

If the respondent’s summary of the crime captured five or six of these elements, their description was rated “Detailed”. Summaries that captured three or four elements were rated “Somewhat Detailed” and those with one or two elements were rated “Not Detailed”. A secondary rating system categorized each summary by accuracy. If all the facts stated in the summary were accurate, the summary was rated “Accurate”. If some statements were accurate and others were not, the summary was rated “Somewhat Accurate” and if all the statements given were incorrect the summary was rated “Not Accurate”. The other questionnaire response that produces a qualitative answer was the description of the suspect’s clothing. The suspect’s clothing description was rated and categorized in the same manner as the summary of the crime that occurred. Each 35 respondent’s description of the suspect’s clothing was rated on detail and accuracy. There were five key elements of the suspect’s clothing, which are:

1. Red and White Flannel Jacket 2. Black T-Shirt under the flannel, with white writing on the front 3. Blue Jeans 4. Black Tennis Shoes with White Soles 5. Black Beanie Cap If the respondent’s clothing description captured four or five of these elements, their description was rated “Detailed”. Summaries that captured two or three elements were rated “Somewhat Detailed” and those with zero or one element were rated “Not Detailed”. A secondary rating system categorized each summary by accuracy. If all the facts stated in the summary were accurate, the summary was rated “Accurate”. If some statements were accurate and others were not, the summary was rated “Somewhat Accurate” and if all the statements given were incorrect the summary was rated “Not Accurate”. Suspect Photo Lineup The suspect who appeared in the video was not one of the six photographs shown in the lineup. This was done intentionally to determine if witnesses would feel compelled to select a suspect. One of the six photographs in the lineup was that of the witness who chased the purse-snatcher during the crime. This witness is the last person seen on the videotape. The intention was to determine if witnesses will select someone that looks familiar or the person they actually believe to have committed the crime. Two of the five participant groups (Groups 2 and 5) were told that the suspect may not be in the lineup and to answer “unsure” if they were not confident in their selection. The other three groups (Groups 1, 3, and 4) were not given this instruction. 36 Chapter 4: Results The first hypothesis in this study was related to the accuracy of crime detail, suspect gender, and suspect ethnicity. Although the crime detail was expected to be accurate, the number of participants that mentioned that the suspect spit on the grass would be considerably low. As described above, the qualitative answers were converted to a combination of Detail and Accuracy, yielding five categories: Detailed/Accurate, Somewhat Detailed/Accurate, Somewhat Detailed/Somewhat Accurate, Not Detailed/Accurate, and Not Detailed/Somewhat Accurate. There were no instances of complete inaccuracy in any respondent’s summaries. The results of the crime detail summaries can be found in Table 4-1.

Crime detail was expected to be accurate. Without taking detail into account, 86% of summaries were accurate. The hypothesis did not reference how detailed each crime detail summary would be. In fact, only 13 participants gave five or six accurate key elements to be rated “Detailed and Accurate,” which accounts for about 8.5% of the total number of participants. The highest percentage of results was “Somewhat Detailed and Accurate” with 42%, followed by 33% of crime detail summaries that were “Not Detailed but Accurate.” Among the demographic groups, the results were mostly consistent with the overall percentages. The two highest percentages were “Somewhat Detailed and Accurate” and “Not Detailed and Accurate.” The notable exception to this is Participant Group 1, of which 31% of their crime detail summaries were “Detailed and Accurate,” followed by 27% that were “Somewhat Detailed and Accurate.” Participant Group 1 was composed of Teachers and Administrators at an elementary school in Southern 37 California. The other participant groups were graduate classes located at the La Jolla campus of National University and were diverse in their occupations. Table 4-1 Questionnaire Results of Crime Detail Summary Participants # Detailed/ Accurate Somewhat Detailed/ Accurate Somewhat Detailed/ Somewhat Accurate Not Detailed/ Accurate Not Detailed/ Somewhat Accurate All 110 13 46 8 36 7 Group 1 26 8 7 1 5 5 Group 2 38 2 17 4 13 2 Group 3 19 3 10 2 4 0 Group 4 23 0 11 1 11 0 Participant Group Group 5 4 0 1 0 3 0 20-29 74 7 30 4 30 3 30-39 24 6 12 2 3 1 40-49 10 0 3 2 2 3 Age 50-59 1 0 1 0 0 0 Unspecified 1 0 1 0 0 0 Female 90 11 40 8 25 6 Gender Male 20 2 6 0 11 1 Asian 14 0 3 0 10 1 Black 6 0 4 0 2 0 Hispanic 23 5 7 2 6 3 White 61 7 30 5 17 2 Ethnicity Other 6 1 2 1 1 1 The second part to this hypothesis was that very few participants would note that the suspect spit on the grass as he walked by the table. The results are in Table 4-2 and, in fact, only 13 of 110 participants mentioned that the suspect spit on the grass, accounting for 12% of the total. When evaluating this data by participant group, 77% of those 13 participants were from Group 1. Although Group 1 accounted for a large majority on participants that mentioned the suspect spitting; still only 38% of the entire 38 group volunteered this information. In the remaining demographic groups, most had less than 12% of the whole that mentioned this in their crime description. The notable exception here is Hispanic participants. Hispanic participants mentioned the suspect’s spit in 26% of the questionnaires.

Table 4-2 Questionnaire Results of Suspect’s Spit mentioned Participant # Mentioned that the suspect spit Did not mention that the suspect spit All 110 13 97 Group 1 26 10 16 Group 2 38 0 38 Group 3 19 2 17 Group 4 23 1 22 Participant Group Group 5 4 0 4 20-29 74 7 67 30-39 24 5 19 40-49 10 1 9 50-59 1 0 1 Age Unspecified 1 0 1 Female 90 11 79 Gender Male 20 2 18 Asian 14 1 13 Black 6 0 6 Hispanic 23 6 17 White 61 5 56 Ethnicity Other 6 1 5 39 The suspect’s gender was presumed to be accurate in all survey responses and this hypothesis was supported as 100% of participants correctly answered that the suspect was male.

The suspect’s ethnicity was also expected to be accurate. Table 4-3 contains the results regarding the suspect’s ethnicity.

Table 4-3 Questionnaire Results of Suspect’s Ethnicity Participants # Hispanic White White/Hispanic Unsure All 110 15 68 2 25 Group 1 26 3 16 0 7 Group 2 38 4 25 1 8 Group 3 19 4 11 0 4 Group 4 23 3 14 0 6 Participant Groups Group 5 4 1 2 1 0 20-29 74 8 48 1 17 30-39 24 5 14 1 4 40-49 10 2 6 0 2 50-59 1 0 0 0 1 Age Unspecified 1 0 0 0 1 Female 90 11 56 2 21 Gender Male 20 4 12 0 4 Asian 14 2 6 0 6 Black 6 0 5 0 1 Hispanic 23 4 14 0 5 White 61 8 38 2 13 Ethnicity Other 6 1 5 0 0 In fact, 62% of participants correctly answered that the suspect was White. Twenty-three percent of participants were unsure of the suspect’s ethnicity, 14% thought 40 the suspect was Hispanic, and 2% thought the suspect was half-Hispanic and half-White. These percentages were consistent among the different demographic groups. White participants correctly answered White as the suspect’s ethnicity in 62% of cases and Hispanic participants incorrectly answered Hispanic as the suspect’s race in 17% of cases. Research (Loftus, 1996) has shown that witnesses are more likely to accurately identify persons of their own ethnicity.

Table 4-4 Questionnaire Results of Suspect’s Age Participants # Under Correct Over Unsure Unspec Adj All 110 64 31 7 4 1 3 Group 1 26 14 6 4 0 0 2 Group 2 38 19 14 1 3 1 0 Group 3 19 15 2 1 1 0 0 Group 4 23 13 8 1 0 0 1 Participant Groups Group 5 4 3 1 0 0 0 0 20-29 74 45 21 2 4 0 2 30-39 24 11 8 4 0 1 0 40-49 10 8 1 1 0 0 0 50-59 1 0 0 0 0 0 1 Age Unspecified 1 0 1 0 0 0 0 Female 90 53 25 6 3 1 2 Gender Male 20 11 6 1 1 0 1 Asian 14 6 6 0 0 0 2 Black 6 4 1 0 1 0 0 Hispanic 23 16 2 4 1 0 0 White 61 33 21 3 2 1 1 Ethnicity Other 6 5 1 0 0 0 0 41 The second hypothesis of the experiment was that age, height, weight and clothing description would greatly vary by participant. The results of the suspect’s age can be found in Table 4-4. Age was expected to be consistently younger than the suspect’s actual age of 33 years old because of the clothing worn during the crime. In fact, 58% of participants underestimated the suspect’s age, 28% gave an accurate range, and 6% overestimated his age. Three participants used adjectives, like “middle-aged,” to describe the suspect, so their results could not be included. Table 4-5 Questionnaire Results of Suspect’s Height Participants # Under Correct Over Unsure Unspec Adj All 110 23 29 34 6 14 4 Group 1 26 6 7 8 0 5 0 Group 2 38 4 12 16 2 4 0 Group 3 19 6 3 3 3 1 3 Group 4 23 6 5 6 1 4 1 Participant Group Group 5 4 1 2 1 0 0 0 20-29 74 12 19 25 4 10 4 30-39 24 5 9 6 2 2 0 40-49 10 6 1 2 0 1 0 50-59 1 0 0 0 0 1 0 Age Unspecified 1 0 0 1 0 0 0 Female 90 18 22 31 4 11 4 Gender Male 20 5 7 3 2 3 0 Asian 14 4 2 4 0 4 0 Black 6 0 1 4 1 0 0 Hispanic 23 6 6 6 2 3 0 White 61 12 19 19 2 6 3 Ethnicity Other 6 1 1 1 1 1 1 42 The expectation of the suspect’s height was that it would be underestimated, especially among female participants. The results of the suspect’s height, which is actually 5’10”, can be found in Table 4-5. In this case, the hypothesis was not supported as 21% of participants underestimated the suspect’s height, but there were 26% of participants that gave the correct range, while 31% overestimated the suspect’s height. Women underestimated the suspect’s height in 20% of cases and males underestimated in 25%, thus also refuting the hypothesis that women are more likely to underestimate the height of male suspects. Twenty five percent of female participants and 35% of male participants correctly answered the suspect’s height and 35% of females and 15% of males overestimated the suspect’s height. Four participants used adjectives, like “average,” to describe the suspect’s height so their data could not be included in the analysis. The expectation of the suspect’s weight, which is actually 190 pounds, is that it will be underestimated. This should be especially true for female participants. The results of the suspect’s weight are found in Table 4-6.

In the analysis of the data collected on the suspect’s weight, more participants that answered this question overestimated the suspect’s weight, thus refuting the hypothesis that witnesses are more likely to underestimate a suspect’s weight. In this analysis, 22% of participants underestimated the suspect’s weight, while 25% overestimated. The amount of participants that answered correctly was 6%, however 39% did not respond to this question., thereby greatly reducing the sample size. 43 For female participants, 19% underestimated the suspect’s weight, 28% overestimated, and 7% answered correctly. With male participants, 35% underestimated the suspect’s weight, 15% overestimated, and 5% answered correctly. The male participants seem to support this hypothesis, however this refutes the hypothesis that females are more likely to follow this trend than males.

Table 4-6 Questionnaire Results of Suspect’s Weight Participants # Under Correct Over Unsure Unspec Adj All 110 24 7 28 5 43 3 Group 1 26 6 1 5 0 13 1 Group 2 38 7 1 13 2 13 2 Group 3 19 5 1 3 2 8 0 Group 4 23 4 3 6 1 9 0 Participant Group Group 5 4 2 1 1 0 0 0 20-29 74 14 4 20 3 31 2 30-39 24 5 3 5 2 9 0 40-49 10 5 0 3 0 1 1 50-59 1 0 0 0 0 1 0 Age Unspecified 1 0 0 0 0 1 0 Female 90 17 6 25 4 36 2 Gender Male 20 7 1 3 1 7 1 Asian 14 3 1 2 0 7 1 Black 6 2 0 1 1 2 0 Hispanic 23 4 1 5 2 11 0 White 61 14 5 18 2 20 2 Ethnicity Other 6 1 0 2 0 3 0 44 Clothing Descriptions were expected to vary considerably. The data regarding clothing descriptions can be found in Table 4-7. Due to misattribution, witnesses may remember the clothing of a witness rather than the suspect. In the crime video, a witness chased the suspect and it was the witness who was the last image on the screen. The results are expected to be somewhat inaccurate due to this fact. Table 4-7 Questionnaire Results of Suspect’s Clothing Description Participants # Detail ed/ Accur ate Detaile d/ Somew hat Accura te Somew hat Detaile d/ Accura te Somew hat Detaile d/ Somew hat Accura te Somew hat Detaile d/ Not Accura te Not Detail ed/ Accur ate Not Detaile d/ Somew hat Accura te Not Detail ed/ Not Accur ate All 11 9 2 27 43 3 6 6 14 Group 1 26 2 0 5 9 2 0 2 6 Group 2 38 4 1 14 13 1 2 2 1 Group 3 19 0 0 3 9 0 2 1 4 Group 4 23 3 1 4 10 0 2 1 2 Participant Group Group 5 4 0 0 1 2 0 0 0 1 20-29 74 6 2 22 26 1 5 3 9 30-39 24 3 0 4 10 1 1 3 2 40-49 10 0 0 1 6 1 0 0 2 50-59 1 0 0 0 1 0 0 0 0 Age Unspeci 1 0 0 0 0 0 0 0 1 Female 90 6 1 24 34 2 5 4 14 Gender Male 20 3 1 3 9 1 1 2 0 Asian 14 0 1 2 5 1 2 1 2 Black 6 1 0 2 2 0 0 0 1 Ethnicity Hispani 23 2 0 4 11 2 1 1 2 45 White 61 5 1 17 23 0 3 3 9 Other 6 1 0 2 2 0 0 1 0 The largest percentage of clothing descriptions were categorized as “Somewhat Detailed and Somewhat Accurate” with 39%. This category means the participants provided two to three key elements pertaining to the suspect’s clothing, but at least one element was incorrect. However, 25% of participants gave a “Somewhat Detailed and Accurate” clothing description, meaning that the two to three key element provided were accurate. Fifteen percent of participant’s clothing descriptions were completely inaccurate. In many cases, the participant stated that the suspect was wearing shorts, a sweatshirt and a baseball cap, which was actually the clothing worn by the witness who chased the suspect. The third hypothesis relates to the photo lineup that was shown to each participant group. A witness, who appears in the crime video and chases the suspect after he steals the purse, was included in the photo lineup, but the suspect was not. Due to misattribution, it is estimated that 40-50% of the participants are going to choose the witness because he looks familiar. The results of the photo lineup are in Table 4-8 and the witness is “Suspect #1.” In fact, Suspect #1 was selected 39% of the time and the next highest photo selected was Suspect #4 with 9%. There were seven respondents that did not participate in this portion of the experiment, as it took place two days after the original video was shown. The participants were not expecting to view a photo lineup. The seven absent participants accounted for 6% of the total. Groups 1 and 2 showed greater misattribution than the other three groups. In Group 1, 65% of the participants selected Suspect #1 and 46 in Group 2, 39% selected the witness as the suspect. The next highest suspect selection was Suspect #2 in Group 1 with 4% and Suspect #4 in Group 2 with 8%. Group 3 and 4 respondents had the same number of participants select Suspect #1 as another suspect in the lineup, Suspect #5 in the case of Group 3 and Suspect #4 in the case of Group 4. Table 4-8 Questionnaire Results of Suspect Lineup Participant # Suspe ct #1 Suspe ct #2 Suspe ct #3 Suspe ct #4 Suspe ct #5 Suspe ct #6 Unsu re Abse nt All 11 43 9 7 10 5 5 24 7 Group 1 26 17 4 1 1 1 1 1 0 Group 2 38 15 0 0 3 0 1 17 2 Group 3 19 4 3 3 0 4 1 2 2 Group 4 23 6 2 3 6 0 2 1 3 Participant Group Group 5 4 1 0 0 0 0 0 3 0 20-29 74 24 4 6 8 5 2 19 6 30-39 24 15 3 0 1 0 2 2 1 40-49 10 3 2 0 1 0 1 3 0 50-59 1 0 0 1 0 0 0 0 0 Age Unspecifi 1 1 0 0 0 0 0 0 0 Female 90 35 8 5 8 3 5 20 6 Gender Male 20 8 1 2 2 2 0 4 1 Asian 14 8 1 3 0 1 0 1 0 Black 6 3 0 0 1 0 0 2 0 Hispanic 23 14 1 0 3 1 2 1 1 White 61 16 7 2 5 3 3 20 5 Ethnicity Other 6 2 0 2 1 0 0 0 1 The fourth hypothesis also related to the photo lineup. Groups 1, 3, and 4 were not given any specific rules regarding photo lineup, while Groups 2 and 5 were told that 47 the suspect might not be present in the lineup, these participants were instructed to answer “unsure” if they are not confident in their answers. This was to determine if participants would feel compelled to select a suspect from the lineup.

Groups 1 and 4 had 4% of the participants answered that they were unsure and Group 3 had 11%. The remaining 96% of participants in Group 1, 80% in Group 3, and 83% in Group 4 incorrectly selected one of the six photographs as the suspect. Group 3 had 11% and Group 4 had 13% of its participants that were absent for the photo lineup portion of the experiment. In Group 2, 45% answered “unsure”. Group 5 had 75% of participants answer unsure, however there were only four participants. 48 Chapter 5: Discussion, Conclusions, and Recommendations Discussion The first hypothesis had four separate elements that were analyzed. First, each participant was asked to specifically describe the crime they had just witnessed. It was expected that the crime details given would be accurate and they were mostly accurate with 86% of respondents providing accurate information. What was not counted on were the details, or lack of details, that were given. Only 12% of participants provided five or six key data elements to the crime, while 42% provided three or four and 33% provided two or less. One limitation in this area was the lack of personal contact with the witnesses to solicit more information. The study would have yielded more detailed responses had it been an interview scenario, which better mirrors an investigator’s attempt to gain details of a crime from witnesses. Another interesting result from the analysis of the crime detail summaries was that Participant Group 1 had 31% of participants accurately provide five or six key data elements in their descriptions. This was the highest percentage in Group 1 when every other demographic group was higher in less detailed, but accurate responses. Group 1 is the only group that was not a diverse population of graduate students at National University. Group 1 was composed of Teachers and Administrators at an elementary school in Southern California. This result could be related to the fact that teaching requires attention to detail, so persons in this occupation may be more naturally thorough and detailed.

In reviewing and analyzing the crime detail summaries, it was very interesting to see how different each summary was regarding the scene. As the focus switched from the family at the picnic table to the suspect walking into the screen, most participants 49 missed the fourth person sitting at the table. This person was mostly hidden from the camera’s view, so that is to be expected. However, many participants stated the family was playing a game or cards and missed the baby sitting on the table, but in the arms on one of the women. In several crime detail descriptions, the victim was described as elderly, even though the witnesses can only see the woman from behind, and one person even mistook the baby stroller for a wheelchair. Aside from the baby, the three people at the victim’s picnic table were all in their early 30s. The fourth person was in her 50s, however she was the person mostly hidden from the camera’s view. Another person mistook the purse for a backpack, which in itself is not interesting because purses sometimes look like backpacks. What was interesting in this case is that this participant described the victim as saying “Hey, he stole my backpack,” rather than “…he stole my purse.” When outlining the six key data elements of the crime, only the suspect’s actions and that of the witness that chased him were included. Had the information regarding the scene prior to the suspect’s arrival been requested, the accuracy of the descriptions would have been dramatically reduced. This could be in support of the Brown-Peterson Distracter Technique previously mentioned, as witnesses became distracted when the suspect entered the scene. There were no instances where the participant included completely accurate information regarding the events occurring at the victim’s picnic table prior to the crime.

The second element to the first hypothesis was the detail regarding the suspect spitting on the grass as he walked by the victim’s table. The suspect actor was asked to spit on the grass to see if witnesses would notice this critical piece of information, as a DNA profile can be obtained from saliva (Saferstein, 2001). Thirteen of the 110 50 participants mentioned that the suspect spit on the grass, however only one noted on the survey that this was potential DNA evidence. One reason so few participants mentioned the spit could be that it may not be commonly known that saliva is a source of DNA. Had the suspect touched a surface that preserves fingerprints, the number of participants that mentioned that detail may be significantly higher. Granted, in the scenario of the purse snatching, the police investigator may not tear up the grass and send the sample to the crime lab for DNA analysis due to the relatively minor aspects of the crime; however, if this had been a murder investigation, this would be a very crucial piece of information for investigators.

The third element of the first hypothesis was the accuracy of the suspect’s gender. As expected, 100% of participants correctly answered that the suspect was male. Had the suspect been a woman, but dressed in similar clothing as the suspect in this experiment, the results may have been different. This could be a subject of further research, but in this experiment, every participant was correct.

The fourth and final element of the first hypothesis was the accuracy of the suspect’s ethnicity. Although more than half of the participants, at 62%, correctly answered White as the suspect’s ethnicity, this was not as high as expected. One reason for this is that the suspect was wearing a long sleeved jacket, blue jeans, tennis shoes and a beanie cap. The only flesh shown by the suspect was on his face and hands. Hispanic, at 14%, and half Hispanic/half White, at 2%, were the other two ethnicities answered for the suspect’s ethnicity and this could be due the slight suntan the suspect had. Had the suspect been wearing a tank top showing a tan line between his natural skin color and his suntan, the results may have been more accurate. White participants were expected to be 51 more accurate than other ethnicity groups that participated. In fact, Black participants and those classified as “Other” correctly answered White as the suspect’s race in 83% and 75% of cases, respectively. White participants were consistent with the overall percentage at 62%, while Hispanic participants were correct 61% of the time. Asian participants were correct 43% of the time. The sample size for Asian participants was 14, Black was 6 and Other was 6. This sample size is too low to interpret this data that Black, Middle Eastern, American Indian and people who do not specify their race on questionnaires are more likely to accurately identify a White suspect. However, there were 23 Hispanic participants and they correctly answered the suspect’s race nearly as often as the 61 White participants. Therefore, the hypothesis that witnesses can more accurately determine the race of a suspect if it is the same as their own was not supported.

The second hypothesis also had four elements, suggesting that suspect’s age, suspect’s height, suspect’s weight, and suspect’s clothing would greatly vary. The first element of the second hypothesis to discuss is age. The age of the suspect was found to be mostly underestimated, at 58%, presumably due to the clothing he was wearing at the time of the crime. This result supports the hypothesis, which was based on the findings of a study (Sporer, 1996), where the target first acted aggressively while wearing a red windbreaker then changed into a nice leather jacket to play the role of a business associate. In the first scenario, the average age given by participants was 23.5, while the second scenario yielded an average age of 27.5. In this experiment, should the attire of the suspect been dressier the results would have most likely been different.

Sporer’s (1996) research also inspired the hypothesis regarding the suspect’s height and weight. These results were also expected to be underestimated, however 52 female participants were more likely than males to underestimate. According to Sporer’s research, witnesses tend to use their own height and weight as an anchor, which makes it difficult for females to judge the height and weight of male suspects and conversely males to judge females. The results in this experiment refuted the hypothesis in both height and weight. Overall, more participants overestimated the height and weight of the suspect, however this is where the similarity of the results ends. The suspect’s height was overestimated in 31% of cases and the next highest percentage was the number that answered correctly with 26%. Only 21% of participants underestimated the suspect’s height. Regarding female participants, the tendency was similar to the overall results, overestimating at 35%, answering correctly at 25%, and 20% underestimating the suspect’s height. The male participants answered correctly 35% of the time, underestimating 25%, and in 15% of cases overestimated the suspect’s height. This difference could support Sporer’s assertion that males use their own height and weight as an anchor when describing a male suspect, however since the participant’s height and weight were not captured in this experiment, it cannot be proven. A critical limitation regarding this analysis of gender-related data is the sample. There were 90 female and 20 male participants, which greatly misrepresents the population. Keeping this limitation in mind, consider the difference in the data collected on the suspect’s weight. While participants were more likely to be correct than underestimate the suspect’s height, very few participants correctly answered the question regarding the suspect’s weight. Only 6% of participants were correct on this question. Overall, 25% of participants overestimated the suspect’s weight and 22% underestimated, thus showing a very slight margin. However, when comparing male and female 53 participants, the margin is considerably more. While neither male nor female participants who accurately answered the questions of the suspect’s weight varied from the trend, with 5% and 7%, respectively, male participant underestimated the suspect’s weight at 35% and only 15% overestimate. Conversely, female participants overestimated the suspect’s weight at 28% and 19% underestimated. The male participants seem to support Sporer’s theory of underestimating weight, which also refutes his theory because he suggests women are more likely to underestimate. Again, the interpretation of the results is hindered because of the sample size. Additionally, 39% of participants did not answer the question of the suspect’s weight, even further reducing the already small sample size. However, Sporer did note in his research that inaccurate results are typical regarding height and weight. The final element of the second hypothesis is the analysis of the clothing description. It was hypothesized that the clothing descriptions would be mostly inaccurate due to misattribution of a witness that was seen chasing the suspect at the end of the video clip. As expected, the highest percentage of participants had crime detail summaries that were “Somewhat Detailed and Somewhat Accurate,” at 39%. This means they only provided two or three key data elements and were mistaken on some of the details. The next highest result, at 25%, was “Somewhat Detailed and Accurate.” As with the other group, they provided two or three key data element, however they were completely accurate in their descriptions of the suspect’s clothing. Most of the participants who gave incorrect details regarding the suspect’s clothing attributed an item of clothing that really belonged to the witness who chased the suspect after the purse was stolen. The witness, who was the last image seen running from the scene, was wearing 54 light colored shorts, a blue sweatshirt and a baseball cap. Many of the descriptions contained part of the suspect’s actual wardrobe combined with articles worn by the witness. One such clothing description said, “tan shorts, red flannel, dark shirt.” The red flannel and the dark shirt were worn by the suspect, however the tan shorts were worn by the witness. Sunglasses were cited as being worn by the suspect, even though no one in the scene, including the suspect, was wearing sunglasses.

The third hypothesis also studies the theory of misattribution. The witness who chased the suspect from the scene, who was also the last image seen on the video, was added to the photo lineup, but the suspect has been excluded. Although there was a 6% attrition rate, as the photo lineup experiment took place two days after the questionnaires were completed, 39% of participants selected the witness, Suspect #1. The suspect photo with the next highest selection percentage was Suspect #4 with 9%. Twenty two percent of participants answered that they were unsure if they recognize any of the photos as the suspect they witnessed steal the purse. This hypothesis was inspired by the Good Samaritan study done by Ainsworth (1995). To recap, two photographs appeared in a local newspaper, one photograph was of the suspect in a crime and the second was a Good Samaritan who thwarted the crime. Three participant groups viewed six-pack photo lineups one week after the newspaper article appeared. The first group’s lineup included the suspect and five other photos and the second group’s lineup included the Good Samaritan’s photo and five other photos, and the third group’s lineup did not include either the photograph of the suspect or the Good Samaritan. In the second group, nearly 50% of participants selected the Good Samaritan as the suspect, when the first group had nearly 40% selected the actual suspect. 55 Participant Group 1 had 65% of participants select Suspect #1 in the lineup. The proxy who administered the experiment to Group 1 is the wife of Suspect #4 in the photo lineup. It was underestimated the amount of participants that would recognize Suspect #4 as their co-worker’s husband. Only one of the 26 participants selected Suspect #4, but this changed the odds of selecting Suspect #1 from a 17% chance to a 20% chance, as Suspect #4 was automatically eliminated as the suspect in the minds of most participants. However, even with this limitation in Group 1, the misattribution theory has been supported in this theory.

The fourth and final hypothesis in this experiment also related to the photo lineup. Two participant groups (Groups 2 and 5) were instructed that the suspect may not be present in the lineup, so please answer “unsure” if they were not confident in their selection. The remaining three groups (Groups 1, 3 and 4) were not given this instruction but simply shown the photo lineup and asked to select the suspect from the video. Astoundingly, 48% of the 42 participants in Groups 2 and 5 answered “unsure”, while only 6%, four out of 68 participants in Groups 1,3 and 4, stated they were unsure of their suspect selection. This supports the theory that witnesses feel compelled to select a suspect, even if they are not confident in their responses. In this experiment, 100% of participants that selected a suspect chose incorrectly, as the suspect was not in the lineup. This illustrates the importance for investigators to properly instruct all witnesses prior to showing them any suspect lineup. The obvious limitation of this lineup is the lack of interaction between the person giving the lineup and the witness. There was no chance to interview the witnesses and judge their confidence level in their suspect selections. 56 Conclusion Witnesses can provide a great deal of information to investigators when a crime has been committed, but it must be understood where their limitations are. The mind is still such a mystery, that it is not easily explained why certain phenomena, like misattribution, could occur. Witnesses can be completely confident that their account of event is accurate, even when it is proven false. What has been learned is that investigators can work with witnesses to get the best and most accurate information without suggestion. There is a fine line between soliciting information and leading a witness along, as seen in the case studies above. For this very reason, guidelines have been set forth by the Department of Justice. Certain information provided by a witness is expected to be somewhat inaccurate, like height and weight estimations. Knowing these limitations can maximize the value of each witness. Also, knowing the limitations in the various methods of suspect lineups. Time and again, studies have proven that mistakes can and do occur. It is important that investigators continue to open-mindedly pursue leads in a case without focusing on one particular individual, simply because they were selected in a lineup. This would be an excellent place to begin further investigation, however it should not be the assumption of guilt without other more substantial evidence. Father Pagano simply had the misfortune of resembling the “Gentleman Bandit.” Most likely, the investigation became narrow focused so he did not have a chance. He was only saved by the confession of the actual robber.

57 Recommendations With advanced technology, it is becoming more apparent that eyewitness identifications are no longer enough. Although it may be difficult in some cases, it is recommended that cases with only eyewitness identifications and circumstantial evidence, which also lack scientific evidence tying the suspect to the crime, should not be tried. Learning from past failures is the only way to improve the system. In 138 cases of prisoners exonerated by The Innocence Project, two-thirds cited mistaken identity as one of the major factors in the original conviction. That means 92 innocent men that served time in a harsh prison for a number of years, partly because of witness’s mistakes. Due to the limitations of the experiment in this project, future research should be done. This project attempted to measure too many variables, but was compared to other studies where one or two variables were the focus. Because many of the hypotheses were not supported, additional studies should be conducted, but focusing on one hypothesis. Results of studies that determine if law enforcement officers make better witnesses have been inconsistent, but there should be additional studies to see if there are other occupations that provide good witnesses. As previously stated, the teachers and administrators in Participant Group 1 provided better crime detail summaries than the other, more diverse groups. A second recommended study would be on the hypothesis relating to the suspect spitting on the grass during the crime. This study should focus on subtle ways a suspect could leave evidence behind, including fingerprints, fibers, and DNA, to determine if witnesses could recognize the evidentiary value of the information. 58 There was another limitation in this project that many of the studies previously cited relied on witnesses actually describing or selecting a correct suspect. Because of the fourth hypothesis, the suspect was not actually in the photo lineup. Additional studies should be completed using photo or live suspect lineups where both the witness and the suspect are included. This could also test the hypothesis that witnesses feel compelled to select a suspect, even though they are not confident. A similar study should be conducted, giving specific information to some groups but not others, however the suspect should be present in the lineup. The more we learn about witness behavior and potential for mistakes, the better we can utilize the information to put only guilty parties in jail. Until we can eliminate wrongful convictions, all studies in this area are only serving to better educate us. 59 Appendix A: Questionnaire Participant Information Name: __________________________________ Age: ____ Gender: M F Race: ________________ Highest Education Level: ___________________ ************************************************************************ Crime Information Please describe the following (if extra space is required, please write on the back):

1. What crime occurred? _______________________________________________ 2. What weapons were used? ____________________________________________ 3. Describe the environment (indoor/outdoor, time of day, weather conditions, etc.) __________________________________________________________________ __________________________________________________________________ __________________________________________________________________ 4. Did the suspect(s) speak, if so, what was said?

__________________________________________________________________ __________________________________________________________________ 5. Did the victim(s) speak, if so, what was said?

__________________________________________________________________ __________________________________________________________________ 6. Suspect(s) Information a. Race ________________________b. Gender_______________________ c. Age ________________________ d. Ht/Wt________________________ e. Hair Color ___________________ f. Eye Color_____________________ g. Clothing Description (be as specific as possible, i.e. red jacket, blue hat) ____________________________________________________________ ____________________________________________________________ ____________________________________________________________ 60 References Ainsworth, P.B. (1995). Turning heroes into villains: The role of unconscious transference in media crime reporting. Paper presented to the 5 th European Conference on Law and Psychology, Budapest, 2 September.

Ainsworth, P.B. (1998). Psychology, law and eyewitness testimony. New York:

John Wiley & Sons.

Ainsworth, P.B., & King, E. (1988). Witnesses’ perceptions of identification parades. In M.M. Gruneberg, P.E. Morris, & R.N. Sykes (Eds), Practical aspects of memory: Current research and issues (Volume 1). Chichester: Wiley.

Atkinson, R.C., & Shiffrin, R.M. (1971). The control of Short Term Memory. Scientific American, 225, 82-90.

Bourtchouladze, R. (2002). Memories are made of this: How memory works in humans and animals. New York: Columbia University Press.

Brigham, J.C., & Cairns, D.L. (1988). The effect of mugshot inspections on eyewitness identification accuracy. Journal of Applied Social Psychology, 18, 1394-1410.

Brown, R. (1986). Social psychology (2 nd Edition). New York: Free Press.

Buckhout, R. (1980). Nearly 2000 witnesses can be wrong. Bulletin of the Psychonomic Society, 16, 307-310.

Chance, J.E., & Goldstein, A.G. (1996). The other-race effect and eyewitness identification. In S.L. Sporer, R.S. Malpass & G. Koehnken (Eds), Psychological issues in eyewitness identification. Hillsdale, NJ: Lawrence Erlbaum.

Cutler, B.L., Berman, G.L., Penrod, S., & Fisher, R.P. (1994). Conceptual, practical and empirical issues associated with eyewitness identification test media. In 61 D.F. Ross, J.D. Read, & M.P. Toglia (Eds), Adult eyewitness testimony: Current trends and developments. Cambridge: Cambridge University Press. Davis, G.M., Shepherd, J., & Ellis, H.D. (1979). Effects of interpolated mughsot exposure on accuracy of eyewitness identification. Journal of Applied Psychology, 94, 232-237.

Dictionary.com (2004). Retrieved June 21, 2004 from http://dictionary.reference.com.

Ellis, H.D. (1984). Practical aspects of face memory. In G.L. Wells & E.F. Loftus (Eds), Eyewitness testimony: Psychological perspectives. Cambridge: Cambridge University Press.

Ellis, H., Deregowski, J., & Shepherd, J. (1975). Descriptions of white and black faces by white and black subjects. International Journal of Psychology, 10, 119-123.

Fisher, R.P., & Geiselman, R.E. (1992). Memory enhancing techniques for investigative interviewing. Springfield, Ill: Charles C. Thomas.

Gay, L.R. (1996). Educational research: Competencies for analysis and application (5 th edition). Upper Saddle River, NJ: Merrill/Prentice Hall. Goldstein, A.G., & Chance, J.E. (1981). Laboratory studies of face recognition. In G.M Davies, H.D. Ellis, & J.W. Shepherd (Eds), Perceiving and remembering faces.

London: Academic Press.

Houts, M (1956). From proof to evidence. Springfield, Ill: Charles C. Thomas.

Light, L.L, Kayra-Stuart, F., & Hollander, S. (1979). Recognition memory for typical and unusual faces. Journal of Experimental Psychology Human Learning and Memory, 5, 212-228. 62 Loftus, E.F. (1996). Eyewitness testimony. Cambridge, Mass: Harvard University Press.

Malpass, R.S. (1981). Training in face recognition. In G.M Davies, H.D. Ellis & J.W. Shepherd (Eds), Perceiving and remembering faces. London: Academic Press.

Ratey, J.J. (2001). A user’s guide to brains. New York: Pantheon Books.

Saferstein, R. (2001). Criminalistics: An introduction to Forensic Science (7 th edition). Upper Saddle River, NJ: Prentice Hall.

Schacter, D.L. (2001). The seven sins of memory. Boston: Houghton Mifflin Company.

Shapiro, P.N., & Penrod, S.D. (1986). Meta-analysis of facial identification studies. Psychological Bulletin, 100, 139-156.

Shepherd, J.W., & Ellis, H.D. (1973). The effect of attractiveness on recognition memory for faces. American Journal of Psychology, 86, 627-629.

Shepherd, J.W. & Ellis, H.D. (1996). Face recall – methods and problems. In S.L.

Sporer, R.S. Malpass, & G. Koehnken (Eds), Psychological issues in eyewitness identification. Mahwah, NJ: Lawrence Erlbaum Associates.

Sporer, S.L. (1992, March). An archival analysis of person descriptions. Paper presented at the biennial meeting of the American Psychology-Law Society in San Diego, CA.

Sporer, S.L. (1996). Psychological aspects of person descriptions. In S.L. Sporer, R.S. Malpass, & G. Koehnken (Eds), Psychological issues in eyewitness identification.

Mahwah, NJ: Lawrence Erlbaum Associates.

Wells, G.L. (1988). Eyewitness identification. Toronto: Carswell. 63 Woodhead, M.M., Baddeley, A.D., & Simmonds, D.C.V. (1979). On training people to recognise faces. Ergonomics, 22, 333-343.

Wrightsman, L.S. (1991). Psychology and the legal system (2 nd edition). Pacific Grove, CA: Brooks/Cole.

Yuille, J.C., & Cutshall, J.L. (1986). A case study of eyewitness memory after a crime. Journal of Applied Psychology, 71, 291-301.