WITNESS TESTIMONY: THE COMMUNICATION 0F REMEMBERED INFORMATION Thesis for ”10 Degree of M. A. MICHIGAN STATE UNIVERSITY David Morgan Hanson I977 V ,.; *: “I 3‘0 ‘ . \7/ I Ii) "(I ABSTRACT WITNESS TESTIMONY: THE COMMUNICATION OF REMEMBERED INFORMATION BY David Morgan Hanson The legal system relies on many sources of infor- mation to ascertain fact. One important source is the eyewitness. The eyewitness provides information through the structure of the rules of evidence. The rules of evidence define who is capable of testifying, how ques- tions may be asked of the witness, what statements of the witness are relevant, etc. Underlying the rules are psychosocial assumptions made about the eyewitness. The literature of psychosocial critiques focuses on three important abilities of the eyewitness: obser- vation, retention, and articulation. Although observa- tion and articulation studies indicate weaknesses in these respective functions, the legal system cannot improve upon eyewitness capabilities without introducing fundamental changes. Alternately, retention abilities are under the direct control of the system through the deposi- tion procedure. David Morgan Hanson Retention studies provide lengthy albeit incom- plete data on witness retention. In studies for legal critiques which attempt to closely parallel witness func- tions, little control over the retention interval is provided. Psychological studies do provide greater con- trol of the retention interval but an important concern of generalizability must be faced. From the psychological studies on retention, three hypotheses are forwarded about the witness. These hypoth- eses examine decay and distortion of relevant witness information as a function of retention interval and camera distractions. The experiment is designed to test the relationships under restrictions of time control and external validity. The design requires subjects to witness an event and respond to a battery of recognition and reCall items at a designated time. The examination environment of the subjects parallels the deposition environment excluding the camera distraction condition. The subject responses are recorded on magnetic tape. Trained coders analyze the responses under a variety of tests. The tests include information quality, mental accuracy, response quality, and some basic measures of response content. Each test is conducted on each sub- ject response. David Morgan Hanson The first hypothesis states that as the retention interval increases, relevant witness information decays. An analysis of variance found that decay is significant at the .001 level. Hence, the results suggest that the witness does lose information as theories of memory would suggest. From the tests conducted, subjects could give 60 percent recall approximately two to twenty minutes after observing the stimulus, while at two or four weeks, recall was down to less than 40 percent. The second hypothesis states that as the reten- tion interval increases, witness testimonial distortion increases. Again an analysis of variance of the data finds that witness distortion increase over time is sig- nificant at the .012 level suggesting that witnesses progressively distort their recollections of relevant past events. At the immediate condition, 19 percent of the information is distorted whereas at two or four weeks distortion was up to 30 percent of the total information. Finally the third hypothesis states that as the retention interval increases, the hindering effects on total testimony by camera increase. Hindering effects are operationalized in two ways: by witness response quality decay and by response rate increase. An analysis of variance using the witness response quality decay gave a significance level of .999. Hence, the camera had David Morgan Hanson no effect on the way in which the witness responded. The second operationalization gave results contradictory to the hypothesis. The camera facilitated the rate of response and the facilitation was found to be significant at the .012 level. WITNESS TESTIMONY: THE COMMUNICATION OF REMEMBERED INFORMATION BY David Morgan Hanson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 1977 ACKNOWLEDGMENTS I wish to thank all the members of my guidance committee, Drs. G. R. Miller, R. V. Farace, D. C. Ralph, and M. L. Moore. Their counsel and support greatly aided me in completing this work. I especially wish to thank Dr. G. R. Miller for his trust and confidence in my ability to function as a researcher on his NSF grant. None of this would have been possible without access and use of the grant resources. I also wish to thank Dr. Farace for his interest in the development of my career. Thanks also goes to Dr. Norm Fontes for his wise counsel. Graduate students unwittingly fall prey to the "system" without proper guidance. His suggestions on maximizing the returns on my efforts are greatly appre- ciated. My deepest gratitude and love are extended to my family. My parents and brothers have been supportive of me throughout my education in many tangible and intan- gible ways. I wish to give special thanks to Care. Her love and prayers made this accomplishment more possible and meaningful than I alone ever could. ii TABLE OF CONTENTS LI ST OF TABLE S O O O O O O O I O O O I O O O O O 0 LIST OF FIGURES O O O O O O O I O O O O O O O O 0 Chapter I. RESEARCH PROBLEM . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . Theory of Law on Testimonial Evidence Rules of Evidence: Witness Qualifications Observation . . . . . . . . . . . . . Narration . . . . . . . . . . . . . . Recollection . . . . . . . . . . . Witness Recollection and the Theory of Memory . . . . . . . . . . . . . . . Witness Retention: Forgetting . . . . Witness Information Distortion . . . . Witness Recollection With Distractions II 0 METHODS AND PROCEDURES 0 O O O O O O O 0' 0 Overview . . . . . . . . . Independent Variables . . The Sample . . . . . . . . Procedure . . . . . . . . . . . . . The Examination Questionnaire . . Dependent Variables: Verbal Response Data Analysis . . . . . . . . . . . III 0 RESULTS I O O O O O O I O I O I O O O O 0 Introduction . . . . . . . . . . . . Witness Retention: Forgetting . . . Witness Testimonial Distortion . . . The Witness and External Distractions Camera Effects . . . . . . . . . . . . iii Page vi \D\IU'IU‘|Nl-‘ I-" 14 19 21 24 24 25 25 26 29 33 35 35 35 39 44 Chapter Page IV. DISCUSSION 0 O O O I O O O O O O O O O O O I 52 Summary . . . . . . . . . . . . . . . . . 52 Conclusions . . . . . . . . . . . . . . . 52 Discussion . . . . . . . . . . . . . . . . 55 Research Extension . . . . . . . . . . . . 58 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . 61 APPENDICES . . . . . . . . . . . . . . . . . . . . . 67 A. MATHEMATIC DERIVATION: MEMORY DECAY . . . . 68 B. MATHEMATIC DERIVATION: MEMORY DISTORTION . . 72 C. STIMULUS TRANSCRIPT . . . . . . . . . . . . 76 D. EXAMINATION QUESTION AND DEBRIEFING STATEMENT O O O O O O C O O O C O O O O O O 81 E. DEMOGRAPHIC QUESTIONNAIRE . . . . . . . . . 85 F O CODING SHEETS O O O O O O O O O O O O O O O 89 iv Table l. 2. 10. 11. 12. 13. 14. 15. 16. 17. 18. LIST OF TABLES Scale Description . . . . . . . . . . . . . . Summary Statistics: Relevant Witness Information . . . . . . . . . . . . . . . . ANOVA: Relevant Witness Information (3x2). . ANOVA: Relevant Witness Information (2x2). . Multiple Regression: Relevant Witness Information . . . . . . . . . . . . . . . . Summary Statistics: Witness Information Distortion . . . . . . . . . . . . . . . . ANOVA: Witness Information Distortion (3x2). ANOVA: Witness Information Distortion (2x2). Multiple Regression: Witness Information Distortion O O O O O I O O O O O O O O O 0 Summary Statistics: Quality of Witness Response 0 O O O O O O O O O O O O O O O 0 ANOVA: Quality of Witness Response . . . . . Multiple Regression: Quality of Witness Response 0 O O O O O I O O O O O O O O O 0 Summary Statistics: Total Witness Words/ Time 0 O O I O O O O I O O O O O O I O O 0 ANOVA: Total Witness Time . . . . . . . . . ANOVA: Total Witness Words . . . . . . . . . Summary Statistics: Witness Word/Time Ratio . ANOVA: Witness Word/Time Ratio . . . . . . . Multiple Regression: Witness Word/Time Ratio 0 O O O O O O C C O C O O O O O O I O V Page 30 36 37 38 39 41 42 42 43 45 46 46 47 48 48 49 50 51 LIST OF FIGURES Figure ‘ Page 1. Retention Curves (After Woodworth and Schlosberg, 1961) . . . . . . . . . . . . . 18 2. Schematic of the Testimony Room . . . . . . . 28 3. Information Scores vs. Retention Interval . . 37 4. Regression Analysis: Relevant Witness Information 0 O O O I O O O O O O O O O O 0 4o 5. Information Distortion vs. Retention Interval O O O O O I O O O O O I O I O O O 41 6. Regression Analysis: Witness Information DiStortion O O O I I O O O O O O O I O O I 44 7. Quality of Witness Response vs. Retention Interval O O O O O O O O O O O O O O O O O 45 8. WOrd/Time Ratio vs. Retention Interval . . . 50 vi CHAPTER I RESEARCH PROBLEM Introduction Recent publications question assumptions of eye- witness accuracy in testimony. Buckhout (1974), in a review of relevant literature, points out that subject error may cast doubts on eyewitness reliability reports. Eyewitness testimony provides an important infor- mation base for many judicial proceedings. If the infor- mation is suspect, the error potential in proceedings is greatly enhanced and questions of justice may follow. The purpose of this study is to examine the func- tions of the eyewitness. Specifically, eyewitness memory functions are examined. This study tests hypotheses of witness information stability and integrity over time. As this study looks at a well-defined role, the research design must conform to the restrictions imposed by the system. Eyewitness roles are found in the rules of testimonial evidence. Assumptions of witness capa- bilities are derived from the rules and are identified in this chapter. Next, a review of literature on previous empiri- cal critiques of testimonial evidence is presented. It focuses on the examination of the witness functions of observation, recollection, and articulation. The studies begin in the late 1800's and sporadically continue to the present. A representative literature is given. The weaknesses of the witness qualification of recollection are examined in a theoretical framework. The witness recollection limitations of forgetting, assimilation, and retrieval are examined in a communication/ information-processing mode. From research on these areas three hypotheses are proposed. Limitations of the previous research are given as part of a rationale for the present research to test these hypotheses. Theory of Law on Testimonial Evidence A popular view of the U.S. judiciary system is that the adversary arena exists to ascertain truth so that jus- tice may be rendered. Steinberg (1957) contends that the American common-law system of jurisprudence best discerns truth when partisan adversaries battle each other for vic- tory. When the eyewitness is involved, truth is arrived at from testimony. It is through the testimony that participants reason to conclusions for a verdict. Wigmore (1935) sug- gests that the search and use for testimony is facilitated by the rules of evidence. These rules provide the neces- sary structure for the proceedings. The structured notion of testimony is not univer- sal. The Romanesque system (complementary to the American) permits the witness the freedom to explain the event in his own way (Weinstein, 1957). Rules of evidence: an overview.--The American law system relies on an elaborate network of evidence rules (Wigmore, 1935). The network was developed as a reaction to the use of the jury system. It was thought that pro- tection against irrelevancies, biases, or fraudulent tes- timony was required to prevent jurors from being misled (Wigmore, 1935). Rules of evidence, then, provide the judicial system with criteria on which to oversee witness testi- mony. The rules pragmatically attempt to maximize the law's resources for the procurement of truth (Weinstein, 1957). A quantification of the rules is presented in Wigmore's Code of Evidence (1942). Categories of quali- fication, impeachment, testimonial rehabilitation, and parties' admission provide the basis for testimonial rules. Legal assessments of the witness' communication of remem- bered information are found in specific witness qualifi- cations of observation, narration, and recollection. Rules of evidence and psychosocial research.--The "pragmatic" attempts to maximize witness capabilities were seldom based on empirical studies. Wigmore's Code of Evidence, first published in 1910, neglected research findings presented in On the Witness Stand (Munsterberg, 1908). The legal community also showed little interest in Hutchins and Slesinger's (1928) review of literature that seriously questioned assumed witness capabilities. Although there has been a small but continuous flow of research on the psychology of the witness (Morris and Fishman, 1957), the rules of evidence appear to be impervious to change. Comparing The Revised Rules of Evidence (U.S. Congress, House Special Subcommittee on the Reform of Criminal Laws, 1973) to Wigmore's Code of Evidence (1942) shows little change in underlying assump- tions on eyewitness capabilities. The neglect of response to psychosocial critiques has been attributed to external validity problems of the research. It is argued that with lawyers seldom con- sulted on witness-related research (Weinstein, 1957), behavioral scientists don't have the means to properly design studies. Summarizing, rules of evidence were created in response to the need of controlling the witness from con- fusing the jurors. The rules were based on common- knowledge assumptions of human capabilities. Social science research has tested various assumptions and found them to be of questionable applicability. The legal community's response to these data has been less than favorable. The next section examines specific assumptions on witness abilities as found in the rules of evidence. Cor- re3ponding research findings are given and compared to the assumptions. Finally, for each category of assumptions, the witness function is examined to see if manipulation of the legal structure could improve witness abilities. Rules of Evidence: Witness Qualifications Observation Several qualifications of witness observation are indicated in the rules of evidence. If the observa- tion is arrived at through personal observation by rational means, then it is assumed that the observation is valid. "Rational" means may include impressions, and negative knowledge, i.e., testimony in denial (Wigmore, 1942). Physical presence is the necessary and sufficient condi- tion for personal observation. Credibility of witness' observation can be brought to question by an examination of sanity, reputation, and presence during the event (Wigmore, 1942). In other words, impeachment is arrived at by demonstrating that the wit- ness is biased, of bad character, or of strong disposi- tion for lying (Tierney, 1970). If it can be shown that the witness of sufficient character and intellectual stability is proximate to the occurrence, eyewitness requirements of the rules are satisfied. Research findings suggest alternate criteria may be of equal or greater importance. Whipple (1917) in review of European findings about witness functions found that the exactness of the observation depends not only on personal aspects of the witness (physical and mental condition), but also on spacial and temporal conditions above and beyond proximity. Specifically, Gardener (1933) found that size perception of objects is distorted when those objects are placed with objects distorted in size. Gardener also finds that the perception of velocity is subject to much distortion. Emotionally distorting factors are also seen to cause difficulty in the observation function. Munsterberg (1923) reports that those most upset by a staged incident involving a shooting were least accurate. Those less emotionally involved had a much clearer perception of the situation. This research suggests that the rules of evidence are insufficient to guard against observational limita- tions. The rules do not address the fact that reality may be distorted externally by perceptual ambiguity or internally by emotional stress. In actuality, rules of evidence take a contrary view to stress. Stress, according to the rules, facilitates information processing rather than impairs it. The question of applicability of research on observational limitations is an important concern. The structure of the legal system is almost totally indepen- dent of the eyewitness observation. Especially when the observation is spontaneous, the eyewitness function is situation Specific and although perceptual difficulties may very well hinder observation, little can be done through the manipulation of the system. Narration The witness function requires narration of what has been observed. The rules of evidence take into account the three ways the communication may take place: free narration, direct examination, and cross-examination. These ways describe the means by which the encoding pro— cess may be stimulated. The rules of evidence declare that the witness, through proper memory stimulation, can effectively commu- nicate the necessary information to the judge and jury; if not through free narration and direct examination, then through cross-examination: parsing out fact from fiction. The assumption is that if the memory has been stimulated by any means, the witness will provide useful information. However, the rules of evidence stipulate that the wintess' testimony will not include opinion (Wigmore, 1942). This rule assumes that either the wit- ness can separate inference from fact, or if opinion is given, it will be perceived as such by the opposing coun- sel and objected to accordingly. Central to the theme of narration is the impor- tance of cross-examination where the witness gets a true test of his knowledge. Contradictory findings by Marston (1924) indicate that the direct examination approach is more complete and accurate than the cross-examination. Jones (1969) also finds that the best method of interro- gation is the directive nonevaluative questioning (the direct examination). It is also assumed that manipulative questions are easily recognized by the opposition. Loftus and Zanni (1975) find that subtle manipulations of wording of a question put to a person about a recently witnessed event can affect a person's answers to that question. Finally, research data suggest that narration is a function of witness' verbal abilities (not independent of them as the rules would suggest). Marshall (1966) finds that the more education the subject has, the more verbal the response. Whipple (1917), in review of Euro- pean literature, finds that correctness of report of the observation depends on a facility with words and the num- ber of times the testimony was given. Rules of evidence develop an elaborate structure of examination procedures but fail to accurately assess the witness' capabilities of narration. Contrary to rules assumptions, direct examination is found to be superior to cross-examination. Further, the rules of cross-examination developed to safeguard against witness manipulation are ineffective in dealing with subtle manipulations by the proficient examiner. Research also shows weaknesses in the Opinion rule. Opinion, the subject evaluation of the witness, is often very difficult to separate from observed fact. With narration a function of witness abilities, a skillful orator could blend Opinion and fact to such an extent that parsing out fact would be almost impossible. As to system changes, the function of narration, unlike observation, is under control of the legal system. However, changing this part of the rules of evidence would change the entire adversary nature of the American system. From past hesitance by the legal system to incor- porate innovation, it is very questionable that sugges- tions of fundamental change would meet with any support. Recollection Several memory assumptions are made in the rules of evidence about qualifications of witness' recollection. Memory is assumed to be a faculty. Any recollection, including circumstances unconnected with the case, is 10 subject to scrutiny. An examination of this sort is com- mon and acceptable practice to measure the witness' capacity for memory (Wigmore, 1935). If it can be shown through any test of memory, no matter how irrelevant to the case, that there is some faultiness in a particular memory, then all memories of the witness become suspect. A further assumption about memory is that it may lie dormant but can be refreshed or stimulated for recol- lection. Refreshing takes place when the memory is jogged into recall by any stimulus, relevant or irrelevant to the case. Stimulation of the witness to nervous excitement is assumed to produce spontaneous and sincere responses (Wigmore, 1942). The rules assume that memory functions indepen- dently of whatever caused the memory to be refreshed and the chain of associations that follow, though previously forgotten, will lead to the apprOpriate memory (Hutchins and Slesinger, 1928). Rules of evidence also make assumptions about the manipulation of the memory. Due to the adversarial nature of the proceedings, the legal system provides safe- guards against memory manipulations by either side. One rule specifically prohibits partisan suggestions or instruction before the case comes to court (Wigmore, 1942). However, explorations of stories by witness-attorney ll interactions prior to the trial are both permissible and common (Weinstein, 1957). As previously noted, misleading or intimidating questions are not permissible in that the witness' memory could be misled. Misleading questions are defined as questions implying testimony of unfavorable matters not yet admitted by the witness. Intimidating questions are those which seek to disconcert the witness to make the testimony appear less trustworthy through confusion. Beyond these manipulations of the partisan adver- saries, few, if any, limitations are placed on random or other external distortions. The rules assume that if no manipulations were attempted by the interested parties, the memory may decay but not distort. Research into witness functions again arrives at contradictory findings. Memory, as a faculty, should be able to be developed. Slight (1911), in a related study, found that practicing the learning of one set of syllables in no way improved the ability to memorize in other cases. Psychology long ago abandoned the notion of memory as a faculty due to lack of empirical justification. Memory manipulation within the legal process is assumed to be only in danger when adversaries attempt to manipulate. Other influences lack importance. Bird (1927) found that if a distorted account was provided to the witness after the event, the witness would distort 12 his testimony accordingly. Loftus (1975) in a similar procedure could manipulate the memorial representations by introducing new knowledge through the use of leading questions given immediately after an event. The recollection procedure also requires that only facts (not inferences) will be provided. Marshall (1966) found that when subjects were asked to make recol- lections, they made a great number of inferences to the number of correctly recalled items, with ratios ranging from 40 percent to 70 percent. The rules of evidence make several unsubstan- tiated assumptions about witness recollection. The assumption of memory as a faculty has long ago been abandoned with the notion of refreshment by psychology. The rules neglect memory manipulation by anything other than the adversaries discounting important albeit random external distorting factors. Unlike the witness functions of observation and narration where the legal system either cannot or will not change its structure for the potential improvement of witness capabilities, the recollection function is amenable to system manipulation. When the retention interval (time between observation and testimony) is under a great deal of control by the court, it becomes a practical concern to study the effects of time on the quality of testimony. 13 The issue of retention over time is a central concern of the psychology of memory. As this area pre- sents a great deal of empirical and resultant theoreti- cal work, it is important to identify findings which correspond to witness memory functions. As testimonial integrity is a key concern, the major focus of the literature review is on memory decay, distortion, and manipulation through distraction. From this literature review, hypotheses on witness retention are generated. Witness Recollection and the Theory of Memory The witness can be seen as a source of informa- tion. It has been argued that one of the basic purposes of communication is the transmission of information. The sender may rely on several sources of information for latter transmission: externally through immediate sense experience or internally through the mind. The mind may act as a storer of information (memory) or as an initiator of information (ideas). When the communicator acts as a witness, he is required to rely on the mind only for stored information. Mortensen (1972) identifies this aspect of commu- nication as part of the intrapersonal system. In this conceptualization, primary communication does not exist with the dyadic interaction. Communication may go on 14 within the individual. The intrapersonal communication is concerned with the memory function. The intrapersonal system is now examined through theory and empirical findings of psychology on memory. The examination will be limited to two areas of change within the long-term memory: (1) forgetting and (2) reor- ganization (Posner, 1973) and the process of bringing forth material for recall: retrieval. Witness Retention: Forgetting A crucial hindrance to the witness' performance results from forgetting. When information cannot be retrieved, the information has been forgotten. Central to the notion of forgetting is the types of memory stores for information. Broadbent (1958) and Brown (1958) suggest that memory may be broken down into two systems: short-term memory (STM) and long-term memory (LTM). A structural-functional viewpoint would identify them as structural features (Klatzky, 1975). A system approach would identify them as control processes (Atkinson and Shiffrin, 1968). Keele (1973) presents strong evidence for the identification of two memories. Differences in the sys- tems are suggested through brain injury cases (Corkin, 1968; Shallice and Warrington, 1970). Peterson, Hillner, and Saltzman (1962) present other evidence showing 15 different effects from spacing and practice. Finally, from studies of Underwood (1965), Kolers (1966), and others, evidence suggests that the nature of the repre- sentation of the information differs. Short-term memory is described by Posner (1973) as the system containing items in the active (conscious) state. Only a few items can be maintained due to a limited capacity. Information may be maintained in the STM indefinitely through rehearsal. However if distrac- tions occur, information is lost within twenty seconds (the unrehearsed duration of the STM) (Keele, 1973). With such limited capacity and duration, this store is of little use to the witness. The more permanent store for information (the store relied on by the witness) is the long-term memory (LTM). It has been suggested by Posner (1973)-that the STM acts as a filter where information is made compatible for long-term storage. Klatzky (1975) suggests that this filtering provides for comprehension: for pattern recog- nition, sensory register information must be matched with information in the LTM. This process requires a great reduction in information so that the LTM takes in only a fraction of the original perception. In terms of capacity, however, Penfield (1959) and others suggest that every piece of information stored will always remain. 16 Tulving (1972) suggests that storage may take place in two different ways: semantically (meanings, rules, etc.) or episodically (autobiographical informa- tion). The episodic store, the most relevant store for the eyewitness, provides both an acoustic and visual representation (Klatzky, 1975). Episodic storage was found by Atwood (1969) to be most successful when images had an organizational framework. In other words, episodic storage necessitates associations. Keele (1973), from a study by Reed (1970), notes that abstraction of information in LTM appears to be common to both verbal and visual material. With much of the original input lost, the resultant product greatly economizes processing demands. Long-term memory for both semantic and episodic storage requires associations which make storage of the material possible. The concept central to long-term storage then is association. LTM, as a network of associated bundles of information, is compatible with S-R theory (Klatzky, 1975) and classical conditioning by paired response (Pavlov, 1927). Many information-processing theorists (Anderson and Bower, 1973; Quillian, 1968; Rumelhart, Lindsay and Norman, 1972) have also adopted association as the best representation for memory (Klatzky, 1975). The associated bundles, "memory cells" by Posner (1973), are the products of past experience. Slamecka 17 (1966) finds that these old associations are highly resistant to anything but superficial change. Thus, Adams (1967) and Paivio (1969) find that internal and external perceptual events and the memories they acti- vate are associated together into what Posner and Keele (1968) call the prototype of the central or average tendency of past and present related stimuli. It is from this framework that long-term for- getting takes place. Previously it was stated that theorists such as Penfield believe that all the informa- tion stored in the LTM is still there. Forgetting in the LTM occurs when information is misplaced (not lost). Klatzky (1975) suggests that information in episodic memory becomes easily inaccessible with new information continuously entering. Ebbinghaus (1913) found that information dis- placement is systematic over time. It was through his pioneering work in memory loss that the original "curve of forgetting" was surmised. This curve suggests a curvilinear decay of information over time. Replicated findings are indicated in Figure l. The "curve of forgetting" is compatible with the previously discussed theories of memory. The transitory nature of the short-term store and the great information reduction from short-term to long-term memory suggests fast decay rate. As the information is associated, it 18 loses its individual nature and with new information entering, the decay continues but at a less dramatic rate. (For a theoretic derivation of this process, see Appendix A.) OBoreas. 1930. poems 0 Boreas. 1930. syllables AFlnkenblnder. 1913 x Luh. 1922 E ‘13 60 _ OEbbmghaus. 1885 a S no 540- E: \ ‘0 \\ 2 . mun mm "' m _ ~ 0 C .. 205 E . m m vie >. a; '9 :33'?‘ O 5 E E L s. L. 5‘: m mm (D U'U'DU \ E g o o .c 1: 1: EN o ‘00 u ”do \ o .— N H N v (ID—l... NM sol ~NM8 \.... N 0L 1 1 1 1 11 L11 1 111 11 \t l ‘4 2 3 4 5 6 n l a time in minutes Figure l.--Retention curves (after Woodworth and Schlosberg, 1961). Properly functioning in the role, the witness provides testimony through tapping memory of stored information about the event. According to the associa- tive theory of memory, witness information is subject to the same processes as any other information. As the witnessed information is sent to the long-term memory, large quantities of information are lost. The informa— tion that does get stored in the LTM is subject to decay through the dynamic process of continual assimilation. The external validity criticism against the information decay findings by the legal community is, to 19 a certain extent, shared by psychology theorists. Klatzky (1975) questions the approach of the testing the LTM through the use of long lists of verbal material. It can be argued that remembering lists of words is fundamentally different than remembering human interac- tion. To find if the generality holds to witness memory capabilities, this study will test the following hypothesis: H1: As the retention interval increases, rele- vant Witness information decays. Witness Information Distortion The previous section indicated that forgetting takes place through assimilation. The witness then has only a limited account of the interaction. This limi- tation leads to a fundamental discrepancy with the wit- ness role. The witness role requires strong mental representations of past witnessed events. In actuality, the mental representation is incomplete. Marshall (1966) contends that this discrepancy is reduced by assimilation to provide a complete picture. The process of discrepancy reduction is facili- tated by the unconscious nature of assimilation. Assimi- lation goes unnoticed because the memory is spontaneously active and changing (Koffka, 1935). Distortion through assimilation then is the second important limitation to 20 be considered. Distortion exists when there is any variance of the report to the facts. When the witness provides information which does not correspond to the actual event, the witness is giving a distorted report. It is suggested that assimilation is subcon- scious and natural because it plays an integral part on making sense out of the environment. Klatzky (1975) considers learning to be the process of adding to and modifying the memory system. Learning to cope with the environment requires assimilation. It is found that assimilation takes place in both the semantic and episodic store. Ceraso (1967) finds that delay in the retention interval causes material to lose its separate identity and merge with other informa- tion. Posner (1973) suggests that there is an inability to maintain boundaries for learned information. A study which appears to have a fair amount of external validity is found in the work of Bartlett (1932). Bartlett had subjects read a short essay, "The War of the Ghosts." Subjects gave accounts which tended to rearrange the story into a more normal pattern. It appeared that the mental representation of the story was assimilated into the subject's knowledge structure (Klatzky, 1975). Summarizing, information is processed and stored in memory through assimilation. Any information is sub- ject to distortion from associations of past experience 21 as well as any future related associations. As the information bundles develOp, they become resistant to change. Associations strengthen and stabilize. (For a theoretic derivation of this process, see Appendix B.) As noted earlier, the rules of evidence command the witness to only give factual testimony. The theory of memory on association would suggest that this is an impossible task. According to the theory, information can only be stored if it is associated with information bundles. As isolation boundaries are not substantiated in research findings, witnesses cannot accurately main- tain the facts free from distortion. As associations continue over time, the witness' information should dis- tort accordingly. The following hypothesis will be tested to deter- mine the effects of time to witness testimonial distor- tion: H : As the retention interval increases, witness testimonial distortion increases. Witness Recollection With Distractions Finally, when the witness is required to recall information, he must search through his memory and revive (retrieve) the necessary information through the organi- zation (Eagle and Leiter, 1964). Retrieval may be defined as the process of long-term memory activation of 22 stored information for reorganization and generation of an output (Posner, 1973). Retrieval may vary from almost automatic responses requiring no conscious effort to responses requiring great effort (sometimes with no results when the infor- mation is lost). Effort is most commonly experienced by the TOT (tip-of-the—tongue) phenomenon. TOT research by Brown and McNiel (1966) shows that peOple go through a systematic search for materials. This search requires both effort and concentration. When retrieval requires effort, distractions may make the necessary attention for recall difficult, if not impossible. Various studies (Fitts and Seeger, 1953; Fitts and Jones, 1961) have found that retrieval is a function of S-R compatibility. In the case of the wit- ness, the compatibility is between the interrogation question to the witnessed event. With retention inter- val increases, information assimilation increases and thus the S-R compatibility weakens. The legal community has shown interest in the use of video tape in testimonial procedures. The use of the camera to take the video portion of the testimony may hamper witness response. The camera may be distract- ing enough to hinder the retrieval process. 23 The final hypothesis tests for the effects of the camera as a distraction on witness response: H - As the retention interval increases, the 3 hindering effects on total testimony by camera increase. CHAPTER II METHODS AND PROCEDURES Overview The experimental design for testing the hypotheti- cal relations was a 2x3 factorial, independent groups design: two conditions of camera/no camera by the three retention interval treatments. The retention interval was trichotomous with an immediate condition, a two-week condition and a four-week condition. The subjects viewed a video taped marital argu- ment. A deposition of the stimulus was taken immediately or at the appropriate delayed time. When the deposition was taken, approximately half the subjects testified in front of a camera (the camera treatment). The deposi- tion was taken in a classroom devoid of many distrac- tions. After the testimony was given, the subjects were debriefed and then filled out a demographic ques- tionnaire. The testimony of the subjects was recorded on cassette or video tape. This permanent record of the responses was content analyzed. Each answer, requiring recognition or recall, was processed through a battery of analyses by trained coders. 24 25 Independent Variables Previously it was noted that retention interval was of hypothetically critical importance to testimony. The retention interval, the time between the event and the testimony, was manipulated by increasing the duration between these two occurrences at specified times. It was also hypothesized that with recall requir- ing attention, distractions would hinder accounts. An important way distractions can take place is through the use of video cameras for recording the testimony. Video procedures are of great interest to the legal community as a means of expediting the various processes. If, however, the testimony is unduly restrained, the camera is of little benefit. To test for the effects of camera on the witness, the camera condition was defined Operationally as the situation in which the witness is close to (ten feet) and directly in front of the camera. This effect was iso- lated with the testimony room containing few other dis- tracting features readily visible to the testifying witness. The Sample One hundred three subjects were drawn from the greater Lansing area. The subjects were primarily mem- bers of various P.T.A. organizations. Ten subjects were 26 'not P.T.A. members. A ten or fifteen dollar donation (depending upon the condition of immediate or delay) was made to the organization. The independents were personally paid. The subjects were randomly assigned to the various conditions. The mean age of the entire sample was thirty-five. Sixty-two percent of the sample was female. Most of the subjects were married. Procedure Three rooms were used in the study: the waiting room, the stimulus room, and the testimony room. After arriving, the subjects were told to be seated in the waiting room. When the stimulus was ready, they were escorted to the stimulus room, a classroom with a student capacity of thirty-five. As many as ten and as few as one viewed the stimulus. The following instructions were given: Today you will be seeing a marital argument. Please do not be offended by any language used. It is important that realism exists and, let's face it, some peOple do talk this way when they are angry. For the purpose of this study, pretend that you are the (brother/sister) of the husband. You just dropped in unexpectedly and are able to watch the fight unnoticed. At this point the subjects viewed the video tape on a monochromatic monitor. The stimulus involved a marital argument. Walter, a heavy-set man in his forties, comes home at 8:00 in the morning to a waiting wife, Louise. An argument 27 develops and such topics as money, the baby, and each other's socializing provide the volatile ingredients for Louise to push and hit Walter. The stimulus was presented on video tape to control for variations which would have occurred with repetitive live presentations. (For the transcript, See Appendix C.) Following the completion of the stimulus, further instructions were given: This quarrel proves to be the last straw and your sister-in-law brings charges of assault and bat- tery against her husband. You, of course, are an eyewitness and will be asked to give testimony when called upon. Immediate report: Please return to room 102 (the waiting room). Please do not talk to anyone about what you just saw. After all, it is you and only you who have seen the fight. This is extremely important research and we need to know what you think. Delay condition: (Person organizing the group) will indicate when you should return to complete the study where you will be asked to testify. Please do not talk to anyone who has or will be seeing the tape about this. After all, it is you and only you who have seen the fight. This is extremely important research and we need to know what you think. The subjects (immediate or returning) then waited until called on to go to the testimony room. The waiting room was a classroom with a student capacity of thirty- five. The stimulus room was a carpeted classroom with a seating capacity of fifty. The stimulus room differs depending on the camera condition. In the "no camera" 28 condition, the camera, the T.V. monitor, and the VTR were not present. (See Figure 2.) \g—fii 'm T__1 30' r--—-‘ subject \LWV a x _—_| 10' fi/Zwa § '--L— examiner. camera JL -11—11r11-1g:1r11-jgr-:L k———23' —-l Figure 2.--Schematic of the testimony room. The testimony room was controlled for other dis- tracting stimuli. The subject entered the room and after passing by four experimental personnel, was seated adjacent to the examiner and the camera (for the camera condition). After the subject sat down, a microphone was placed around the subject's neck. In the camera condition, the examiner informed the subject that s/he was being video taped. 29 The examiner then proceeded through the examina- tion question and then debriefed the subject. (See Appendix D for the examination questions and the debrief- ing statement.) The subject returned to the waiting room, filled out a questionnaire, and then was dismissed. (See Appendix E for the questionnaire.) The Examination Questionnaire Questions for eliciting eyewitness testimony should be brief and to the point for the procurement of facts (Heller, 1968). Facts of sequence are of primary importance to cases of argument and assault. It is very relevant to find out who spoke or struck first (Marshall, 1966). The questionnaire was designed to determine who said what; or where and when the particular interaction took place. Every attempt was made to keep the examina- tion free from leading questions. Specific information was ascertained through a sequence of contingent questions. The first question in the set tested for recognition: did the witness have any memory about the occurrence. If the witness failed to respond affirmatively, no other questions of clarifi- cation were asked (about unknown facts). If the witness responded affirmatively, Open-ended clarifying questions were asked producing recall and/or assimilation. 30 Dependent Variables: Verbal Response Retention interval and distracting camera effects were manipulated to test for various differences in tes- timony. The variety of analyses on the witness report included the total amount of correctly recalled informa- tion and the total amount of assimilated information present in the testimony. All testimony was recorded on cassette tapes or transcribed on cassette tapes from video tape. From this record of the proceedings, coders assessed each answer following the scale description scheme contained in Table 1. Table l.--Scale description. Part A: Information quality The information quality of the response will be judged on a five-point scale. This Likert scale will look at positive and negative aspects of information. Positive information will be defined as that information which conforms to the actual events. Negative informa- tion will be defined as that information that does not conform to the objective events but originates from a different source. Information quality will be operationally defined in the following manner: L l I T -2 -1 0 +1 +2 31 Table 1.--Continued. Description of the scale points: -2: a wrong "manufactured" statement having no relation to any part of the stimulus -1: discussion of an aspect of the stimulus not directly relevant to the question 0: stating a lack of knowledge or no report +1: a vague but correct assessment of the answer +2: a strong assessment of the objective facts Part B: Accuracy of mental ability Accuracy of mental ability, when positive, will be defined as a true assessment of the mental state as it interacts with objective reality. Inaccurate mental ability is defined as a false assessment of the facts where apprOpriately no response should have been given. Accuracy of mental ability is operationally defined the same as information quality except that a statement of lack of knowledge (being a correct or safe assessment) is given a +2. ‘ Part C: Response quality Response quality will also be judged on a five- point scale. This Likert scale will be used to judge the quality in the response from a low quality response of "one" to a high quality response of "five." Response quality will be defined as those vocal behaviors both verbal and nonverbal which would indicate how the source responds independently of the information conveyed (i.e., the face value). Response quality will be operationally defined as follows: 32 Table 1.--Continued. +1: low quality of response--long pauses, very unsure, contradictory or statement of lack of knowledge +2: long pauses, confused, appears to have little con— fidence ‘ +3: midpoint of quality--exercising caution, choosing words +4: characteristics of "+5" but not as quick or soft spoken +5: high quality--strong, domineering, confident, clear, crips, concise, sure, direct, quick to respond Miscellaneous Other measures will be given including: the time of response, the number of words used in the response, the number of nonverbal vocal utterances, the number of negative references to the cognitive state, and the num- ber of positive references to the cognitive state. These measures are operationally defined in the following manner: The time of response--the number of seconds it takes from the end of the examiner's question until the end of the response by the witness The number of words used in the response--anything that was said including questions of clarification The number of nonverbal vocal utterances--any utterances such as "um" or "ah" etc. The number of negative references to the cognitive state-- the total number of times during the testimony where the witness made the statement such as "I don't know" or "I don't remember" The number of positive references to the cognitive state-— the total number of times during the testimony where the witness made statements such as "As I remember" or "I think" etc. 33 Data Analysis The dependent variables of witness testimony (relevant eyewitness testimony, witness testimonial dis- tortion, and total disclosed information) were Opera- tionalized through previous coder scaling procedures. The source of witness testimonial data is limited to particular questions. Questions testing recognition do not elicit much information. The primary use of these questions was for introduction to questions of recall that test for knowledge specifics. Therefore, only questions eliciting recall were used in the analysis. Relevant eyewitness testimony (the test of retained knowledge) was determined through a sum of posi- tive scores on code scheme A: "Information Quality." With distortion identified as a type of forgetting, nega- tive scores were recoded to zero for this test. Witness testimonial distortion (the total amount of learned assimilated information) was found through a sum of negative scores on recall items from code scheme A. With correct information showing no distortion, positive scores were recoded as zero for this test. Total disclosed information (the "Openness" of the witness) was found by the total number of words used in testimony. Tests for memory accessibility may be found through a ratio of the number of words per unit 34 time and through the coders' evaluation of the quality of the response to recall items (code scheme C). Hypothesized curvilinear relationships of rele- vant eyewitness testimony and witness testimonial dis- tortion to retention interval were tested by regression procedures. The correlation and significance level are reported in conjunction with the first two hypotheses. . Analysis of variance procedures were used to test the two-way design (camera by retention interval). The analysis of variance tests were used on each of the three dependent variables. The third hypothesis was also tested through this procedure. CHAPTER III RESULTS Introduction The results have been divided into three sections. The first section is concerned with the first hypothesis testing for the effects of retention interval on witness information retrieval (forgetting). The second section examines the second hypothesis: the effects of retention interval on distortion. The third section is concerned with the effects of the distracting stimulus of the camera. The hypothesis suggests that witness content is hindered by the effects of the camera and that these effects become more pronounced as the retention interval increases. Witness Retention: Forgetting H1 states that as the retention interval increases, relevant witness information decays. The hypothesis was tested by the measure of the correct information con- veyed with respect to the six recall items. With a completely correct item worth two, and distortion or lack of knowledge coded as zero, the scores could range from zero to twelve. The range of scores reported gave a maximum of twelve at the immediate recall condition to a 35 36 value of one at the four-week delay condition. A sum- mary of the results is shown in Table 2. Table 2.--Summary statistics: relevant witness information. Treatment (Retention Interval) Conditions 1/2 hour Two Weeks Four Weeks (n=20) (n=20) (n=12) ié7.75 §é4.35 253.67 No camera S=1.9l6 S=l.725 S=2.146 +%=65.0 +%=36.3 +%=30.5 (n=l9) (n=l4) (n=l4) ié7.65 ié4.86 354.86 camera s=2.43 S=l.46 S=2.80 +%=58.8 +%=4o.5 +%=4o.4 (n=39) (n=34) (n=26) Both 2e7.41 ié4.56 224.31 +%=60.0 +%=38.0 . +%=36.0 The decay of witness information over time is evidenced in Figure 3 on means of information retention. An analysis of variance on witness information indicates that the decay is significant in support of the first hypothesis. (See Table BJ .Approximately 33 percent of the variance was explained by the retention interval. 37 no camera camera both 8 I I I r Mean Relevant 6 r Information fi.___l 1_.o “ Scores 4 :::::;_____;:_D x (% correct) A a 2 l + 0 A I. 0 H 2 3 21 5 Retention Interval (Weeks) Figure 3.--Information scores vs. retention interval. Table 3.--ANOVA: relevant witness information (3x2). Source of Sum of df Means of F Signif. Variation Squares Squares of F Main effects 209.1 3 69.7 15.68 .001 Time 207.6 2 103.8 23.36 .001 Camera 1.1 l 1.1 .24 .999 Two—way interactions 15.0 2 7.5 1.68 .190 Residual 413.34 93 4.5 Total 617.4 98 6.5 Variance explained: R2 = .328. 38 It is also important to note that with the decay, there was relatively little difference between the two long-term delay conditions at two and at four weeks. By collapsing the retention interval of the long-term delay conditions to immediate/delayed, a twoéway interaction of time and camera showed significance at the .09 level. (See Table 4.) Table 4.--ANOVA: relevant witness information (2x2). Source of Sum of df Means of F Signif. Variation Squares Squares of F Main effects 207.99 2 103.99 23.70 .001 Time 206.51 1 206.51 47.10 .001 Camera .86 l .86 .20 .999 Two-way interactions 12.58 1 12.58 2.82 .090 Residual 416.84 95 4.39 Total 637.41 98 6.504 It has been shown previously that age and sex may influence witness report. To determine the effects of these factors, a multiple regression was run with the inclusion of these two variables. The criterion variable was the relevant witness information. The predictor variables were age, sex, time, and camera. From this 39 analysis, a regression model was drawn to show the coef- ficients on the witness information. (See Table 5.) Table 5.--Mu1tiple regression: relevant witness information. Variable B Std. Error B F/Significance Beta Age -.0297 .0187 2.52/.ll6 -.137 Sex -.621 .458 l.84/.l79 -.118 Camera -.310 .440 .50/.482 .051 Time -.862 .138 38.76/.000 .543 (Constant) 8.42 .863 The coefficients give weightings for the four predictor variables. The demographic variables of age and sex showed weights of -.137 and -.188, respectively. The distracting variable "camera" showed a weight of -.061. The only predictor variable of any magnitude was the retention interval (time) with a weight of -.543. (See Figure 4.) Witness Testimonial Distortion H2 states that as the retention interval increases, witness testimonial distortion increases. The hypothesis was tested by the measure of the distorted information conveyed with respect to the six recall items. A com- pletely distorted response (information bearing no resem- blance to the stimulus) was worth a -2. Information given 40 which was relevant to the stimulus but of no applicabil- ity to the recall question was worth -1. Other responses (or lack of response) were scored as 0. AGE (-.l37) SEX (-.188) ““-~..\* WITNESS KNOWLEDGE / (-.06l) CAMERA l..—~——"”' (-.543) TIME Figure 4.--Regression analysis: relevant witness information. The scores could range from a minimum of zero (no distortion) to a maximum of -12. The range of scores reported gave a minimum of 0 in the immediate condition, and a maximum of ~11 in the four-week delay condition. (See Table 6.) The increase of witness distortion over time is evidenced in Table 6 of the means. The analysis of variance indicated that the increase was significant at 41 no camera camera both Mean Witness Testimonial Distortion (% distorted) -2 w 01' 1 ll 2 ' 3 4 5 Retention Interval O I‘d-1r Figure 5.--Information distortion vs. retention interval. Table 6.--Summary statistics: witness information distortion. Treatment (Retention Interval) Conditions 1/2 hour Two Weeks Four Weeks (n=20) (n=20) (n=12) —=-2.35 §=-3.35 §=-2.75 N° camera 3: 1.83 s: 1.73 S: 1.29 -%= 20 -%= 28 —%= 23 (n=l9) (n=14) (n=14) §=-2.21 22-3.71 §=-4.36 camera s= 2.37 s= 2.23 s= 2.53 -%= 18 -%= 31 -%= 36 (n=39) (n=34) (n=26) Both i=-2.28 §=-3.so ’2-3.62 -%= 19 -%= 29 -%= 30 42 the .003 level in support of the second hypothesis. (See Table 8.) Approximately 10 percent of the variance was explained by the retention interval. Table 7.--ANOVA: witness information distortion (3x2). Source of Sum of df Means of F , Signif. Variation Squares Squares of F Main effects 44.1 3 14.70 3.52 .018 _Time 38.6 2 19.30 3.62 .012 Camera 5.91 1 5.92 1.42 .235 Two-way interactions 12.1 2 6.03 1.43 .240 Residual 388.6 93 4.18 Total 444.7 98 4.54 R = .315 R2 = .099 Table 8.--ANOVA: witness information distortion (2x2). Source of Sum of df Means of F Signif. Variation Squares Squares of F Main effects 44.072 2 22.00 5.31 .007 Time 38.60 1 38.60 9.30 .003 Camera 6.07 .1 6.07 1.46 .227 5W°“waY . 6.50 1 6.50 1.57 .211 interactions Residual 394.2 95 4.15 Total 444.7 98 4.54 2 R = .315 R = .099 43 Again, age and sex have been reported to be fac- tors involved in witness distortion. To determine this effect, a multiple regression was run. The criterion variable was the witness information distortion. The predictor variables were age, sex, time, and camera. (See Table 9.) From this analysis, a regression model was drawn to show the coefficients on the witness infor- mation distortion. Table 9.--Mu1tiple regression: witness information distortion. Variable B Std. Error B F/Significance Beta Age -.0307 .0176 3.04/.085 -.170 Sex .3148 .4310 .533/.467 .0718 Camera —.5157 .4140 l.55/.216 .121 Time -.3428 .1300 6.92/.010 -.259 (Constant)-1.304O .8120 The coefficients give weightings for the four predictor variables. The demographic variables of age and sex showed weights of +.l70 and —.072, respectively. The distracting variable "camera" showed a weight of +.121. The retention interval predictor variable (time) has a stronger weight than any of the other predictor variables, but is much weaker in effect than when used as a predictor for information decay. (See Figure 6.) 44 AGE \ (+.l70) (-.072) __g WITNESS DISTORTION CAMERA / (+. 121) (+.259) TIME / Fiture 6.--Regression analysis: witness information distortion. The Witness and External Distractions: Camera Effects The final hypothesis tested, H3, states that as the retention interval increases, the hindering effects on total testimony by camera increase. Hindering effects were Operationalized in two ways. The first way examines verbally discernible effects of the witness testimony (quality of witness response). The second way tests for the effects in a more precise (and less noticeable) way through word and time counts, and word/time ratios. The verbal response was coded on a 0 to 5 scale. The maximum value for the witness would be 100. The 45 minimum value would be 0, indicating no confidence what- soever in the response. Range of scores reported gave a minimum of 27 at the four-week camera condition and a maximum score of 98 at the immediate camera condition. That there was no relationship with camera presence or absence of is evidenced by the means in Table 10. Table 10.--Summary statistics: quality of witness response. Treatment (Retention Interval) Conditions 1/2 hour Two Weeks Four Weeks (n=20) (n=20) (n=12) NO camera i572.85 2553.10 i555.42 S=16.46 S=13.67 S=17.75 (n=l9) (n=14) (n=14) Camera §E67.95 §566.29 i553.95 S=13.30 S=18.31 S=15.83 o no camera A camera 80— l ’1 ”F '£2:= ‘_.o r 60" \ \ F Quality of — A A _ Witness 4o- - Response - _ 20" l- 0 g - 0 2 5 21 5 Retention Interval Figure 7.--Quality of witness response vs. retention interval. 46 Table 11.--ANOVA: quality of witness response. Source of Sum of df Means of F Signif. Variation Squares Squares of F Main effects 4722.9 3 1574.30 6.38 .001 Time 4619.6 2309.80 9.36 .001 Camera 107.3 1 107.3 .44 .999 Two-way interactions 1573.0 2 786.52 3.19 .045 Residual 22960.0 93 248.88 Total 29255.9 98 298.53 Variance explained: R2 = .161. Table 12.--Multiple regression: quality of witness response. Variable B Std. Error B F/Significance Beta Age -.1122 .139 .647/.423 -.0765 Sex -l.245 3.415 .133/.716 -.0350 Camera 2.316 3.278 .499/.482 .0673 (Constant)73.240 6.429 129.780/.000 Age and sex may be factors involved in the qual- ity of witness response. A multiple regression with the criterion variable of testimony quality and predictor variables of age, sex, time, and camera was run. AS seen 47 in Table 12, the only significant predictor variable was time. The second test requires the analysis of two variables and a ratio of the two creating a rate variable. Witness time was the total time in secOnds of all the witness statements. "Witness words" is simply a count of all the words used by the witness. The witness ratio of the number of words per second indicates the rate of response. Table 13.--Summary statistics: total witness words/time. Treatment (Retention Interval) Conditions 1/2 hour Two Weeks Four Weeks 22193/136 22199/120 22118/91.3 NO camera - S=65.6/47.8 s=69.7/5o.o S=58.4/27.4 2:193/125 §é139/98.l 22157/97 Camera s=95.9/44.2 s=70.5/32.o s=57.3/41.7 As indicated in Tables 14 and 15, the time and word variables show decay over time but are not affected by the camera condition. An analysis of variance was run for both and the effect of retention time in witness time and words was significant at the .005 and the .003 levels, respectively. Table 14.--ANOVA: 48 total witness time. Source of Sum of df Means of F Signif. Variation Squares Squares of F Main effects 23577 3 7859 4.29 .007 Time 20736 2 10368 5.66 .005 Camera 2498 l 2498 1.36 .244 Two-way interactions 2733 2 1366 .746 .999 Residuals 170387 93 1832 Total 196697 98 g 2007 R = .346 R2 = .120 Table 15.--ANOVA: total witness words. Source of Sum of df Means of F Signif. Variation Squares Squares of F Main effects 70885 3 23628 4.53 .005 Time 67677 2 33839 6.49 .003 Camera 2679 1 2679 .514 .999 Two-way interactions 7318 2 3659 .702 .999 Residual 484674 93 5211 Total 562851 98 5743 2 R = .355 R = .126 49 Caution is advised in examining the variables of witness time and words as they apply to the predictor variables. Both variables are highly dependent on the witness' lack of recognition. pendent of this, is the rate of recall. A better measure, inde- When these variables were combined in a ratio indicating rate of response, the camera and less notice- ably the retention interval had an effect. (See Figure 8 .) In all three interval treatments, witness response was facilitated by the camera (in contradiction to the third hypothesis). An analysis of variance was run and the camera effect was shown to be significant at the .012 level. A beta value of .263 was found. Table 16.—-Summary statistics: witness word/time ratio. Treatment (Retention Interval) Conditions 1/2 hour Two Weeks Four Weeks xe1.464 211,170 251.296 NO camera 8: .362 S: .299 s= .458 i=1.549 221.393 iél.666 Camera s= .510 5: .425 S: .347 50 0 camera A no camera 24 1.5N / - \ 0/ . A ' Word/Time 1- ~ Ratio .5" l- o - 0 l 2 3 4 5 Retention Interval Figure 8.--Word/time ratio vs. retention interval. Table 17.--ANOVA: witness word/time ratio. Source of Sum of f Means of F Signif. Variation Squares Squares of F Main effects ].323 3 .774 4.772 .004 Time 1.073 2 .537 3.306 .040 Camera 1.044 1 1.044 6.433 .012 Two-way interaction .318 2 .159 .981 .999 Residual 15.094 93 .162 Total 17.735 98 .181 Variance explained: R2 = .131. Table 18.—-Multip1e regression: witness word/time ratio. Variable B Std. Error B F/Significance Beta Age -.0013 .00360 .133/.716 -.036 Sex .0157 .08812 .032/.859 .018 Camera .2228 .08459 6.93/.010 .263 Time -.0128 .02662 .230/.633 -.048 (Constant) 1.3730 .16590 68.5/.000 CHAPTER IV DISCUSSION Summary This study was initiated to investigate various aspects of witness retention. Questions of applicability of previous research needed answers through a study which was more closely directed to actual witness functions. It was hypothesized that the retention interval affected recall in three ways: internally with loss and distor- tion, and externally with distractions. Results indicate that the retention interval is a strong determinant of remembered material (especially material relevant to witness testimony). Results from tests of the first two hypotheses provide support for the generalizing assumptions of psychological theory on memory. Results from tests of the last hypothesis show contradictory results (of the retrieval phenomenon) sug- gesting that the camera may have been something other than a distraction. Conclusions Hypothesis 1: Figure 3 illustrates the decay process. A few minutes after observing the stimulus 52 53 (immediate recall condition) the average witness response was about 60 percent of the total recall information. At two or four weeks, the average witness response was at 37 percent. This difference was significant at the .001 level. Thus, Hypothesis 1, which predicted decay of witness information over the retention interval, was supported. Further, a multiple regression including predictor variables of age, sex, camera, and time indi- cated that the variable of time (retention interval) was the only significant cause of witness knowledge decay, explaining about 30 percent of the variance. Hypothesis 2: Figure 5 shows the distortion growth over time. Almost immediately after the obser- vation of the stimulus, the average witness testimonial response is up to 20 percent distorted. This distortion appears to level out at the delay condition (at two and four weeks) with the mean distortion at about 30 percent. The difference in distortion between the immediate and delayed conditions was significant at the .003 level. Thus Hypothesis 2, which predicted distortion growth of witness testimony with retention interval, was supported. A multiple regression again including predictor variables of age, sex, camera, and time indicated that the variable of time (retention interval) was the main determinant of witness distortion (with a beta coeffi- cient of .259). The only other variable of apparent 54 effect was age. With a regression coefficient of .17 at a significance level of .085, age also contributes to distortion in a nontrivial way: the older the witness, the more the witness is prone to distortion. Hypothesis 3: The final hypothesis testing for camera effects was addressing the issue of aggregate verbal response independent of relevant information con- veyed. This concept was operationalized in two ways: through a coder evaluation of the response quality and through the recall rate (words/second). The first measure indicated that retention inter- val (and not camera distracting effects) hindered the verbal response. Although the witness with lack of knowledge may supplement testimony with assimilated material, the testimony does not possess the verbal quality that testimony which is fresh in the witness' mind possesses. As to the camera effects, the witness' verbal behavior showed the same general appearance of quality within the apprOpriate interval. Results from the second measure of the recall rate (Table 16) indicate that the camera does have an effect on the witness response. In all three retention interval treatments, witness response was faster in the camera condition than with no camera. An analysis of variance was run and the rate difference was significant at the .012 level. 55 This result contradicts the hypothesized concept of camera hindering effects. The camera seems to facili- tate response. For the retention interval delay condi- tion, the difference is almost three times greater than the immediate condition (.056 Aword/second vs. .157 Aword/ second). Hypothesis 3 requires hindering effects on tes- timony by camera to increase as the retention interval increases. Over time the response was facilitated. Discussion Many contend that the end goal of all court proceedings is to provide justice. When practitioners or critics discover that this goal is difficult or impossible to reach, new goals are established. Several shortcomings provide a strong basis for redefinition of goals where means become ends. The adversary system becomes the end when the strongest attorney can manipu- late the witness with greatest dexterity. Ends may also arise from tangential concerns where perceived justice is more important than actual justice. Limitations of basic legal assumptions and prac- tices may, for the present, deny the realization of jus- tice. However, legal understanding of limitations is incomplete. Experience and common sense, which often provide the main data base for those who define and use the procedural structure, may unnecessarily limit the possibility for innovation. A more sound data base is 56 provided through social science research. Using such knowledge may clarify the issues so that change will not be counterproductive. In terms of eyewitness potential, the judicial membership may appreciate some of the most apparent effects of time delay on the witness, but this under- standing is at best incomplete and therefore implica- tions are ill-founded. This study supported some basic psychological work on retention. Support for the two hypotheses on retention indicates that witness response is subject to loss and distortion over time. After only two weeks, there is a 30 percent reduction from the average amount in the immediate recall treatment. This, taken with a 50 percent rise in information distortion (from 20 percent to 30 percentL,makes the taking of tes- timony at the earliest time imperative. It must be added that the witnessing and testify- ing were staged under favorable conditions. The stimulus was pre-prompted so that all knew they would be asked about the scene. The interactions in the stimulus were comprehensible and straightforward. Instructions follow- ing the stimulus were fairly clean of any biasing agents. Between the stimulus presentation and the testi- mony, there was no partisan practice of seeing the wit- ness before testimony to thoroughly explore the story. Nonrandom distorting influences by others (besides 57 the attorney) did not exist as they do in many pro- ceedings. Although rules of evidence theoretically safe- guard the articulation phase of witness testimony, dis- torting effects are nonetheless present. The experiment, on the other hand, did not have an abundance of these factors. There was comparatively little pressure or emotionalism introduced. Further, the interrogation was free from leading questions (which often occur in direct examinations) and misleading or intimidating questions (which Often occur in cross—examinations.) For all these safeguards against witness trauma, the witness' performance was less than typically attribut- able to him. After only two weeks, less than half the story was given correctly. The rest of the information (about one-third) is either irrelevant or immaterial. Although the information loss and distortion curves appear to level out, the difference between two weeks and two years may make a substantial difference in testimony. (It is doubtful that the curves ever completely level out.) The taking of testimony at two years appears to be an exercise in futility. The only major environmental disturbance for the witness was the camera. The camera is seldom used in deposition and trial procedures. This manipulation proves to be of little effect. In terms of information 58 loss, distortion, and response quality decay, there were no significant differences between the camera and no camera conditions. The only significant difference was found in the response rate. Although this rate may be impossible to differentiate by the listener (the maximum difference of .2 words per second), the difference does exist in all three retention treatments. With the opposite relationship being found for the camera, its environmental addition appears to result in something other than a response inhibitor. Apparently, the camera brings about a focusing of attention: the witness subconsciously becomes more aware of the situa- tion and the information retrieval mechanisms become more effective. Research Extension The present research was highly focuSed dealing with some of the basic concerns of witness testimony. From a multitude of areas in the rules of evidence rele- vant for social science research, only the specific wit- ness qualification of witness recollection was examined. Through this examination many issues have been alluded to and, as such, many questions still need to be answered. Extension of research on witness retention to supplement and complement the present study can address the issues of information for storage, environmental 59 conditions for the witness, variations of the camera stimulus, and variations on retention interval. The information for storage (the stimuli data) may be varied along a multitude of dimensions. Central to the notion of association (and therefore memory) is that of familiarity. Unfamiliar events may only be comprehensible if redefined in a different but sensible pattern having marked effects on distortion. Attention to the stimuli is also of theoretical importance to memory. Arousal may be varied along the dimensions of degree and sign. A bland stimulus (low degree of arousal) may be difficult to remember. A positive arousal elicit- ing empathy or joy may affect the witness quite differ- ently than a negative arousal eliciting anger or disgust. Closely tied to the event (stimulus) is the envi- ronment of the witness. An unfamiliar environment may do much to distract the witness at the time of perception and therefore make retention difficult. The environment at the time of recollection is also important. A deposi- tion taken with an audience of less than a half dozen may facilitate many more responses than a trial situation with an audience of 50 or 100 pe0ple. A new part of the environment of interest was the camera. In this experiment, the camera was isolated to be the only distracting agent. Camera effects on response rate may not be significant when the camera is 60 one of a multitude of distractions (as in the court room.) Finally the retention interval, found to be of critical importance, had only three data points. Data at short-term intervals (every hour fOr four or five hours) and long-term intervals of two, four, and six months would provide important data to better define the hypothetical growth and decay relationships. BIBLIOGRAPHY 61 BIBLIOGRAPHY Adams, J. A. Human Memory. New York: McGraw-Hill, 1967. Anderson, J. R., and Bower, G. H. Human Associative Memory. Washington, D.C.: V. H. Wenston and Sons, 1973. Atkinson, R. C., and Shiffrin, R. M. "Human Memory: A Proposed System and Its Control Processes." In The Psychology of Learning and Motivation: Advances in Research and Theory, Vol.'II, pp. 89- 195. Edited by R: W. Spence and J. T. Spence. New York: Academic Press, 1968. Atwood, G. E. "Experimental Studies of Mnemonic Visual- ization." Ph.D. dissertation, University of Oregon, 1969. Bartlett, Sir F. C. Remembering: A Study in Experimental and Social Psychology. Cambridge, England: The University Press, 1932. Bird, C. "The Influence of the Press on Accuracy of Report." Journal of Abnormal and Social Psy- chology 22 (1927): 123-29. Broadbent, D. E. Perceptions and Communication. New York: Pergamon, 1958. Brown, J. "Some Tests of the Decay Theory of Immediate Memory." Quarterly Journal of Experimental Psy- chology 19 (1958): 289-99. Brown, R. W., and McNiel, D. "The 'Tip of the Tongue' Phenomenon." Journal of Verbal Learning and Verbal Behavior 5 (1966): 325-37. Buckhout, Robert. "Eyewitness Testimony." Scientific American 231 (1974). Ceraso, M. "The Interference Theory of Forgetting." Scientific American 217 (1967): 117-25. 62 63 Corkin, S. "Acquisition of Motor Skills After Bilateral Media Temporal Lobe Excision." Neuropsychologia 6 (1968): 255-65. Eagle, M., and Leitner, E. "Recall and Recognition in Intentional and Incidental Learning.“ Journal of Experimental Psychology 68 (1964): 58-63. Ebbinghaus, H. Memory,.A Contribution to Experimental Psychology. Translated by H. A. Ruger and C. E. Bussenius. New York: Columbia University Press, 1913. Fitts, P. M., and Jones, R. E. "Analysis of Factors Contributing to 460 'Pilot Error' Experiences in Operating Aircraft Controls." In Selected Papers on Human Factors in Design and Use of Control7 Systems. Edited by H. W. Sinaiko. New York: Dover, 1961. Fitts, P. M., and Seeger, C. M. "S.R. Compatibility: Spatial Characteristics of Stimulus and Response Codes." Journal of Experimental Psyghology 46 (1953): 199-210. Gardener, T. "The Perception and Memory of Witnesses." Cornell Law Quarterly 18 (1933): 391-98. Heller, Louise B. Do You Solemnly Swear? Garden City, New York: Doubleday and Company, 1968. Hutchins, Robert M., and Slesinger, D. "Some Observations on the Law of Evidence-Memory." Harvard Law Review 41 (1928): 860-73. Jones, S. E. "Directivity vs. Nondirectivity: Implica- tions of the Examination of Witnesses in Law for the Fact Finding Interview." Journal of Communi- cation 19 (1969): 64-75. Keele, Steven W. Attention and Human Performance. Pacific Palisades, California: Goodyear Publish- ing Company, Inc., 1973. Klatzky, Roberta L. Human Memory Structures and Processes. San Francisco: W. H. Freeman and Co., 1975. Koffka, K. Principles of Gestalt Psychology. New York: Harcourt, Brace Jovanovitch, Inc., 1935. 64 Kolers, P. A. "Interlingual Facilitation of Short-Term Memory." Journal of Verbal Learning and Verbal Behavior 5 (1966): 314-19. Lindsey, P. H., and Norman, D. A. Human Information Processing. New York: Academic Press, 1972. Loftus, Elizabeth F. "Leading Questions and the Eye- witness Report." Cognitive ngchology 7 (1975): 560-72. , and Zanni, G. "Eyewitness Testimony: The Influence of the Wording of a Question." Bulle- tin Of the Psychonomic Society 5 (1975): 86-88. Luh, M. "The Conditions of Retention." Psychology Monographs 142 (1922). Marshall, J. Law and Psychology in Conflict. New York: Bobbs-Merrill Co., Inc., 1966. Marston, William M. "Studies in Testimony." Journal of Criminal Law and Criminology 15 (1924): 5-31. Morris, Rudolph E., and Fishman, J. A. "Witness and Testimony: A Social Problem In Need of Social Research." Journal of Social Issues 13 (1957). Mortensen, C. D. Communication: The Study of Human Inter- action. New York: McGraw-Hill, 1972. Munsterberg, H. On the Witness Stand. New York: Double- day Page, 1908. . "On the Witness Stand." Essay on Psychology and Crime 50 (1923). Pavio, A. "Mental Imagery in Associative Learning and Memory." Psychology Review 76 (1969): 241-63. Pavlov, I. Conditioned Reflexes. Translated by G. V. Anrep. London: Oxford University Press, 1927. Penfield, W. "The Interpretive Cortex." Science 129 (1959): 1719-25. Peterson, L. R.; Hillner, K.; and Saltzman, D. "Supple- mentary Reports: Time Between Pairings and Short- Term Retention." Journal of Experimental Psychology 64 (1962): 550-51. 65 Posner, M. I. Cognition: An Introductigp. Glenview, Illinois: Scott, Foresman and Co., 1973. , and Keele, S. W. "On the Genesis of Abstract Ideas." Journal of Experimental Psychology 77 (1968): 353-63. Quillian, M. R. "Semantic Memory." In Semantic Infor- mation Processing. Edited by M. Minsky. Cambridge, Mass.: MIT Press, 1968. Reed, S. K. "Decision Processes in Pattern Classifica- tion." Ph.D. dissertation, U.C.L.A., 1970. Rumelhart, D. E.; Lindsay, P. H.; and Norman, D. A. "A Process Model for Long-Term Memory." In Or ani- zation and Memory. Edited by E. Tulving ang W. Donaldson. New York: Academic Press, 1972. Shallice, T., and Warrington, E. K. "Independent Func- tioning of Verbal Memory Stores: A Neuropsycho- logical Study." Quarterly Journal of Experimental Psyghology 22 (1970): 261-73. Slamecka, N. J. "Differentiation Versus Unlearning of Verbal Associations." Journal of Experimental Psychology 71 (1966): 822-28. Sleight, T. "Memory and Formal Training." British Jour- nal of Psychology 4 (1911). Steinberg, Harris B. "The Practitioner Speaks: Witness Performance as Seen by a Trial Attorney." Jour- nal of Social Issues 13 (1957). Tierney, Kevin. Courtroom Testimony, A Policeman's Guide. New York: Funk’and Wagnalls, I970. Tulving, E. "Episodic and Semantic Memory." In Organi- zation of Memory,_pp. 282-402. Edited by E. Tulving and W. Donaldson. New York: Academic Press, Inc., 1972. ' Underwood, B. J. "False Recognition Produced by Implicit Verbal Responses." Journal of Experimental Psy- chology 70 (1965): 122-29. U.S. Congress. House. Committee on the Judiciary. Rules of Evidence. Hearing before the Special Subcom- mittee on Reform of Federal Criminal Laws, House of Representatives, 93 Cong., lst sess., 1973. 66 Weinstein, Jack B. "The Law's Attempt to Obtain Useful (Testimony."' Journal Of Social Issues 13 (1957). Whipple, Guy M. "Psychology of Testimony." Psychology Bulletin 14 (1917): 234-36. Wigmore, John H. Wigmore on Evidence. Brooklyn: The Foundation Press, 1935. ‘ . Wigmore's Code of Evidence. Boston: Little, Brown and Co., 1942} WOodworth, R. S., and Schlosberg, H. Experimental Psy- chology. New York: Holt, Rinehart, and Wifiston, 1961. APPENDICES 67 APPENDIX A MATHEMATIC DERIVATION: MEMORY DECAY 68 APPENDIX A MATHEMATIC DERIVATION: MEMORY DECAY The system: short-term and long-term memory Initial conditions (at t=0): h=ho=dimension of new information in the sensory store t=0 (the step input to the STM) h A= "volume“ of assimilatable information in the sensory store A= area of the relevant information bundle R= resistance to information loss--the ability to process meaningful units into which the sensory information may be transformed . Variables: h= perceived dimension of new information = incoming (to STM) flow of information information loss (out of STM and LTM) ("when the information exceeds the amount that can be trans- formed or rehearsed with later decay" Keele, 1973) Conservation of information bit flow: "The rate of information loss (after the step input) equals the rate of decay in stimulus representation over time." =_d(Ah) qout dt (1) 69 70 for A=constant (past associations fairly resistant to change): _- 9.11 aout_ Adt (2) .Resistance is the potential change (in information) required to cause a unit change in flow: R = 3h (3a) qout for laminar resistance: R=h (3b) qout =2 qout R (4) Substituting (4) into (2): %:= -1;%% (5) %+Ag—:‘-=0 (5) A 3% + g = o (7) divide by A: %+%h=0 m This is a first order linear homogeneous differential equation and may be solved in the following manner: let h = ke-(l/RA)t (9) dh _ ll “(l/RA)t 5E - 1k§Ee ‘1°’ Substitute (9) and (10) into (8): 71 1 -(1/RA)t / ]. -(l/RA)t _ -—e +kEA-e —0 _kRA e— (lRA)t Initial conditions at t=0: h=hO h0 = k Substitute (14) into (12): h = hoe(-1/RA)t perceived dimension of inputted info. (11) (12) (13) (14) time APPENDIX B MATHEMATIC DERIVATION: MEMORY DISTORTION 72 APPENDIX B MATHEMATIC DERIVATION: MEMORY DISTORTION The system: short-term and long-term memory Initial conditions (at t=0) Q = the amount of directly related information bits in the bundle of associations V= volume of the bundle of associations Variables: n.= number of information bits per volume of associated bundle r.= rate of assimilation (information associations/ time) Conservation of information bits of flow in and out of the static control volume system: "The rate of change of the concept association at time t, (Q'(t)) must equal the rate at which new information enters the associated bundle (R1) minus the rate at which information leaves the bundle (R2). R = rate at which new information enters l = (n1) (ri) R2= rate at which information is displaced = (Q(t)/V)(ri) Therefore: ' — - Q (t)—R1 R2 0' (t)=(niri)‘-% Q(t) Rearranging the terms: I r.j_ _ Q (t)-l-T’_7Q(t) — niri 73 74 This is a first order linear nonhomogeneous differential equation. It may be solved in the following manner: a. multiply both sides of the equation by e(ri/V)t: (e'ri/V’t10'(t> + (eri/V’t)(§})(oxt>> = (e‘ri/V’t)