Estimator Variables And Eyewitness Identification

This sample POSTNAME Research Paper is published for educational and informational purposes only. If you need help writing your assignment, please use our research paper writing service and buy a paper on any topic at affordable price. Also check our tips on how to write a research paper, see the lists of criminal justice research paper topics, and browse research paper examples.


Eyewitness identifications can be either an asset or a detriment to a criminal investigation. Thus, it is vital that the individuals who are involved in the criminal justice system (e.g., police officers, attorneys, judges, and jurors) are aware of the myriad variables that affect eyewitness identification accuracy. These variables can be categorized as either system or estimator variables. The police have control over system variables, such as lineup presentation, biased instructions, post-identification feedback, lineup fairness, and wording of questions. In contrast, police have no control over estimator variables, which can, nonetheless, influence identification accuracy. This research paper will examine estimator variables associated with the criminal event (presence of a weapon, crime seriousness, stress, exposure duration, and distance), perpetrator (race and disguise), and witness (age and psychological impairment). The impact of these variables on eyewitness memory and identification decisions should be considered when evaluating the reliability of eyewitness evidence.


Gary Wells (1978) revolutionized researchers’ approach to eyewitness identification issues by creating a distinction between system and estimator variable research. Both categories of variables can influence eyewitness identification accuracy; however, he distinguished between variables that are under the control of the criminal justice system (system variables) and those that are not (estimator variables). System variables include how the lineup is constructed, how the lineup is presented to the witness, and the instructions that accompany the lineup. This type of research informs recommendations regarding the current best practices for obtaining eyewitness identification evidence. Estimator variables include factors related to the criminal event, witness, and perpetrator, such as how long the perpetrator was in view, the age of the witness, and whether or not a weapon was present. Wells labeled these estimator variables because their impact in an actual case can only be estimated post hoc.

The courts generally rely on five factors to evaluate the reliability of eyewitness identification evidence (Neil v. Biggers, 1972; Manson v. Braithwaite, 1977): (a) opportunity to view the criminal during the event, (b) degree of attention, (c) accuracy of prior description of the criminal, (d) level of certainty demonstrated at confrontation/ identification, and (e) time between crime and identification. The reliance on these limited variables is problematic, especially given the robust finding that providing confirming feedback to a witness (e.g., “Good, you got the guy!”) significantly bolsters reports of the opportunity to view the criminal, degree of attention, and certainty in the identification decision (Wells and Bradfield 1998). Thus, relying on the Biggers/Manson criteria alone may lead triers of fact to make erroneous conclusions regarding the reliability of the evidence.

In a revolutionary decision, the New Jersey Supreme Court in State v. Henderson (2011) ruled that in order to determine the suggestiveness of an identification procedure, the court must consider all factors that may have influenced the witness’s decision (e.g., blind administration, lineup construction, feedback). If the evidence is still admissible, the court should then consider the Biggers/Manson criteria and any additional applicable estimator variables (e.g., stress, weapon presence, exposure duration, age).

System variable research improves police procedures; estimator variable research, on the other hand, contributes to the understanding of how factors at the time of the crime influence eyewitnesses. When evaluating the reliability of identification evidence, both system and estimator variables should be considered. For instance, an identification made by a 3-year-old who viewed the perpetrator for 15 s in a dark alley from a distance of 250 ft should not be given as much weight as an identification made by a sober young adult who was face-to-face with the perpetrator for 30 min in a well-lit room. One cannot definitively state that the former identification is wrong and the latter is correct; however, awareness of the impact of age, exposure duration, lighting, and distance on identification accuracy should lead one to deem the latter identification as likely to be more reliable than the former.

State Of The Knowledge Of The Field

This research paper maintains Wells’ (1978) organization of estimator variables with regard to the division of characteristics associated with the criminal event, perpetrator, and witness. This nonexhaustive discussion reviews the most commonly researched variables in each category. Specifically, the section entitled “Characteristics of the Criminal Event” examines encoding conditions, the presence of a weapon, crime seriousness, and stress; “Characteristics of the Perpetrator” focuses on race and disguise; and “Characteristics of the Witness” covers age and psychological impairment.

Characteristics Of The Criminal Event

Encoding Conditions

The memory process comprises three steps: encoding, storage, and retrieval. Any factor that reduces the information available for encoding reduces performance at retrieval – information (e.g., details of the person’s face) cannot be stored nor later retrieved (e.g., used to distinguish between members of a lineup) if it was never encoded. This section addresses two of the conditions that restrict the availability of information at encoding: exposure duration and distance. Readers are directed to Narby, Cutler, and Penrod (1996) and Shapiro and Penrod (1986) for a review of additional encoding-related variables, such as perceptual salience, knowledge of recognition task, and number of faces presented during exposure.

Exposure Duration

Exposure duration refers to how long the witness is exposed to the target’s face. In general, studies have found that as exposure time increases, identification accuracy increases. As discussed in Bornstein, Deffenbacher, Penrod, and McGorty’s (2012) meta-analysis, shorter durations in research range from a few tenths of a second to 12 s, and longer durations typically hover around 45–60 s (although a few studies examined durations longer than 180 s). This updated metaanalysis confirmed findings from Shapiro and Penrod’s (1986) meta-analysis. Specifically, longer durations produce higher hit rates and lower false identification rates compared to shorter durations, but exposure has a greater impact on hit rates than false alarms. In comparing the effect sizes from both meta-analyses, Bornstein et al. concluded that exposure duration is even more influential than researchers originally thought.

In real crimes, exposure durations may be longer than those typically used in eyewitness and facial recognition studies; however, exposure duration has a nonlinear effect, meaning that additional exposure beyond a certain point (approximately 30 s) has minimal impact on accuracy. Further evidence of the generalizability of laboratory findings to the real world is that research paradigm (facial recognition vs. eyewitness identification) did not significantly moderate the relationship; exposure duration had similar effects regardless of whether or not the witness’s memory was studied under more ecologically valid conditions (Bornstein et al. 2012).

Exposure duration does not operate in isolation, but instead interacts with, and is affected by, other variables. For instance, as discussed in the next section, the presence of a weapon (or unusual object) is most detrimental to identification accuracy at intermediate durations, with a smaller effect at shorter and longer durations (Fawcett et al. 2013). One system variable that affects witnesses’ reports of exposure duration is whether or not the police provide confirming post-identification feedback to the witness (“Good, you got the guy!”). This feedback leads witnesses to overestimate how long the perpetrator was in view (e.g., Wells and Bradfield 1998). Thus, in real cases, police and attorneys must be cautious when attempting to determine length of exposure. When possible, police should gather this information before any feedback is provided and attempt to corroborate the witness’s report through other means.


When considering the distance between the eyewitness and the perpetrator of a crime, two factors need to be taken into account: (a) the actual distance, which affects identification accuracy, and (b) the witness’s perceived distance, which impacts people’s perceptions of that evidence. In a thorough investigation, Lindsay, Semmler, Weber, Brewer, and Lindsay (2008) examined the relationship between actual and perceived distance and how both influence identification accuracy. They found a relatively high level of correspondence between perceived and actual distances, but this relationship deteriorated significantly after an ecologically valid delay of one day, at which point witnesses underestimated the actual distance. Furthermore, participants’ tendency to underestimate distance varied according to actual distance, with the largest underestimates at 36–41 m ( 39–44 yards). Neither perceived nor actual distance influenced witnesses’ choosing rates, but longer distances did reduce identification accuracy. The authors hesitated to state a distance beyond which an identification should be inadmissible in court, but instead argued that distance should be used to evaluate the reliability of the evidence.

Presence Of A Weapon

Exposure to a perpetrator brandishing a weapon significantly impairs eyewitness identification accuracy because rather than focusing on the perpetrator, the witness’ attention is drawn to the weapon itself – this phenomenon is known as the weapon focus effect. Original research in this area sought to understand how the presence of a weapon influenced witnesses’ cognitive processes, which ultimately led to the development of the arousal/threat hypothesis. Easterbrook’s (1959) cue-utilization hypothesis contends that arousal and attention interact to affect the cues people utilize. In a state of high arousal, people’s attention is limited, and they can only utilize cues that are the focus of their attention. Meanwhile, all non-central, peripheral cues are underutilized. The arousal/threat hypothesis applies cueutilization to the eyewitness scenario: The weapon increases physiological arousal due to its threatening nature and induces fixation as the source of that arousal. The weapon, then, becomes a central cue and the perpetrator’s face becomes a peripheral cue, leaving the witness with a stronger memory for the weapon than the perpetrator. Tests of the arousal/threat hypothesis include manipulating threat or arousal, or examining the amount of arousal experienced in the presence of the weapon; these studies have produced mixed results suggesting that arousal/threat is not solely responsible for the weapon focus effect (Fawcett et al. 2013).

In contrast, other researchers have posited that the weapon focus effect is caused by the presence of an unusual or unexpected item (e.g., Pickel 1998). Supporters of the unusual item hypothesis contend that a witness’s attention is drawn to an object that is unusual in a given context. An isolated object, like a pair of scissors, would be unusual in a yoga class, but not at all unexpected in a hair salon. Similarly, a whole, uncooked chicken would not be unexpected in a butcher shop, but extremely unusual in a hair salon. Pickel sought to separately examine the role of threat and unusualness by contrasting the two components. Unusualness, but not threat, significantly influenced witnesses’ memory for nonobject details. Specifically, compared to the low-unusualness object, a high-unusualness object significantly impaired witnesses’ memory for peripheral details. The unusual item hypothesis has received considerable empirical support. The shifting view of the weapon focus effect is substantiated by the definition used in a recent meta-analysis: “weapon focus shall be defined as an object-related decrease in memory performance.. .for those elements of an event or visual scene coinciding with the presence of a weapon or unusual object” (Fawcett et al. 2013, p. 2).

Although the investigation into the underlying theoretical explanation of the effect continues, the effect itself is not in question. Fawcett and colleagues’ (2013) meta-analysis reported an overall moderate effect of weapon presence on memory performance, with retention interval and exposure duration moderating the effect. Weapon presence impairs memory for both event-related details and identification accuracy; however, effect sizes are greater for recall than recognition, potentially because of measurement sensitivity. Both the arousal/threat and unusual item hypotheses received partial support in the meta-analysis. The weapon focus effect was greater when the stimuli were threatening rather than non-threatening. Likewise, compared to non-unusual objects, unusual objects reduced memory performance.

Crime Seriousness

The seriousness of a crime is generally evaluated by the level of violence, the level of risk for those involved, and/or the value of property stolen or damaged. In one of the earliest investigations, Leippe, Wells, and Ostrom (1978) isolated crime seriousness from physiological arousal by specifically staging a low-arousal criminal event in which a confederate stole a paper bag ostensibly left by a previous participant. Crime seriousness was determined by the contents of the bag (expensive item = highseriousness). Knowledge of the bag’s contents was also manipulated, such that half were aware of the contents of the bag before the crime was committed and half learned this information afterward. Eyewitness identification accuracy was significantly higher when the crime was serious than not, but only when participants had prior knowledge of the bag’s contents (and thus, the seriousness of the theft). Leippe et al. suggested that prior knowledge led the witnesses to pay more attention to the event and potentially devote higher levels of cognitive processing necessary to encode the culprit’s facial features. Without prior knowledge, witnesses may not attend to features necessary to later accurately identify the target. Interestingly, and surprisingly, research has shown similar identification accuracy rates regardless of whether the witness was the victim of the crime or an uninvolved bystander to the event (e.g., Hosch et al. 1984).


The early literature examining stress and eyewitness memory is rife with speculation regarding whether increased stress would improve memory (up to a certain point, as in an inverted-U curve) or impede encoding (due to a reduction in available cognitive resources). Recently, in a metaanalysis, Deffenbacher, Bornstein, Penrod, and McGorty (2004) concluded that increased stress negatively affects eyewitness recall and identification accuracy. In terms of recall memory, stress has a greater impact on interrogative recall (specific questions) than narrative (free) recall. In terms of identification accuracy, target presence was a significant moderator. Increased stress reduces the likelihood that the witness will correctly identify the perpetrator from a targetpresent lineup, but does not increase the likelihood that the witness will falsely identify an innocent suspect from a target-absent lineup. Witnesses who did not select the perpetrator from the target-present lineup instead incorrectly chose a known-innocent lineup member, thereby discrediting them as reliable witnesses. Thus, although stress does not increase the likelihood that an innocent suspect will be selected, it does decrease the likelihood that the actual perpetrator will be prosecuted.

The type of research paradigm also significantly moderated the relationship between stress and eyewitness memory. Two research paradigms are generally used to examine memory for a face: facial recognition and eyewitness identification studies. Researchers examining theoretical questions about memory processes commonly use facial recognition studies in which participants are presented with numerous faces to remember, and then later asked to indicate whether each face from a larger pool of faces is old (previously seen) or new. Eyewitness identification studies, on the other hand, are more ecologically valid and generally used to examine applied issues regarding factors that could affect real-world eyewitness identification performance; typically, the witness is presented with one “perpetrator” and then later asked to identify him from a lineup. With regard to stress, participants’ performance in the facial recognition paradigm was minimally affected by increased stress, but it did negatively impact their eyewitness identification accuracy. In addition, age was a significant moderator between stress and recall; increased stress negatively affected adults’ recall, but had minimal impact on children’s recall. Age, however, did not moderate the relationship between stress and eyewitness identification accuracy – both children and adults’ recognition memory was debilitated by stressful conditions (Deffenbacher et al. 2004).

Related to stress is the physiological arousal associated with physical exertion, a common occurrence in the day-to-day activities of police, soldiers, and other emergency respondents. Physical exertion, however, is not limited to certain occupations – victims or witnesses may need to either fight with, or flee from, perpetrators. Hope, Lewinski, Dixon, Blocksidge, and Gabbert (2012) used a simulation study with police officers to determine how physical exertion influenced participants’ memory. Compared to the control group, officers who physically exerted themselves made fewer correct identifications and recalled fewer details from information encountered before and after the physical exertion. Physical exertion, however, did not impact participants’ reports of the presence of weapons. According to Hope et al., physically exerted participants had to allocate their cognitive resources differently than the control group. During the scenario, the physically exerted officers likely prioritized monitoring the immediate environment for anything that could be harmful (supported by the finding that their memory for presence of weapons was unaffected), while other environmental information was a lower priority. Notably, physical exertion not only dampened memory for the information encountered after the physical exertion, but also had a retroactive impact on memory for pre-exertion information, potentially due to an interruption in memory consolidation.

Taken together, the evidence suggests that there is no “optimal” level of stress when it comes to eyewitness memory; rather as stress increases, eyewitness memory decreases. The debilitating effects of stress are predicted to be even greater in real-world criminal situations, which are generally considered unethical to recreate in the laboratory.

Characteristics Of The Perpetrator


The race of a witness or perpetrator – in isolation – does not determine eyewitness recall or identification accuracy; however, varied combinations of witness and perpetrator race can significantly influence witnesses’ identification decisions (Narby et al. 1996). In particular, individuals are more accurate when identifying someone of their own race rather than someone of a different race. This pattern of results is referred to as the own-race bias (also known as the cross-race or other-race effect). The majority of the own-race bias research has utilized facial recognition paradigms rather than eyewitness identification tasks; however, the effect persists across both paradigms (Meissner and Brigham 2001).

Most investigations into the own-race bias (ORB) seek to examine the possible socialcognitive processes responsible for the effect. Thus far, researchers have generally considered four theoretical explanations for the own-race bias (see Meissner and Brigham 2001 for more detailed explanations and additional citations). First, physiognomic homogeneity may play a role in the ORB. Group members of a certain race may have less physiognomic variability (and thus, greater facial feature similarity), which may reduce the ability of members of one race to distinguish between other-race faces. This theoretical explanation has received little support. Second, individuals with prejudiced attitudes may be less likely than non-prejudiced individuals to even attempt to differentiate other-race faces from one another. Racial attitudes are unrelated to memory performance when response bias is taken into account; however, prejudice is relevant to the extent that people who are more prejudiced report less contact with people of other races. This leads to the third theoretical explanation: interracial contact. That is, the more an individual interacts with someone from another race (in terms of quality or quantity), the less likely they are to demonstrate the ORB. In the most current meta-analysis, Meissner and Brigham found that interracial contact is a significant (but small) mediator of the ORB, explaining 2 % of the variability across participants. Fourth, perceptual learning may contribute to the ORB such that individuals use appropriate facial cues to discriminate between own-race faces; however, the appropriateness of these cues may not generalize to other-race faces. Using this general framework, researchers are exploring several possibilities including training people to discriminate between other-race faces, examining differences in processing the faces as a whole rather than featurally, and investigating the importance of how the face is represented in one’s memory. Despite the unresolved questions regarding the social-cognitive mechanisms underlying the own-race bias, there is no question that this is a robust phenomenon.

The implications of the own-race bias have not gone unnoticed by the courts. Evaluating the reliability of cross-race identification evidence is an especially challenging task because there are a number of unanswered questions about how the ORB interacts with other variables. There is, however, evidence that factors such as short encoding times and long retention intervals exacerbate the own-race bias (Meissner and Brigham 2001).


Although one might expect that disguise would always impair identification accuracy, the evidence is more equivocal. A few inconsistent findings notwithstanding (e.g., O’Rourke et al. 1989), disguises are generally detrimental to accuracy in eyewitness identification and facial recognition studies. Disguise may impair accuracy through four distinct (or possibly intertwined) processes. First, disguise prevents the witness from seeing, and thus encoding, specific and undistorted facial features. Second, disguise may alter the witness’s metacognitions regarding the difficulty of the identification task. Third, disguise affects encoding specificity; that is, the discrepancy between exposure (encoding a disguised face) and test (recognizing an undisguised face) may interfere with identification accuracy. Finally, disguise may influence a witness’s allocation of attention to the dangerousness of the situation, the disguise itself, or to the criminality of the perpetrator, thereby leaving fewer cognitive resources available to encode the undisguised features. At this point, researchers are unclear about the specific mechanism through which disguise functions but some or all of these mechanisms likely work in combination.

Research regarding different types of disguise has typically focused on a hat, sunglasses, or a stocking; in general, all three types impair identification accuracy. Until recently, investigations have focused on the presence or absence of disguise rather than varying degrees of disguise. To address this limitation, Mansour and colleagues (2012) systematically examined eyewitness identification accuracy across four levels of disguise (ranging from no disguise to fully disguised). Their first experiment investigated the impact of a hat and/or sunglasses, whereas the second examined a stocking disguise. The general pattern across both experiments was that correct identifications decreased as disguise increased; this supports Brewer, Weber, and Semmler’s (2005) assertion that disguise affects the amount of information available for encoding. Notably, covering the face with a stocking to just above the mouth was as effective as covering the entire face; this supports the contention that disruptions to global facial configurations also play a role. Moreover, eyewitnesses’ willingness to choose from the lineup decreased at a slower rate than their accuracy, suggesting that witnesses were not sensitive to the quality of their memory (ideally, a poorer memory resulting from greater disguise should be associated with lower choosing rates). At present, further research is needed to tease apart the different mechanisms through which disguise is effective.

Characteristics Of The Witness


Eyewitness researchers use adults (typically, between 18 and 59 years of age) as the baseline against which to compare the identification accuracy of children and older adults. Age has little impact on correct identifications – when presented with a target-present lineup, all age groups are equally likely to choose the target. Age becomes a concern, however, when the target is absent from the lineup – children and older adults have a greater tendency to choose, therefore, compared to younger adults, they are more likely to choose an innocent lineup member. Of course, researchers can manipulate whether the target is present or absent, but the police do not know if their suspect is guilty or innocent. Thus, the age of the eyewitness is a critical estimator variable that should be considered by triers of fact.

A substantial amount of research has examined children’s eyewitness identification decisions from different lineup procedures. When presented with a simultaneous lineup, children under the age of 12 produce correct identification rates comparable to adults, but significantly fewer correct rejections than adults. The sequential lineup tends to reduce choosing for adults; however, the same reduction is not found with children, particularly when presented with a target-absent lineup (Steblay et al. 2011). Recent evidence suggests that children’s decreased identification accuracy can be further exacerbated by an additional (and common) variable – presentation of the lineup by an officer in uniform. In this research, children were nearly six times more likely to make a false identification from a target-absent simultaneous lineup when it was administered by a uniformed officer compared to when the officer wore casual clothes (Lowenstein et al. 2010).

Researchers are working to devise new procedures to reduce children’s tendency to choose regardless of whether the lineup actually contains the perpetrator. A promising (and on-going) line of research involves the development and investigation of the elimination lineup (Pozzulo and Lindsay 1999). Unlike the simultaneous or sequential lineup, the elimination lineup involves two distinct types of judgments. In the first judgment, the eyewitness is presented with all of the lineup members and is directed to select the lineup member who is most similar to the culprit. In the second judgment, the selected lineup member is isolated from the others, and the eyewitness must decide whether or not the selected person is in fact the culprit. This two-decision process is thought to be effective because the first decision maintains correct identification rates and the second increases correct rejection rates. At this point, the elimination lineup is the most encouraging solution to increasing children’s identification accuracy.

At the other end of the age spectrum are older adults – normally considered to be 60 years of age and older. Like children, compared to young adults, older adults generally produce comparable correct identification rates and higher false identification rates from simultaneous lineups. When used with older adults, the sequential lineup decreases target-absent choosing rates, but also produces significantly lower correct identification rates compared to young adults (Steblay et al. 2011). As with children, researchers are examining ways to reduce older adults’ tendency to choose from target-absent lineups. One line of investigation utilized a preidentification questionnaire, which had previously increased young adults’ correct rejection rates. Before showing participants the lineup, the administrator asked participants the following questions: “How clear a memory do you have for the face of the criminal? How confident are you that you will be able to select the criminal if you see a photograph of him in a lineup? How confident are you that you will realize that the guilty person is not in the lineup if you are shown a lineup with only innocent people in it?” (p. 159). Unfortunately, the benefits of these questions did not generalize to older adults (Memon and Gabbert 2003). Another attempt involved modifying unbiased lineup instructions in order to increase the saliency that the lineup may not actually contain the culprit (Wilcock et al. 2005). Although the enhanced instructions made the older adults more aware that the culprit may or may not be present, they did not affect correct rejection rates. Taken together, the research demonstrates that children and older adults share the propensity to make more false identifications, and thus age should be considered by triers of fact.

Psychological Impairment

Police, judges, lawyers, and jurors are likely to question the credibility and reliability of a witness with a psychological impairment, whether it is temporary (e.g., intoxication) or a long-term medical disorder (e.g., schizophrenia). Our discussion centers on the most studied impairment in the eyewitness area: alcohol. For a thorough consideration of how mental retardation, autism, schizophrenia, and post-traumatic stress disorder affect memory see Soraci et al. (2007).

Alcohol may be a common estimator variable for crimes that occur at bars, parties, or late at night. Like many of the other estimator variables, alcohol can affect a person’s attention, thereby influencing the cues encoded during the crime and those relied upon during recall and recognition. Yuille and Tollestrup (1990) found that, compared to sober witnesses, intoxicated witnesses recalled less information immediately after the crime, as well as in a follow-up interview 1 week later. While both groups had similar correct identification rates, intoxicated witnesses made more false identifications than sober witnesses. A similar pattern was found in a field study using showups (when the suspect is the only individual presented to the witness). Correct identification rates were similar regardless of blood alcohol level; in contrast, witnesses with higher blood alcohol levels were more likely to choose from a target-absent showup (Dysart et al. 2002). The authors attributed these results to the intoxicated witness’s use of salient cues (e.g., hairstyle, clothing). These cues are appropriate when the suspect is the perpetrator; however, when the suspect is innocent, an intoxicated witness – instead of correctly rejecting the lineup – is more likely to choose an innocent lineup member if he/she is a close enough match to the witness’ flawed memory.

Controversies In The Courtroom

Estimator variable research not only furthers people’s understanding of the factors that influence witnesses’ memory, but it also guides triers of fact who evaluate the reliability of eyewitness evidence. An examination of the opinions and beliefs of key players in the courtroom (i.e., judges, lawyers, and potential jurors) highlights the importance of this research. Take the weaponfocus effect as an example: 97 % of researchers surveyed indicated that the presence of a weapon can impair an eyewitness’s ability to accurately identify the perpetrator’s face (Kassin et al. 2001); in comparison, 68 % of judges (Magnussen et al. 2010), 50 % of prosecutors, 88 % of defense attorneys (Wise et al. 2009), and 53 % of jurors (Desmarais and Read 2011) agreed with this statement. Interested readers should refer to the aforementioned publications to better understand the discrepancies between empirical evidence and courtroom opinions.

Open Questions

Countless unanswered questions exist on the topic of estimator variables despite decades of eyewitness identification and facial recognition research. A few questions and controversies were discussed throughout this research paper, most of which focused on theoretical explanations for various effects (e.g., own-race bias, weapon focus effect, disguise). Now, a few additional questions are highlighted.

Does the weapon focus effect occur in the real world? Neither archival analysis of actual crimes nor field studies have reported a significant effect of weapon presence on recall or recognition (see Fawcett et al. 2013 for a discussion). At least three differences between laboratory research and real-world conditions are relevant to discrepancy: (a) the lack of experimental control, (b) actual exposure conditions related to factors such as exposure duration and retention interval, and (c) exclusion of poor witnesses. The lack of experimental control is problematic because the archival researcher is unaware of what actually occurred (known as “ground truth”), and this could impact decisions regarding accuracy. For instance, it is flawed to evaluate description accuracy by examining the appearance of the suspect because the police may have picked up that suspect precisely because he fit the description provided by the witness. Moreover, time frames in laboratory studies do not adequately capture the real-world conditions in which a witness may be exposed to the weapon for several minutes or more, and may not have the opportunity to provide details to an officer until hours have passed, and worse, not be able to attempt an identification for days or weeks while a suspect is located (Fawcett et al. 2013). Finally, in real cases, witnesses who experience the weapon focus effect may realize that they did not get a good view of the perpetrator and may inform the police that they would not be able to identify him in a future lineup. Thus, witnesses with the poorest memory are likely never considered in the archival analysis. Further research is clearly needed to understand the role of weapon presence in the real world, and to resolve the current schism between laboratory and field research.

Questions about witness age remain unanswered: Why does age moderate the relationship between stress and recall, but not identification accuracy? Are similar cognitive processes responsible for increased choosing by children and older adults? If so, then perhaps the success of the elimination lineup with children will generalize to older adult eyewitnesses. With respect to crime seriousness, how does it interact with increased arousal? The limited research on crime seriousness has only examined low-arousal situations, but in the real world, a serious crime is likely to be associated with increased arousal. This area of research will be aided by the development of novel (and ethical) ways to investigate crime seriousness under high-arousal conditions. Finally, future research should investigate how alcohol and other drugs interact with system variables to influence the reliability of witness testimony.


Estimator variable research examines how the presence or absence of various factors – separately – influences identification accuracy. Minimal research, however, has examined the interactive effects of multiple estimator variables (a few notable exceptions: alcohol and cross-race identifications, exposure duration and weapon focus, and stress and age). Given the large number of variables that are likely to exist in any given criminal event, an examination of all possible main effects and interactions results in an unfathomable research design (Wells 1978). At present, the utility of estimator variable research is two-fold: (1) to further understanding of the factors that influence eyewitness identification accuracy, which will in turn help triers of fact evaluate the reliability of that evidence, and (2) to inform system variable research (e.g., although police cannot control the age of the eyewitness, they can reduce a child’s tendency to choose from a lineup by altering how that lineup is presented).

Wells’ (1978) distinction between system and estimator variables not only revolutionized how researchers and courts alike thought of the role of different variables, but it also (perhaps inadvertently) directed the field away from estimator variables and toward system variable research. This shift in the field led to major developments regarding best-practice identification procedures and provides the empirical support for widespread recommendations regarding the collection of eyewitness evidence. Perhaps the time has come for another shift – one toward a more balanced investigation of estimator and system variables. After all, estimator variables are always present and their impact is unlikely to be entirely eliminated even by best-practice identification procedures.

Finally, in this research paper, and in the field in general, more attention is being paid to theory-driven estimator variable research. Greater reliance on theoretical frameworks is encouraged because it allows for integration of the extant body of work while informing and guiding future lines of research.


  1. Bornstein BH, Deffenbacher KA, Penrod SD, McGorty EK (2012) Effects of exposure time and cognitive operations on facial identification accuracy: a metaanalysis of two variables associated with initial memory strength. Psychol Crime Law 18:473–490
  2. Brewer N, Weber N, Semmler C (2005) Eyewitness identification. In: Brewer J, Williams KD (eds) Psychology and law: an empirical perspective. The Guilford Press, New York, pp 177–221
  3. Deffenbacher KA, Bornstein BH, Penrod SD, McGorty EK (2004) A meta-analytic review of the effects of high stress on eyewitness memory. Law Hum Behav 28:687–706
  4. Desmarais SL, Read J (2011) After 30 years, what do we know about what jurors know? A meta-analytic review of lay knowledge regarding eyewitness factors. Law Hum Behav 35:200–210
  5. Dysart JE, Lindsay RCL, MacDonald TK, Wicke C (2002) The intoxicated witness: effects of alcohol on identification accuracy from showups. J Appl Psychol 87:170–175
  6. Easterbrook JA (1959) The effect of emotion on the utilization and organization of behavior. Psychol Rev 66:183–201
  7. Fawcett JM, Russell EJ, Peace KA, Christie J (2013) Of guns and geese: a meta-analytic review of the ‘weapon focus’ literature. Psychol Crime Law 19:35–66
  8. Hope L, Lewinski W, Dixon J, Blocksidge D, Gabbert F (2012) Witness in action: the effect of physical exertion on recall and recognition. Psychol Sci 23:386–390
  9. Hosch HM, Leippe MR, Marchioni PM, Cooper DS (1984) Victimization, self-monitoring, and eyewitness identification. J Appl Psychol 69:280–288
  10. Kassin SM, Tubb V, Hosch HM, Memon A (2001) On the “general acceptance” of eyewitness testimony research: a new survey of the experts. Am Psychol 56:405–416
  11. Leippe MR, Wells GL, Ostrom TM (1978) Crime seriousness as a determinant of accuracy in eyewitness identification. J Appl Psychol 63:345–351
  12. Lindsay RCL, Semmler C, Weber N, Brewer N, Lindsay MR (2008) How variations in distance affect eyewitness reports and identification accuracy. Law Hum Behav 32:526–535
  13. Lowenstein JA, Blank H, Sauer JD (2010) Uniforms affect the accuracy of children’s eyewitness identification decisions. J Investig Psychol Offender Profiling 7:59–73
  14. Magnussen S, Melinder A, Stridbeck U, Raja AQ (2010) Beliefs about factors affecting the reliability of eyewitness testimony: a comparison of judges, jurors and the general public. Applied Cognitive Psychology 24:122–133
  15. Mansour JK, Beaudry JL, Bertrand MI, Kalmet N, Melsom EI, Lindsay RCL (2012) Impact of disguise on identification decisions and confidence with simultaneous and sequential lineups. Law Hum Behav 36:513–526
  16. Meissner CA, Brigham JC (2001) Thirty years of investigating the own-race bias in memory for faces: a meta-analytic review. Psychol Publ Policy Law 7:3–35
  17. Memon A, Gabbert F (2003) Improving the identification accuracy of senior witnesses: Do prelineup questions and sequential testing help? J Appl Psychol 88:341–347
  18. Narby DJ, Cutler BL, Penrod SD (1996) The effects of witness, target, and situational factors on eyewitness identifications. In: Sporer S, Malpass RS, Koehnken G (eds) Psychological issues in eyewitness identification. Lawrence Erlbaum Associates, Hillsdale, pp 23–52
  19. O’Rourke TE, Penrod SD, Cutler BL, Stuve TE (1989) The external validity of eyewitness identification research: generalizing across subject populations. Law Hum Behav 13:385–395
  20. Pickel KL (1998) Unusualness and threat as possible causes of “weapon focus”. Memory 6:277–295
  21. Pozzulo JD, Lindsay RCL (1999) Elimination lineups: an improved identification procedure for child eyewitnesses. J Appl Psychol 84:167–176
  22. Shapiro PN, Penrod SD (1986) Meta-analysis of facial identification studies. Psychol Bull 100:139–156
  23. Soraci SA, Carlin MT, Read JD, Pogoda TK, Wakeford Y, Cavanaugh S, Shin L (2007) Psychological impairment, eyewitness testimony, and false memories: Individual differences. In: Lindsay R, Ross D, Read D, Toglia M (eds) Handbook of eyewitness psychology: memory for events. Lawrence Erlbaum and Associates, Mahwah, pp 261–297
  24. Steblay NK, Dysart JE, Wells GL (2011) Seventy-two tests of the sequential lineup superiority effect: a meta-analysis and policy discussion. Psychol Publ Policy Law 17:99–139
  25. Wells GL (1978) Applied eyewitness-testimony research: system variables and estimator variables. J Pers Soc Psychol 36:1546–1557
  26. Wells GL, Bradfield AL (1998) “Good, you identified the suspect”: feedback to eyewitnesses distorts their reports of the witnessing experience. J Appl Psychol 83:360–376
  27. Wilcock RA, Bull R, Vrij A (2005) Aiding the performance of older eyewitnesses: enhanced non-biased line-up instructions and line-up presentation. Psychiat Psychol Law 12:129–140
  28. Wise RA, Pawlenko NB, Safer MA, Meyer D (2009) What US prosecutors and defence attorneys know and believe about eyewitness testimony. Appl Cognit Psychol 23:1266–1281
  29. Yuille JC, Tollstrup PA (1990) Some effects of alcohol on eyewitness memory. J Appl Psychol 75:268–273

See also:

Free research papers are not written to satisfy your specific instructions. You can use our professional writing services to buy a custom research paper on any topic and get your high quality paper at affordable price.


Always on-time


100% Confidentiality
Special offer! Get discount 10% for the first order. Promo code: cd1a428655