Role of Eyewitness Testimony in The Conviction of a Suspect

Eyewitness identification can play an important role in the conviction of a suspect, but research has also shown that unreliable eyewitness evidence is a key reason for wrongful convictions (Brandon & Davies,1973; Huff et al, 1976). For example, Wells et al. (1998) found that of 40 exonerated cases, 90% involved mistaken eyewitness evidence. Consequently, there has been substantial research into how both police practice and policy may affect the reliability of eyewitness testimony (Steblay et al, 2003; Wells et al, 1995). This has led to the development of guidelines that aim to encourage fair police policies and reduce mistaken eyewitness evidence (Wells et al, 2016). For example, the American Psychology-Law Society (APLS) put forward recommendations about line-ups, a police procedure used to gauge whether a witness can identify and distinguish the suspect from other innocent individuals (fillers). One of the guidelines states that line-ups should be fair, such that no one should stand out due to distinctive features. Second, it states that the person who conducts the line-up should not know who the culprit is (Wells et al, 2016). In the case of R v Anderson, these two guidelines, and other suggestions for fair line-ups are not adhered to, and so pose as important issues regarding the quality of evidence leading to the conviction of Stephen Anderson. These issues and their implications will be further discussed throughout this essay.

There are many issues concerning how the line-up was conducted. First, there is a clear large discrepancy between the descriptions of the perpetrator and Anderson. According to the victim, the man was said to be clean-shaven whereas Anderson had a mustache. Furthermore, the man was 5’9, whereas Anderson was 6’2. Lastly, the man was said to have a gap between his teeth whereas Anderson had a chipped tooth. Overall, the descriptions of these men largely vary. Nonetheless, as stated in the case, the detective arrested Anderson and included him in the line-up as he thought that he resembled the description of the perpetrator. This points to a larger issue in police procedures: police do not generally need tangible evidence that a suspect is involved in a crime to place them in a line-up. Some detectives include suspects in a line-up based on a mere hunch (Wells, 2006). In a national survey in the USA, more than one-third of law enforcement agencies stated that their officers did not need any evidence at all to place someone in a line-up. Therefore, including innocent people in a line-up increases the risk of a wrongful conviction. (Wise, Safer & Maro, 2011).

A fair line-up is one whereby no single individual stands out due to distinctive features, and this is a guideline put forward by the APLS (Wells et al, 2016). When an individual does stand out, it has implications for the functional size of the line-up. This refers to the number of viable line-up members who resemble the culprit, whereas the nominal size is the overall amount of people in the line-up (Wells et al, 1979). Ideally, these should match. However, in R v Anderson, Anderson was the only one with defects to his teeth, therefore arguably making him stand out against the others who do not have teeth defects. Therefore, the functional size does not match the nominal size.

Research has repeatedly found that when a suspect stands out from the others in a line-up, people are more likely to pick them as the suspect (Clark, 2012; Doob & Kirshenbaum, 1973; Wogalter, Marwitz &Leonard, 1992). Collof, Wade, and Strange (2016) conducted a large study looking at line-ups involving those with distinctive features. In this study, the distinctive feature was hidden for some participants. This was done either by replicating it onto others in the line-up, using pixilation, or blocking over it. For other participants, the distinctive feature was left to stand out. Findings showed that regardless of the method used to cover the distinctive feature, participants performed with similar accuracy. Yet, when the distinctive feature was left to stand out, participants' accuracy decreased, and false identifications increased. For example, when the culprit was absent but included a person with a similar distinctive feature, participants identified this person who is an innocent suspect 35% of the time, compared to only 9% of the time when the distinctive feature was replicated across all fillers in the line-up. This is in line with the argument that people with distinctive features are likely to be picked (Clark, 2012).

So why do unfair line-ups make us more likely to pick the individual that stands out? Wells (1998) explained this by suggesting that when selecting from a line-up, people tend to use a relative judgment strategy. This consists of choosing the person that most closely matches the suspect, even if the overall match is not close. So, when an individual with a distinctive feature similar to the perpetrator is included, they are more likely to be chosen as they are more similar to the perpetrator relative to the others included in the line-up (Zarkadi, Wade & Stewart, 2009). In R v Anderson, Anderson was the only one with teeth defects, even though he had a chipped tooth rather than a gap in his teeth. This is the closest match to the perpetrator, relative to the other individuals in the line-up who did not have any teeth defects at all. Conversely, Wixted and Mickes (2014) argue that when a line-up is fair, the victim has the ability to deduct the distinctive feature and draw on other cues and therefore are more likely to make an informed decision.

The next issue also centers around a rule set forward by the APLS: the person who conducts the line-up should not know who the culprit is (Wells et al, 2016). However, in R v Anderson the same detective who included Anderson in the line-up also conducted it. Research suggests that this is problematic, as they may be at risk of biasing the victim (Wells,2006). For example, Clark (2012) states that if police know who the suspect is, they may unknowingly encourage the victim to select that person. Phillips et al (1999) suggest this may be done through various verbal and non-verbal cues. For example, statements such as ‘carefully look at number…’ could cause extra attention to that individual. Instead, research suggests that a detective who does not know who is the suspect should conduct the line-up. Canter, Hammond, and Youngs (2013) found that when an informed administrator conducted the line-up, the target member was selected more than twice as often than when a blind administrator conducted it. In the case of R v Anderson, the detective may have unconsciously directed attention to Anderson, increasing the likelihood of the victim choosing Anderson as the perpetrator.

After the victim had made her selection of the suspect, the detective said “Well done, that’s our suspect!” This is an example of post identification feedback, which, research suggests can have negative consequences (Wells & Bradfield, 1998). For example, post identification confirmation feedback consequently inflates rates of how confident someone felt when they made their selection, which can be termed retrospective certainty (Bradfield, Wells & Olson, 2002). A meta-analysis, which looks at a large range of studies confirmed this effect across the US, Canada, Europe, and Australia (Steblay, Wells & Douglas, 2014). Wells and Bradfield (1998) demonstrated this effect in their study. Participants witnessed a video of a gunman and were asked to select who they thought was the suspect from a line-up. All of the participants made false identifications as the gunman was not included in the line-up. One group of participants were given ‘confirming feedback’ that they identified the correct suspect, whilst another group was given ‘negative feedback’, that they identified the wrong suspect. The control group was given no feedback. The results were in line with the idea of retrospective certainty. Those who were given confirming feedback gave an average certainty rate of 5.4 which was substantially higher than the average certainty of a control group which was 4.0. In R v Anderson the victim received similar feedback to those in this study, and so this may have caused an inflated certainty in her selection.

One explanation for this effect comes from a well-known social psychology theory- Bem’s self-perception theory (Bem, 1972). After hearing confirmation that participants picked the right suspect, they may undergo a self-perception process whereby they may think ‘I must have been certain when I made my choice because I picked the right suspect’. This may explain inflated post-decision judgment rates. Conversely, Steblay, Wells, and Douglas (2014) put forward the cue- accessibility interpretation as an alternate explanation. They suggest that participants use the feedback given as a cue to how confident they were at the time of their selection. For example, if they received negative feedback, they would use this as a cue to decide that they were not that confident when making their decision.

Arguably, this effect can cause a problem because cross-culturally, people tend to think confidence is an indicator of accuracy (Deffenbacher & Loftus,1982; McConkey & Roche,1989; Noon & Hollin,1987). However, Bradfield, Wells, and Olson (2002) found that receiving confirmation feedback decreases the relationship between certainty and accuracy. The practical implications of this are profound. In court, prosecutors are more likely to sentence suspects whereby eyewitnesses are confident in their decision than when they are uncertain (Bradfield, Wells & Olson, 2002). As the victim in the present case received post-confirmation feedback, she may be more confident in her decision, and this confidence may increase the likelihood of Anderson's conviction.