Shaw, John S. III; Zerr, Tana K. (2003).  Extra effort during memory retrieval may be associated with increases in eyewitness confidence.  Law & Human Behavior, 27, 315-329.

 

INTRODUCTION:

Eyewitnesses often try to work hard to perform well when testifying in court and their motivation may be selfish or unselfish.  Some may try to work hard because they understand the importance of their testimony.  Others may try to work hard to impress their peers and themselves, which could improve their self-esteem.    Most witnesses believe that extra effort on their part will result in increased accuracy.  However, simply trying harder is not sufficient to produce more accurate memories.  Poor encoding can alter memories so that any type of retrieval will have no effect on the accuracy of memories.  Also, there is very little any type of retrieval can do to recover the effects the delay between the event and trial has on memory.  

 

While extra effort may not increase accuracy, it can however increase confidence.  Most witnesses are likely to conclude that because the worked harder at being a “good” witness, they will produce accurate testimony.  According to the self-perception theory, this will result in high confidence ratings.  Shaw, Woythaler, et al. (2001) found evidence to support the theory that extra effort leads to higher confidence on a memory test.  Participants were informed, just before the memory test, some test items “counted” and some did not.  Those that did not count were part of a future study.  Participants reported higher confidence on the items that “counted” than those that did not.  Response accuracy was the same, regardless of confidence ratings.

 

OVERVIEW:

This study was designed to add to Shaw, Woythaler, et al. (2001) in three ways.  Shaw, Woythaler et al. (2001) did not measure effort so this study was designed to measure whether extra effort increased eyewitness confidence.  Another aspect the authors wanted to measure the effect of selfish motivation to use more effort on witness confidence.  

 

The event to be tested was a presentation by a visitor to an intro to psych class.  A surprise memory test was given five days later with questions about the visitor’s presentation and appearance (eyewitness and ear-witness memory).  The subjects were tested over incidental, as opposed to intentional, learning.  Half of the subjects could potentially earn prizes (such as candy or money) which constituted the high-motivation condition.  The other half were not eligible for prizes (low-motivation condition).  While this kind of selfish motivation may not directly parallel the types of selfish motivation in the courtroom, it was designed to provide selfish motivation to cause the witnesses to expend more effort.  Also, witnesses were not rewarded for expending more effort but for higher levels of performance that might come about as a result of this extra effort.

 

Using the self-perception theory (Bem, 1967, 1972), the authors hypothesized that confidence would be higher in the high-motivation condition than in the low-motivation condition.  Participants should give higher confidence ratings because they believe that extra effort correlates with increased accuracy.  However, the authors did not expect an actual effect of effort on accuracy.  They also believed that there would be a positive correlation between confidence and effort and that there would be a negative correlation between difficulty of questions and confidence.  Another hypothesis was that the confidence-accuracy and effort-accuracy correlations would be larger in the low-motivation condition rather than the high motivation condition.  This prediction was made on the basis that the authors expected individual differences to play a part in the motivation manipulation.  (The presence of reward would cause some participants to work harder while it would not have any effect on those who naturally work hard.)

 

METHODS:

During the third week of classes, 75 undergrad students in an intro to psych class heard a short, 9 minute talk given by the director of the counseling center.  She described the roles of counselors and discussed educational programs.  Five days later, the students were given a surprise memory test about the speaker.  There were 12 multiple choice questions about the speaker’s appearance and presentation.  There were three possible answers for each of the questions.  Before working on the questions, students were told that half of the class was eligible for prizes and the other half was not.  A random drawing determined which half (based on the color of paper the questionnaire was printed on) was eligible for the prizes ($25 for first place and candy bars for the next 10 highest).  After going through the questionnaire at their own pace, the students received a form for indicating amount of effort and confidence ratings, both of which were on 11-point scales.  Students also indicated how many questions they believed the answered correctly.

 

RESULTS:

The motivation condition in which students were placed did not affect accuracy, which was expected (M=46.6% correct for high, M=46.5% correct for low).  However, motivation had no effect on confidence ratings, which was not expected (M=3.46 for high, M=3.44 for low).  Motivation also did not affect the mean accuracy estimate.  Response accuracy was significantly correlated with confidence and effort in the low-motivation condition, but was completely unrelated in the high-motivation condition (Table 1, p. 321).  Consistent with the authors’ hypothesis, confidence was positively correlated to effort and negatively correlated with difficulty.  Also, accuracy estimate was positively correlated with effort and negatively correlated with difficulty in the low-motivation condition.  (The same pattern seemed to appear in the high-motivation as well, but was not statistically reliable.)  Finally, confidence and accuracy estimate were highly correlated for both conditions.  

 

Figure 1 (p. 323) shows the mean confidence ratings as a function of motivation and effort and Figure 2 (p. 324) shows the data for accuracy estimate.  There was a main effect of effort for both confidence and accuracy.  And in Figure 2 there was an interaction between motivation and effort for the accuracy estimate.  

 

DISCUSSION:  

This study shows that although witnesses may expend extra effort and report higher confidence, it does not result in higher accuracy.  The high-effort group had much larger confidence ratings and accuracy estimates than the low-effort group.  However, accuracy did not differ between the groups.  These results follow the self-perception explanation (Bem, 1967, 1972) where students may believe that extra work will result in higher accuracy which increased confidence even though there was no change in accuracy.  

 

Another explanation is that post event questioning or reflective thought can lead to higher confidence (Wells and Bradfield, 1999).  Another possible explanation is that participants may have subjected to demand characteristics and thought that since they put more effort into the memory test that they should report higher confidence.  

 

The fact that motivation did not affect witness confidence may have been due to the fact that the motivation manipulation simply was not powerful enough.  However, this does not take away from the importance of the finding that witness effort was significantly correlated with witness confidence.

 

IMPLICATIONS:

Jurors typically consider eyewitness confidence as an indicator of accuracy (Cutler, Penrod, & Dexter, 1990; Cutler, Penrod, & Stuve, 1998; Wells, Lindsay, & Ferguson, 1979).  The results from this study provide troubling implications that even though an eyewitness may be extremely confident about their testimony; it may simply be due to extra effort, not to accuracy.  This situation does not occur only in the courtroom but also during police interviews and while viewing lineups.  There is a lot of pressure on the witness to “get it right,” and because of this, the witnesses will put extra effort into their testimony which in turn, will result in high confidence.  But as this study shows, increased accuracy may not go along with this increased confidence.

 


 

University of Arkansas

Department of Psychology

Graduate Program in Experimental Psychology

Lampinen Lab

False Memory Reading Group

False Memory Reading Group Spring 2004