Jurors believe interrogation tactics are not likely to elicit false confessions: will expert witness testimony inform them otherwise?

Abstract

Situational factors – in the form of interrogation tactics – have been reported to unduly influence innocent suspects to confess. This study assessed jurors' perceptions of these factors and tested whether expert witness testimony on confessions informs jury decision making. In Study 1, jurors rated interrogation tactics on their level of coerciveness and likelihood that each would elicit true and false confessions. Most jurors perceived interrogation tactics to be coercive and likely to elicit confessions from guilty, but not from innocent suspects. This result motivated Study 2 in which an actual case involving a disputed confession was used to assess the influence of expert testimony on jurors' perceptions and evaluations of interrogations and confession evidence. The results revealed an important influence of expert testimony on mock-jurors decisions.

Keywords:

• interrogations
• confessions
• jury
• decision making
• expert testimony

Notes

1. All materials from this study are available from the first author.

2. This survey was conducted before the widely publicized controversy of tactics used by the CIA to interrogate suspected terrorists.

3. The pattern of results was the same when gender was used as a variable. The exception was for coerciveness ratings in the Request and Presentation of Evidence category; overall males (M=2.43, SD = 1.29) found the tactics in this category to be more coercive than females (M=1.96, SD = 0.97), t(115) 2.25, p<0.05, r=21. This indicates that, for the most part, males and females perceived interrogation tactics and confessions in similar ways.

4. Some of the distributions of coerciveness ratings showed significant Kurtosis and/or Skewness. Specifically, the Actual or Threat of Violence subscale showed significant Kurtosis (5.50, SE = 43) and positive Skewness (−2.47, SE = 0.22). Likewise, False Evidence showed significant Kurtosis (1.81, SE = 0.43) and positive Skewness (−1.50, SE = 22). Finally, the Request and Presentation of Evidence subscale showed a significant degree of negative Skewness (0.95, SE = 22). These non-normal distributions of coerciveness ratings are not surprising given that a high percentage of jurors gave extreme coerciveness ratings (4/5) on tactics comprising the Actual or Threat of Violence (M=85.33%) and Fabricated Evidence subscales (M=76.95%), and only a small percentage gave the extreme ratings on the Request and Presentation of Evidence subscale (M =17.47%). The likelihood ratings on eliciting true and false confessions were normally distributed, except for one distribution: the Request and Presentation of Evidence subscale showed a degree of negative Skewness (0.96, SE = 0.22) on likelihood of eliciting a false confession.

5. A table with inter-correlation is available from the first author.

6. The pattern of results was the same when race (white vs non-whites) was used as a variable. This indicates that race of participant did not influence perceptions in this case.

7. Only the scalar (guilty-confidence) measure showed significant departures from normality. Pre-expert testimony (M =3.35, Med = 4.00, SD = 2.18) showed a significant degree of Kurtosis (3.56, SE = 0.40) and positive Skewness (−2.13, SE = 0.20). Post-expert testimony (M =1.87, Med = 3.00, SD = 3.20) only a significant positive Skewness was revealed (−0.93, SE = 0.21). These results are not surprising given that the majority of participants found the defendant guilty both pre and post expert testimony. Because of this departure from normality, a Wilcoxon Sign Rank test was used to test for differences between means, and confidence intervals for correlations are provided where relevant (see Howell, 2002, pp. 691–718).

8. It is important to note that we do not want to imply that truthful and voluntary processes always co-occur; in many cases a confession is truthful but involuntary as when elicited because of coercion.

9. Only tactics numbers 1–4 and 7–10 were included in the ANOVA because of missing data on tactics numbers 5 and 6. This was due to a printing error that led to the omission of these two tactics from some post-expert questionnaires and as a result only partial data were obtained.

10. We thank an anonymous reviewer for suggesting this point.