‘But can you prove it?’ – examining the quality of innocent suspects' alibis

Abstract

Despite an assumption in the legal system that innocent people can generate accurate alibis, little research has examined the process of alibi generation. The current study examined this process with respect to three theoretical reasons why innocent suspects may fail to generate convincing alibis: they may lack the necessary memory, generate mistaken alibis, and generate weak or uncorroborable alibis. Undergraduates (N=255) were asked to report four initial alibis – each for a different time – along with corroborating physical and person evidence. Participants attempted to corroborate that evidence before returning 48 hours later. Upon return, participants reported their investigated alibis for the same four time periods. Results indicated that, despite participants' willingness to generate initial alibis, a substantial proportion of these alibis (36%) were mistaken, requiring either a change in narrative or a change in corroborating evidence. The majority of investigated alibis relied on evidence that evaluators would consider weak. Distant-past alibis were more likely to be mistaken than near-past alibis. Results indicate that innocent alibi providers may find convincing alibis difficult to generate, and we explain these results within a quantity–accuracy trade-off framework.

Keywords:

• alibi
• alibi generation
• autobiographical memory
• alibi witness
• corroborator

Notes

1. The large amount of variability in the distant-past condition is due to our desire to keep the date of the time period for which participants had to account constant. Thus, participants who participated later in the study had a longer time delay than participants who participated earlier in the study. Because this would confound time delay in the distant past with participants if we examined differences in latency among distant-past alibis (i.e. participants who signed up later in the semester may have differed from participants who signed up earlier in the semester), we examined latency by simply comparing near-past alibis with distant-past alibis, which is a non-confounded comparison due to the within-subjects design.

2. This very likely overestimates the actual accuracy rate of alibis, but is consistent with our overall ‘best case scenario’ treatment of alibis.

3. The percentages equal more than 16.4% because some alibis grew weaker due to a loss of both person and physical evidence.

4. It is likely, however, that college students are less aware than actual investigators of the types of strong physical evidence that could exist to support their alibis. For example, students may not realize just how often they are captured on videotape.

5. It should be noted that if the reader is interested in the alibis provided by people who are able to consult external sources, we do have data that speak to that issue – our participants' investigated alibis are, essentially, their alibis after having consulted such sources (see Table 1).