Attribution of crime motives biases eyewitnesses’ memory and sentencing decisions

ABSTRACT

In court, the basic expectation is that eyewitness accounts are solely based on what the witness saw. Research on post-event influences has shown that this is not always the case and memory distortions are quite common. However, potential effects of an eyewitness’ attributions regarding a perpetrator’s crime motives have been widely neglected in this domain. In this paper, we present two experiments (N = 209) in which eyewitnesses were led to conclude that a perpetrator’s motives for a crime were either dispositional or situational. As expected, misinformation consistent with an eyewitness’ attribution of crime motives was typically falsely recognised as true whereas inconsistent misinformation was correctly rejected. Furthermore, a dispositional vs. situational attribution of crime motives resulted in more severe (mock) sentencing supporting previous research. The findings are discussed in the context of schema-consistent biases and the effect of attributions about character in a legal setting.

KEYWORDS:

• Misinformation effect
• eyewitness memory
• attribution
• crime motives
• sentencing
• mock-jurors

Acknowledgements

We are indebted to Lina-Maraike Nitz and Pia-Renée Kobusch for their assistance in data collection and coding. We also thank two anonymous reviewers and Jens H. Hellmann for valuable comments on earlier versions of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Fourteen participants already knew the movie we had chosen for the experimental manipulation (see Materials and procedure) or correctly guessed the purpose of the experiment in a post-experimental suspicion check. Their data were not included in the analyses, resulting in the sample described above. Excluded datasets were equally distributed to the experimental conditions, χ²(2, N = 144) = 0.118, p = .943.

2. All distractor items of Experiments 1 and 2 were tested regarding their suitability with a separate sample of N = 25 participants who were familiar with the basics of attribution theory. They first read one of the film reviews used in the main experiments implying a situational (N = 12) vs. dispositional (n = 13) attribution of crime motives. Next, they rated each distractor item with respect to Weiner’s (1986) attribution dimensions stability (‘To what extent do you evaluate this statement as something that is variable or stable with respect to time?’; 1 = variable to 9 = stable), controllability (‘To what extent do you evaluate this statement as something that is uncontrollable or controllable?’; 1 = uncontrollable to 9 = controllable), and locus (‘To what extent do you evaluate this statement as something that is due to the respective situation or the person?’; 1 = situation to 9 = person). Additionally, each item was rated regarding its valence (‘To what extent do you evaluate this statement as something negative or positive?’; 1 = negative to 9 = positive). The distractor items were presented in randomised order. Four separate 2 (attribution: dispositional vs. situational; between-subjects) × 2 (distractor type: dispositional vs. situational; within-subjects) mixed factorial MANOVAs were conducted. As intended, dispositional distractors were attributed more to the person (M = 6.06, SD = 1.28) and situational distractors were attributed more to the situation (M = 4.64, SD = 1.38), F(1, 23) = 9.80, p = .005,  = .30. Concurrently, there was no main effect of attribution condition, F(1, 23) < 1, and no interaction between attribution condition and distractor type, F(1, 23) = 1.00, p = .327. Furthermore, no main effects and no interactions were detected with respect to the other attributions dimensions, all F(1, 23) < 4.11, all ps > .05. However, dispositional distractors were regarded as more negative (M = 2.76, SD = 1.24) than situational distractors (M = 4.98, SD = 1.12), F(1, 23) = 41.81, p < .001,  = .65. No effect of attribution condition, F(1, 23) = 2.84, p = .106, and no interaction, F(1, 23) < 1, revealed. Potential limitations of the chosen distractor items are addressed in the ‘General discussion’ section.

3. Data of eight participants who already knew the movie ‘Monster’ that was central to our experimental manipulation, or who held reasonable suspicions regarding the cover story of the experiment were excluded. Excluded datasets were equally distributed to the experimental conditions, χ²(1, N = 87) = 0.255, p = .614.