Brief report: narrative qualities and perceptions of generated alibis.
One way to study alibis is to ask participants to state where they were at a particular time (called generated alibis as opposed to fictitious alibis; Stromwell, Granhag, & Jonsson, 2003). The alibi generation research has yet to address the importance of narratives. Narratives are people's personal stories which include external actions, as well as private thoughts and emotions (Bruner, 1992). Although alibi researchers have failed to study narratives, research conducted in other domains demonstrated that narrative influenced perceptions of credibility. Coherent and well-organized speech positively affected perceptions of prosecutors (Voss, Wiley, & Sandak, 1999), and speech containing indications of uncertainty (hedges) negatively affected views of defendants (Erickson, Lind, Johnson, & O'Barr, 1978; Voss & Van Dyke, 2001).
The current study examined generated alibis in which individuals displayed a large variety of speech patterns. While past studies of narratives have demonstrated differences in speech styles and narrative, one limitation has been that most of the studies have used "all or nothing" manipulations of narrative. In the current study, the use of generated alibis allowed for more insight as to how moderate levels of narrative variables affected believability. In Experiment 1, individuals were asked to recall where they were for dinner three nights prior. Some were given three minutes to remember the details of their whereabouts (warned group), while others were asked to respond immediately (no warning group). We chose to limit the preparation time to three minutes because most participants did not have cell phones with them and were not given access to computers. A lengthy preparation time would have given them more time to think but would not have helped them in terms of verifying their alibis or discovering if they had evidence to back up their claims. The content of participants' alibis was then examined to determine what kinds of narrative variables they included and to discover whether there were differences between the warning and no warning groups. In Experiment 2, a second group of participants provided their perceptions of the interviews. The results were analyzed to see how their perceptions might have been affected by the narrative qualities in the alibis and the warning manipulation.
We predicted that participants who were not warned that they would be supplying an alibi would use narrative features that were indicative of hesitancy and uncertainty at higher rates and would have lower word counts and interview times than those who were warned. We also hypothesized that the effects would be stronger in participants' free recall because by the time the interview would come to the cued recall portion, the interviewee may have had time to completely "straighten out" his/her story. We hypothesized that the participants in the no warning group would also be perceived as less credible by participants who evaluated their alibis (Experiment 2).
The methodology used in this study was approved by the University's Institutional Review Board.
Participants. Twenty-seven undergraduate psychology students, eight males and nineteen females (M age = 18.59 years, SD = .75), participated in the first experiment. We did not gather other demographic information. Individuals participated in order to receive credit for their introductory psychology classes.
Materials and Procedure. After obtaining informed consent, participants were interviewed individually concerning their whereabouts and details surrounding their dinner three nights earlier. Participants were randomly assigned to either a warning or no warning condition. All interviews were audio-recorded with the participants' knowledge and permission. After being asked to provide a free recall of their experience (i.e., "I want you to recall what you did for dinner three nights ago. Try to be as detailed as you possibly can."), participants were asked several cued recall questions regarding their alibi (e.g., "Do you remember where everyone was sitting in relation to one another?"). Following the interview, participants were debriefed and asked if they had lied in the interview. Individuals who had time to prepare their alibi were also asked whether they used any resources outside of their own memory to recall any details (e.g., called a friend while alone in the room). No participants stated that they intentionally lied or used external sources to help them remember their whereabouts. Finally, participants were asked to sign a release form regarding the use of their interview for future research (i.e., analysis of the interview for narrative qualities, showing other participants a transcript of the interview).
After the collection of the alibis was complete, the interviews were transcribed and analyzed by the first and second author for eight narrative variables: negative qualifiers (e.g., "I'm not sure"), positive qualifiers (e.g., "He was definitely wearing a t-shirt"), verbal hedges (e.g., "Um," "Uh"), logical connections (e.g., "but," "because"), temporal connections (e.g., "and," "then"), elaborations (e.g., "it was unusual 'cause we were talking about Italy"), spatial order (e.g., "next to"), and uses of personal pronouns (e.g., "I," "me"). The first author (who was blind to condition) was trained in discourse analysis and had used similar methods in an earlier study (Allison, Brimacombe, Hunter, & Kadlec, 2006). The operational definitions were modified based on the Allison et al. study (available from the first author). Percent agreement (# agree / #agree + #disagree) was used to calculate inter-rater reliability for each narrative variable on 10% of the interviews. Agreement ranged from 80.7% to 100% for all variables.
For each narrative feature, two rate measures were calculated: a rate based on the entire interview (e.g., total number of narrative variables used in free and cued recall/ total word count) and a free recall rate (e.g., number of variables used in free recall / free recall word count).
Additionally, total word count, interview length (in seconds), and pauses were calculated for both the free recall portion and the total interview. Pauses were also calculated during the first 10 words of free recall.
Narrative Features. A 2-way between-subjects Multivariate Analysis of Variance (MANOVA) with warning as the between-subjects factor was conducted on the narrative variables. Using Wilk's lambda, the multivariate test was not significant, F(2, 23) = 1.80, p > .05. The univariate results revealed that, as predicted, individuals who were warned that they would have to give their alibi used positive qualifiers at a higher rate (M = .31, SD = .28) over the course of the entire interview than those who were given no warning of the impending alibi interview (M = .11, SD = .18), F (1, 24) = 4.88, p < .05, [[eta].sup.2] = .17. Participants who were given no warning were more likely to pause at the beginning of their alibi statement (i.e., within the first ten words of their recall; Warning: M = .17, SD = .39; No warning: M = .71, SD = .83), F (1, 24) = 4.42, p < .05, [[eta].sup.2] = 16. Contrary to our predictions, none of the other variables differed significantly between groups for the entire interview or for the free recall (all p's > .05).
Experiment 1 demonstrated that when individuals were given warning that an alibi was about to be taken, they used language that was more confident (positive qualifiers) and less hesitant (pauses). While there were no significant differences between the groups in all of the variables predicted, the effects were in the predicted directions. We believe that one reason for this lack of significance was the size of the standard deviations. For example, although the word count for the free recall portion was 107.17 for the warned group and 75.21 for the no warning group (a difference of about 32 words) the standard deviations of these groups were 67.27 and 42.57 respectively. So while these effects were in the predicted directions, the large variance may have prevented significant effects for some of the variables (Tabachnik & Fidell, 2001). Similarly, the effect sizes were quite small. Had we obtained a larger sample size, it is possible that some of these effects would have become significant. We were unable to collect more data because once the analysis was complete, the interviewer/second author had moved to another state.
Participants. Fifty-four undergraduate students (19 males and 35 females) at the same small liberal arts college participated in this experiment in exchange for credit in their introductory psychology course. The mean age was 18.9 years (SD = 1.11).
Materials and Procedure. Each participant was given one transcript from Experiment 1 to read carefully. They were told that afterwards, they would be asked to provide their perceptions of the interviewee. Each transcript was read by two different participants who had no knowledge of the warning/no warning manipulation. After reading the transcripts, participants completed a questionnaire. The first section included demographics of the participant such as age and gender. The second section consisted of 15 questions on an 11-point Likert scale where the participants rated their perceptions of the interviewee and the alibi on items such as alibi believability, cohesiveness of the alibi, honesty of the interviewee, and strength of the alibi evidence (based on Allison et al., 2006). The perceived ratings on the Likert scaled questions of the two participants reading each transcript were then averaged to create one score for each credibility variable. Finally, participants were thanked and debriefed.
Pearson's correlations were computed in order to examine the relationships between the narrative variables (Experiment 1) and the perceived credibility of the alibis (Experiment 2). Results demonstrated significant correlations between several narrative variables and credibility ratings in the predicted direction. Higher frequencies of verbal hedges were associated with higher ratings of hesitancy of the interviewee (r = .47, p < .05). Higher frequencies of pauses (r = -.38, p < .05) and higher frequencies of hedges (r = -.41, p < .05) led to lower ratings of the strength of physical alibi evidence. Interviewees who hedged more were, understandably, seen as being more hesitant. Hedges also led to perceptions of weaker physical evidence. For interviewees, pausing gave the impression that it would be more difficult to find physical proof in support of their alibi claim.
There were also significant correlations that were not a part of our initial hypotheses. Greater use of spatial order (i.e., where things were located in space like "to my left") was associated with lower ratings of the credibility of the alibi (r = -.46, p < .05), the credibility of the corroborator (r = -.45, p < .05), and the sequencing of the alibi (r = -.39, p < .05). Also, high frequencies of personal pronouns were associated with lower ratings of hesitancy (r = -.40, p < .05). More negative qualifiers also led to lower interest ratings (r = .66, p < .001).
Although there were no hypotheses regarding the perceptions of the alibis being significantly affected by the no warning/warning manipulation, we conducted a 2-way between-subjects Multivariate Analysis of Variance (MANOVA) with warning as the between-subjects factor on all credibility ratings. Using Wilk's Lambda, there was a significant multivariate effect of warning, F (9, 16) = 3.36, p < .05, [[eta].sup.2] = .857, power = .86, however, the univariate between-subjects ANOVAs revealed no significant effects p's > .05. This indicates that the independent variable affected the multivariate combination of the dependent variables; the warning manipulation had no independent effect on the dependent measures.
Overall, there was some support for the hypotheses regarding narrative and perceptions of the alibi and interviewee. The results concerning verbal hedges and pauses were consistent with previous research (Voss & Van Dyke, 2001). If the interviewees used more pauses when describing their physical evidence and how easy it would be to produce said evidence, the evaluator may have inferred that the interviewees would not be able to find the evidence easily. The results on personal pronouns, spatial order, and negative qualifiers were more difficult to interpret. While previous research demonstrated that greater use of personal pronouns was detrimental to a defendant's case (Barry, 1991), it is possible that interviewees who used pronouns frequently may have been the same people who were more confident in themselves and their speaking ability. This confidence may have led raters to feel that the individuals were less hesitant. In contrast to our initial predictions, the warning manipulation had no significant effect on any of the perception variables in Experiment 2. The lack of expected results in Experiment 1 led us to believe that there would be no significant results regarding the warning manipulation in Experiment 2 and this was confirmed.
This study used alibis generated by participants to examine the narrative qualities of alibis, differences that occur in the generation of alibis between people who know they will have to provide an alibi and those who do not, and finally how the qualities of alibis affected perceptions of the alibi itself and the people presenting them. While the Experiment 1 results did not display all expected effects, there were differences between warning and no warning groups on the frequency of positive qualifiers and pauses at the beginning of the interview. Experiment 2's results demonstrated that different speech patterns can affect perceptions to some degree.
As is always the case when conducting laboratory research, certain real-world factors were not present. Although we were attempting to interview people in order to simulate a police interrogation, the Experiment 1 participants clearly knew they were not under any kind of investigation. As a result, their responses may have been different from individuals who were actually suspected of a crime. Future research should continue to focus on alibi generation and narratives, attempting to create scenarios that would closer more closely mimic how alibis are gathered in the real world.
Allison, M., Brimacombe, C. A. E., Hunter, M. A., & Kadlec, H. (2006). Young and older adult eyewitnesses' use of narrative features in testimony. Discourse Processes, 41, 289-314. doi: 10.1207/s15326950dp4103_3
Barry, A. K. (1991). Narrative style and witness testimony. Journal of Narrative and Life History, 1, 281-293.
Bruner, J. S. (1992). Acts of meaning. Cambridge: Harvard University Press.
Erickson, B., Lind, E. A., Johnson, B. C., & O'Barr, W. M. (1978). Speech style and impression formation in a court setting: The effects of "powerful" and "powerless" speech. Journal of Experimental Social Psychology, 14, 266-279.
Olson, E. A., & Wells, G. L. (2004). What makes a good alibi? A proposed taxonomy. Law and Human Behavior, 28, 157-176. doi:10.1023/B: LAHU.0000022320.47112.d3
Stromwell, L. A., Granhag, P. A., & Jonsson, A.-C. (2003). Deception among pairs: "Let's say we had lunch and hope they will swallow it!" Psychology, Crime & Law, 9, 109-124.
Tabachnik, B. G., & Fidel, L. S. (2001). Using multivariate statistics. Boston: Allyn & Bacon.
Voss, J. F., & Van Dyke, J. A. (2001). Narrative structure, information certainty, emotional content, and gender as factors in a pseudo jury decision-making task. Discourse Processes, 32, 215-243.
Voss, J. F., Wiley, J., & Sandak, R. (1999). On the use of narrative as argument. In Goldman, S. R., Graesser, A. C., & van den Broek, P. (Eds.), Narrative comprehension, causality, and coherence: Essays in honor of Tom Trabasso (pp. 235-252). Mahwah: Lawrence Erlbaum Associates Publishers.
Wells, G. L., Small, M., Penrod, S., Malpass, R. S., Fulero, S. M., & Brimacombe, C. A. E. (1998). Eyewitness identification procedures: Recommendations for lineups and photospreads. Law and Human Behavior, 22, 603-647. doi:10.1023/A:1025750605807
Meredith Allison, Stephen W. Michael, Kyla R. Mathews, & Amy A. Overman
Author info: Correspondence should be sent to: Dr. Meredith Allison, Dept. of Psychology, CB 2337, Elon University, Elon, NC 27244-2010.
|Printer friendly Cite/link Email Feedback|
|Author:||Allison, Meredith; Michael, Stephen W.; Mathews, Kyla R.; Overman, Amy A.|
|Publication:||North American Journal of Psychology|
|Date:||Nov 1, 2011|
|Previous Article:||Culture and gender differences in boredom proneness.|
|Next Article:||Students' first day of class preferences: factor structure and individual differences.|