Date on Master's Thesis/Doctoral Dissertation


Document Type

Doctoral Dissertation

Degree Name

Ph. D.


Psychological and Brain Sciences

Degree Program

Experimental Psychology, PhD

Committee Chair

Lyle, Keith

Committee Co-Chair (if applicable)

DeCaro, Marci

Committee Member

DeCaro, Marci

Committee Member

Danovitch, Judith

Committee Member

Pani, John

Committee Member

Hughes, Thomas

Author's Keywords

eyewitness memory; metamemory; interviews


Eyewitnesses are important for criminal investigations. When interviewing witnesses, police first ask witnesses to describe what they observed. This initial statement is a free narrative (FN) and is usually highly accurate. Next, police ask follow-up questions (FQs), which generate additional information but often at the cost of accuracy. This dissertation aims to examine factors that may contribute to the effect of interview format on accuracy and whether FQ instructions change detail quality provided in follow-up question responses (FQRs). Additionally, individual differences in responding to FQ instructions were explored. Subjects watched a movie clip depicting a crime. Next, subjects filled out personality questionnaires and then typed their FNs. Subjects were then given one of five FQ instructions that encouraged accuracy (Accuracy+), encouraged informativeness (Informative+), encouraged “I don’t know” responses if needed (IDK+), discouraged “I don’t know” responses unless absolutely necessary (IDK-), or only told them they would be asked FQs (Control). Ten variables were used to measure interview quality (e.g., accuracy, precision, quantity). FN quality was compared to FQR quality in the Control condition to determine the effect of interview format, without influence of instructions. FQR quality was compared across FQ conditions to determine the effect of FQ instructions. FQR quality was correlated with conscientiousness and social desirability scores to explore individual differences. Analyses were conducted once using all FQRs and once using only responses if the corresponding FQ topic had been mentioned during the preceding FN. Interview format affected all variables except overall accuracy. This null effect on accuracy does not align with previous interview literature and limited the interpretation of other effects. IDK+ instructions were the only instructions that affected FQR quality. IDK+ subjects said “I don’t know” significantly more frequently, and provided significantly fewer inaccurate details, in FQRs than Control subjects. Yet, IDK+ subjects were not significantly more accurate than Control subjects. Thus, it remains unclear whether IDK+ instructions selectively filter inaccurate details or reduce total detail output. The effects of IDK+ instructions were only significant when using all FQRs. Significant individual differences emerged in Accuracy+, Informative+, and Control conditions when using all FQRs.