What is it about?

Research that aims to improve and test the ways eye witnesses produce face composites have led to a range of new technologies. The current research shows how the assessments of these technologies have been potentially misleading because they have not considered the variability of the witnesses creating them. The current research explains how this can lead to misleading conclusions about the effectivness of eye-witness supportive technologies. A range of suggestions for improving how face-composite research is carried out are made.

Featured Image

Why is it important?

The reserach considers an arcane but important statistical point. The variability of both the participants recognising composites and those constructing the composities need to be included within the statistical model in order to safely test whether a change to the face-composite method is significant. A review of previous research on face-composite methods suggest that few studies provide this level analysis and so their significance could have been overstated. Methods that can resolve this issue are explained.


Helping a witness to provide an image of an offender is an important pursuit and new technologies are able to support this. Finding the best way to do this requires careful attention to detail and statistical analysis of the available date. The current research provides details on how researchers should approach these topics in order to ensure that their conclusions are statistically valid.

Michael Lewis
Cardiff University

Read the Original

This page is a summary of: Fixing the stimulus-as-a-fixed-effect fallacy in forensically valid face-composite research., Journal of Applied Research in Memory and Cognition, June 2023, American Psychological Association (APA),
DOI: 10.1037/mac0000128.
You can read the full text:



The following have contributed to this page