To investigate the relative recognisability of internal and external features of a facial composite.
1 of 13
Describe the methodology
3 laboratory experiments
2 of 13
Who were the participants in experiment 1?
30 staff and students from Stirling University. Paid £2 to sort composites. 15 male, 15 females. Mean age of 29.2
3 of 13
Who were the participants in experiment 2?
48 undergraduates at Stirling University. 21 males, 27 females. All volunteers.
4 of 13
Describe the procedure of experiment 1
Stimuli were photos of 10 celebrities and 40 composite images produced by E-FIT, PRO-fit, Sketch and EvoFIT. Face was clean shaven, no glasses.
5 of 13
Describe the 3 sets of composites used
A complete set, a set containing internal features and a set containing external features.
6 of 13
What were participants tested on and what were they asked to do?
Tested individually on one of the 3 conditions. Independent measures. Asked to place each composite in front of each celebrity face until completed.
7 of 13
Describe what happened in experiment 2
Used a photo array/line-up with distracter faces/foils making the task more difficult. Faces+foils made easy (all different) or hard (all similar to each other).
8 of 13
How were the composites presented and what did the participants do?
Composites presented 1 at a time along with the photo array. Participant had to pick out of celebrity faces from array which matched composite. Either internal/external composites.
9 of 13
What were the results from experiment 1?
Whole composites and those of external features were sorted at roughly 35% correct. Internal feature composites were only 19.5% correct.
10 of 13
What were the results from experiment 2?
Composites of external features were identified correctly 42% of time compared to 24% for internal features. Consistent across array.
11 of 13
What was concluded from these experiments?
Participants performed just above chance with internal features. Performed equally as well with external features/whole faces.
12 of 13
What do these results show?
Show that there is something about internal features of a face that don't work well when trying to create a reconstruction even when face is familiar.
Comments
No comments have yet been made