Skip to main content

Table 2 Literature measuring correlation for lineup accuracy

From: The importance of decision bias for predicting eyewitness lineup choices: toward a Lineup Skills Test

Paper

Predictor

Outcome

r

N

CI lower

CI upper

Andersen et al., 2014

CFMT

CP simultaneous lineup

0.26a

119

0.09

0.42

CFMT

CA simultaneous lineup

0.28a

119

0.1

0.44

CFMT

CP sequential lineup

nsb

119

  

CFMT

CA sequential lineup

0.27a

119

0.09

0.43

Bindemann et al., 2012

Hit rate, Bruce 1-in-10 as memory task

Probability of being a good witness (choosers)

0.7

37

0.49

0.83

Hit rate, Bruce 1-in-10 as memory task

Probability of being a good witness (choosers)

0.83

86

0.75

0.89

FA rate, Bruce 1-in-10 as memory task

Probability of being a good witness (nonchoosers)

0.49

43

0.22

0.69

FA rate, Bruce 1-in-10 as memory task

Probability of being a good witness (nonchoosers)

0.38

99

0.2

0.54

Deffenbacher et al., 1978

Y/N face recognition overall accuracy

4-person simultaneous lineup of class exam administrators

−0.28

45

−0.53

0.01

Hosch, 1994

BFRT

Single lineup of experimenter (half CP)

0.54

32

0.24

0.75

BFRT

Single lineup of experimenter (half CP)

0.39

38

0.08

0.63

BFRT

Single lineup of experimenter (half CP)

0.41

27

0.04

0.68

Y/N face recognition sensitivity

Single lineup of experimenter (half CP)

−0.07

33

−0.4

0.28

Y/N face recognition sensitivity

Single lineup of experimenter (half CP)

−0.21

36c

− 0.5

0.13

Y/N face recognition response bias

Single lineup of experimenter (half CP)

0.5

33

0.19

0.72

Y/N face recognition response bias

Single lineup of experimenter (half CP)

0.28

36c

−0.05

0.56

Kantner & Lindsay, 2014

Y/N face recognition response bias

1 CP and 4 CA lineups

0.29

65

0.06

0.5

  1. BFRT Benton Facial Recognition Task, CA culprit absent, CFMT Cambridge Face Memory Test, CI confidence interval, CP culprit present, FA False alarm
  2. aChi-squared values converted to correlation coefficients at campbellcollaboration.org/escalc/html/EffectSizeCalculator-R5.php
  3. bNon-significant chi-squared value not reported in manuscript
  4. cSample sizes not reported, but are inferred based on reported p-values