Extraordinary face recognition performance in laboratory and online testing

Abstract

The Cambridge Face Memory Test Long (CFMT+) is used to investigate extraordinary face recognition abilities (super‐recognizers [SR]). Whether lab and online presentation of the CFMT+ lead to different test performance has not yet been investigated. Furthermore, we wanted to investigate psychometric properties of the CFMT+ and the Glasgow face matching test – short (GFMT‐S). We report item difficulties, Cronbach's Alpha, and T norms for the CFMT+ and GFMT‐S depending on the presentation mode. We analyzed variations of CFMT+ and GFMT‐S performance by means of presentation mode, age, and gender. The results showed significant better CFMT+ performance for online presentation, women, and participants between 26 and 35 years of age. For the GFMT‐S, significantly higher test performance was shown for online participants exclusively. Overall, we discuss the modulating effects of participants' motivation for online assessments, the necessity of differential norms (e.g., for personnel decisions), and the application of T norms.

Face Memory and Face Matching

“…Moreover, the low MICs for the GFMT-S cannot be easily explained because the stimulus material is more homogeneous than in the CFMT+. Possibly, the low item difficulties (see Petersen & Leue, 2021) distort the variance of the items because some items had a variance of zero. The reported test-retest reliability of the GFMT-S corresponds to values of other face matching tests like the KENT face matching test with test-retest reliability of r (28) = 0.67 using a seven-day interval (Fysh & Bindemann, 2018).…”

“…For example, participants were excluded when they used smartphones or reported technological problems. Similar criteria have been used in Petersen and Leue (2021). Thus, N = 72 participants (49 females, 68.06%, M = 44.38 years, SD = 12.40, range 21-69 years) were included for data analysis.…”

“…The CFMT+ has no time limit. Test properties, for example, item difficulties and Cronbach’s α, of the adapted online CFMT+ are reported in Petersen and Leue (2021) for laboratory ( N = 109) and online settings ( N = 1,435; Cronbach’s α = .92).…”

“…The GFMT-S has no time limit. Some test properties, for example, item difficulties and Cronbach’s α, of the adapted online GFMT-S conducted in the laboratory ( N = 109) and online settings ( N = 1,435; e.g., Cronbach’s α = .71) are reported in Petersen and Leue (2021).…”

“…The Cambridge Face Memory Test Long Form (CFMT+) measuring short-term face memory (Russell et al, 2009) and the Glasgow Face Matching Test Short Form (GFMT-S) measuring face matching (Burton et al, 2010) are frequently used in the laboratory and online studies (Table 1 in Ramon et al, 2019). Currently, there are only a few reports of psychometric properties for the CFMT+ and the GFMT-S (e.g., Cronbach’s α in Petersen & Leue, 2021; Verhallen et al, 2017). To our knowledge, there are no test-retest reliability reports for the CFMT+ and only one for the GFMT-S (Stantic et al, 2021).…”