Detection of insincere grips: multivariate analysis approach

Abstract

BACKGROUND Smith and Chengalur were successful in using grip test and the cut-off criterion method to detect fake grips in healthy subjects and subjects with hand injuries. The purpose of this study is to test if methods other than the cut-off method, i.e. discriminate analysis and logistic analysis methods, cn be used to provide a more accurate detection of fake grips, with the use of sustain grip test. METHOD Two groups of subjects were recruited. Group one consisted of 40 healthy subjects and group two consisted of 80 subjects with history of various types of hand injuries. For each subjects they were randomly assigned a hand to perform fake grips. ANALYSIS Analyses were performed separately in Group 1 and Group 2, and then combining group 1 and 2. For each group, we tried to estimate the percentage of accuracy of detection of fake grip by the discriminate analyses method and the logistic analysis method. RESULTS For group 1, the percentage of accuracy in detecting fake grips were 66.7% in male and 39.5% in female by discriminate analysis, and 77.8% in male and 84.2% in female by logistic regression. Both methods can detect 90%-100% sincere grips. For group 2, the percentage of accuracy in detecting fake grips were 58.3% in male and 58.0% in female by discriminate analysis, and 88.3% in male and 77.0% in female by logistic regression. Both methods can detect 90%-100% sincere grips. Further analysis by testing the relationship between degree of faking and accuracy of fake detection, it was found that the higher degree o faking, the more accurate in the detection. For subjects who performed 20% or 50% of their maximum grips, we can detect 92% faking in male and 86.5% in female by using logistic regression method.

More information

Identifiers

Catalogue record

Export record