Abstract
Within-individual standardization is a method of eliminating individual differences and habituation effects of physiological responses in the Concealed Information Test. In this study, effects of within-individual standardization on the following tests were investigated. (1) A paired t-tests between the critical item and the average of non-critical items. (2) A paired t-tests between the critical item and a single non-critical item. (3) An ANOVA with one within-subject factor (each item). (4) The effect size of each of the above statistical tests. Results indicated that larger was the differences of mean of population distribution between the response to the critical and non-critical item, more overestimated were within-individual standardized data statistics. However, with the exception of Cohen's d, this overestimation did not cause a serious error in the interpretation of the statistical test results or the effect size. In addition, it was indicated that the overestimation of each statistic might have resulted from the overly small standard deviation for each item of within-individual standardized data.
