Development and Initial Validation of an Inconsistent Responding Scale for the Youth Psychopathic Traits Inventory

Differential Item Functioning of the Youth Psychopathic Traits Inventory Across Race/Ethnicity and Gender Among a Sample of Justice-Involved Youth: An Item Response Theory Analysis

Research is yet to examine whether the items of the Youth Psychopathic Traits Inventory (YPI) function equally well across race/ethnicity and gender. The current study applies an item response theory analysis to detect differential item functioning (DIF) of the YPI subscale across White, Black, and Hispanic youth and males and females among a sample of justice-involved youth. Significant DIF was detected for several items between Black youth and White youth and Black youth and Hispanic youth. Few incidences of DIF emerged between White and Hispanic youth and between males and females. The findings suggest that subscales of the YPI provide more information for White and Hispanic youth compared with Black youth. They also suggest that while there was significant DIF in the difficulty of items, the direction of DIF did not substantially favor one group or another. Thus, the findings suggest that the YPI produces comparable estimates of psychopathic traits for females and males and for White and Hispanic youth. However, the results raise concerns about comparing YPI subscale scores between White and Black youth and Hispanic and Black youth. The findings have important implications for the use of the YPI subscales among diverse samples.

Detecting Inconsistent Responding on the Youth Psychopathic Traits Inventory-Short Form

The Youth Psychopathic Traits Inventory-Short Form (YPI-S) is a convenient measure for assessing psychopathy in settings with constraints on resources. However, the YPI-S does not contain a means of detecting careless or random response styles. The present study describes the development and evaluation of an inconsistent responding scale for the YPI-S using five archival samples that vary in language (English, German, Italian, Dutch) and other participant characteristics (juvenile offenders, adolescent students). Inconsistency scores resulting from the new scale effectively distinguished genuine participant responses from randomly generated cases (area under the ROC curve [AUC] = .85-.90) and from cases in which 50% of original responses were replaced with random data (AUC = .75-.82). The associations between the YPI-S and theoretically relevant correlates were reduced among participants exceeding proposed cutoff scores for profile validity compared with associations among more consistent respondents.