Psychometric properties of the FTCD and Brief WISDM: support for validity in a legal-system-involved sample.

Abstract

The Fagerstrom Test of Cigarette Dependence (FTCD) and Brief Wisconsin Index of Smoking Dependence Motives (WISDM) are widely-used measures of smoking dependence. The FTCD was previously found to have 1- and 2-factor structures and Brief WISDM has been found to have an 11-factor and 11-factor hierarchical structure. As such, the current study sought to further investigate the psychometric properties of the FTCD and Brief WISDM with a novel criminal-legal system-involved sample using both a factor-analytic and an Item Response Theory (IRT) approach. Data from 517 criminal-legal system-involved adults (i.e., 18 years of age or older) who smoke from Alabama, USA were analyzed. Confirmatory factor analyses (CFA) were conducted on 1-factor and 2-factor structures of the FTCD and 1-factor, 11-factor, and 11-factor hierarchical structures of the Brief WISDM. IRT analyses investigating item discrimination and threshold parameters were also conducted on the Brief WISDM. The CFA showed poor fit for a single-factor structure, and mixed results for two 2-factor results for the FTCD. CFA also showed poor fit for a single-factor, and mixed results for the 11-factor model. Initial IRT investigations using the 11-factor model showed strong item discrimination, but non-ordered threshold parameters. Two-factor structures for the FTCD and the 11-factor model for the Brief WISDM were partially supported in a criminal-legal population, suggesting continued support for the multidimensional structure of the measures. Additionally, exploratory IRT analyses suggested good discrimination across the use spectrum for the Brief WISDM. The Fagerstrom Test of Cigarette Dependence (FTCD) and Brief Wisconsin Index of Smoking Dependence Motives WISDM are two widely used measures of nicotine dependence, though previous research has shown mixed results for their internal consistency and factor structure. The current study used a unique sample of criminal-legal-involved participants who generally have moderate to high levels of nicotine dependence. The current study found that the FTCD displayed poor internal consistency, poor fit for a single-factor model, but mixed support for two two-factor models. The Brief-WISDM was found to have strong internal consistency, poor fit for a single-factor model, but mixed fit for an 11-factor model and good item discrimination.