A study of the psychometric properties of the Fagestrom Test for Nicotine Dependence.

A factor analysis of 772 Fagestrom Test for Nicotine Dependence (FTND) was performed in a population of French workers with STATISTICA software, utilizing tetrachoric correlations to account for the dichotomous responses of the FTND. Confirmatory factorial analysis found the two-factor solution recently proposed for the FTND to be unsatisfactory. An exploratory factorial analysis concluded that the best solution was monofactorial when Item 3 was deleted. A confirmatory factorial analysis of the FTND provided support for this monofactorial solution. With the deletion of Item 3, the FTND achieved a high internal consistency (Cronbach's alpha=.86). Thus, a revision of the FTND with the exclusion of Item 3 is proposed. This revised form of the FTND appears to assess a unidimensional construct indicating robust construct validity.