Online

Geist, Thomas

Psychology

2021

M.A.

Though rates of smoking have declined for decades, the tobacco epidemic persists. During this time, denormalization of smoking and the tobacco industry emerged as viable strategies to decrease smoking rates. However, researchers have become increasingly concerned with denormalization strategies' impact on the stigmatization of smokers, and have questioned their ethics and continued utility. Unfortunately, current research lacks a theoretically and psychometrically sound measure to represent smokers' experience of stigma and help explore these issues. This study sought to develop and validate such a measure, the Smoker Self-Stigma Questionnaire (SSSQ). Initial items for the SSSQ were developed to cover three domains of self-stigma (enacted, felt, and internalized) as proposed by Bos et al. (2013). To construct the items, a panel of three judges synthesized the sematic representations of self-stigma reflected by a) the items of existing addiction-related, self-stigma measures, b) the findings of qualitative research investigating smokers' stigma in diverse populations, and c) the descriptions of denormalization strategies and denormalization consequences outlined in prominent literature reviews addressing denormalization. A total of 88 items were rated by 13 tobacco-research investigators, who classified each item into one of the three stigma domains, and rated each item's smoking-related face validity. This process led to the selection of 45 items (15/self-stigma domain), which were administered to 592 participants recruited through Amazon's Mechanical Turk (MTurk). Using confirmatory factor analyses (CFA) and one half of the dataset, I tested a hypothesized three-factor structure and then eliminated items with the goal of selecting at least 5 items per dimension for further psychometric testing and validation. Via inspection of fit indices, factor loadings, and modification indices, I retained an 18-item measure with 6-items per self-stigma dimension, to which I applied CFA using the second half of the data set, as well the full data set through a series of invariance, cross-validation tests. With a pre-specified three-factor structure, the 18-item measure yielded excellent fit indices, adequate and significant factor loadings, and minimal problematic modification indices, Correlation analyses provide evidence of high convergent validity with measures of nicotine dependence and motivation to quit smoking, as well as divergent validity by showing a low correlation with social desirability.