Development of the PROMIS Social Motivations for Smoking item banks.

Abstract

Smoking behavior is influenced by social motivations such as the expected social benefits of smoking and the social cues that induce craving. This paper describes development of the PROMIS Social Motivations for Smoking item banks, which will serve to standardize assessment of these social motivations among daily and nondaily smokers. Daily (N = 4,201) and nondaily (N =1,183) smokers completed an online survey. Item factor analyses, item response theory analyses, and differential item functioning analyses were conducted to identify a unidimensional set of items for each group. Short forms (SFs) and computer adaptive tests (CATs) were evaluated as tools for more efficiently assessing this construct. A total of 15 items were included in the item banks (9 items common to daily and nondaily smokers, 3 unique to daily, 3 unique to nondaily). Scores based on full item banks are highly reliable (reliability = 0.90-0.91). Additionally, the item banks are strongly unidimensional and perform similarly across gender, age, and race/ethnicity groups. A fixed SF for use with both daily and nondaily smokers consists of 4 items (reliability = 0.80). Results from simulated CATs showed that, on average, fewer than 5 items are needed to assess this construct with adequate precision using the item banks. A new set of items has been identified for assessing the social motivations for smoking in a reliable, standardized manner for daily and nondaily smokers. In addition to using the full item banks, efficient assessment can be achieved by using SFs, employing CATs, or selecting items tailored to specific research or clinical purposes.