Psychometric Validity of the Minnesota Tobacco Withdrawal Scale in Smokers Attempting to Quit: Longitudinal Invariance across Smokers and Abstainers

Tobacco withdrawal symptoms vary during smoking cessation increasing relapse risk; therefore, a longitudinal invariant measure seems necessary to validly assess withdrawal changes. This study aimed to examine the psychometric properties of the 7-item Minnesota Tobacco Withdrawal Scale (MTWS) during cessation, and to analyze longitudinal invariance across smokers and abstainers. We conducted a longitudinal study with 309 Spanish smokers (56.2 +/- 9.9 years; 52.4% women). We assessed withdrawal at three occasions: pre-treatment (T1), week-12 (T2), and week-24 (T3). Reliability, validity, and invariance analyses were performed. MTWS reliability remained adequate over time (omega(T1) = 0.78; omega(T2) = 0.68; omega(T3) = 0.80) in both smokers and abstainers, with satisfactory temporal stability (ICC = 0.73). MTWS correlations with anxiety, depression, and nicotine dependence (r(s) > 0.3) supported convergent and concurrent validity. MTWS showed no correlation with craving at T2 (r(s) < 0.24), suggesting discriminant validity. Unifactorial structure proved partial scalar invariance across time (chi(2) = 246.009; CFI = 0.91; IFI = 0.91; SRMR = 0.09), yet longitudinal invariance between abstainers and smokers was not supported. Across groups, partial scalar invariance was only achieved at T2. Noninvariance at T3 was mainly due to item irritability (d(MACS) = 0.93). The MTWS is reliable and valid measure to assess withdrawal changes during cessation. Long-term MTWS comparisons between smokers and abstainers should be taken with caution since different withdrawal patterns may exist.