Establishing Construct Validity of a Stages-of-Change Algorithm for Physical Activity

Abstract

Purpose.

Examine the construct validity of a modified stages-of-change (SOC) scale among adults with physical disabilities by investigating the relations among the transtheoretical model (TTM) constructs, physical activity (PA), and SOC. Additionally, the SOC algorithm was compared with previously recommended SOC scales.

Design/Setting.

Cross-sectional online survey.

Subjects.

271 adults with physical disabilities (x̄_age = 49.25; 73.4% females).

Measures.

Standardized questionnaires for the TTM constructs and PA.

Analyses/Results.

Based on the multivariate analysis of variance, the hypothesized relations between the SOC and the TTM constructs were supported. According to the analysis of variance, PA levels increased almost linearly across the SOC. Precontemplators (x̄ = 8.53 metabolic equivalent [MET]-h/d) had lower PA scores than contemplators (x̄ = 14.27 MET-h/d) and preparators (x̄ = 17.64 MET-h/d), who in turn had lower scores than those in action (x̄ = 23.03 MET-h/d). Based on the direct discriminant function analysis, all TTM constructs were important predictors of the SOC (e.g., r = .85 for behavioral processes; r = .69 for self-efficacy; and r = .58 for cognitive processes). The most accurately predicted SOC were action (76.9%), precontemplation (70.6%), contemplation (44.6%), and preparation (14.9%), with an overall classification accuracy of 56.8%. The study participants were more evenly distributed across the SOC algorithm compared with previous studies.

Conclusion.

The construct validity of the SOC scale was supported. The scale differentiated almost linearly the levels of PA and most TTM constructs, and participants were well-represented across SOC. Health promoters should feel confident in using the proposed SOC scale and the TTM to implement stage-matched PA motivational programs.