Stage Stability and Test-Retest Reliability of the Green Eating Survey

Abstract

To evaluate test-retest reliability of the Green Eating (GE) survey including Stage of Change (SOC), Decisional Balance (DB), and Self-Efficacy (SE) towards pro-environmental eating behaviors. GE survey was developed and validated in Spring 2012 and a strong relationship was demonstrated between SOC and GE constructs. Eighty-five college students (age 18 to 24) completed the GE items as part of a larger survey at baseline and post-intervention. The 6-week whole grain/low-fat milk intervention was not expected to directly impact constructs in GE survey. The sample was restricted to subjects who remained in the same stage from baseline to post (n=40). SOC was measured using a single-item. Coefficient alphas were calculated at both time points for GE-behavior (6-items), DB-Pro (5-items), DB-Con (5-items), and SE (5-items) scales. Chi-Square analyses assessed SOC and paired t-tests and correlations compared scales at baseline and post. Distribution of subjects in pre-action stages was similar to previous research. Precontemplation and maintenance were the most stable stages with 50 to 54% of subjects remaining in stage while other stages demonstrated greater stage movement (χ2=47.50, p<0.05). GE behavior (baseline: α =.89; post: α =.93), DB-Pro (α =.87; α =.93), DB-Con (α =.63; α =.85), and SE (α=.88; α =.86) with baseline/post scale scores correlated significantly (r=.88, r=.63, r=.34, r=.65, respectively; p<0.05). There were no differences between baseline and post mean scores. Stage movement suggests there may have been an intervention effect. The consistently high coefficient alphas and test-retest reliability of the measures among stage-stable subjects provide validation of the GE survey as a measure to assess pro-environmental eating behaviors in the college population.