The Journal of

Introduction: To introduce Bayesian method in validation of transtheoretical model’s Self-Efficacy and Decisional Balance for nutritional behavior improvement among Prediabetes with ordinal data.

Methods: This is an Experimental trial with parallel design and sample was included 220 Prediabetes who Participated in screening program and had over 30 years old, fasting blood glucose ranged 100-125 and at least elementary Education. We used OpenBugs 3.2.3 to fit Bayesian ordinal factor analysis to achieve validation of TTM’s decisional balance and self-efficacy.

Results: All of the factor loadings corresponded to mentioned constructs was significant at α= 0.05%. That support validation of the Constructs. Correlation between Pros and Cons was not significant(-0.076, 0.007).Furthermore a specific statistical model for ordinal data created that can estimate odds ratios and marginal Probabilities for each choice of any item in questionnaire.

Conclusion: Thanks to benefits of Bayesian method in use of prior information such as Meta-analysis and other resources, In comparison to similar studies that used standard or other factor analysis for ordinal data, our results had good accuracy(with aspect to standard deviation) even with lower sample size.so the results can be used  in future clinical researches.