2.6 Data management and statistical analysis
Data were managed in Microsoft Excel 2007 version spreadsheets (Microsoft Corporation, Redmond, WA, USA). Analytic and descriptive statistics were used for data analysis in Epi Info 7 version 10. Associations between the drivers for antimicrobial misuse in farms and outcome responses were first subjected to univariable analysis using Chi-square tests (Dohoo et al., 2009). Factors found to be significant statistically at this stage were then fitted to likelihood stepwise backward multivariable logistic regression models to control for confounders and test for effect modifiers. The Goodness of Fit of the final model was assessed with Hosmer and Lemeshow test and found to be good. In all analyses, p<0.05 was considered statistically significant. Frequencies and proportions were the outputs of this study’s descriptive analysis. Relationships of both occupational farmers’ response proportions on practices and risk pathways were determined by Pearson’s Chi-square test or Fischer’s exact test, as appropriate.