3.3.7 Data analysis
Flea data were recorded in Microsoft Excel 2016 and prepared in a readable data set format, data set were imported in R programming language (R Core Team version 4.2.2 of 2022) from which all statistical analyses were conducted with a significance level (α) of 0.05. Data were tested for normality before analyzed and were found not normally distributed (Shapiro-Wilk test p <.05) leading us to use non-parametric test for inferential statistics. Rodent’s flea load on examined rodents were evaluated by assessing total flea indices and specific flea indices (Zimba et al ., 2012). Percentage prevalence of rodents infested with fleas (proportion of examined rodents positive for fleas) were computed with 95 % confidence interval for proportion and multiplied by 100%.
Chi-square χ2 test of association was applied to examine the relationship of abundance of species of fleas with rodent species, plague and non-plague foci villages, season, habitat type and rodent sex. The nature of association was further analyzed to find the contribution of level of factors to the Chi-square result using absolute standardized residuals and the relationship were presented in balloon chart (Kassambara, 2022). Flea biased parasitism on rodent’s sex and weight was analyzed by assessing the variation of flea abundance between sexes using Wilcoxon rank sum test and through evaluation of correlation between flea abundance and rodent’s weigh using Spearman rank correlation test (rho) respectively; these factors were further analyzed in generalized linear model (GLM) to assess their influence on flea abundance. The GLM was used to establish the model variables that described the influence of predictors of abundance of flea. Model was selected based on AIC in which model with possible lowest AIC was selected with help of step-AIC function in MASS package of R programming language (Venables and Ripley, 2002). Predictor variables were both numerical and categorical variables that include, host characteristics (rodent species, sex and weight), habitats, season and plague locations.