2.3.3. Data Analysis

Mortality was calculated as the total number of individuals knocked down or dead in a test unit as a percentage of the total number of individuals in the test unit at the end of the 24 h scoring period. If the mortality in the negative control test unit was <20% but ≥5%, then the observed mortality in each treatment test unit was corrected using Abbott's formula [13]. Where the control mortality was ≥20%, the results were discarded and the test replicate repeated.

Both datasets (age and number of mosquitoes per tube) were screened using a binomial generalized linear model (GLM), a binomial generalized linear mixed model (GLMM) with a random effect for biological replicates to account for any inter-assay variation, a binomial GLMM with a random effect for biological replicates, and a nested random effect for technical replicate to account for the intra-assay variation using the glmmTMB package in R [14]. For each analysis, the variable was treated as a factor, with 5 days used as a reference for age and 25 mosquitoes used as a reference for the number of mosquitoes in the tube. The negative controls were excluded from all analyses, as we were only testing for the influence of these factors on the deviation from the reference. A likelihood ratio test (LRT) was conducted to identify the best-fitting GLMM.
