*2.3. Data Analyses*

The recorded interviews were transcribed verbatim into English. The questionnaires were cleaned, coded, and subsequently inputted into a statistical computer software, SPSS (version 20.0), for analysis. Crosstabulations were performed to examine the relationship between variables across the study communities. To determine uniformity between respondents in the different study communities, a Chi-square (χ 2 ) test of homogeneity analysis was performed. *p* < 0.05 was used as a criterion for statistical significance. The Friedman

Test was performed to compare the mean rank of the factors responsible for the 2020 minor season's maize production failure in the communities. When the test showed overall statistical significance, a post hoc test, a Wilcoxon signed-rank test with Bonferroni-corrected alpha level, i.e., an alpha level divided by the number of comparisons, was used to isolate factors that were significantly different. Relevant quotes extracted from the transcripts are used to emphasize key quantitative descriptions in the results and discussion sections. Rainfall, temperature, and maize grain yield data were analyzed in Microsoft Excel. The frequencies of the length of the dry spell and the longest dry spell in the major and minor rainy seasons were computed using Instat+ version 3.36. The dry spell was defined as four or more consecutive days without rain or with precipitation of less than 1 mm. The longest dry spell was computed according to Gbangou, et al. [10] as the largest number of consecutive days during which rainfall was less than 1 mm in the season. For each year, the major season was defined as April–July, and the minor season was September–October.
