Modelling association between two Multiple Response Categorical Variables (MRCV) using generalized log-linear models: the case of urban agriculture in Mutare City

Urban Farm DB summary

Using survey data from 115 household farmers collected in December 2020 across urban and peri-urban Mutare City, Zimbabwe, including the Weirmouth Park and Fern Valley areas, this study applied generalized log-linear modelling to model associations between multiple response categorical variables (MRCV) on urban agriculture. Simultaneous Pairwise Marginal Independence tests revealed associations, and log-linear tests found a strong positive association between field agricultural activities and herbicide use, and a strong negative association between municipal vacant places and field agricultural activities; the study could not model associations among three or more MRCVs because the required computations exceeded 2GB of memory.