Multinomial Logistic Regression to Estimate and Predict the Influence of Gentrification and Urban Renewal on Residential Choices in Kisumu City, Kenya

Residential location choices involve making trade-offs between housing status and dwelling quality because no single residential area can provide all of its housing needs. In this work, we used multinomial logistic regression to predict the effects of gentrification and urban renewal on residential choices. A mixed-method research design was employed in the study. The Multinomial logistic regression’s likelihood ratio test showed that the relationship between the component gentrification and residential choices was significant (p = 0.022) at the p<0.05 level, while that of urban renewal was significant (p=0.000) at the p<0.05. Hence, gentrification and urban renewal significantly influenced choice. The odds ratio for gentrification in a low-income neighborhood was found to be 0.672 (<1). This meant that the probability of selecting a home from a low-income neighborhood fell by 33% relative to selecting a home from a high-income neighborhood. Residents are more likely to consider gentrification when selecting residents from a high-income area than a low-income one. In the middle-income neighborhood, the odds ratio for gentrification was found to be 0.714 (<1). This meant that the probability of selecting a home from a low-income neighborhood decreased by 29% relative to selecting a home from a high-income neighborhood. Residents are more likely to consider gentrification when selecting residents from a high-income area than a middle-income one. In the study to predict how the effect of urban renewal on residential choices varied between neighborhoods, the odds ratio was found to be 2.829(>1), implying that the probability of selecting a home from a low-income neighborhood increased by 182% as compared to choosing from a high-income neighborhood. Residents are more likely to consider urban renewal when selecting residents from low-income areas than high-income areas. As in the case of the middle-income neighborhood, the odds ratio for gentrification was found to be 2.419 (>1). This meant that the probability of selecting a residence in a middle-income neighborhood decreased by 142% relative to selecting a home in a high-income neighborhood. Residents are more likely to consider urban renewal when selecting a residence from a middle-income area than from a high-income one. These findings will help with policy formulation in housing provision by the national government, the county government and property developers.