Climate-smart dairy technology adoption profiles and multidimensional food security among smallholder dairy households in Northwest Ethiopia

This study examines associations between climate-smart dairy technologies and practices (CSDTPs) adoption profiles and multidimensional food security (MDFS) among 734 smallholder dairy households in Northwest Ethiopia using cross-sectional survey design. Latent class analysis (LCA) identified three profiles: selective (19.5%), intermediate (45.9%), and intensive-diversified adoption (34.6%). Augmented-inverse probability weighting (AIPW) estimated the adjusted counterfactual contrast, with inverse-probability weighting regression adjustment (IPWRA) and inverse-probability weighting (IPW) estimators as a robustness checks. Compared with selective adoption, both adoption profiles had lower food insecurity deprivation scores across the four MDFS dimensions. The intensive-diversified adoption generally showed larger estimated differences relative to the selective profile mean: 23.46% for quantity, 10.35% for quantity, 27.22% for acceptability, and 16.62% for stability. Sensitivity analysis broadly supported the findings, although the quality contrast of intermediate profile was not robust to kebele-clustered standard errors. For binary MDFS, compared to selective adoption, intermediate and intensive-diversified adoption profiles were associated with 11.70% and 12.20% lower probabilities of food insecurity, respectively; their difference however, was not statistically significant. These findings highlight the relevance of context-specific combinations of CSDTPs in relation to MDFS. However, the cross-sectional design limits causal interpretation, longitudinal and experimental studies are needed to establish causal effects and investigate underlying mechanisms.