Scholarlink

<p>Food security is a critical pillar of household welfare, yet modeling its determinants is often complicated by over-dispersed data and an excess of food-secure households (zeros). This study identifies the sociodemographic, economic, and institutional drivers of food security among households engaged in bean and maize farming in Uganda, providing evidence for targeted resilience-building policies. Using secondary data from the National Agricultural Research Organization (NARO) Impact Evaluation Survey (<i>n</i> = 1,428), food security was measured as the Household Food Insecurity Access Scale (HFIAS), a nine-item experiential tool (0–27). To address the high proportion of zeros representing both “structurally” secure and “period-secure” households, a multivariate Zero–Inflated Negative Binomial (ZINB) regression was employed. This two-part framework technically justifies the distinction between the drivers of permanent food security and the drivers of insecurity severity. Although HFIAS is a bounded sum, post-estimation diagnostics confirmed the model’s appropriateness for this sample, with 100% of scores falling within the 0–22 range, avoiding “boundary leakage.” Model robustness was further validated using the Likelihood Ratio tests of alpha, the Vuong test, and Rootogram diagnostics. Results are reported as Incidence Rate Ratios (IRRs) for the count process and coefficients for the inflation process. The count model revealed that higher education level (Secondary IRR = 0.68; Tertiary IRR = 0.63, (<i>p</i> &lt; 0.01) and larger farm size (IRR = 0.98, (<i>p</i> = 0.012) significantly reduced the intensity of food insecurity. Conversely, being widowed or divorced (IRR = 1.41, <i>p</i> = 0.043) and residing in the Western region (IRR = 1.27, <i>p</i> = 0.008) significantly increased insecurity severity. The inflation model demonstrated that tertiary education and farm size were primary predictors of “structural food security,” significantly increasing the probability of a household being consistently food secure. Food security in Uganda is driven by a two-stage process where the level of education and land assets govern both long-term resilience and the intensity of hunger shocks. Policymaking should pivot toward region-specific, literacy-sensitive interventions and social protection for widowed heads of households. Strengthening secondary education and addressing land fragmentation are essential to moving smallholders beyond transient security toward structural resilience.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Modeling food security status among beans and maize farming households in Uganda

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月2日

查看全文
  • Derrick Watsala,
  • Felix Wamono,
  • John Bosco Asiimwe

摘要

Food security is a critical pillar of household welfare, yet modeling its determinants is often complicated by over-dispersed data and an excess of food-secure households (zeros). This study identifies the sociodemographic, economic, and institutional drivers of food security among households engaged in bean and maize farming in Uganda, providing evidence for targeted resilience-building policies. Using secondary data from the National Agricultural Research Organization (NARO) Impact Evaluation Survey (n = 1,428), food security was measured as the Household Food Insecurity Access Scale (HFIAS), a nine-item experiential tool (0–27). To address the high proportion of zeros representing both “structurally” secure and “period-secure” households, a multivariate Zero–Inflated Negative Binomial (ZINB) regression was employed. This two-part framework technically justifies the distinction between the drivers of permanent food security and the drivers of insecurity severity. Although HFIAS is a bounded sum, post-estimation diagnostics confirmed the model’s appropriateness for this sample, with 100% of scores falling within the 0–22 range, avoiding “boundary leakage.” Model robustness was further validated using the Likelihood Ratio tests of alpha, the Vuong test, and Rootogram diagnostics. Results are reported as Incidence Rate Ratios (IRRs) for the count process and coefficients for the inflation process. The count model revealed that higher education level (Secondary IRR = 0.68; Tertiary IRR = 0.63, (p < 0.01) and larger farm size (IRR = 0.98, (p = 0.012) significantly reduced the intensity of food insecurity. Conversely, being widowed or divorced (IRR = 1.41, p = 0.043) and residing in the Western region (IRR = 1.27, p = 0.008) significantly increased insecurity severity. The inflation model demonstrated that tertiary education and farm size were primary predictors of “structural food security,” significantly increasing the probability of a household being consistently food secure. Food security in Uganda is driven by a two-stage process where the level of education and land assets govern both long-term resilience and the intensity of hunger shocks. Policymaking should pivot toward region-specific, literacy-sensitive interventions and social protection for widowed heads of households. Strengthening secondary education and addressing land fragmentation are essential to moving smallholders beyond transient security toward structural resilience.