School of Statistics and Planning (SSP) Collections

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 5 of 447
  • Item
    Impact of rural electrification on household poverty in Uganda: Propensity score matching approach.
    (Makerere University, 2026) Irinatwe, Derrick
    Rural electrification is a central pillar of Uganda’s development agenda for reducing rural poverty, yet only about 9–11% of rural households have electricity. This study examined the impact of rural electrification on household poverty in Uganda using the 2023/24 Uganda National Household Survey (UNHS). Of the 15,762 households surveyed nationally, 9,891 rural households headed by persons aged 18 years and above were analysed, comprising 1,795 households that have electricity and 8,096 that do not. Propensity Score Matching (PSM), a quasi-experimental method that pairs households that have electricity with similar households that do not, was applied to reduce selection bias, yielding a matched sample of 1,741 households in each group (3,482 observations) within a caliper of 20% of the standard deviation of the propensity score. Covariate balance was assessed using the standardised mean difference (SMD), with values below 10% indicating acceptable balance. Logistic regression within the PSM framework was then used to estimate the impact of rural electrification on household poverty, controlling for household size, age, sex, marital status, education level, occupation and region, and including interactions of electricity status with region and education level. Robustness to unobserved confounding was assessed using Mantel–Haenszel sensitivity analysis. Households that have electricity were 72% less likely to be poor than those that do not (OR = 0.28, p < 0.01). Households with more than five members were 2.9 times more likely to be poor (OR = 2.88, p < 0.01), and households in Eastern Uganda were 2.5 times more likely to be poor than those in Central Uganda (OR = 2.51, p < 0.01). Higher education reduced poverty, with post-secondary education lowering the odds of poverty by 80% (OR = 0.20), while households headed by persons in non-agricultural informal and semi-skilled occupations were 53% less likely to be poor (OR = 0.47, p < 0.01). The interactions showed that the poverty-reducing impact of electricity was weaker in Western Uganda than in Central Uganda, and weaker among households headed by persons with primary education than among those with no formal education. Having electricity was associated with non-youth headship, region, marital status, education and occupation. The study concludes that rural electrification significantly reduces household poverty in Uganda, although its impact is not uniform across households and regions. Policy recommendations are to expand and target rural electrification, especially in Eastern and Northern Uganda; to strengthen education and skills training so that households can use electricity productively; and to direct social protection and regional development support to large households and lagging regions.
  • Item
    Modelling anaemia severity among women aged 15–49 years in Uganda: a comparative analysis of ordered logistic and ordered probit regression models.
    (Makerere University, 2026) Lamwaka, Brenda
    Anaemia remains a major public health challenge among women of reproductive age, particularly in low-resource settings, where it poses significant risks to maternal health and wellbeing. This study examined the determinants of anaemia severity among women aged 15–49 years in Uganda and compared the performance of ordered logistic and ordered probit regression models. The analysis was based on complete data from 6,363 women drawn from the 2016 Uganda Demographic and Health Survey (UDHS). Descriptive statistics were used to summarise the characteristics of the study population and the prevalence of anaemia severity. Subsequently, Ordered Logistic Model (OLM) and Ordered Probit Model (OPM) were fitted to examine the associations between anaemia severity and selected socio-demographic, obstetric and maternal, nutritional, and preventive factors. Model performance was evaluated using the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), deviance and log-likelihood, with statistical significance assessed at the 5% level. The findings indicated that 1.0% of the women had severe anaemia, 12.9% had moderate anaemia and 15.9% had mild anaemia. Model comparison favoured the OLM, which demonstrated better overall fit (lower AIC and BIC and higher log-likelihood) and comparatively stronger predictive discrimination, particularly for severe anaemia. Consequently, the OLM was selected for the final analysis. The results revealed that women aged 45 years and above (OR = 1.317; 95% CI: 1.047, 1.656), those in the middle (OR = 1.259; 95% CI: 1.009, 1.571) or fourth (OR = 1.299; 95% CI: 1.044, 1.616) wealth quintiles, and women with secondary (OR = 1.283; 95% CI: 1.049, 1.568) or higher education (OR = 1.357; 95% CI: 1.003, 1.807) had higher cumulative odds of being in more severe anaemia categories. Conversely, pregnancy status was associated with lower cumulative odds of greater anaemia severity (OR = 0.422; 95% CI: 0.341– 0.521). Gravidity was positively associated with anaemia severity (OR = 1.047; 95% CI: 1.023, 1.071), while overweight status (OR = 1.596; 95% CI: 1.135, 2.244) and its interaction with iron supplementation (OR = 0.597; 95% CI: 0.413, 0.865) also showed significant associations. The study concludes that anaemia severity among women of reproductive age in Uganda is shaped by socio-demographic, obstetric, and nutritional factors. Effective control demands context-specific interventions spanning education, nutrition, iron supplementation, and screening services.
  • Item
    Women’s empowerment and fertility preferences in Uganda: a multi-domain analysis (2006-2022)
    (Makerere University, 2026) Kirunga, Shania
    Ugandan women continue to portray relatively high fertility preferences of 5 children despite ongoing improvements in women’s empowerment and access to reproductive health services. This study examined the relationship between domains of women’s empowerment and fertility preferences in Uganda, and the changes in the relationship over time. The study used secondary data from the Uganda Demographic and Health Surveys for 2006, 2011, 2016, and 2022. Fertility preferences were operationalized as women’s ideal number of children, and women’s empowerment was operationalized as a multidimensional construct comprising agency, economic resources, access to information, gender norms, and sexual autonomy. Bivariate and ordered logistic regression analyses were conducted to assess the relationship between empowerment domains and fertility preferences. A substantial proportion of women consistently preferred large and moderate family sizes, reflecting a slow fertility transition. Between 48.0% and 52.6% of women preferred four to five children in the survey period. Women who preferred six or more children decreased from 34.8% in 2006 to 30.3% in 2016 and increased to 35.8% in 2022. Women’s empowerment improved over the period; however, improvements were uneven across domains. The influence of women’s empowerment on fertility preferences varied across domains. Agency and gender norms showed a consistent relationship with the ideal number of children; women with high agency had 27%, 23%, 25%, and 26% lower odds of preferring more children in 2006, 2011, 2016, and 2022, respectively. The economic resources domain was not significant in any of the survey years. Women’s empowerment plays a significant role in shaping fertility preferences in Uganda. The multidimensional approach adopted in this study provides deeper insight into the pathways through which empowerment influences reproductive decision-making and highlights the importance of adopting holistic approaches that strengthen women’s empowerment across multiple domains. The findings further underscore the need for integrated interventions that strengthen women’s autonomy and promote informed reproductive choices. Such an approach to policy formulation and implementation would facilitate Uganda’s advancement on Sustainable Development Goals 3 and 5.
  • Item
    Multi-dimensional poverty measurement and its determinants in peri-urban areas of Uganda.
    (Makerere University, 2026) Miiro, Samuel
    This study estimated the incidence and intensity of multidimensional poverty among peri-urban households in Uganda, examined its associated factors, and compared multidimensional and monetary poverty estimates. Secondary data from the 2016/17 Uganda National Household Survey were analysed. Multidimensional poverty was constructed using the Alkire–Foster counting approach and a predefined weighted deprivation threshold. Given by household i, the weighted deprivation score was cᵢ = Σⱼwⱼgᵢⱼ, and the binary outcome was ρₖ(cᵢ)=1(cᵢ≥0.40). The aggregate adjusted headcount was M₀ = H×A, where H is the weighted proportion classified as multidimensionally poor and A is their weighted mean deprivation intensity The analysis applied descriptive statistics, pairwise correlations, bivariate tests and Determinants were estimated using log[pᵢ/(1−pᵢ)] = β₀ + Xᵢβ. A survey-weighted linear probability model and design-adjusted Wald (F)-test were additionally used to compare multidimensional-poverty headcounts between peri-urban and gazetted urban households. The survey design incorporated sampling weights, 15 strata and 279 primary sampling units. Of the 1,945 peri-urban households in the analytic sample, 486 (25.0%) were multidimensionally deprived and 1,459 (75.0%) were not deprived. At the bivariate level, deprivation was significantly associated with the household head’s age, marital status, educational attainment and industry of employment, as well as distance to the nearest telephone service. Sex of the household head, geographical region and borrowing goods on credit were not significantly associated with deprivation. Correlations among deprivation indicators were generally weak to moderate, with the strongest association observed between housing and toilet deprivation ((r=0.578)), indicating no evidence of excessive inter-indicator correlation. In the adjusted logistic model, age, sex, region and borrowing status were not statistically significant. Households located 5–<8 kilometres from the nearest telephone service had 4.80 times the odds of multidimensional poverty compared with households located less than 3 kilometres away (OR=4.80; 95% CI: 2.47–9.31; (p<0.001)). Each additional income source increased the odds of deprivation by approximately 18% (OR=1.179; 95% CI: 1.019–1.363; (p=0.026)), suggesting that multiple income sources may reflect fragmented, seasonal or poorly remunerated livelihood activities rather than effective diversification. Relative to agriculture, forestry and fishing, employment in construction (OR=0.044; (p<0.001)), professional, scientific and technical activities (OR=0.163; (p=0.006)), administrative and support activities (OR=0.173; (p=0.040)), and education (OR=0.059; (p=0.002)) was associated with significantly lower odds of deprivation. The national survey-weighted comparison showed that the multidimensional-poverty headcount was 23.7% among peri-urban households (95% CI: 16.8%–32.4%) and 14.1% among gazetted urban households (95% CI: 11.9%–16.7%). The 9.7-percentage-point difference was statistically significant (95% CI: 1.4–12.1%; (F(1,264)=5.34), (p=0.022)); therefore, the null hypothesis of equal headcounts was rejected. The MPI also produced higher poverty estimates than the national monetary poverty-line approach, although MPI intensity and the monetary poverty gap measure conceptually different quantities and are not directly comparable. The study concludes that peri-urban poverty is strongly associated with infrastructure access and livelihood quality and recommends integrating multidimensional measures into poverty monitoring while prioritizing essential services and stable, adequately remunerated employment.
  • Item
    Comparing the frequentist and Bayesian estimation approaches for cox regression in modelling cervical cancer patient outcomes in Uganda.
    (Makerere University, 2026) Anena, Daisy. Barbara
    The objective of this paper was to compare the Frequentist and Bayesian estimation approaches for Cox regression in modelling cervical cancer patient outcomes in Uganda by evaluating the predictive performance of the two approaches, and assessing the association of clinical and sociodemographic factors with patient survival. The Cox PH, Exponential PH, and Weibull PH models were run under both frameworks, and their model performance and predictive accuracy was compared to determine which framework best fits survival analysis for cervical cancer patients in Uganda. The models were applied on dataset containing 454 women diagnosed with Cervical Cancer which was primarily collected by the Uganda Cancer Institute (UCI) between June 2019 to February 2020. The study also observed how different covariates such as age at diagnosis, morphology, initial disease stage, region of treatment, and other clinical and sociodemographic factors influence the time until death for the cervical cancer patients involved in the study. The non-parametric Cox PH model, and the Exponential PH model demonstrated the best overall performance based on model fit statistics under the Frequentist and Bayesian approaches respectively. Both models yielded nearly identical predictive outcomes when evaluated against each other with C-index ≈ 0.8 and AUC ≈ 0.79. Further findings based on the Bayesian Exponential model revealed that, diagnosis at stage 3 (HR=5.259; CrI: 1.446, 25.798) and stage 4 (HR=6.678; CrI: 1.63, 35.196), treatment by palliation (HR=4.940; CrI: 1.473, 20.503), and unspecified morphology (HR=2.174; CrI: 1.231, 3.785), were significantly associated with lower chances of survival. While women with negative metastases (HR=0.505; CrI: 0.284, 0.944), and those who were treated from the Eastern (HR=0.115; CrI: 0.029, 0.36) and Central (HR=0.498; CrI: 0.249, 0.985) regions, exhibited reduced hazard of death. Nevertheless, age at diagnosis, occupation and marital status had no significant effect on the survival outcomes of the women with involved in this study. In conclusion, the study underscores the critical role of stage at diagnosis in determining cervical cancer survival, it therefore calls upon the Ministry of Health to strengthen routine screening initiatives to promote early detection. The findings of the study also showed that both the Frequentist and Bayesian approaches perform effectively and can be applied in survival analysis. The study therefore recommends that future research should evaluate Frequentist and Bayesian frailty survival models to account for unobserved heterogeneity among cervical cancer patients.