Awande07-tk/sa-child-labour-sampling-analysis

Statistical comparison of four probability sampling designs (SRS, SRS with replacement, stratified, and cluster sampling) using South Africa's 2019 Survey of Activities of Young People, evaluating which design yields the most precise national estimates of school attendance, child labour, and hours worked, implemented in SAS. # Comparing Probability Sampling Designs: Evidence from South Africa's Survey of Activities of Young People (2019) A statistical sampling study comparing four probability sampling designs — **Simple Random Sampling (SRS)**, **SRS with Replacement**, **Stratified Sampling**, and **Cluster Sampling** — on a national dataset of South African children, to determine which design yields the most precise and reliable estimates for key wellbeing and child labour indicators. **Tools:** SAS (PROC SURVEYSELECT, SURVEYMEANS, SURVEYFREQ) · Excel/xlsx data source · South African Department of Statistics microdata --- ## Table of Contents - Overview - Aim of the Study - Data - Methodology - Results - Visual Comparison - Discussion & Conclusion - Repository Structure - Reproducing the Analysis - Author --- ## Overview Governments and researchers routinely rely on sample surveys — rather than a full census — to estimate national indicators such as school attendance and child labour prevalence. **Which sampling design to use matters**: it directly affects how much sampling error the resulting estimates carry, and therefore how much confidence policymakers can place in them. This project uses microdata from Statistics South Africa's **Survey of Activities of Young People (SAYP) 2019** (13,336 children aged 5–17) to design and execute four probability sampling schemes in SAS, and compares the resulting point estimates and 95% confidence intervals for four indicators of child wellbeing. ## Aim of the Study The aim of this study is to estimate key parameters concerning the status of South African children aged 5–17 years — specifically the **mean age**, the **proportion currently attending school**, the **proportion involved in child labour**, and the **total hours worked**. A critical secondary aim is to evaluate the performance of four different sampling methods to determine which provides the most reliable estimates for these parameters, as indicated by the narrowest confid …