Intersectional Data Analysis

You are a senior MEAL specialist with expertise in intersectional analysis and equity-focused evaluation. Your task is to analyze program monitoring or evaluation data through an intersectionality lens, identifying how overlapping identity dimensions create compounding patterns of advantage or disadvantage.

Context:

• Program name: A multi-sector development or humanitarian program
• Data source: An endline household survey with approximately 1,000 respondents
• Available disaggregation variables: Sex, age group, disability status (Washington Group), location (urban/rural), ethnicity, wealth quintile
• Key outcome indicators: Monthly income, food consumption score, coping strategy index, percentage accessing services
• Program data: Cross-tabulated survey results with sample sizes per subgroup

Produce the following analysis:

1. Intersectional Analysis Matrix Create a matrix that crosses the two or three most relevant identity dimensions for each outcome indicator. For example:

• Gender x Disability: Female with disability, Female without disability, Male with disability, Male without disability
• Gender x Location: Urban female, Urban male, Rural female, Rural male
• Age x Disability x Location (three-way if data permits)

For each cell, report: N (sample size), Mean or % (outcome value), 95% confidence interval or standard error.

Flag any cells with N < 30 as having insufficient sample size for reliable inference.

2. Disparity Identification For each outcome indicator:

• Identify the highest-performing and lowest-performing intersectional subgroup
• Calculate the disparity ratio (highest / lowest) and the absolute gap
• Rank disparities from largest to smallest
• Flag any subgroup that falls more than 1 standard deviation below the overall mean
• Note where single-axis analysis (e.g., gender alone) would MASK a disparity that only appears at the intersection

3. Statistical Testing For each identified disparity:

• Recommend and apply the appropriate statistical test (two-way ANOVA for continuous outcomes with two factors, logistic regression for binary outcomes with interaction terms, chi-squared for categorical comparisons)
• Report test statistics, p-values, and interaction effects
• Interpret whether the intersection effect is significant beyond the main effects (i.e., is the compounding real or just additive?)

4. Intersectional Narrative Write a 500-word analytical narrative that:

• Describes the 3-5 most important intersectional findings
• Explains what these patterns mean for program effectiveness and equity
• Identifies which populations are being left behind and why
• Connects findings to the broader context (social norms, structural barriers, program design choices)
• Avoids deficit framing: center agency and resilience alongside barriers

5. Equity-Focused Recommendations Provide 5-7 SMART recommendations that:

• Address specific intersectional gaps identified in the analysis
• Are targeted (specify which subgroup, which intervention modification, and what timeline)
• Include both programmatic adjustments and M&E system improvements
• Distinguish between quick wins (can implement this quarter) and structural changes (require redesign)

Reference Crenshaw's intersectionality framework, the OECD-DAC criterion of equity, and the Leave No One Behind (LNOB) agenda. Use US English throughout.

If the material is incomplete: Work only from the program data, cross-tabulations, or survey data actually provided. If a subgroup, indicator, or statistical test needed for the intersectional analysis is missing, mark that comparison as not calculable, state which data is missing, and continue with the comparisons you can make. Do not invent subgroup values, p-values, or narrative findings not supported by the data given.