Practical Exercises for South Asian Development Research
Workshop: Bivariate Analysis 101 Duration: 90-120 minutes Format: Hands-on exercises with real development data Materials needed: Calculator, graph paper (optional)
Learning Objectives
Apply bivariate analysis techniques to development research questions
Interpret correlation coefficients and regression results in context
Identify appropriate methods for different variable types
Practice communicating statistical findings to policy audiences
Context: You are analyzing factors affecting school enrollment in rural Bangladesh. You have collected data from 20 villages on distance to nearest primary school and enrollment rates.
Calculate standard deviations: sx = _____, sy = _____
Calculate correlation coefficient: r = _____
Your calculations:
Task 2: Interpret Results
1. What does the correlation coefficient tell us about the relationship?
2. Is this relationship strong, moderate, or weak?
3. What might explain this relationship?
4. What policy implications does this suggest?
Exercise 2: Simple Linear Regression - Healthcare Access
Context: A health NGO in rural India wants to understand how distance to health centers affects maternal healthcare utilization. They collected data from 15 districts.
Given Results:
Using the regression: Healthcare Utilization (%) = 94.2 - 8.5 × Distance (km)
R² = 0.68
Standard error of slope = 1.4
p-value for slope = 0.001
95% CI for slope: [-11.2, -5.8]
Task 1: Interpret Coefficients
1. What does the intercept (94.2) mean in this context?
2. What does the slope (-8.5) tell us?
3. How much of the variation in healthcare utilization is explained by distance?
Task 2: Make Predictions
Calculate predicted healthcare utilization for: a) A district where health center is 2 km away: _____ %
b) A district where health center is 5 km away: _____ %
c) A district where health center is 8 km away: _____ %
Task 3: Statistical Significance
1. Is the relationship statistically significant? How do you know?
2. What does the confidence interval tell us?
3. Would you recommend this model for policy decisions? Why or why not?
Exercise 3: Chi-Square Analysis - Gender and Financial Inclusion
Context: A microfinance organization in Pakistan wants to understand if there are gender differences in loan repayment patterns among their clients.
Observed Data:
Repaid on Time
Late/Defaulted
Total
Male Borrowers
320
80
400
Female Borrowers
540
60
600
Total
860
140
1000
Task 1: Calculate Percentages
Male borrowers on-time repayment rate: _____ %
Female borrowers on-time repayment rate: _____ %
Overall on-time repayment rate: _____ %
Task 2: Set Up Hypothesis Test
Null Hypothesis (H₀):Alternative Hypothesis (H₁):
Task 3: Chi-Square Test
Given: χ² = 15.84, df = 1, p-value = 0.0001
1. What is your conclusion about the relationship between gender and repayment?
2. What might explain this pattern?
3. How would you present this finding to the microfinance organization?
4. What additional data would strengthen your analysis?
Exercise 4: Comparing Groups - Urban vs Rural Income
Context: You are studying income differences between urban and rural households in Sri Lanka for a poverty reduction program.
Sample Data:
Urban households: n = 150, mean income = ₹45,000, standard deviation = ₹12,000
Rural households: n = 200, mean income = ₹32,000, standard deviation = ₹9,500
Task 1: Independent Samples T-test
Given test results: t = 10.85, df = 348, p-value < 0.001
1. What does this t-test tell us about urban vs rural incomes?
2. Calculate the mean difference and interpret its practical significance:
3. What factors might explain this urban-rural income gap?
Task 2: Effect Size
Calculate Cohen's d (effect size):
d = (Mean₁ - Mean₂) / Pooled Standard Deviation
Your calculation: d = _____
Interpretation: (small = 0.2, medium = 0.5, large = 0.8)
Task 3: Policy Implications
Based on your analysis, what would you recommend for poverty reduction efforts?
Workshop Reflection
Key Takeaways
1. Which analysis method was most appropriate for each research question?
2. How did the South Asian context influence your interpretation of results?
3. What challenges did you encounter in interpreting statistical significance vs. practical significance?
4. How would you explain these findings to a non-technical policy audience?
Sample bias: Ensure your sample represents the population of interest
Outlier influence: Check if extreme values are driving your results
Missing context: Statistical results need cultural and economic context
Next Steps
To deepen your bivariate analysis skills:
Practice with your own data: Apply these methods to a development question in your context
Learn software tools: Excel, R, or SPSS for larger datasets
Study visualization: Create clear charts and graphs for your findings
Prepare for multivariate: Consider how multiple factors interact
Bivariate Analysis Workshop | ImpactMojo Knowledge Series
Licensed under CC BY-NC-ND 4.0 | For educational use with attribution
Part of the OpenStacks initiative for development education
