RCT are widely recognised as the golden standard in causal inference. But due to several limitation in social science research, say you hypothesise that education has an effect on usage of mobile money services, it might not always be possible to conduct RCTs. My questions are:
1. Apart from RCT, what design(s) can you use?
2. What would be the best statistical analysis to achieve causal inference?
You can have a look at the book ´´Causal Inference in Statistics: A Primer´´, there you find description of methods for doing statistical inference using observational data only.
One of my favorite papers in this area is attached. Good luck. David Booth
A book published by World Bank is attached - Impact Evaluation in practice.
Which method to select? How? Pros and cons are answered.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis