Andrew discusses the challenges of measuring the effectiveness of early childhood intervention programs due to small sample sizes and changing conditions. He emphasizes the importance of theory and observational data in making decisions and suggests that research should be more real-world focused.