Why use a map? How to use the map 5 key questions for choosing a fit-for-purpose method 4 Seven routes to experimentation Experimental Methods Observational Methods Combining Methods Pros and cons of the 7 methods BOXES Box 3.1. Observational and experimental methods Box 3.2. Internal and External Validity Box 3.3. Experiments, quasi-experiments, and natural experiments Box 3.4. Designing for Generalisability Box 3.5. Sample size, power, and effect size SEVEN ROUTES TO EXPERIMENTATION: © OECD 2023 Box 3.6. Labelling the methods Box 4.1. A Randomised controlled case study: Behavioural prompts to increase early filing of tax returns: a population-level randomised controlled trial of 11.2 million taxpayers in Indonesia Box 4.2. An A/B Testing case study: Enhancing the design of a webpage to increase the number of consumer product safety reports in Canada Box 4.3. A diff-in-diff case study: Energy efficiency in Switzerland Box 4.4. A before-after case study: Digital health in the UK Box 4.5. A longitudinal case study: Physical and mental health in the US Box 4.6. A correlational case study: The role of political ideology in energy conservation in the US Box 4.7. A qualitative case study: Drought management in Australia Box 4.8. A mixed methods case study: Transparency on online platforms in four European countries