Due to a short timeframe and narrow scope for this research, we adopted a globally collaborative approach to data collection:
- Interviews with Open Contracting Partnership country managers. Firstly, we asked managers what they knew about commitment implementation and identified partners in each country to help with data collection. To keep the scope narrow, we sought to reach out to one governmental team and one CSO in each country. 
- Questionnaires completed by country partners. We then designed and shared a questionnaire with country partners, clearly laying out relevant IMF commitments, and asking what progress had been made against each. We also asked for any documentary evidence to support each claim.
- Interviews with country partners. Finally, we conducted interviews with country partners to fill out any missing pieces of information and to add narrative depth to our analysis. We also sought to test our emerging recommendations with interviewees to determine if they were suitable within the country’s political environment.
Limitations, bias and incompleteness risks, and mitigation
The most significant limitation we faced concerned our small sample size. For the most part, we partnered with just one government team and one CSO team in each country. This meant that we could not reasonably expect to reach a saturation point in our data collection, wherein additional interviews offer no new information.
To an extent we were able to mitigate against the limitations of a small sample by focussing on verifiable facts and events, such as the publication of data or new legislation. However, we also relied on narrative judgements to help identify obstacles and design our recommendations, meaning that we have had to take care to contextualise interviewees’ statements and consider alternative interpretations of the fact pattern. Where possible, we have worked to verify our findings through desk research and follow-up with research partners.
 Despite outreach, we were unable to carry out formal interviews with government officials in Honduras, Bolivia and Malawi. To mitigate the impact this had on our findings, we carried out extra desk research on these countries, focussing on verifiable events.