Difference between revisions of "Structuring a Survey Pilot"
|Line 34:||Line 34:|
== Follow-up Surveys ==
== Follow-up Surveys ==
It is a good idea to '''pilot''' even in case of '''follow-up''' surveys
It is a good idea to '''pilot''' even in case of '''follow-up''' surveys are conducted after a '''baseline''' (round) . If the [[Impact Evaluation Team|research team]] has reduced the sample size from previous rounds, they can use '''pre-loaded''' data of respondents who were not part of the actual baseline survey, but for whom data is available because of the pilot rounds. However, if there is no pre-loaded data available, the '''field team''' can conduct '''pre-interviews''' before the actual pilot to collect data for '''pre-loading'''. While this can be challenging in practice, it is useful for improving the quality of the actual '''follow-up''' survey.
== Related Pages ==
== Related Pages ==
Revision as of 17:31, 16 April 2020
The process of structuring a survey pilot involves agreeing upon the logistics (or practical aspects) of a survey pilot, including duration, approvals, review sessions, and training. In order to conduct a successful pilot, the impact evaluation team (or research team) must discuss each of these aspects in detail to ensure that the survey pilot is conducted smoothly. A good pilot helps to improve the quality of the data collection process and the survey protocols.
- Survey pilot involves carrying out interviews and tests on different components of a survey, including content and protocols.
- The research team should discuss the structure with all participants of a pilot, and resolve any issues that might come up.
- While finalizing the structure, keep the timeline of the survey pilot in mind.
- A pilot should be conducted before enumerator training to avoid enumerator effects, which arise when different enumerators ask the same question differently to different respondents.
Duration of the Pilot
It is important for the members of research team to agree on the expected time the entire process will take from start to finish. They should also keep extra buffer time in case some issues come up during the survey pilot. The average duration of the pilot will depend on:
- Extent of changes. If a lot of changes are made in the instrument design, translations or protocols, the pilot will take longer to complete. Ideally, the pilot process ends when everyone in the research team agrees that there are no more changes to be made.
- Printing facilities. In case of pen-and-paper pilots, availability of printing facilities in the area where the pilot is being conducted will also affect the duration.
- Pilot location. The location of the area where the pilot is being conducted will also affect duration. For example, if the location of the pilot area is far from home base (the location of the research team), the duration of the pilot will increase. In such cases, it is better for the field teams to stay in the pilot area, and conduct the pilot on consecutive days to complete the piloting on time.
- Budget. The organizing funding the study has to approve the survey budget. This includes approving the budget for the pilot, including expenses for transport, equipment, communication, and other operational costs.
- Institutional review boards (IRBs). In some cases, the research team will need to obtain an IRB approval even for the pilot stage. It is useful to consult the principal investigators (PIs), since they will know when this might need to be done for a pilot. Consider a case where the research team wants to publish certain aspects of the study at the piloting stage itself. Since the research team is handling personally identifiable information (PII), they will need an IRB approval for the survey pilot as well.
- Government agencies. While an approval from government agencies is not always necessary for a pilot, a letter of support (LOS) from the relevant ministry or implementing agency can help conduct the pilot smoothly.
It is important to hold review sessions regularly during the survey pilot. These sessions allow the research team to discuss feedback from the early stages of a pilot, and use it to improve components like instrument design, translations, and survey protocols. Keep in mind the following guidelines for the review sessions:
- Schedule. Schedule the review sessions frequently. Ideally, the research team should conduct a session with field teams at the end of each day to resolve any issues they might face. Do not schedule these sessions very late in the day. The sessions will be more useful if the field teams are not tired and hungry. If feasible, give a gap of at least one day between two rounds of piloting to allow everyone to recover.
- Listen. Listen to what everyone has to say. Every opinion or feedback can help improve the overall quality of the data collection process. This is particularly important when the field teams vary in terms of age, gender, ethnicity and so on.
- Revise. Ideally, the research team should make the revisions before the next round of piloting, so that the changes can be tested in the next round. This helps get an idea of whether the changes were useful or not.
- Logistics. In the case of pen-and-paper pilots, the research team should have access to printers to share the revisions immediately. It might be a good idea to purchasing a portable printer so that the field coordinators (FCs) can carry it to the area of the pilot.
- Version control. It allows the research team to keep track of all changes that are made throughout the review process. For example, if the survey is in French, while the field coordinator (FC) is only fluent in English, keeping track of the revisions can be tricky. In this case, the research team can hire an assistant who speaks French. Then the hired assistant can edit the French version, while the field coordinator (FC) can edit the English version. Ideally this assistant should be someone other than an enumerator to allow the enumerator enough time to rest and prepare for the next round of piloting.
- Length of training. This depends entirely on the complexity of the survey instrument and the related protocols. It should at least take one full day.
- Feedback. Interviewers can provide useful feedback on the survey instrument and the protocols. The research team should take this feedback into account to improve the training outcomes.
- Gap. There should be a gap of at least 1 day between the end of training and the start of piloting.
- Mock interviews. In case of data-focused pilots, the research team should ensure that the interviewers conduct mock interviews. This will allow them to become familiar with the process and also help to catch any bugs (errors) in programmed instrument.
It is a good idea to pilot even in case of follow-up surveys which are conducted after a baseline (or first round) survey. If the research team has reduced the sample size from previous rounds, they can use pre-loaded data of respondents who were not part of the actual baseline survey, but for whom data is available because of the pilot rounds. However, if there is no pre-loaded data available, the field team can conduct pre-interviews before the actual pilot to collect data for pre-loading. While this can be challenging in practice, it is useful for improving the quality of the actual follow-up survey.