Before Training
Before Training: Preparation Checklist
Training goes wrong in the week before it starts, not in the room. Settle these items first. This page is the checklist; for how to pretest the questionnaire itself, see questionnaire pretest vs pilot.
| Task | Why it matters |
|---|---|
| Final questionnaire with a version number | Trainees practice on the form they will use |
| Enumerator manual, one per trainee | It is the reference in the field when the trainer is not there |
| Devices charged, form loaded, a test submission sent | Finds sync and permission problems early |
| Consent script translated and read back by a second speaker | Catches meaning that was lost in translation |
| Room with space for pairs | Mock interviews need talking space |
| Referral list for distress or safeguarding concerns | Enumerators must know who to call before they meet anyone |
| Selection test and pass mark agreed | Selection after training needs a rule set in advance |
| More trainees invited than posts available | Allows selection on results, with reserves |
Ethics content belongs in every training, not only the consent segment. Pull the points from the evaluation ethics checklist and do no harm.
Training Days
Training-Day Checklist
Three days is a common minimum for a new questionnaire. A team using a form it already knows can shorten the plan; a form with complex sampling or sensitive topics may need a fourth day.
| Day | Cover | Exit check |
|---|---|---|
| Day 1 | Study purpose, ethics and consent, interviewer conduct, reading each question aloud, what each term means | Trainees can explain consent in their own words |
| Day 2 | Tablet use and skip logic, paired mock interviews, the role-plays below, household selection rules, common mistakes | Each trainee completes two full mock interviews |
| Day 3 | Field practice in a village outside the sample, debrief, written test, sign-off, team assignments | Pass mark reached, supervisors know who needs a check visit |
Three habits keep the days useful:
- Read every question aloud together on day 1. Where the room disagrees about meaning, fix the wording or the manual.
- Practice on the device, not on paper, if collection is digital. Skip logic that confuses trainees will confuse respondents. The wiring of the form is covered in skip logic in KoboToolbox forms.
- Keep a notes sheet of repeated errors. These become the topics for supervisors' first check visits.
Role-Play Script
Work in threes: enumerator, respondent and observer. Rotate after each scene. The respondent follows the scene note and does not make it easy. The observer ticks the last column only when the behavior was clearly seen.
| Scene | What happens | Observer looks for |
|---|---|---|
| 1. Opening | Enumerator greets, states name and study, asks for the right person and reads the consent text | Consent not skipped or shortened |
| 2. Sensitive question | Respondent hesitates over an income question. Enumerator repeats once, offers to skip, moves on | Neutral tone, no pressure |
| 3. Unclear answer | Respondent says "a few" when asked for a number. Enumerator probes without suggesting a number | Non-leading probe |
| 4. Interruption | A neighbor sits in. Enumerator asks politely for privacy or arranges another time | Privacy protected |
| 5. Closing | Enumerator thanks the respondent, gives contact details and checks no page was missed | Contact details given |
After each round, the observer gives one thing done well and one to change. The trainer collects the common slips and replays that scene with the whole group. In a pilot run, the same behaviors are what back-checks and spot-checks later look for; see J-PAL's guidance on data quality checks.
Sign-Off and Use
Sign-Off Before Fieldwork
Do not send everyone out because they attended. Sign off against criteria:
- Passed the written test at the agreed pass mark
- Completed two observed mock interviews with no consent errors
- Can use skip logic and correct a wrong entry on the device
- Knows the stop rule for distress and who to call
- Signed the confidentiality statement
Trainees who miss a criterion can take a second observed practice on day 3, or join as a reserve. Supervisors then give extra check visits in the first week to anyone borderline. Data quality checks during fieldwork build on this; see how to conduct a DQA and checking data quality.
Using the Checklist
The preparation list, the three-day plan, the role-play scenes with an observer column and the sign-off table above make up the full checklist. Adapt them for your own survey, then print the role-play table as a handout. More on questionnaire planning is in survey design.
Common Mistakes
A training that is only a read-through. People remember what they practiced. Mock interviews and role-play are the training; reading is the setup.
Training and selecting in one step. Select after you have seen results and observed practice, and tell trainees the rule at the start.
Skipping the field day. A practice interview outside the sample is where tablets, routes and respondents surprise you.
No plan for the first week. Trained is not the same as checked. Schedule supervisor observation and early back-checks before anyone leaves the room.