Most fieldwork errors are older than fieldwork. They’re written into the questionnaire on Day 1, then carried through scripting, translation and testing until real respondents find them. By Day 3, fixing one costs data, budget and a client conversation. On Day 1, it cost a line in a Word doc.
Everyone who has run fieldwork knows the Day 3 email. It usually starts with “Quick question about Q17.”
Why Fieldwork Errors Start in the Questionnaire
Survey methodologists have a name for this. Paul Biemer’s review of total survey error in Public Opinion Quarterly lists “ambiguous questions, confusing instructions, and easily misunderstood terms” as questionnaire design problems that lead to measurement error. He also describes specification error: the question measures something slightly different from what the study needed.
Both start in the document, long before anyone writes code.
We work from one simple rule. Every step after the questionnaire copies it faithfully, mistakes included. The programmer scripts what the document says. Translators translate it. The QA tester checks the script against it, and passes it, because the script matches. A broken route that matches the questionnaire looks exactly like a working one.
So a clean QA round tells you the survey does what the document says. It can’t tell you the document was right. We covered that gap from the programming side in The Script Mistakes That Slip Past a Rushed QA Round.
What Has Already Happened by Day 3
A soft launch helps. Most teams release the survey to a small slice of sample first. aytm suggests 5 to 10% of the target, and recommends checking skip logic, completion times and response recording before going wide.
But a soft launch is still fieldwork. By the time it shows a problem, a lot is already in motion.
| By Day 3 | What It Means |
|---|---|
| Completes are in | Some respondents answered on the wrong route, and you can’t ask them again |
| Quotas are counting | Cells may be filling with people who shouldn’t be in them |
| Sample is spent | Every complete on the wrong route was paid for |
| Translations are live | One fix becomes one fix per market |
| The client has a date | Often a first topline, sometimes a steering meeting |
The team did its job. The error just had three days to collect company.
Your Options on Day 3, and What Each One Costs
Once you confirm a fieldwork error, you have four ways forward. Each one costs something different.
| Option | What It Costs |
|---|---|
| Pause, fix and relaunch | Field days, a revised timeline, and a client update |
| Fix it live and keep going | Two versions of the same survey in one dataset |
| Re-field the affected base | More sample, more time, and a question about the first wave of completes |
| Keep going and footnote it | A caveat in the debrief that the client will remember |
But fixing a live survey has its own risks. A University of Oklahoma guide to editing active Qualtrics surveys warns that “deleting a question or answer also deletes all response data associated with that question or answer.” It recommends hiding items with display logic instead.
Other platforms also have the same problem in different words. SurveyCTO’s guidance on mid-survey updates explains that some structural changes split the data across form versions. Someone then has to merge them by hand during cleaning.
So even the “quick fix” lands on the data processing team two weeks later.
Why Fieldwork Errors Hurt More in Q4
Q4 makes every option on that list more expensive. In November, the programmer who built the survey is already on the next launch. The PM is running three studies at once. So a Day 3 fix doesn’t get a clear afternoon. It gets squeezed between other deadlines, and that’s how one fieldwork error turns into two.
Field windows shrink too, because the December slowdown is coming. A three-day pause in October is an inconvenience. In late November, it can push the final data past the client’s planning meeting.
That’s why we put the effort on Day 1 when the calendar is tight. Fewer fieldwork errors in Q4 means your team spends its capacity on new launches instead of relaunches.
Why Cleaning Can’t Fix Fieldwork Errors
Cleaning is good at removing bad answers. Speeders, straight-liners and duplicates come out. What cleaning can’t do is go back and ask the right question.
If a route sent managers past a question meant for them, those answers don’t exist. If a question measured the wrong thing, every response is a clean, consistent answer to the wrong question. As Kantar’s Steve Wigmore puts it, a design mistake degrades “the accuracy of your data and the subsequent extrapolations” you can make from it.
That’s the cost buyers feel most. Your team can fix the fieldwork error itself in an afternoon. The doubt it leaves in the data lasts until the next wave.
Five Questions to Ask Before You Sign Off on Day 1
Questionnaire sign-off is usually a read for content. These five questions turn it into a read for fieldwork errors:
- Who sees each question? Write the base for every question, in plain words. If you can’t, the route isn’t clear yet.
- Would anyone miss this question? If nobody can say what decision it feeds, cut it before it costs a minute of LOI.
- Where does each respondent end? Trace every termination and quota against the screener, including the rare paths.
- Has someone who didn’t write it read it? The author reads what they meant. A fresh reader reads what’s there.
- What will each market see? Check the master version is final before translation starts, so one fix stays one fix.
Each answer takes minutes on Day 1. Each wrong answer takes days on Day 3.
What a Day 1 Check Can’t Catch
Still, some problems only appear with real respondents. Incidence can come in lower than planned. People can read a clear question in a way nobody expected. That’s what the soft launch is for, and it still earns its place.
A Day 1 check changes what the soft launch finds. With the questionnaire already reviewed, the soft launch tests real behavior. It stops being the first time anyone walks the routes.
Where to Go Next
If your Q4 questionnaires arrive late and messy, start with Your Questionnaire Looked Fine in Word. Then November Happened. It covers the six Word habits behind most fieldwork errors.
If you’d like a structured Day 1 read, ResearchReady reviews the questionnaire in about 15 minutes, before scripting starts. It helps your team catch design flaws early, protect data quality and keep field days for fieldwork. Your researchers still make every call.
The questionnaire doesn’t change between Day 1 and Day 3. Only the price of fixing it does.
