Fieldwork data acceptance workshop / 6 MIN READ
Block impossible entries without blocking credible unusual answers
A strict form can create inaccurate data by refusing a truthful response.
AI-assisted research and analysis for the SNS Creative Fieldbook
Sources checked 2026-11-07
Collection validation should distinguish an impossible representation from an unusual observation. Use hard stops only where the research definition justifies them, and provide a review route for credible exceptions. A plausible-looking dataset is not trustworthy if collectors had to change answers to make the form advance.
What you will leave with: A validation decision register with justified hard stops, reviewable warnings and boundary-case acceptance evidence.
Use this when
- A digital field instrument includes numeric or cross-field checks.
- Collectors report entering substitutes because a truthful answer is rejected.
This is instrument validation design, not statistical outlier deletion, eligibility screening or generic error-message writing. ODK Form Logic documentation checked on 7 November 2026. Research owners approve rules; implementation changes and participant contact require their own authorization.
Explain why a value is invalid before blocking it
Write the meaning, unit and reference period of each checked field. A negative count of completed tasks may be impossible under one definition, while a negative net change is meaningful under another. A value that differs sharply from the team’s expectations is not automatically invalid. Ask the research owner to distinguish definition violations from observations that merely deserve confirmation.
For each proposed rule, identify its consequence: prevent progression, show a warning or permit entry without interruption. Record the evidence behind the choice. Avoid converting an analyst’s convenient range into a collection restriction without justification. If a cross-field total must reconcile, verify that the categories really cover the same scope before forcing them to add up. Otherwise the rule can manufacture agreement between quantities that should differ.
Design the exception route before testing typical values
ODK documents constraints that block progression when false and conditional notes that can act as soft warnings. It also treats blank-response requirements separately. This supports distinguishing blocking, warning and missing-entry rules rather than treating them as one validation switch.
For a credible unusual answer, specify what the collector should confirm and what can be retained after confirmation. Use neutral wording that checks units or interpretation without coaching the respondent toward a preferred range. Keep an authorized route for unresolved cases. Do not instruct collectors to enter a nearby acceptable value, use zero as a placeholder or hide the original report in an unrelated comment field merely to complete the interview.
Exercise boundaries with realistic synthetic cases
Build fixtures for an ordinary valid answer, a genuine definition violation, a plausible unusual value and a permitted missing response. Test values at the boundary as well as just beyond it. Where rules involve several fields, include a case in which each field is individually valid but the combination needs clarification. Record the expected message, whether progression is allowed and what should appear in the resulting data.
Have someone unfamiliar with the implementation follow the collector instruction. Watch for a warning that looks like a prohibition, a hard stop that offers no recovery or a message that asks for information outside the study’s approved scope. Inspect the saved output after the test. Passing the visible form is insufficient if the exported answer has been silently transformed or the explanation needed for an approved exception is missing.
Review rule failures without rewarding tidy data
When fieldwork reveals an unexpected case, preserve the reported value and the rule version through the approved process. Ask the research owner whether the issue is a definition problem, implementation error or legitimate rare observation. Do not weaken every rule because one case was difficult, but do not defend an invalid rule merely because it was approved before launch.
If the instrument changes, document which records encountered the earlier restriction and what can actually be recovered. Earlier answers may have been distorted in ways that cannot be reconstructed. Tell the analyst about that limitation rather than presenting the revised form as a retroactive repair. Acceptance means the rule behaves as intended and respects the measurement definition, not that the resulting data contains no surprises.
Make the reasoning concrete.
A fictional study asks adult service coordinators how many handoffs occurred during a busy shift. A form blocks any answer above the team’s usual operating range, and a coordinator reports a credible count beyond it.
Require an in-range answer
Opening: Ask the collector to reduce the value so the form can advance.
Hypothesis: The dataset stays within the expected range.
Risk: The instrument replaces the reported experience with an invented value.
Review the rule and preserve the report
Opening: Use the approved clarification route to check the unit and retain a credible exception.
Hypothesis: The unusual observation remains available for proper review.
Risk: The team must assess whether earlier respondents were constrained similarly.
The researcher chooses a reviewable warning for this expectation-based boundary while retaining justified checks for genuinely impossible representations.
Take the questions into your next brief.
Use these prompts in your own notes. Nothing is submitted from this page.
Definition
What unit, period and meaning justify the rule?
Consequence
Should failure block, warn or allow entry?
Exception
How can a credible unusual report be preserved?
Fixtures
Which ordinary, boundary, impossible and missing cases were tested?
History
Which records encountered a rule later found unsuitable?
Watch for the wrong conclusion.
- Using a typical range as a universal hard limit.
- Entering zero or a nearby value to escape a constraint.
- Assuming a later fix repairs earlier distorted responses.
Your next action
Review one hard-stop rule and write a credible unusual fixture that tests whether the rule protects or changes the intended measurement.
Sources & scope
Workflows, worksheets and fictional examples are original editorial analysis. Sources support only specifically attributed principles. Platform guidance can change; legal requirements depend on jurisdiction and context. No measured results or named human reviewer are claimed.
- ODK Docs: Form Logic →Primary form-authoring documentation · Supports: Supports blocking constraints, conditional warnings and separate blank-response handling. The validation decision method is original analysis.