ODK Submission Review Checklist

Review ODK submissions in five passes: confirm that records arrived, inspect the submission table and details, assign a review state when your team has made a decision, export the data with repeats and media preserved, and connect OData when the analysis needs to refresh. Keep the original export before cleaning or recoding the dataset.

This checklist covers the Central-to-analysis handoff after field devices have synchronized. For the device workflow, use the ODK Collect offline checklist.

If submissions create or update shared records, review the Entity data alongside the submissions with the ODK Entities and longitudinal workflow.

1. Confirm that the expected records arrived

Start with receipt, not interpretation:

• Open the expected project and form in ODK Central.
• Compare the number of submissions received with the field plan, sample frame, or enumerator log.
• Filter by submitter and date to find gaps or sudden changes in collection activity.
• Check whether expected media attachments are present.
• Confirm that the form version and submission dates match the fieldwork period.

If records are missing, check the device and server workflow before declaring them lost. Finalized forms may still be waiting on a device, and a form version change can affect what appears in an export. Start with ODK Collect troubleshooting when a device has a sending backlog.

For the checks that should happen before a version reaches the field, use the ODK Central form release checklist.

2. Inspect the Central submission table

Open the form's Submissions tab and review a useful sample before downloading everything. Central lets you choose visible columns, search the table, filter by submitter, date, and review status, and open individual submission details.

For each review sample, check:

1. The instance name or instance ID identifies the interview clearly.
2. Required answers and important calculated fields are populated.
3. The form version is the expected version for that field period.
4. Attachments such as images or audio are present when the form requires them.
5. The timing, location, and submitter information fit the study protocol.
6. Any edits or comments have an explanation in the review record.

These checks are prompts for human review. A short duration, unusual location, missing attachment, or edit is not by itself proof that an interview is invalid. Compare the record with the form, assignment, and field context.

3. Use review states consistently

ODK Central includes review states that help a team record what happened to a submission:

State	Use it for
Received	The default state for a newly uploaded submission
Edited	A state Central applies after a submission is edited
Has Issues	A reviewer found a problem that needs follow-up
Approved	The team accepted the record for the next workflow step
Rejected	The team decided the record should not be used under its protocol

Write a short reason when your team changes a state. A review state is a workflow label, not an automatic cleaning operation. Rejected submissions remain in exports unless you filter them out, so document the filter or exclusion rule used by the analysis team.

4. Choose the right data access path

Use the analysis path that matches the job:

Need	Best starting point	Why
Archive or share a fixed dataset	CSV download	Creates a local export that can be preserved with the study files
Keep repeats and media together	CSV or ZIP download	Central includes the root table, repeat tables, and a files folder when those data exist
Refresh a dashboard as submissions arrive	OData	Power BI, Excel, Tableau, and other clients can read a live feed
Review audit evidence and fieldwork signals	DataSnap or audit export	Keeps timing, GPS, identity, and answer-change review connected to the source evidence

For a one-time file, follow How to export ODK Central data. For a refreshable dashboard, use ODK Central with Power BI or Excel.

5. Preserve repeats, media, and audit evidence

Do not assume the main CSV contains every table or file:

• The root CSV contains the main submission table.
• Each repeat group can appear as its own related CSV table.
• A files folder contains submitted media when the form collected attachments.
• An audit-enabled form can include an audit CSV with event evidence from the submissions.
• Keep the instance ID or join columns intact when moving data into an analysis tool.

Open the export and compare its columns, repeat tables, media, and audit files with the form and analysis plan. If a field appears to be missing, check whether the form version or export options explain the difference before changing the source data.

6. Move from review to quality analysis

After the receipt and export checks, review the evidence that supports the study's quality decisions. The ODK Field Data Quality Toolkit covers pre-fieldwork checks, same-day monitoring, audit evidence, and documentation.

For audit data, use How to Analyze Audit Log Data or create a DataSnap report. DataSnap shows available activity, time-use, question-friction, flagged-interview, interview-detail, and GPS views from the source. It helps the team choose records for review; it does not prove fraud or automatically rank enumerators.

SurveyLoopr can host the ODK Central server and provide the DataSnap workflow, but your team still owns the study's review protocol, access decisions, export retention, and final analysis rules. If you need the server layer, see managed ODK hosting.

For repeatable extracts or controlled resource operations, see the ODK Central API and automation guide.

Review ODK data quality with DataSnap

Export ODK submissions correctly