Colquhoun.AI
← All guides

Research Guides

Verify the Scope of a Bulk Research Record Removal

· Colquhoun.shop editorial

A research workspace may offer “Delete selected” and “Delete all matching.” The two can look almost identical on the first page of a filtered list, yet affect very different sets of records. Test the scope shown at confirmation and what can actually be recovered afterward.

This is a worksheet for a fictional research tool, not a description of Colquhoun.shop features. Use a disposable test workspace, not actual research records.

1. Write down what the action promises

W3C's guidance on Error Prevention for important submissions describes alternatives that let people reverse, check and correct, or review and confirm affected actions. For user-controlled data, it highlights avoiding serious loss, not adding a confirmation to every routine save.

Start by naming the action. “Remove from this comparison” may only detach an item; “Delete from workspace” may remove the underlying record. A useful review needs to distinguish them before deciding whether the interface communicates the consequence.

2. Use a fixture where the two scopes disagree

Imagine a fictional workspace with 30 source records. Twelve match a draft filter. Only five are visible on the first page, and the reviewer selects three. Record the three fixture identifiers, the 12 matching identifiers and one nonmatching record to serve as a control.

Run the selected-record operation on one disposable copy and the all-matching operation on another. For the first, the confirmation should reflect three selected records under the example contract. For the second, it should reflect all 12 matching records, including those off the first page. After each operation, compare the actual affected set with the confirmed set. A success notification alone does not establish which records were touched.

Test canceling before final confirmation in both journeys. Verify that it makes no change. If the filter or selected set changes between opening and confirming, test whether the implementation preserves the reviewed set or requires scope to be reviewed again. Pick the behavior for the product and check that the user is not surprised by a silent expansion.

3. Verify recovery as a separate action

Where the interface offers recovery, restore the test records and compare them with their pre-action state. Check the source text, citations and any documented associations. Restoring a title without its evidence link may not restore what the researcher needs.

Review the stated recovery boundary, including any time limit. Test only within a permitted fixture, and mark expiry untested if you did not exercise it. If the action is not reversible, give the review and confirmation step extra attention; do not describe a generic success message as a recovery mechanism.

4. Save a scope-and-recovery checklist

5. Close with observed behavior

A useful note reads: “The selected-record journey removed the three confirmed fixture records and left the other nine matching records intact. Restoration recovered their source links. We did not test expiration.” That conveys the actual evidence more clearly than “bulk deletion works.” It does not certify every data-management flow or establish WCAG conformance.

This guide describes the built-in Web Reader and Keyword Density behavior verified for this publication. Available data depends on the page response; these tools do not measure visits or search rankings.