Keep Conflicting Evidence Visible

Today's Goals

1. Identify date and definition differences behind conflicting numbers.
2. Present incompatible sources side by side instead of averaging them.
3. Write open questions that guide the next research step.

D2 taught us to trace a sentence back to its source. That still leaves a hard case: two credible passages can disagree. The wrong response is to ask the assistant to make them sound consistent at any cost. Today the fictional city shade study gains CITY-05, a later planning slide with a different percentage from CITY-01. The discrepancy is intentional. Neither document is a real city record, and no product run is being reported here. Your task is to preserve the conflict until you know whether dates, definitions, samples, or methods explain it.

Plain-Language Walkthrough

Treat disagreement as a research finding

Suppose CITY-01 says tree cover around stop A is 18 percent in the measured parcel. CITY-05 says 31 percent in the surrounding service area. A shallow summary says “the coverage is about 25 percent.” That invented middle number answers neither source's question. Another shallow summary picks CITY-05 because it is newer. But if the newer slide measures a larger area, its date does not make it the correct replacement for the parcel result. The first job is to copy both claims accurately, with exact units and qualifiers, before trying to reconcile them.

A conflict is a relationship between claims, not a property of one file. Record claim A, claim B, their precise locations, and the dimension on which they differ. They might be contradictory under the same definition, or merely incomparable because they use different definitions. They might describe change over time. They might also reflect an error. Do not choose among these explanations without evidence. A good assistant can help surface possibilities, but possibility is not resolution. Write “unresolved” when the documents do not supply enough detail.

This is why D2's claim ledger matters. Its row for CITY-01 remains valid for the measured parcel and date. CITY-05 creates a new row; it does not silently overwrite the earlier one. A reviewer can ask whether the pilot recommendation changes under either interpretation. Sometimes it does not: both numbers might still point to the need for shade. Sometimes it does: a threshold for grant eligibility may hinge on the exact area. The decision consequence determines how urgently you need a new measurement or clarification.

Sign in and link WeChat to read

Sign in, then link your WeChat mini program account on the account page. Linked accounts share one free article per day. Unlock additional articles with rewarded videos in the mini program.

Scan in WeChat to open Dayward AI. Open the same course and chapter, choose the rewarded video, then refresh this page after completion.

Unlocked — refresh to readSign in / scan with WeChat →