Check a before-and-after claim

Prompt that tests a before-and-after claim against season, chance and other explanations and gives a clear verdict. Paste it into an AI tool of your choice, which then checks claims.

Topic: Deciding & analysing
Type: Prompt
Level: Intermediate
Time: about 8 min

Results depend on the tool, the model and what you enter.

Result preview

Example: written by our editors, not real AI output.

Testable statement: orders per week rose from 410 to 470 (+15%) after the new form, four weeks before versus four weeks after. Season: open, peak season started in week 2 after Chance: open, weeks varied by ±40 before as well Promotion at the same time: supports another explanation Measurement: ruled out, counting unchanged Verdict: plausible, not supported Test: show the form to only half of the visits for two weeks Missing: figures for the same period last year

Fill in

What you type stays in this browser window. We store and send nothing.

ClaimWhat is being claimed?

DataPaste the before and after figures, with periods

ContextWhat else changed?

Your prompt

Check the following claim critically.
Claim: [Claim]
Available figures:
<data>
[Data]
</data>
What else happened in the period: [Context]
Approach: 1. Rephrase the claim as a testable statement with a metric, period and comparison. 2. Check other explanations: season and holidays, random fluctuation, simultaneous changes, changed measurement or tracking, return to normal after an outlier, and a different customer mix. 3. For each explanation, say whether the figures rule it out, support it or leave it open. 4. Give a verdict: supported, plausible or open, with a two-sentence reason. 5. Suggest the simplest clean test, such as a comparison group, switching the change off for a while or a comparison with the same period last year. Use only the figures I give and write "Missing:" for anything the check would need.

Take particular care to take other explanations seriously and calculate only with the figures provided.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

Check the following claim critically.
Claim: [Claim]
Available figures:
<data>
[Data]
</data>
What else happened in the period: [Context]
Approach: 1. Rephrase the claim as a testable statement with a metric, period and comparison. 2. Check other explanations: season and holidays, random fluctuation, simultaneous changes, changed measurement or tracking, return to normal after an outlier, and a different customer mix. 3. For each explanation, say whether the figures rule it out, support it or leave it open. 4. Give a verdict: supported, plausible or open, with a two-sentence reason. 5. Suggest the simplest clean test, such as a comparison group, switching the change off for a while or a comparison with the same period last year. Use only the figures I give and write "Missing:" for anything the check would need.

Take particular care to take other explanations seriously and calculate only with the figures provided.

Before you answer, check: Are all instructions met? If key information is missing or a placeholder in square brackets is still in the text, ask instead of guessing. Mark assumptions as assumptions.

This is an instruction for an AI, not the result itself.

Where this prompt works

These AI tools are a good fit for this prompt:

How to use it

Fill in the fields or tap a suggestion. Leave out names, addresses and other personal data.
Copy the finished prompt into an AI tool of your choice.
Ask someone who did not introduce the change to read the check and suggest other explanations that are missing.

Why does this work?

Before-and-after comparisons look convincing but easily confuse chance, season and cause. The fixed list of other explanations slows down the wish that your own measure worked.

Tips

A verdict of "open" is not a failure. It shows which small test will settle the question next time.

You may use our prompts freely, including at work. Please do not republish them as your own collection. We give no warranty for AI results; check them yourself. Terms of use

Check a before-and-after claim

Result preview

Fill in

Your prompt

Where this prompt works

How to use it

Why does this work?

Tips

Also a good fit

Prepare a decision with a scoring matrix New

What to drop when everything is urgent New

Create a SWOT analysis New