Research and evidence

Free Causal Claim Reviewer.

Causation assumptions, confounders and alternative explanations.

Review causal language in a report and align the conclusion with the evidence actually available.

Chrome extension + account required.
25 free automation tasks/month across tools. Browser-provider limits apply.

WHAT TO HAVE READY

Claim and study description.

YOUR STARTING POINT

A task you can make yours.

Copy it here. Paste it into WebAct when you are ready.

This page prepares your prompt. The task runs in WebAct.

01

Make the task yours.

Review the prompt above. Add your goal and any requirements that matter.

02

Bring it to WebAct.

Open the source page, activate WebAct and select the relevant readable content. Paste the task and provide any additional source material it requests.

03

Check the useful part.

The result covers the readable content you provide or select. Expand the relevant page sections first and check that the output includes the records you intended.

A PRACTICAL GUIDE

How to use the Causal Claim Reviewer.

A causal claim requires more than two things changing together. Identify alternative explanations and what the design can actually establish.

Make the workflow fit your task.

Separate the observed association from the claimed causal effect. Identify timing, comparison groups, concurrent changes and plausible alternative explanations. Describe what the design supports and what additional evidence would help isolate the proposed cause.

What you provide
Claim and study description.
What you get
Causation assumptions, confounders and alternative explanations.

See the input and the result.

Illustrative input and output · a teaching example, not a live WebAct run

Example input

Signups increased after a homepage edit while an advertising campaign also launched.

Completed example

Treat the increase as an observation; both changes and other factors may contribute, so the edit's causal effect remains unresolved.
Download this example

Load this input into the prompt, then copy it to WebAct to try the task. Your result may differ from the illustration.

Decisions and troubleshooting.

Does an increase immediately after a website change prove the change caused it?

Timing alone is insufficient. Campaigns, seasonality, measurement changes and other factors may explain some or all of the increase.

Why can a controlled test still leave causal uncertainty?

Check assignment, adherence, missing data and whether the measured outcome matches the claim. Study design details affect interpretation.

Try it with your own source.

Replace the example with your material in the task prompt. Keep the requirements you need, then copy the task into WebAct.

Customize and copy the task ↑
BEFORE YOU BEGIN

Using WebAct.

The prompt and example are available here. Run your task in the WebAct Chrome extension.

Is this tool free?

You can copy and edit the prompt here for free. To use it with WebAct, install the Chrome extension, create and verify your account, and connect an available browser ChatGPT session. The Free plan includes 25 user-started automation tasks per calendar month, shared across tools. Ordinary browser Chat is separate; provider limits apply.

Learn about shared context in our privacy policy and choosing Chat or Task.

FROM A PROMPT TO YOUR OWN WORK

Bring the task to your page.

Install WebAct, add your source material and make the result yours.

Install WebAct