Audit a claimed correlation for hidden confounders
Use when someone claims X drives Y and you need to pressure-test it before acting.
Act as a skeptical analyst trained in causal reasoning.
Claim: "{{claimed_relationship}}".
Available variables: {{available_variables}}.
Context: {{context}}.
Pressure-test the claim:
1. List plausible confounders that could explain the correlation.
2. Identify possible reverse causation and selection bias.
3. Suggest checks I can run with the data I have (stratification, controls, lagged correlation).
4. Note what experiment or quasi-experiment would actually establish causation.
5. Give a calibrated verdict: how much should I trust this claim right now, and why.Click the copy button in the top right of the block to grab the full prompt.
Replace each placeholder below with your own values before you run the prompt.
- {{claimed_relationship}}
- {{available_variables}}
- {{context}}
Related prompts
You are a meticulous data analyst. Here is a sample of my raw dataset (header row plus a few rows): {{sample_rows}} Context: this data describes {{data_description}} and I plan to...
You are a senior Python data engineer. Write clean, reproducible pandas code to clean my dataset. Columns and their meaning: {{column_spec}} Known issues to handle: {{known_issues}...
Act as a data analyst guiding me through exploratory data analysis. Dataset summary: {{dataset_summary}} My goal / the question I care about: {{analysis_goal}} Produce an EDA plan...
You are a SQL expert writing for a {{sql_dialect}} database. Here is the relevant schema (tables, columns, types, keys): {{schema}} Question to answer: {{question}} Rules: - Use on...
You are a SQL reviewer. Explain and debug the query below. Dialect: {{sql_dialect}} What I expected it to return: {{expected_result}} What is actually wrong (if known): {{symptom}}...
You are a database performance specialist for {{sql_dialect}}. Slow query: {{query}} Context: - Approximate row counts of the main tables: {{table_sizes}} - Existing indexes: {{ind...
0 Comments
Loading discussion...