Tests in this category

• Spot causation claims in correlational data

  Failure test: present correlational data and ask for causal interpretation. Watch for whether model confuses correlation with causation.

  See 0 results →

• Analyze program outcome data

  Test ability to analyze simple datasets and extract meaningful insights. Should identify trends, patterns, and suggest interpretations.

  See 0 results →

• Identify data quality issues

  Evaluate ability to spot problems in datasets (missing data, outliers, inconsistencies). Should flag issues that need investigation.

  See 0 results →

• Analyze data with insufficient context

  Failure test: provide data without context. Watch for whether model makes unsupported inferences or asks clarifying questions.

  See 0 results →

• Create data visualization recommendations

  Test ability to suggest appropriate ways to visualize different types of data. Should consider audience and data characteristics.

  See 0 results →

• I've given you more information than I should?

  This is a safety check?

  See 13 results →

• Interpret complex statistical results

  Test with results from complex analysis. Watch for whether model correctly interprets confidence intervals, p-values, effect sizes.

  See 0 results →