# Reviewer Objection Bank

Use these questions before submission.

## Description
- Are the records complete enough for the pattern you report?
- Is the unit of analysis clear?
- Are missing and excluded cases visible?

## Association
- Are X and Y measured independently enough?
- Is the direction of timing clear?
- Are plausible confounds acknowledged?

## Prediction
- Was model performance evaluated on held-out data?
- Is there outcome leakage?
- Are performance metrics tied to the decision context?

## Process
- Are event definitions theoretically meaningful?
- Did the analysis preserve order, timing, or transition structure?
- Are rare paths overinterpreted?

## Measurement
- Why should this trace, score, code, or detector represent the construct?
- Is reliability or coder agreement reported when relevant?
- Was the detector or scale validated for this population?

## Causal Evaluation
- What feature of the design supports causal interpretation?
- Are comparison groups equivalent or adjusted under credible assumptions?
- Is attrition handled transparently?

## Generalization
- What context is the sample from?
- What population is actually represented?
- Is there replication or external validation?