========================================================================
ANALYSIS PLAN
========================================================================

1. Data Cleaning:
   - Build panel from JSON response files (one per respondent).
   - Apply exclusion criteria.
   - Compute derived variables: MmryErr, ExpRev, OthExpRev, NormShift.

2. Descriptive Statistics:
   - Means, medians, SDs by condition (Truth vs Deception).
   - Correlation matrix of all return perception variables.
   - PCA of pre-reveal variables.

3. Replication Regressions (R1--R12):
   - Run on Truth-condition respondents first (clean replication).
   - Run on full sample with deception indicator as control.
   - Compare coefficients and R2 to original study.

4. Deception Treatment Analysis (N1--N5):
   - Compare means across conditions (t-tests).
   - Regression with deception indicator and controls.
   - Heterogeneous treatment effects by debrief_reaction.

5. New Variable Analysis (N6--N7):
   - Add optimism and risk aversion to existing regression specifications.
   - Assess incremental R2 beyond existing controls.

6. Exploratory (E1--E9):
   - As described above, conditional on data availability and sample size.

Multiple Testing Adjustment:
  We do not pre-commit to a specific adjustment method for the replication hypotheses (R1--R12), as these are directional replications of known effects.  For the new hypotheses (N1--N7), we will report both unadjusted and Bonferroni-adjusted p-values.