Teaching Norms: Direct Evidence of Parental Transmission

Replication of "Teaching Norms: Direct Evidence of Parental Transmission" · I4R reference

Execution Method: Translated Python
Note: Results were obtained via Stata-to-Python translation. Numerical differences may be translation artifacts rather than genuine reproducibility failures. Running with native Stata would provide a more accurate assessment.

Overall: Partially Reproducible

26 output CSVs produced, but coefficient matching is poor. 5 tables compared with scores ranging from 0% to 72%. Variable name formatting differences between Stata and Python (e.g., interaction terms) complicate matching.

Table Comparisons

Table 2 major differences Score: 65.0%

12 coefficients with paper values compared. Many interaction terms show sign reversals or large discrepancies.

RowPaperCodeStatusRel Diff
ViolationHelp-0.1-0.0974within tolerance2.6%
(unlabeled)-0.04-0.0477within tolerance19.3%
(unlabeled)-0.04-0.0453within tolerance13.3%
(unlabeled)-0.04-0.0429within tolerance7.3%
Child x ViolationHelp-0.120.1362major diff213.5%
Child0.12-0.1223minor diff202%
ViolationHelp-0.040.11minor diff375%
Child0.22-0.224major diff202%

Table A7 minor differences Score: 72.2%

Best-matching table. Several interaction coefficients within tolerance, others with sign or magnitude issues.

RowPaperCodeStatusRel Diff
Child0.110.1113within tolerance1.2%
ViolationHelp-0.11-0.133within tolerance20.9%
Child x ViolationHelp-0.11-0.0814within tolerance26.0%
Child0.180.3654minor diff103%
Child x ViolationHelp-0.13-0.0642minor diff50.6%

Table A10 major differences Score: 66.7%

Large appendix table with many specifications. Coefficients near zero amplify relative differences.

RowPaperCodeStatusRel Diff
ViolationHelp-0.039-0.0387within tolerance0.8%
(unlabeled)-0.057-0.0651within tolerance14.2%
Child0.2230.3684minor diff65.2%
ViolationHelp-0.0960.2227major diff332%
Child0.206-0.1313major diff164%

Table A9 major differences Score: 61.7%

Heterogeneity analysis with many interaction terms and subgroup specifications.

RowPaperCodeStatusRel Diff
children[1]0.190.1891within tolerance0.5%
Rain-0.08-0.1035within tolerance29.4%
children[3]0.260.0218minor diff91.6%
ViolationHelp-0.11-0.224minor diff104%

Table A6 no comparison data Score: N/A

Output produced but no I4R reference coefficients available for comparison.

Summary

The study examines how parents transmit social norms to children through direct observation of parental behavior in helping and punishment scenarios. The replication produced 26 output CSV files covering Tables 1-2 and Appendix Tables A2-A14. While some coefficients match well (particularly simple main effects and standard errors), many interaction terms (Child x ViolationHelp) show sign reversals or magnitude discrepancies. This is likely due to differences in how Stata and Python handle categorical variable interactions and the eststo/esttab coefficient capture pipeline.

Replication package: Stata .do files with experimental data. 26 output CSVs produced. Key blockers: eststo/esttab coefficient capture for multi-column regression tables.