§ ExampleSample review · simulated output

A real PeerPanel review.

A complete review of a published carbon-tax meta-analysis. Click any finding on the left to highlight the exact quote on the right. Keyboard ↑ ↓ to navigate findings.

§ 01 · Dashboard

Carbon-tax meta-analysis review

Score
34 / 100
Decision
▲ MAJOR REVISIONS
Findings
23
Run time
7:42
Dr. Methods
5
Dr. Statsshifted
7
Dr. Literature
4
Dr. Writing
3
Dr. Baselineshifted
4
§ 02 · Data

What the review processed

Pages parsed28
References checked62 / 71 verified (87%)
Tables extracted4
Figures detected3
Hidden text scannedclean
Author-blindenforced
§ 03 · Major findings

Six items the editor escalated

Click any finding to jump to the source passage.
majorDr. Baseline · consensusp. 6
No-policy counterfactual is undefined.

The reported emissions reduction is computed against an unspecified baseline. The methodology section refers only to "trend extrapolation" without naming the model, parameters, or counterfactual horizon.

majorDr. Stats · consensusp. 11
No multicollinearity diagnostics for the panel regression.

VIF or tolerance values are not reported for the seven covariates in Table 3. Carbon price, GDP per capita, and energy intensity are likely to share substantial variance.

majorDr. Stats · consensusp. 12
Standard errors not clustered at the country level.

With repeated observations per country, residuals are almost certainly correlated within country. Conventional standard errors will understate uncertainty in the policy effect.

majorDr. Literature · contestedp. 3
Novelty claim overstated relative to Andersson (2019).

The introduction characterises this work as "the first cross-jurisdictional synthesis." Andersson (2019) covers 40 jurisdictions over a similar window. Recommend softening to "an updated cross-jurisdictional synthesis."

majorDr. Methods · consensusp. 8
Control jurisdictions selected post-hoc.

Section 3.2 describes the selection of seven non-tax jurisdictions as "matched on emissions intensity." Matching appears to have been performed after the treatment effect was first computed. This is a degree of freedom that should be disclosed.

majorDr. Writing · uniquep. 18
Conclusion overstates causal direction.

"Our findings demonstrate that carbon taxes cause substantial emissions reductions" is stronger than the panel design supports. Reframe as "are associated with" pending an instrumental-variable robustness check.

§ 04 · Minor & contested

Eleven minor items

minorDr. Stats · consensusp. 13
Confidence intervals missing on Figure 2.

The bar chart shows point estimates only. Add 95% CIs or note their absence.

minorDr. Writing · uniquep. 9
Inconsistent terminology: "carbon levy" vs. "carbon tax."

Used interchangeably throughout. Pick one and apply consistently.

minorCitation check · systemp. 24
Reference 41 (Lin & Park 2021): DOI not resolved.

Verify DOI or replace with a reachable citation.

contestedDr. Stats vs. Dr. Methodsp. 14
Whether log-transform of GDP is appropriate.

Dr. Stats: "log is conventional and improves residual normality." Dr. Methods: "transformation should be motivated, not assumed." Editor: not deductive, note in revision.

§ 05 · Individual review excerpts

What each agent said before deliberation

minorDr. Stats · initialround 1
Started with 9 findings, retained 7 after deliberation.

Two retracted under Dr. Methods’ rebuttal: a complaint about heteroskedasticity (the paper does report a Breusch-Pagan test on p. 13) and a complaint about missing weights (Section 3.4 specifies inverse-variance weighting).

minorDr. Baseline · initialround 1
Started with 2 findings, escalated to 4 after deliberation.

Two findings escalated after reading Dr. Methods’ p. 8 critique. The connection between weak baselines and post-hoc control selection produced two additional major findings.

§ 06 · Strengths

What the panel agreed was strong

+

Pre-registered hypothesis, time-stamped before data collection began.

+

Reproducibility package on Zenodo with complete code and intermediate data.

+

Reference list passes citation verification at 87%, above panel average of 81%.

+

Clear separation of treatment and identification sections.

§ 07 · Evidence trail

How to verify any finding

Every finding above is anchored to a passage in your manuscript. Click on any finding card to scroll the right pane to the source page and highlight the exact quote. The PDF export carries the same anchoring as inline footnotes.

Carbon Pricing and Emissions Outcomes: A Cross-Jurisdictional Meta-Analysis · pp. 1-28

Abstract

We synthesise empirical estimates of emissions response to carbon pricing across 34 jurisdictions and a fourteen-year window. As the first cross-jurisdictional synthesis of this scope, we report a pooled treatment effect of −5.4% on per-capita emissions. Sub-group analyses by sector and pricing instrument follow.

1. Introduction (p. 3)

Carbon pricing has become a central instrument in climate policy. The empirical literature on its emissions effect remains fragmented, with single-jurisdiction studies dominating. This paper consolidates published estimates into a single estimator…

3. Methods (p. 6)

We adopt a difference-in-differences specification, comparing treatment jurisdictions to non-treated comparators over the policy window. The counterfactual is constructed from a trend extrapolation of pre-treatment emissions. Robustness checks vary the pre-period bandwidth.

Section 3.2, control jurisdictions. Seven non-tax jurisdictions were selected as matched on emissions intensity within ±0.4 standard deviations. Matching balances are reported in Appendix B.

3.4 Estimation (p. 9)

We use “carbon levy” and “carbon tax” interchangeably throughout the analysis, following the OECD convention. Inverse-variance weighting is applied to study-level estimates.

4. Results (p. 11)

Table 3 reports the panel regression of log per-capita emissions on log carbon price, log GDP per capita, log energy intensity, urbanisation, manufacturing share, electricity mix, and time-invariant fixed effects. Multicollinearity diagnostics are not reported.

Standard errors. We report conventional heteroskedasticity-robust standard errors throughout. Country-level clustering is not used in the main specification, though robustness in Appendix C explores cluster-robust variants.

4.2 Figure 2 (p. 13)

The bar chart on this page summarises the per-instrument effect. Point estimates are shown without confidence intervals; the underlying values are tabulated in Appendix A. A log-transform was applied to GDP. The log-transform is conventional in the carbon-pricing literature and improves residual normality.

6. Discussion (p. 18)

Our findings demonstrate that carbon taxes cause substantial emissions reductions across the studied jurisdictions. The magnitude is consistent with the upper end of single-jurisdiction estimates and supports the policy recommendation in Section 7.

References (p. 24)

[41] Lin, J. & Park, S. (2021). “Sectoral responses to carbon pricing in East Asia.” J. Climate Policy 14(2): 211-235. DOI: 10.1041/jcp.2021.0411. [42] Andersson, J. J. (2019). Carbon taxes and CO2 emissions: Sweden as a case study. American Economic Journal…

Round 1: initial review excerpts

[Dr. Stats, initial round, before deliberation] Concern raised about absent heteroskedasticity testing and missing weighting scheme. Both subsequently retracted after Dr. Methods quoted the Breusch-Pagan test (p. 13) and the inverse-variance weighting (Section 3.4).

[Dr. Baseline, initial round] Two new findings escalated after reading Dr. Methods' p. 8 critique. The connection between weak baselines and post-hoc control selection produced two additional major findings.

Try it on your own paper.

Free during Research Preview. The same report format you just navigated.

Start your first review →