REVIEW 4 major objections 5 minor 48 references
Global p-Values in Multi-Design Studies
T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read The paper proposes a global p-value, the g-value, that aggregates all p-values from a multi-design study by rescaling the maximum p-value by a calibration constant, so the test keeps asymptotic size α with power close to a single valid…
desk verdict A neat rescaling idea for aggregating p-values, but the flagship analytic example only works for a specific error distribution, and the general algorithm is unvalidated. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the calibration constant $\alpha^*(\alpha, \mathcal{A})$, the value $\gamma \in [\alpha, 1]$ at which the asymptotic null probability that all p-values in $\mathcal{A}$ exceed $1-\gamma$ equals $\alpha$. Equivalently, it inverts the null distribution of the largest p-value. In the paper's worked example, the two test statistics have an asymptotic bivariate normal distribution with correlation $\sqrt{3/\pi}$, so the null distribution of the maximum p-value has a closed form and $\alpha^*$ can be computed by a one-dimensional root search. For settings without a closed form, the paper proposes a resampling algorithm—bootstrap resampling of observations followed by permutation of the variable of interest—to approximate the null distribution of the maximum p-value. A quantile variant, replacing the maximum by a $\gamma$-quantile of the p-values, is offered as a robustness adjustment.
What would settle it
Take a small sample, run two analyses on the same variable where one uses a t-test and the other uses a different model specification or a transformation of the outcome, and apply the proposed bootstrap-then-permutation algorithm to compute $\alpha^*$. Under $H_0$ with $\mu = 0$, simulate the empirical rejection rate of $G_n(\mathcal{A}) \le \alpha$ over many replications; if the rate exceeds $\alpha$ by more than simulation error, the claimed size control fails. A second check applies the same simulation to the quantile variant, since the paper does not prove that the max-based $\alpha^*$ remains valid when the statistic is replaced by a quantile.
Extended reading notes
Core claim
At the paper's center is the decision rule $G_n(\mathcal{A}) = (\alpha/\alpha^*(\alpha,\mathcal{A}))\,\max_{A\in\mathcal{A}} P_n(A)$, where $\alpha^*(\alpha,\mathcal{A})$ solves $\lim_{n\to\infty} \Pr(\max_{A\in\mathcal{A}} P_n(A) \le \gamma \mid H_0) = \alpha$ for $\gamma$. Because the unadjusted maximum p-value is conservative—it rejects only when all tests happen to be significant—the paper rescales it by $\alpha/\alpha^*$, which is equivalent to comparing the largest p-value against the larger threshold $\alpha^*$. The paper proves in a two-test example (one-sample t-test and Wilcoxon rank-sum test) that the asymptotic joint distribution of the test statistics is bivariate normal with correlation $\sqrt{3/\pi}$, computes $\alpha^*(0.05,\mathcal{A}) \approx 0.0601$, and shows by simulation that empirical rejection rates under $H_0$ sit near $0.05$ while power tracks the individual tests rather than the conservative bound.
Load-bearing premise
Everything rests on being able to compute or reliably approximate the null distribution of the maximum p-value over the full set $\mathcal{A}$; if the resampling scheme does not produce a valid null for every analysis strategy simultaneously (for example, when the strategies use different models or transformations), the calibrated constant $\alpha^*$ will not deliver the claimed error control.
Editorial extensions
If this is right
- A multi-design study can report a single g-value; rejecting $H_0$ when $G_n(\mathcal{A}) \le \alpha$ gives an asymptotic type I error of $\alpha$ regardless of how many strategies are in $\mathcal{A}$.
- The g-value's power is close to that of the individual valid tests, so the method avoids the large power loss of the most conservative p-value.
- Because the set $\mathcal{A}$ must be stated in advance, selective reporting of only the most favorable p-value is structurally discouraged.
- The framework extends to regression and maximum-likelihood settings whenever estimators admit an influence-function representation; for complex cases the bootstrap-permutation algorithm provides a numerical route.
- The quantile version lets researchers downweight a few poorly fitting analysis choices while retaining the g-value interpretation.
Reading between the lines
- A natural extension not developed in the paper is to apply the same calibration logic to any class of asymptotically normal test statistics, so the g-value could serve as a template for aggregating specification-level effect sizes or confidence bounds.
- Because the paper deliberately leaves dependent observations out of scope, extending the bootstrap-permutation scheme to time series would require a block-bootstrap or sieve-based null, which is a next step the paper does not take.
- The requirement to enumerate $\mathcal{A}$ in advance suggests a practical workflow: researchers register the set of analyses before looking at data, then report the g-value; this could be tested as an intervention on reporting behavior.
- The quantile variant may control size only under stronger assumptions than the maximum variant; a simulation comparing type I error across quantile levels would clarify when it is safe.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a 'g-value' for multi-design studies: for a pre-specified set A of analysis strategies, one computes the p-value P_n(A) for each strategy, takes the largest p-value, and rescales it by α/α*, where α* is the α-quantile of the null distribution of the maximum p-value. The authors argue that this construction controls the asymptotic type I error at α while avoiding the conservatism of simply reporting the largest p-value, and that it reduces selective reporting by forcing researchers to enumerate A. The method is illustrated with a two-test example involving the one-sample t-test and the Wilcoxon signed-rank test, supported by a small Monte Carlo study, and the paper discusses practical computation via a bootstrap-plus-permutation algorithm and a quantile-based robust variant.
Significance. Conceptually, the g-value is an elegant and appealing idea: instead of aggregating p-values through an arbitrary combination rule, one calibrates the maximum p-value against its true null distribution, thereby turning the most conservative multi-design summary into an exact-level test. If the joint null distribution of the p-values were available, the construction would deliver much of what the abstract promises. The paper also usefully distinguishes its setting from classical multiple testing and selective inference. However, the current manuscript does not establish the required null distribution in any general setting: the analytic example is valid only under additional distributional assumptions, the general resampling algorithm is unvalidated, and the quantile variant as written does not control size. The contribution is therefore a promising framework rather than a demonstrated method.
major comments (4)
- [Section 3.1, Eq. (3) and Eq. (4)] The covariance between the normalized t-statistic and the Wilcoxon signed-rank statistic is not universal. From the influence-function representation in Eq. (3), the covariance equals (√3/σ) E[|X_i| G(|X_i|)], which depends on F through G and σ. The value √(3/π) used in Eq. (4) is correct for Gaussian F but not for every symmetric F. For instance, if X_i ~ Uniform(-1,1), then σ=1/√3 and G(x)=x, so the covariance equals 1; the two tests are asymptotically equivalent and the null distribution of max P is Uniform(0,1), so α* should be 0.05, not 0.0601. Consequently, the claim that Example 1 demonstrates asymptotic size α for arbitrary symmetric F is false, and the Section 3.2 simulation, which feeds the normal-based α* into t15 data, does not provide the advertised evidence. The framework itself survives if α* is computed from the correct joint null distribution, but the paper's only analytical demonstration needs either a restriction to a specific F or a derivation of the F-dependent correction.
- [Section 4.1 (quantile variant)] The robust g-value G_n(A;γ) replaces max_A P_n(A) by Quantile_γ but keeps the same α* defined in Eq. (2) from the null distribution of the maximum. This does not control type I error: because Quantile_γ ≤ max_A P_n(A) pointwise for γ<1, the event {Quantile_γ ≤ α*} contains {max_A P_n(A) ≤ α*}, so the asymptotic rejection probability is at least α, and generically strictly larger. The quantile variant requires its own calibration constant computed from the null distribution of the quantile statistic.
- [Section 4.2 (computation)] The bootstrap-with-replacement followed by permutation proposal is the only general route to α* for arbitrary A, but it is not accompanied by any formal result or simulation. The statement that permutation 'ensures that the null hypothesis holds for all resampled datasets' addresses the marginal validity of each test, not the joint null distribution of the entire vector of p-values under the original data-generating process; for analyses that differ in variable transformations, model specification, or preprocessing, the resampling scheme can alter the dependence structure. As written, the paper provides no evidence that the resampled maximum-p distribution consistently estimates the quantity needed in Eq. (2). A theorem with explicit conditions or a simulation study is required before the general claims in the abstract and conclusion can be accepted.
- [Section 2 and Section 5 (scope of claims)] The paper states in the abstract and conclusion that the g-value 'rigorously controls type I error rates' and 'maintains strong statistical power.' These claims are not established for the general setting: no theorem is stated for the existence and uniqueness of α* under the stated regularity conditions, and the power statement is only illustrated in the two-test example. Please restate the guarantees as asymptotic and conditional on a correctly specified null distribution of max_A P_n(A), or prove them under explicit assumptions.
minor comments (5)
- [Section 3.1 and Section 3.2] The one-sample procedure used in Example 1 is the Wilcoxon signed-rank test, not the Wilcoxon rank-sum test; the terminology should be corrected.
- [Section 3.1] The notation G_n(A) for the g-value collides with the CDF G(·) used in Eq. (3); consider renaming the CDF (e.g., F_{|X|}).
- [Section 3.1, Eq. (2)] The notation 'argzero' is non-standard; define it as the unique γ satisfying the displayed equation, and state conditions for existence and uniqueness.
- [Section 3.2] With 10^5 Monte Carlo replications, the Monte Carlo standard error of a rejection proportion near 0.05 is about 0.0007; reporting the estimated type I error with this precision or a confidence interval would make the 'close to 0.05' claim quantitative.
- [Section 4.2] The 'algorithm' is described only in prose; a step-by-step pseudocode and a discussion of how α* is estimated from the resampled distribution would improve reproducibility.
Circularity Check
No significant circularity: the g-value is a calibration rule, and its size control follows from the definition of alpha* rather than from any fitted parameter or load-bearing self-citation.
full rationale
The g-value construction in Eq. (2) is a calibration rule: alpha* is defined as the null quantile of the maximum p-value, so the statement that the rule has asymptotic size alpha is true by construction rather than by an independent theorem. This is normal statistical calibration, not a fitted-input/prediction circularity, because alpha* is not estimated from data and the hard part—computing the joint null distribution of the p-values—is addressed separately (analytically in Example 1, via resampling in Section 4.2). The only self-reference is the citation of Boulaguiem et al. (2024a,b) for the general idea of adjusting conservative procedures; that citation is illustrative, not load-bearing, since the g-value's definition and calibration do not rest on any result from those papers. The analytic example's use of a covariance matrix that is not distribution-free for all symmetric F is a correctness concern about the example, not a circularity: the alpha* used in the simulation is computed from the stated null distribution and then evaluated on t15 data rather than being fit to the t15 simulation. The quantile variant and the permutation/bootstrapping algorithm are under-verified assumptions, but they are not circular: they do not assume the conclusion. Hence no step reduces to its own input in the sense of the circularity checklist.
Assumptions & free parameters
free parameters (1)
- quantile level γ in robust variant
assumptions (5)
- domain assumption The set A of analysis strategies is fixed in advance and chosen independently of the data, and every test in A is asymptotically valid and powerful.
- domain assumption The limiting joint distribution of the p-values (equivalently test statistics) across A under H0 is known or can be approximated by resampling.
- standard math The maximum p-value statistic has a continuous, strictly increasing null distribution, so the argzero defining α* exists and is unique.
- ad hoc to paper Bootstrap with replacement followed by permuting the variable of interest yields a valid null distribution for all A∈A simultaneously.
- standard math Estimators in regression settings admit an influence function representation √n(β̂_j−β_j)=Σφ_j(Y_i,x_i,β)+o_p(1).
invented entities (1)
-
g-value (global p-value)
Cite this review
Pith. "Pith review of Global p-Values in Multi-Design Studies." pith.science (2026). https://pith.science/paper/PGIXWWGI
@misc{pith2026250703815,
author = {Pith},
title = {Pith review of: Global p-Values in Multi-Design Studies},
year = {2026},
howpublished = {\url{https://pith.science/paper/PGIXWWGI}},
note = {Machine review of arXiv:2507.03815}
}
read the original abstract
Replicability issues -- referring to the difficulty or failure of independent researchers to corroborate the results of published studies -- have hindered the meaningful progression of science and eroded public trust in scientific findings. In response to the replicability crisis, one approach is the use of multi-design studies, which incorporate multiple analysis strategies to address a single research question. However, there remains a lack of methods for effectively combining outcomes in multi-design studies. In this paper, we propose a unified framework based on the g-value, for global p-value, which enables meaningful aggregation of outcomes from all the considered analysis strategies in multi-design studies. Our framework mitigates the risk of selective reporting while rigorously controlling type I error rates. At the same time, it maintains statistical power and reduces the likelihood of overlooking true positive effects. Importantly, our method is flexible and broadly applicable across various scientific domains and outcome results.
Figures
Reference graph
Works this paper leans on
-
[1]
Allison, P. D. (2009). Missing data. The SAGE handbook of quantitative methods in psychology , 23:72--89
work page 2009
-
[2]
Banks, G. C., Rogelberg, S. G., Woznyj, H. M., Landis, R. S., and Rupp, D. E. (2016). Evidence on questionable research practices: The good, the bad, and the ugly. Journal of Business and Psychology , 31:323--338
work page 2016
-
[3]
Benjamin, D. J., Berger, J. O., Johannesson, M., Nosek, B. A., Wagenmakers, E.-J., Berk, R., Bollen, K. A., Brembs, B., Brown, L., Camerer, C., et al. (2018). Redefine statistical significance. Nature Human Behaviour , 2(1):6--10
work page 2018
-
[4]
and Hochberg, Y
Benjamini, Y. and Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B , 57(1):289--300
1995
-
[5]
Boateng, E. Y., Abaye, D. A., et al. (2019). A review of the logistic regression model with emphasis on medical research. Journal of Data Analysis and Information Processing , 7(04):190
work page 2019
- [6]
-
[7]
Boukerche, A., Zheng, L., and Alfandi, O. (2020). Outlier detection: Methods, models, and classification. ACM Computing Surveys , 53(3):1--37
work page 2020
-
[8]
Boulaguiem, Y., Insolia, L., Victoria-Feser, M.-P., Couturier, D.-L., and Guerrier, S. (2024a). Multivariate adjustments for average equivalence testing. arXiv Preprint , (2411.16429)
work page Pith review arXiv 2024
Show all 48 references
-
[9]
N., Victoria-Feser, M.-P., Guerrier, S., and Couturier, D.-L
Boulaguiem, Y., Quartier, J., Lapteva, M., Kalia, Y. N., Victoria-Feser, M.-P., Guerrier, S., and Couturier, D.-L. (2024b). Finite sample corrections for average equivalence testing. Statistics in Medicine , 43(5):833--854
2024
-
[10]
Brodeur, A., Cook, N., and Heyes, A. (2020). Methods matter: P-hacking and publication bias in causal analysis in economics. American Economic Review , 110(11):3634--3660
2020
-
[11]
Brodeur, A., L \'e , M., Sangnier, M., and Zylberberg, Y. (2016). Star wars: T he empirics strike back. American Economic Journal: Applied Economics , 8(1):1--32
2016
-
[12]
F., Dreber, A., Forsell, E., Ho, T.-H., Huber, J., Johannesson, M., Kirchler, M., Almenberg, J., Altmejd, A., Chan, T., et al
Camerer, C. F., Dreber, A., Forsell, E., Ho, T.-H., Huber, J., Johannesson, M., Kirchler, M., Almenberg, J., Altmejd, A., Chan, T., et al. (2016). Evaluating replicability of laboratory experiments in economics. Science , 351(6280):1433--1436
2016
-
[13]
F., Dreber, A., Holzmeister, F., Ho, T.-H., Huber, J., Johannesson, M., Kirchler, M., et al
Camerer, C. F., Dreber, A., Holzmeister, F., Ho, T.-H., Huber, J., Johannesson, M., Kirchler, M., et al. (2018). Evaluating the replicability of social science experiments in nature and science between 2010 and 2015. Nature Human Behavior , 2:637--644
2018
-
[14]
Christensen, G., Freese, J., and Miguel, E. (2019). Transparent and reproducible social science research: How to do open science . University of California Press
2019
-
[15]
J., Gopalan, R., Berlin, J., and Matthews, D
Easterbrook, P. J., Gopalan, R., Berlin, J., and Matthews, D. R. (1991). Publication bias in clinical research. The Lancet , 337(8746):867--872
1991
-
[16]
Elliott, G., Kudrin, N., and W \"u thrich, K. (2022). Detecting p-hacking. Econometrica , 90(2):887--906
2022
-
[17]
and Li, R
Fan, J. and Li, R. (2001). Variable selection via nonconcave penalized likelihood and its oracle properties. Journal of the American Statistical Association , 96(456):1348--1360
2001
-
[18]
Gasparin, M., Wang, R., and Ramdas, A. (2025). Combining exchangeable p-values. Proceedings of the National Academy of Sciences , 122(11):e2410849122
2025
-
[19]
S., Parker, T
Gould, E., Fraser, H. S., Parker, T. H., Nakagawa, S., Griffith, S. C., Vesk, P. A., Fidler, F., Hamilton, D. G., Abbey-Lee, R. N., Abbott, J. K., et al. (2025). Same data, different analysts: Variation in effect sizes due to analytical decisions in ecology and evolutionary bi...
2025
-
[20]
Hansen, L. P. (1982). Large sample properties of generalized method of moments estimators. Econometrica , 50(4):1029--1054
1982
-
[21]
R., Burli, P., Chen, N., Grieco, P., Ekpe, G., Pugatch, T., et al
Huntington-Klein, N., Arenas, A., Beam, E., Bertoni, M., Bloem, J. R., Burli, P., Chen, N., Grieco, P., Ekpe, G., Pugatch, T., et al. (2021). The influence of hidden researcher decisions in applied microeconomics. Economic Inquiry , 59(3):944--960
2021
-
[22]
Ioannidis, J. P. (2005). Why most published research findings are false. PLoS Medicine , 2(8):e124
2005
-
[23]
K., Loewenstein, G., and Prelec, D
John, L. K., Loewenstein, G., and Prelec, D. (2012). Measuring the prevalence of questionable research practices with incentives for truth telling. Psychological Science , 23(5):524--532
2012
-
[24]
Johnson, V. E. (2013). Revised standards for statistical evidence. Proceedings of the National Academy of Sciences , 110(48):19313--19317
2013
-
[25]
King, G. (1995). Replication, replication. PS: Political Science & Politics , 28(3):444--452
1995
-
[26]
Lehmann, E. L. and Romano, J. P. (2022). Testing S tatistical H ypotheses, 4th Edition . Springer
2022
-
[27]
and Sevestre, P
M \'a ty \'a s, L. and Sevestre, P. (2013). The econometrics of panel data: Handbook of theory and applications , volume 28. Springer Science & Business Media
2013
-
[28]
and Michaillat, P
McCloskey, A. and Michaillat, P. (2025). Critical values robust to p-hacking. Review of Economics and Statistics , Forthcoming:1--35
2025
-
[29]
J., Dreber, A., Holzmeister, F., Huber, J., Johannesson, M., Kirchler, M., Neus \"u ss, S., Razen, M., Weitzel, U., Abad-D \' az, D., et al
Menkveld, A. J., Dreber, A., Holzmeister, F., Huber, J., Johannesson, M., Kirchler, M., Neus \"u ss, S., Razen, M., Weitzel, U., Abad-D \' az, D., et al. (2024). Nonstandard errors. Journal of Finance , 79(3):2339--2390
2024
-
[30]
Miao, X., Wu, Y., Chen, L., Gao, Y., and Yin, J. (2022). An experimental survey of missing data imputation algorithms. IEEE Transactions on Knowledge and Data Engineering , 35(7):6630--6650
2022
-
[31]
M., Gerber, A., Glennerster, R., Green, D
Miguel, E., Camerer, C., Casey, K., Cohen, J., Esterling, K. M., Gerber, A., Glennerster, R., Green, D. P., Humphreys, M., Imbens, G., Laitin, D., Madon, T., Nelson, L., Nosek, B. A., Petersen, M., Sedlmayr, R., Simmons, J. P., Simonsohn, U., and der Laan, M. V. (2014). Promot...
2014
-
[32]
and Roehrkasse, A
Muller, C. and Roehrkasse, A. F. (2025). Falling racial inequality and rising educational inequality in us prison admissions for drug, violent, and property crimes. Proceedings of the National Academy of Sciences , 122(4):e2418077122
2025
-
[33]
A., Alter, G., Banks, G
Nosek, B. A., Alter, G., Banks, G. C., Borsboom, D., Bowman, S. D., et al. (2015). Promoting an open research culture. Science , 348(6242):1422--1425
2015
-
[34]
Shrout, P. E. and Rodgers, J. L. (2018). Psychology, science, and knowledge construction: Broadening perspectives from the replication crisis. Annual Review of Psychology , 69(1):487--510
2018
-
[35]
L., Martin, D
Silberzahn, R., Uhlmann, E. L., Martin, D. P., Anselmi, P., Aust, F., Awtrey, E., Bahn \' k, S ., Bai, F., Bannard, C., Bonnier, E., et al. (2018). Many analysts, one data set: Making transparent how variations in analytic choices affect results. Advances in Methods and Practi...
2018
-
[36]
P., Nelson, L
Simmons, J. P., Nelson, L. D., and Simonsohn, U. (2021). Pre-registration: Why and how. Journal of Consumer Psychology , 31(1):151--162
2021
-
[37]
D., and Simmons, J
Simonsohn, U., Nelson, L. D., and Simmons, J. P. (2014). P-curve: A key to the file-drawer. Journal of Experimental Psychology: General , 143(2):534
2014
-
[38]
P., and Nelson, L
Simonsohn, U., Simmons, J. P., and Nelson, L. D. (2020). Specification curve analysis. Nature Human Behaviour , 4(11):1208--1214
2020
-
[39]
Stanley, T. D. (2005). Beyond publication bias. Journal of Economic Surveys , 19(3):309--345
2005
-
[40]
and Tibshirani, R
Taylor, J. and Tibshirani, R. J. (2015). Statistical learning and selective inference. Proceedings of the National Academy of Sciences , 112(25):7629--7634
2015
-
[41]
Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society: Series B , 58(1):267--288
1996
-
[42]
and Wang, R
Vovk, V. and Wang, R. (2020). Combining p-values via averaging. Biometrika , 107(4):791--808
2020
-
[43]
J., and Hammad, M
Wang, H., Bah, M. J., and Hammad, M. (2019). Progress in outlier detection techniques: A survey. IEEE Access , 7:107964--108000
2019
-
[44]
Wilson, D. J. (2019). The harmonic mean p-value for combining dependent tests. Proceedings of the National Academy of Sciences , 116(4):1195--1200
2019
-
[45]
Xie, Y., Wang, K., and Kong, Y. (2021). Prevalence of research misconduct and questionable research practices: A systematic review and meta-analysis. Science and Engineering Ethics , 27(4):41
2021
-
[46]
Yoon, S., Baik, B., Park, T., and Nam, D. (2021). Powerful p-value combination methods to detect incomplete association. Scientific Reports , 11(1):6980
2021
-
[47]
and Kumbier, K
Yu, B. and Kumbier, K. (2020). Veridical data science. Proceedings of the National Academy of Science , 117(8):3920--3929
2020
-
[48]
Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American Statistical Association , 101(476):1418--1429
2006
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.