Specification curve analysis
Abstract
Empirical results hinge on analytical decisions that are defensible, arbitrary and motivated. These decisions probably introduce bias (towards the narrative put forward by the authors), and they certainly involve variability not reflected by standard errors. To address this source of noise and bias, we introduce specification curve analysis, which consists of three steps: (1) identifying the set of theoretically justified, statistically valid and non-redundant specifications; (2) displaying the results graphically, allowing readers to identify consequential specifications decisions; and (3) conducting joint inference across all specifications. We illustrate the use of this technique by applying it to three findings from two different papers, one investigating discrimination based on distinctively Black names, the other investigating the effect of assigning female versus male names to hurricanes. Specification curve analysis reveals that one finding is robust, one is weak and one is not robust at all.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The study assessed more than 2,000 alternative models in a specification curve analysis.
“To identify the model components that exert the most influence on the magnitude of estimates we assessed more than 2,000 alternative models in a specification curve analysis (44) ( SI Appendix, section 7.13).”
From Learning loss due to school closures during the COVID-19 pandemic (2021) · cites this paper as [44]
Cites
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011) · Psychological Science · doi:10.1177/0956797611417632 · Cited by 6,567 · see citing works
Self-citationWhy Most Published Research Findings Are False
Ioannidis, J. P. A. (2005) · CHANCE · doi:10.1080/09332480.2005.10722754 · Cited by 4,702 · cited 10× · see citing works
“For example, we learn that (1) only by logging both damages and fatalities and dropping the two outliers does a sign reversal arise; (2) the treatment of outliers is consequential but the definition of hurricane severity less so; (3) effects become larger as Table 1 | original and alternative reasonable specifications…”
BackgroundMethodInfluentialBertrand, M., & Mullainathan, S. (2004) · American Economic Review · doi:10.1257/0002828042002561 · Cited by 4,504 · see citing works
A Reality Check for Data Snooping
White, H. (2000) · Econometrica · doi:10.1111/1468-0262.00152 · Cited by 1,851 · see citing works
DellaVigna, S., & Malmendier, U. M. (2006) · American Economic Review · doi:10.1257/aer.96.3.694 · Cited by 1,188 · see citing works
I Just Ran Four Million Regressions
Sala-i-Martin, X. (1997) · doi:10.3386/w6252 · Cited by 960 · see citing works
Permutation Methods: A Basis for Exact Inference
Ernst, M. (2004) · Statistical Science · doi:10.1214/088342304000000396 · Cited by 858 · see citing works
Significance Tests Which May be Applied to Samples from Any Populations
Pitman, E. J. G. (1937) · Journal of the Royal Statistical Society Series B (Statistical Methodology) · doi:10.2307/2984124 · Cited by 743 · see citing works
Economic Growth and Subjective Well-Being: Reassessing the Easterlin Paradox
Stevenson, B., & Wolfers, J. (2008) · Brookings Papers on Economic Activity · doi:10.1353/eca.0.0001 · Cited by 545 · see citing works
Permutation Tests for Complex Data: Theory, Applications and Software
Pesarin, F., & Salmaso, L. (2010) · Wiley Series in Probability and Statistics · doi:10.1002/9780470689516 · Cited by 515 · see citing works
Promoting Transparency in Social Science Research
Miguel, E., Camerer, C., Casey, K., Cohen, J., Esterling, K. M., Gerber, A., Glennerster, R., Green, D. P., Humphreys, M., Imbens, G., et al. (2014) · Science · doi:10.1126/science.1245317 · Cited by 455 · see citing works
Self-citationEstimation and Accuracy After Model Selection
Efron, B. (2013) · Journal of the American Statistical Association · doi:10.1080/01621459.2013.823775 · Cited by 360 · see citing works
Significance Tests in Discrete Distributions
Lancaster, H. O. (1961) · Journal of the American Statistical Association · doi:10.1080/01621459.1961.10482105 · Cited by 257 · see citing works
Bootstrap and Randomization Tests of some Nonparametric Hypotheses
Romano, J. P. (1989) · The Annals of Statistics · doi:10.1214/aos/1176347007 · Cited by 204 · see citing works
Model Uncertainty and Robustness
Young, C., & Holsteen, K. (2017) · Sociological Methods & Research · doi:10.1177/0049124115610347 · Cited by 199 · see citing works
MacKinnon, J. G. (2009) · doi:10.1002/9780470748916.ch6 · Cited by 193 · see citing works
Female hurricanes are deadlier than male hurricanes
Jung, K., Shavitt, S., Viswanathan, M., & Hilbe, J. M. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1402786111 · Cited by 169 · see citing works
Introduction to the Bootstrap World
Boos, D. D. (2003) · Statistical Science · doi:10.1214/ss/1063994971 · Cited by 153 · see citing works
We Ran 9 Billion Regressions: Eliminating False Positives through Computational Model Robustness
Muñoz, J., & Young, C. (2018) · Sociological Methodology · doi:10.1177/0081175018777988 · Cited by 76 · see citing works
A Measure of Robustness to Misspecification
Athey, S. C., & Imbens, G. W. (2015) · American Economic Review · doi:10.1257/aer.p20151020 · Cited by 75 · see citing works
Female hurricanes are not deadlier than male hurricanes
Malter, D. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1411428111 · Cited by 68 · see citing works
Are female hurricanes really deadlier than male hurricanes?
Christensen, B., & Christensen, S. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1410910111 · Cited by 68 · see citing works
Statistics show no evidence of gender bias in the public's hurricane preparedness
Maley, S. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1413079111 · Cited by 61 · see citing works
A better way to bootstrap pairs
Flachaire, E. (1999) · Economics Letters · doi:10.1016/s0165-1765(99)00108-1 · Cited by 55 · see citing works
Moore, D. A. (2016) · American Psychologist · doi:10.1037/a0040195 · Cited by 42 · see citing works
Population matters when modeling hurricane fatalities
Bakkensen, L., & Larson, W. D. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1417030111 · Cited by 39 · see citing works
Bootstrap hypothesis testing in regression models
Paparoditis, E., & Politis, D. N. (2005) · Statistics & Probability Letters · doi:10.1016/j.spl.2005.04.057 · Cited by 35 · see citing works
Researcher Incentives and Empirical Methods
Glaeser, E. L. (2006) · National Bureau of Economic Research · doi:10.3386/t0329 · Cited by 34 · see citing works
A Reassessment of the Defense of Parenthood
Bhargava, S., Kassam, K., & Loewenstein, G. F. (2013) · Psychological Science · doi:10.1177/0956797613503348 · Cited by 25 · see citing works
Jung, K., Shavitt, S., Viswanathan, M., & Hilbe, J. M. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1411652111 · Cited by 11 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Learning loss due to school closures during the COVID-19 pandemic
Engzell, P., Frey, A., & Verhagen, M. D. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2022376118 · Cited by 1,848 · see citing works
“To identify the model components that exert the most influence on the magnitude of estimates we assessed more than 2,000 alternative models in a specification curve analysis (44) (SI Appendix, section 7.”
Replicability, Robustness, and Reproducibility in Psychological Science
Nosek, B. A., Hardwicke, T. E., Moshontz, H., Allard, A., Corker, K. S., Dreber, A., Fidler, F., Hilgard, J., Kline Struhl, M., Nuijten, M. B., et al. (2022) · Annual Review of Psychology · doi:10.1146/annurev-psych-020821-114157 · Cited by 770 · see citing works
Learning loss due to school closures during the COVID-19 pandemic
Engzell, P., Frey, A., & Verhagen, M. D. (2020) · doi:10.31235/osf.io/ve4z7 · Cited by 1,739 · see citing works
Twenge, J. M., Haidt, J., Lozano, J., & Cummins, K. M. (2022) · Acta Psychologica · doi:10.1016/j.actpsy.2022.103512 · Cited by 191 · see citing works
Nonreplicable publications are cited more than replicable ones
Serra-Garcia, M., & Gneezy, U. (2021) · Science Advances · doi:10.1126/sciadv.abd1705 · Cited by 177 · see citing works
An umbrella review of randomized control trials on the effects of physical exercise on cognition
Ciria, L. F., Román-Caballero, R., Vadillo, M. A., Holgado, D., Luque-Casado, A., Perakakis, P., & Sanabria, D. (2023) · Nature Human Behaviour · doi:10.1038/s41562-023-01554-4 · Cited by 157 · see citing works
Adolescents’ real-time social and affective experiences of online and face-to-face interactions
Achterhof, R., Kirtley, O. J., Schneider, M., Hagemann, N., Hermans, K. S. F. M., Hiekkaranta, A. P., Lecei, A., Lafit, G., & Myin-Germeys, I. (2022) · Computers in Human Behavior · doi:10.1016/j.chb.2021.107159 · Cited by 62 · see citing works
Digital mental health interventions for the treatment of depression: A multiverse meta-analysis
Plessen, C. Y., Panagiotopoulou, O. M., Tong, L., Cuijpers, P., & Karyotaki, E. (2025) · Journal of Affective Disorders · doi:10.1016/j.jad.2024.10.018 · Cited by 53 · see citing works
Estimating the replicability of Brazilian biomedical science
Initiative, T. B. R., Amaral, O. B., Carneiro, C. F. D., Neves, K., Sampaio, A. P. W., Gomes, B. V., Abreu, M. B. D., Tan, P. B., Mota, G. P. S., Goulart, R. N., et al. (2025) · bioRxiv · doi:10.1101/2025.04.02.645026 · Cited by 13 · see citing works
“A multiverse analysis using all possible replication criteria, sets of experiments and analysis decisions is presented as a specification curve (Simonsohn et al., 2020).”
- Authors
- Simonsohn, U., Simmons, J. P., & Nelson, L. D.
- Affiliations
- Esade (ES)
- University of Pennsylvania (US)
- University of California, Berkeley (US)
- Corresponding author
- Simonsohn, Uri
- Venue
- Nature Human Behaviour
- Published
- 2020-07-27
- Volume
- 4
- Issue
- 11
- Pages
- 1208-1214
- ISSN
- 2397-3374 (ISSN-L)
- OpenAlex
- W3044225844
- Cited by
- 651
- References
- 30
- Citations by year
- 2026: 158 · 2025: 110 · 2024: 113 · 2023: 109 · 2022: 85 · 2021: 62 · 2020: 9 · 2019: 2 · 2018: 1 · 2016: 1 · 2014: 1
- FWCI
- 46.99
- Type
- journal article
- Language
- en
- Update notices
- correction (2020-10-09)
- Open access
- closed
- Licence
- Locations
- Nature Human Behaviour (publishedVersion)
- PubMed (publishedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
- Keywords
- specification curve analysis, robustness of results, specification bias, name discrimination, research reproducibility
- MeSH terms
- Data Visualization, Adult, Data Interpretation, Statistical, Humans, Models, Theoretical, Research, Research Design, Models, Statistical
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
