Inverse Probability Weighting: From Survey Sampling to Evidence Estimation
Pub. online: 13 April 2026
Type: Methodology Article
Open Access
Open Access
Area: Statistical Methodology
Accepted
3 February 2026
3 February 2026
Published
13 April 2026
13 April 2026
Abstract
We consider the class of inverse probability weight (IPW) estimators, including the popular Horvitz–Thompson and Hájek estimators used routinely in survey sampling, causal inference and for Bayesian computation. We focus on the ‘weak paradoxes’ for these estimators due to two counterexamples by Basu (1988) and Wasserman (2004) and investigate the two natural Bayesian answers to this problem: one based on binning and smoothing: a ‘Bayesian sieve’ and the other based on a conjugate hierarchical model that allows borrowing information via exchangeability. We compare the mean squared errors for the two Bayesian estimators with the IPW estimators for Wasserman’s example via simulation studies on a broad range of parameter configurations. We also prove posterior consistency for the Bayes estimators under missing-completely-at-random assumption and show that it requires fewer assumptions on the inclusion probabilities. We also revisit the connection between the different problems where improved or adaptive IPW estimators will be useful, including survey sampling, evidence estimation strategies such as Conditional Monte Carlo, Riemannian sum, Trapezoidal rules and vertical likelihood, as well as average treatment effect estimation in causal inference.
References
Basu, D. Statistical information and likelihood: a collection of critical essays by Dr. D. Basu. (J. K. Ghosh, ed.) Lecture Notes in Statistics 45. Springer Science & Business Media (1988). MR0953081
Bhadra, A., Datta, J., Polson, N. G. and Willard, B. Lasso meets horseshoe. Statistical Science 34(3) 405–427 (2019). https://doi.org/10.1214/19-STS700. MR4017521
Chauvet, G. A note on the consistency of the narain-horvitz-thompson estimator (2014). arXiv preprint arXiv:1412.2887.
Chen, Q., Elliott, M. R., Haziza, D., Yang, Y., Ghosh, M., Little, R. J., Sedransk, J. and Thompson, M. Approaches to improving survey-weighted estimates. Statistical Science 32(2) 227–248 (2017). https://doi.org/10.1214/17-STS609. MR3648957
Chopin, N. and Robert, C. P. Properties of nested sampling. Biometrika 97(3) 741–755 (2010). https://doi.org/10.1093/biomet/asq021. MR2672495
Datta, J. and Polson, N. G. Quantile importance sampling (2025). arXiv preprint arXiv:2305.03158. https://doi.org/10.1214/25-bjps638. MR5007225
Delevoye, A. and Sävje, F. Consistency of the horvitz–thompson estimator under general sampling and experimental designs. Journal of Statistical Planning and Inference 207. 190–197 (2020). https://doi.org/10.1016/j.jspi.2019.12.002. MR4066130
Diaconis, P. Bayesian numerical analysis. Statistical decision theory and related topics IV 1. 163–175 (1988). MR0927099
Firth, D. On improved estimation for importance sampling. Brazilian Journal of Probability and Statistics 25(3) 437–443 (2011). https://doi.org/10.1214/11-BJPS155. MR2832895
Ghosh, J. K. Weak paradoxes and paradigms. In Statistical Paradigms: Recent Advances and Reconciliations 3–12. World Scientific (2015). https://doi.org/10.1142/9789814343961_0001. MR3308074
Hahn, P. R., Carvalho, C. M., Puelz, D. and He, J. Regularization and confounding in linear regression for treatment effect estimation. Bayesian Analysis 13(1) 163–182 (2018). https://doi.org/10.1214/16-BA1044. MR3737947
Haziza, D. and Beaumont, J.-F. (2017). Construction of weights in surveys: A review. https://doi.org/10.1214/16-STS608. MR3648956
Hesterberg, T. C. Advances in importance sampling. PhD thesis, Stanford University (1988). MR2637036
Horvitz, D. G. and Thompson, D. J. A generalization of sampling without replacement from a finite universe. Journal of the American statistical Association 47(260) 663–685 (1952). MR0053460
Hájek, J. Comment on “an essay on the logical foundations of survey sampling” by basu. (V. P. Godambe and D. A. Sprott, eds.) In Foundations of Statistical Inference 236–242. Holt, Rinehart and Winston (1971). MR0423625
Isaki, C. T. and Fuller, W. A. Survey design under the regression superpopulation model. Journal of the American Statistical Association 77(377) 89–96 (1982). MR0648029
James, W. and Stein, C. Estimation with quadratic loss. In Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1: Contributions to the Theory of Statistics 361–379. University of California Press (1961). MR0133191
Khan, S. and Ugander, J. Adaptive normalization for ipw estimation. Journal of Causal Inference 11(1) 20220019 (2023). https://doi.org/10.1515/jci-2022-0019. MR4545024
Kim, C., Tec, M. and Zigler, C. M. Bayesian nonparametric adjustment of confounding (2022). arXiv preprint arXiv:2203.11798. https://doi.org/10.1111/biom.13833. MR4680719
Kong, A., McCullagh, P., Meng, X.-L., Nicolae, D. and Tan, Z. A theory of statistical models for monte carlo integration. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 65(3) 585–604 (2003). https://doi.org/10.1111/1467-9868.00404. MR1998624
Li, L. Are bayesian inferences weak for wasserman’s example? Communications in Statistics—Simulation and Computationő 39(3) 655–667 (2010). https://doi.org/10.1080/03610910903576540. MR2784548
Linero, A. R. In nonparametric and high-dimensional models, bayesian ignorability is an informative prior. Journal of the American Statistical Association 119(548) 2785–2798 (2024). https://doi.org/10.1080/01621459.2023.2278202. MR4833915
Little, R. J. Weighting and prediction in sample surveys. Calcutta Statistical Association Bulletin 60(3–4) 147–167 (2008). https://doi.org/10.1177/0008068320080301. MR2553424
Madrid-Padilla, O.-H., Polson, N. G. and Scott, J. A deconvolution path for mixtures. Electronic Journal of Statistics 12(1) 1717–1751 (2018). https://doi.org/10.1214/18-ejs1430. MR3806437
Narain, R. On sampling without replacement with varying probabilities. Journal of the Indian Society of Agricultural Statistics 3(2) 169–175 (1951). MR0045354
Owen, A. B. Comment: Unreasonable effectiveness of monte carlo. Statistical Science 34(1) 29–33 (2019). https://doi.org/10.1214/18-STS676. MR3938960
Philippe, A. Processing simulation output by riemann sums. Journal of Statistical Computation and Simulation 59(4) 295–314 (1997). https://doi.org/10.1080/00949659708811863. MR1623186
Philippe, A. and Robert, C. P. Riemann sums for mcmc estimation and convergence monitoring. Statistics and Computing 11(2) 103–115 (2001). https://doi.org/10.1023/A:1008926514119. MR1837130
Polson, N. G. and Scott, J. G. Shrink globally, act locally: Sparse Bayesian regularization and prediction. Bayesian Statistics 9. 501–538 (2010). https://doi.org/10.1093/acprof:oso/9780199694587.003.0017. MR3204017
Polson, N. G. and Scott, J. G. Local shrinkage rules, lévy processes and regularized regression. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 74(2) 287–311 (2012). https://doi.org/10.1111/j.1467-9868.2011.01015.x. MR2899864
Polson, N. G. and Scott, J. G. (2014). Vertical-likelihood monte carlo. arXiv preprint arXiv:1409.3601.
Raftery, A. E., Newton, M. A., Satagopan, J. M. and Krivitsky, P. N. Estimating the integrated likelihood via posterior simulation using the harmonic mean identity (2006). MR2433201
Ramakrishnan, M. An alternative proof of the admissibility of the horvitz-thompson estimator. The Annals of Statistics 1(3) 577–579 (1973). MR0408061
Rao, J. N., Chaudhuri, A., Eltinge, J., Fay, R. E., Ghosh, J., Ghosh, M., Lahiri, P. and Pfeffermann, D. Some current trends in sample survey theory and methods (with discussion). Sankhy: The Indian Journal of Statistics, Series B 1–57 (1999). MR1720726
Ritov, Y., Bickel, P. J., Gamst, A. C. and Kleijn, B. J. K. The bayesian analysis of complex, high-dimensional models: Can it be coda? Statistical Science 29(4) 619–639 (2014). https://doi.org/10.1214/14-STS483. MR3300362
Särndal, C.-E., Swensson, B. and Wretman, J. Model assisted survey sampling. Springer Science & Business Media (2003). https://doi.org/10.1007/978-1-4612-4378-6. MR1140409
Shortreed, S. M. and Ertefaie, A. Outcome-adaptive lasso: variable selection for causal inference. Biometrics 73(4) 1111–1122 (2017). https://doi.org/10.1111/biom.12679. MR3744525
Skilling, J. Nested sampling for general bayesian computation. Bayesian analysis 1(4) 833–859 (2006). https://doi.org/10.1214/06-BA127. MR2282208
Stigler, S. M. The 1988 neyman memorial lecture: a galtonian perspective on shrinkage estimators. Statistical Science 147–155 (1990). MR1054859
Trotter, H. F. and Tukey, H. Conditional monte carlo for normal samples. In Proc. Symp. on Monte Carlo Methods 64–79. John Wiley and Sons (1956). MR0079825
Wang, C., Dominici, F., Parmigiani, G. and Zigler, C. M. Accounting for uncertainty in confounder and effect modifier selection when estimating average causal effects in generalized linear models. Biometrics 71(3) 654–665 (2015). https://doi.org/10.1111/biom.12315. MR3402601
Wang, C., Parmigiani, G. and Dominici, F. Bayesian effect estimation accounting for adjustment uncertainty. Biometrics 68(3) 661–671 (2012). https://doi.org/10.1111/j.1541-0420.2011.01731.x. MR3055168
Wasserman, L. Bayesian inference. In All of Statistics 175–192. Springer (2004). https://doi.org/10.1007/978-0-387-21736-9. MR2055670
Yakowitz, S., Krimmel, J. and Szidarovszky, F. Weighted monte carlo integration. SIAM Journal on Numerical Analysis 15(6) 1289–1300 (1978). https://doi.org/10.1137/0715088. MR0512700
