Corrigendum

Corrigendum to ``Augmented Inverse Probability Weighting and the Double Robustness Property'' by Kurz CF. Medical Decision Making. February 2022; 42(2):156--167. First published July 2021. doi:10.1177/0272989X211027181

The article was updated and republished after its original publication in OnlineFirst to correct a typo in the AIPW equation on page 4 (now page 159 in the final print version). The control-group augmentation term in the original publication of the article was incorrect.

$$\begin{array}{c}{\hat{C}}_{\mathrm{AIPW}}(X_n)=(\underset{\mathrm{IPW}}{\underbrace{\frac{A_n{Y}_n}{e(X_n)}}}-\underset{\mathrm{Augmentation}}{\underbrace{\frac{A_n-e(X_n)}{e(X_n)}f(1,X_n)}})\hfill \\ \multicolumn{1}{c}{-(\underset{\mathrm{IPW}}{\underbrace{\frac{(1-A_n)Y_n}{1-e(X_n)}}}-\underset{\mathrm{Augmentation}}{\underbrace{\frac{A_n-e(X_n)}{1-e(X_n)}f(0,X_n)}}),}\\ {\hat{\tau }}_{\mathrm{ATE}}^{\mathrm{AIPW}}=\frac{1}{N}\sum _{n=1}^N{\hat{C}}_{\mathrm{AIPW}}(X_n).\hfill \end{array}$$

The weight of the outcome-model prediction has been corrected to the following:

$$\frac{A_n-e(X_n)}{1-e(X_n)}$$

The author has assured the journal that the calculations and the code referenced in the article used the correct specification, and that the incorrect formula in the original article was only a typo and the overall results in the article are correct.