Skip to main content
. Author manuscript; available in PMC: 2021 Jul 1.
Published in final edited form as: J R Stat Soc Series B Stat Methodol. 2020 May 13;82(3):719–747. doi: 10.1111/rssb.12372

Figure 4:

Figure 4:

Observed coverage of pointwise 95% confidence intervals using θn and the plug-in method (top row), θn and eht doubly-robust method (second row), the local linear estimator and associated intervals (third row), and the sample splitting estimator (bottom row), considered for different values of a and computed empirically over 1000 simulated datasets of different sizes. Columns indicate whether μn and gn is consistent or not. Black dashed lines indicate the nominal coverage rate.