TY - JOUR
T1 - Model checking for general linear regression with nonignorable missing response
AU - Guo, Xu
AU - Song, Lianlian
AU - Fang, Yun
AU - Zhu, Lixing
N1 - Funding Information:
The research described herein is supported by the National Natural Science Foundation of China NSFC (11601227, 11701034, 11671042, and 71602089), the National Social Science Fund of China (NSSFC-16BTJ013 and NSSFC-16ZDA010), Shanghai Normal University Research Fund (KF201905), Sichuan project of Science and technology (2017JY0273), the Natural Science Foundation of Jiangsu Province (BK20160785), and a grant from the University Grants Council of Hong Kong.
PY - 2019/10
Y1 - 2019/10
N2 - Model checking for the general linear regression model with nonignorable missing response is studied. Based on an exponential tilting model, two estimators are proposed for the unknown parameter in the regression model. Then, two empirical-process-based tests are constructed. The asymptotic properties of the proposed tests are investigated under the null and local alternative hypotheses in different scenarios. It is found that the two tests perform identically in the asymptotic sense. In addition, a nonparametric Monte Carlo test procedure is performed to obtain the critical values. Further, simulation studies are conducted to assess the performance of the proposed tests and compare them with other possible approaches. Finally, a real data set is analyzed for illustration.
AB - Model checking for the general linear regression model with nonignorable missing response is studied. Based on an exponential tilting model, two estimators are proposed for the unknown parameter in the regression model. Then, two empirical-process-based tests are constructed. The asymptotic properties of the proposed tests are investigated under the null and local alternative hypotheses in different scenarios. It is found that the two tests perform identically in the asymptotic sense. In addition, a nonparametric Monte Carlo test procedure is performed to obtain the critical values. Further, simulation studies are conducted to assess the performance of the proposed tests and compare them with other possible approaches. Finally, a real data set is analyzed for illustration.
KW - Inverse probability weighting
KW - Model checking
KW - Nonignorable missing response
UR - https://www.scopus.com/pages/publications/85063759783
U2 - 10.1016/j.csda.2019.03.009
DO - 10.1016/j.csda.2019.03.009
M3 - Journal article
AN - SCOPUS:85063759783
SN - 0167-9473
VL - 138
SP - 1
EP - 12
JO - Computational Statistics and Data Analysis
JF - Computational Statistics and Data Analysis
ER -