Empirical likelihood approach for treatment effect in pretest-posttest trial

Abstract

The empirical likelihood approach is suggested to the pretest-posttest trial based on the constrains, which we construct to summarize all the given information. The author obtains a log-empirical likelihood ratio test statistic that has a standard chi-squared limiting distribution. Thus, in making inferences, there is no need to estimate variance explicitly, and inferential procedures are easier to implement. Simulation results show that the approach of this paper is more efficient compared with ANCOVA II due to the sufficient and appropriate use of information.

Keywords

Empirical likelihood / Pretest-posttest trial / Treatment effect

References

[1] Chen S. X.. Empirical likelihood confidence intervals for linear regression coefficients. J. Multivariate Anal., 1994, 49: 24-40

[2] Chen S. X., Hall P.. Smoothed empirical likelihood confidence intervals for quantiles. Ann. Staatist., 1993, 21: 621-637

[3] Loen S., Tsiatis A. A., Davidian M.. Semiparametric estimation of treatment effect in a pretest-protese study. Biometrics, 2003, 59: 1046-1055

[4] Owen A. B.. Empirical likelihood ratio confidence intervals for single functional. Biometrika, 1988, 75: 237-249

[5] Owen A. B.. Empirical likelihood ratio confidence region. Ann. Statist., 1990, 18: 90-120

[6] Owen A. B.. Empirical likelihood for linear model. Ann. Statist., 1991, 19: 1725-1747

[7] Qin J., Lawless J. F.. Empirical likelihood and general estimating equations. Ann. Statist., 1994, 22: 300-325

[8] Thomas D. R., Grunkemeier G. L.. Confidence interval estimation of survival probabilities for censored data. J. Amer. Statist. Assoc., 1975, 70: 865-871

[9] Yang L., Tsiatis A. A.. Efficiency study of estimators for a treatment effect in a pretest-posttest trial. Amer. Statist., 2001, 4: 314-321