Uzun Süreli Çalışmalarda Ma(1) ve Arma(1,1) Otokorelasyonlu Hatalara Sahip Doğrusal Karma Etkiler Modelleri için Örneklem Büyüklüğünün Belirlenmesi

Abstract

The aim of the study is to determine sample sizes for the number of subjects and the number of repeated measures for a Linear Mixed Effects Model (LMM) with random intercept and slope in the analysis of longitudinal data set with first order moving average (MA(1)) or first order autoregressive moving average (ARMA(1,1)) autocorrelated errors. In this study, for this reason, a Monte Carlo (MC) simulation approach, which compute statistical power for LMM with random intercept and slope in the analysis of longitudinal data set with MA(1) or ARMA(1,1) autocorrelated errors, was developed by means of SAS 9.1 statistical program and according to assumptions in simulation studies, for the number of subjects and the number of repeated measures, sample sizes based on statistical power were calculated. At the end of the study, it was concluded that the conditions controlled by researchers, like the structure of LMM with random intercept and slope, the number of subjects, the number of repeated measures, the autocorrelation coefficient and a hyphothesis test on parameter interested in model, affected the statistical power and sample sizes calculated based on statistical power were changing.