Comparative Study of Generalized Estimating Equations and Logistic Regressions on Different Sample Sizes and Correlation Levels

Önder H.

COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION, cilt.45, sa.10, ss.3528-3533, 2016 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 45 Sayı: 10
  • Basım Tarihi: 2016
  • Doi Numarası: 10.1080/03610918.2015.1010000
  • Dergi Adı: COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.3528-3533
  • Anahtar Kelimeler: Autocorrelation, Generalized estimating equations, Logistic regressions, 62G05, 62G08, LONGITUDINAL DATA-ANALYSIS, POWER, GEE
  • Ondokuz Mayıs Üniversitesi Adresli: Evet

Özet

In this study, it was aimed to determine accuracy of generalized estimating equations versus logistic regressions on different correlation levels and sample sizes. For this aim, two methods were compared with different sample sizes 10, 25, 50 and 100 and correlation levels 0.0, 0.3, 0.5 and 0.8. Result of this study showed that using generalized estimating equations could be preferred versus logistic regression when the sample size is over than 25 and correlation level is higher than 0.3 on data taken from studies with repeated measurements, but logistic regression could be better when the autocorrelations do not exist.