Nisa, Khoirin and Usman, Mustofa and Warsono, Warsono and Hamsyiah, Nurmaita (2018) ANALISIS VARIANS UNTUK DATA TAK LENGKAP PADA RANCANGAN STRIPPLOT MENGGUNAKAN PENDEKATAN SATTERTHWAITECOCHRAN. In: Seminar Nasional Hasil Hasil Penelitian 2018, 13 November 2018, Universitas Lampung (Bukit Randu Hotel).

Text
Artikel Semnas LPPM 2018 Lengkap.pdf Download (1MB)  Preview 
Abstract
Missing data in strip plot design cause unbalanced design. As a result, several problems occur in analysing the data. This paper discusses the estimation toward missing data and the hypotheses test in strip plot design. The estimation of missing data is done by using Yates approach. SatterthwaiteCochran approximation is used to omit the bias caused by value estimation of missing data. Then, the simulation for power of test is done to compare the power of test resulted from direct tests to the imputed data using Yates approach estimation and the modified test using SatterthwaiteCochran approximation (adjustedtest) toward standard test for complete data. Analytically, the result shows the estimation of missing data using Yates approach give an unbiased mean square error in strip plot design. However, the mean squares of treatment is over biased (positive). Using SatterthwaiteCochran approach, the adjusted mean squares of treatments can be formulated such that the expected mean squares obtained are the same as the expected mean squares of strip plot design from complete data. The simulation study result shows that the adjustedtest using SatterthwaiteCochran approximation give better result than the directtest based on their power of test from 1000 replicated samples.
Item Type:  Conference or Workshop Item (Paper) 

Subjects:  H Social Sciences > HA Statistics 
Depositing User:  DR. KHOIRIN NISA 
Date Deposited:  02 May 2019 04:17 
Last Modified:  02 May 2019 04:17 
URI:  http://repository.lppm.unila.ac.id/id/eprint/11095 
Actions (login required)
View Item 