- Gazi University Journal of Science
- Vol: 25 Issue: 3
- The Power of Four Tests in Two Way Models with Interaction and No Replication
The Power of Four Tests in Two Way Models with Interaction and No Replication
Authors : Hülya Olmus, Semra Erbas
Pages : 677-688
View : 12 | Download : 4
Publication Date : 2012-07-19
Article Type : Other
Abstract :This article presents tests for interaction in a two way table with one observation per cell. The conventional linear model theory cannot be used to check for interaction when there is exactly one observation per cell (no replication) in two way ANOVA models. For this, two different approaches are presented in the literature. In the first of these approaches is to assume a specific functional form for the interaction terms. In the second approach is not to assume a specific functional form for the interaction terms. In this article, we present four additivity tests by proposed Tusell (1990), Boik (1993a), Piepho (1994), Kharrati-Kopaei and S adooghi-Alvandi (2007) for second approach. We compared their performance by means of simulation studies with respect to the power. This article presents tests for the interaction in a two way table with one observation per cell. The conventional linear model theory cannot be used to check for the interaction when there is exactly one observation per cell (no replication) in two way ANOVA models. For this purpose, two different approaches are presented in the literature. The first of these approaches is to assume a specific functional form for the interaction terms. The second approach is not to assume a specific functional form for the interaction terms. In this article, we present four additivity tests by proposed Tusell (1990), Boik (1993a), Piepho (1994), Kharrati-Kopaei and Sadooghi-Alvandi (2007) for second approach. We compared their performance by means of simulation studies with respect to the power. Key Words : Two-way ANOVA models, fixed effects, interaction, no replication, power.Keywords : Two-way ANOVA models, fixed effects, interaction, no replication, power