Univariado model applied data longitudinal
with sugarcane

Edjane Gonçalves de FREITAS [1]

Décio BARBIN 1

Geraldo Veríssimo de Souza BARBOSA [2]

Monalisa Sampaio CARNEIRO [3]

Antônio Ismael BASSINELLO 3

§    ABSTRACT: In the test phase of a program for genetic improvement of sugarcane, are implanted several experiments. The interest of the search is assess the profile of the productive genotypes (clones and varieties) on different occasions (Years). Among researchers of agricultural sciences is frequent use of the split plot designs for analysis of data longitudinal, due to ease of analysis and interpretation of the results. It ignored the existence of some form the relationship between the observations taken at different times, and the analysis is performed considering the univariate model with the split plot designs, which requires a strong restriction on the variance - covariance matrix of the data. The objective of this study is to verify whether there are differences between the results the F test in relation to the subplot (years and interaction genotypes $\times$ years) when the analysis is performed with and without the correction of the numbers of degrees of freedom (if the condition of sphericity not satisfied), highlighting a more safe and correct application of the split plot designs for the analysis of experiments of sugarcane. The results of the analysis showed that the significance the tests for intra-subjects factors was unchanged after the correction of figures the degrees of freedom.

§    KEYWORDS: Repeated measures in time; genetic improvement; sugarcane; univariate model; split plot.

[1] Departamento de Ciências Exatas, Escola Superior de Agricultura “Luiz de Queiroz” – ESALQ, Universidade de São Paulo – USP, CEP: 13418-900, Piracicaba, São Paulo, Brasil. E-mail: efreitas@esalq.usp.br / debarbin@esalq.usp.br

[2] Departamento de Fitotecnia, Universidade Federal de Alagoas – UFAL, Campus Delza Gitaí, BR 104 Norte, Km 85, CEP: 57100-000, Rio Largo, Al, Brasil. E-mail: gvsb@fapeal.br

[3] Departamento de Biotecnologia Vegetal, Universidade Federal de São Carlos – UFSCar, Caixa Postal 173, CEP: 13600-970, Araras, SP, Brasil. E-mail: monalisa@cca.ufscar.br / bassinello@cca.ufscar.br