The use of a multivariate statistical procedure in analysing the germination process of two bean cultivars, compared with a univariate approach

• Autor(a) principal: Bianconi, André [UNESP]
• Data de Publicação: 2011
• Outros Autores: Govone, José Silvio [UNESP], Manly, Bryan Frederick John, Watts, Michael J.
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://www.iaees.org/publications/journals/piaees/articles/2011-1(2)/1-Bianconi-Abstract.asp; http://hdl.handle.net/11449/140806
• topic: Germination rate; Hotelling’s T2 Synchronisation; T test

Several studies on plant physiology are aimed at describing or assessing seed germination processes under laboratory conditions. With respect to seed germination of crop species, some statistical complexities have been discussed, but they have not been developed much in practice. That is, such discussions are not as common as in other areas of plant biology. Additionally, the current literature that is concerned directly with the application of statistics in seed germination indicates that simple and well-known statistical procedures still merit further consideration. Regarding the use of multivariate statistical methods in agriculture, several field studies have used such procedures as a means of clarifying some underlying ecological principles that govern crop production. Nonetheless, multivariate tests have not been widely employed in germination experiments. Therefore, in the present study a simple multivariate statistical procedure (Hotelling’s T2 statistic) was utilised in order to compare two common bean cultivars, using germination indices as variables. The outcome derived from the multivariate approach was compared with that obtained from the utilisation of the univariate t test. The simultaneous application of both methods (that is, the classical univariate t test and the multivariate T2 test) showed that the outcomes may well depend on the approach utilised.