Sequential path analysis: what does "sequential" mean?

• Autor(a) principal: Kozak,Marcin
• Data de Publicação: 2014
• Outros Autores: Azevedo,Ricardo Antunes
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Scientia Agrícola (Online)
• Texto Completo: http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0103-90162014000600012
• topic: interpretation; relationships; models; statistics

Studying relationships among plant and crop traits is crucial for crop scientists to understand complex biological systems that occur in plants and the field. Such knowledge constitutes the basis for more practical information on how to manage breeding and production to provide better or more suitable cultivars, higher yields, lower yield gaps, and resistance to pests etc. To acquire such knowledge, however, representative models of associations between plant and crop traits must be constructed. In path analysis - one of the major methods for analyzing multivariate relationships between quantitative traits - it is important to decide on an appropriate model for these associations, a model that is representative of the corresponding biological phenomena that are of interest to crop researchers. Adopting this "point of view", we asked various questions relating to such model building: (i) how should sequentiality in sequential path analysis be understood? (ii) how should it be interpreted? (iii) how should such sequential models be formulated? We discussed these issues in the context of crop science. Differences in simple and complex (sequential) models of path analysis are presented. Based on crop science examples, we show how important it is to correctly represent the biological relationships for a path analysis model.