A discussion on yield component analysis is presented in the paper. The discussion regards such issues as inappropriateness of a linear model in application to yield component analysis, correct measurement of yield and its components, sequential and nonsequential development of the components, and a choice of a statistical analysis method to provide a view on a relationship between yield and its components.
Authors:Marcin Kozak, Dariusz Gozdowski, and Zdzisław Wyszyński
A problem of decomposing a response variable into a sum of other variables (so-called additive components) is presented in the paper. We propose how to analyze a relationship in which the response variable is affected by its additive components. The approach and interpretation proposed are based on and quite similar to path analysis. Therefore, the approach provides a commonly known view on inferring from causal systems. Its application is presented for grain yield of two spring barley cultivars, which is studied as an effect of grain yields from plants with different number of spikes.