Structural equation modeling (SEM): An alternative for assessing causal relationships in safflower

Abstract

Structural equation modelling (SEM) offers a theoretical basis for developing an understanding of relationship between characters and behaviour. This paper imposes the SEM framework focusing on seed yield and oil yield of safflower. We use a data set of IVT safflower 2010-2011 grown at multi-locations under AICRP program. The model includes one latent variable 100 seed weight (SW) for seed yield (SY), oil content (aC) and oil yield (OY) of safflower and determining ex.ogenous factors were no of capsules (CAP), No. of seeds per capitulum (SEEDNo), final plant stand(FPS), plant height (PHT). The result indicated that the oil yield was significantly correlated with exogenous variables. The predictors of OY explained 93 per cent of its variance and the error variance was only 7%. All the path coefficients leading to OY from the four exogenous variables were significant. A highest regression weight estimate was obtained for the loading from SW to SY, indicating that when SW goes up by 1, SY goes up by 98.18± 4.17 Similarly highest regression weight estimate was obtained for the loading from SW and OC to OY, indicating that when SW and OC goes up by 1, OY goes up by 16.88±. 6.4. The chi-square statistic for the model was non- significant which revealed that the model fits the data well indicated that estimated variances and covariances did not differ from the observed ones.