Questions to be addressed

This case study involves the use of mixed models of fixed and random effects and addresses a number of questions.

    What is a mixed model?

    How does one use REML to fit a mixed model and how can one interpret the output?

    Finally, having fitted the model how can one deduce whether there are significant random effects of ram and ewe on offspring weaning weight?

  • In addressing these questions the case study first considers the statistical model with just the fixed effects developed in Case Study 3 and compares the outputs obtained by the conventional method of general least squares with that using REML.

  • The case study then dwells at some length on the meaning of a mixed model, how it can incorporate units of observation at different layers and how the data structure framework can be diagrammatically sketched in the form of a ´mixed model tree´.

  • Finally, having fitted the mixed model, the case study describes how to interpret the findings.

Table of content  Back     next