Virtual StatsPD@Waite meeting

Every month, the professional development meetings of statisticians and data scientists at Waite, known as StatsPD@Waite, bring together specialists in various aspects of data sciences in agriculture from Waite, Roseworthy and Adelaide.

The next StatsPD@Waite meeting will take place on 14th July where Richard Jarrett will present on his paper about to be published.

Email Beata Sznajder for details of the Zoom meeting.

Getting your error variances right: a salutary lesson

Richard Jarrett

This talk is based on a paper “Random effects models for complex designs” which is about to be published in “Statistical Methods in Medical Research”. The paper reanalyses a “plaid design” (a bit like a strip-plot design) previously analysed by my co-authors.  I recognised this as a two-phase design and proceeded to reanalyse it on that basis.  We first identified that one of the main effects needed to be tested against a different variance and ultimately realised that there were a number of additional variance components that needed to be added to the model.  When this was done, it was found that all the treatment effects in this 2^3 factorial needed careful testing.  I will go through the model and show that it can be analysed in ASReml – so long as you can properly identify the terms needed for the model!