Dr Simon Diffey


Apex Biometry Pty Ltd


Dr Simon Diffey is a consulting statistician / biometrician who specialises in collaborating with agricultural researchers in companies, state based agricultural departments and universities.

As Director of Apex Biometry he provides an independent, modern and efficient statistical service based on providing clients the latest in statistical design, analysis and reporting technology.

Ensure that the statistical design, analysis and reporting component of your research project is a strength rather than a weakness by collaborating with a specialist.



Successful research begins with a valid statistical design. Designs relevant to your research could be block designs, incomplete block designs, unbalanced designs, partially replicated designs, embedded designs, cross-over designs, multi-phase designs, designs for multi-environment trials and designs which incorporate pedigree information.


Get the most out of your data by analysing it in an appropriate manner using the latest statistical technology. In a plant breeding context an appropriate analysis might need to consider spatial variability within field trials, whether or not there is inter-plot competition, estimating GxE and incorporating pedigree or molecular marker information.


Make important decisions using quality information. Timely information delivery is through your choice of spreadsheets, written reports and password authenticated web tools.


Testimonials from previous and current clients.

Dr Jo Stringer

Sugar Research Australia began working with Apex Biometry in 2018 for the analysis of some research trials. In particular Simon analysed data from stem and top borers from PNG and Indonesia. The data from these trials are counts, have spatial variation within the trial and measurements are taken repeatedly over time. Simon used generalised linear mixed models with ASReml to analyse the data. He wrote a very comprehensive report on the analyses of the trials which enabled the chief investigator of the project to easily meet project milestones.

I have worked with Simon since the early 2000’s and know he has a thorough understanding of experimental design and analysis of agricultural field trials. We would happily engage Simon in future work when the need arises.


I operate a Research Management Consulting company specialising in the co-ordination of large scientific research projects for industry and government clients. A crucial part of the business is the integrity of the results produced, which hinges greatly on the design and analysis of experiments. Simon Diffey is an essential collaborator for the business. I have worked with Simon for more than 10 years. He has improved greatly the quality of experiments and the value of results.

Simon has a special talent for devising alternative ways to achieve the most robust results through creative experimental designs. These designs will frequently minimise the resources needed while achieving the best scientific outcomes.

We have collaborated in the writing of several scientific papers outlining the importance of statisticians in the whole research process and the need for them to be part of the research team from the inception of a project.

Our collaboration has been across research for the animal industries and the analysis of experiments related to the preservation of the unique 50,000-year-old Aboriginal petroglyphs on the Burrup Peninsula in northwest Western Australia.

Craig Choice

Recently we had our guest speaker Dr Simon Diffey address our National Sales Team & Business to discuss many facets of product evaluation.

Simon addressed our team and we discussed in depth – not just statistics, but trial design, randomisation, plot size, interpretation, proper trial and experimental conduct, and importantly quality information delivery to have higher levels of engagement and trust around our data.

Demonstration trials was also a key area of interest to us at Pioneer as we do many of these across the production areas we are involved in.

Thank you Simon for a great presentation and a deeper understanding of your services.

Kate Light

Dow AgroSciences Australian Canola Breeding program began working with Apex Biometry in 2016 for the design and analysis of advanced Canola inbred and hybrid trials. It is important for a plant breeding program to have expert people supporting the program in different scientific disciplines. For design and analysis of field trials we chose to work with Simon and never looked back.

Simon worked with our team to design partially replicated yield trials for test hybrids, advanced hybrids and open-pollinated canola across a number of locations in Eastern Australia. These designs enabled the program to trial new test entries that would have been excluded in previous years purely based on seed availability which resulted in improved prediction of combining abilities of parental lines. Whilst accurate analysis of yield results of test entries is imperative, the accurate analysis and reporting of parent line general combining ability (GCA) is critical for a hybrid breeding program. Without well analysed GCA a hybrid program cannot make the most informed test-cross decisions and focus resources on where the best potential gains can be made. There was a clear improvement in breeding program efficiencies simply by working with Simon to ensure field trials were designed to enable analysis of yield performance of trial entries and parental GCA.

As a collaborator Simon was easy to work with due to his thorough understanding of plant breeding stages and how field trialling is conducted. He became an integral member of our small breeding team and made a significant contribution to the rapid success of the program.

Example web tools

Client web tools are developed and deployed using the R package Shiny with password authentication.

Example web tool: hybrid plant breeding program selection tools

A web tool for reporting results from a factor analytic mixed model (incorporating pedigree information) applied to multi-environment trial (MET) data.

Recent Publications

A new REML (parameter expanded) EM algorithm for linear mixed models

In the cases where the AI algorithm failed, a REML PX‐EM based on the new incomplete data specification converged in 28% to 30% fewer …