SSA SOUTH AUSTRALIAN BRANCH MEETING
The South Australian Branch of the Statistical Society would like to invite you to the August meeting of the 2022 program.
17th 24th August (Wednesday) 2022
Presentation Time: 6:00 – 7:00 pm (Adelaide Time)
Physical Venue: 182 Victoria Square, Flinders University, Adelaide, Level 10, Room 10.3.
Virtual Venue: Zoom link:
https://unisa.zoom.us/j/86271481011?pwd=cjR6YWd1Qy9WN2p5Rm9ZLzZJWWsyUT09 (Password: 850322)
To adhere with SA Health Covid Safe guidelines and maximum capacity of available facility, please inform us by email to firstname.lastname@example.org if you are planning to attend the meeting in person.
5.30 pm: Refreshments and networking
6.05 pm: Seminar Presentation
7:30 pm: A dinner will be held for those attending in person. Please RSVP to email@example.com.
Topic: The design and analysis of a two-phase experiments involving human subjects: a case study
Speaker: Chris Brien, Adjunct Associate Professor, UniSA STEM, The University of South Australia; Senior Biostatistician, The Australian Plant Phenomics Facility, University of Adelaide
Two-phase experiments were introduced in 1952 by the Australian statistician George McIntyre. Their use has been most prevalent in agriculture experiments, especially plant breeding experiments. However, there is the potential for their application to be much more widespread. This potential is not being realized because of a lack of awareness of multiphase experiments within the statistical community.
A potted history of two-phase experiments will be given and an explanation of what constitutes a two-phase experiment provided, using a scenario in a sports science context as an introductory example. The use of the anatomy of a design for understanding the confounding in the experiment will be outlined.
The case study involves a pain-rating experiment reported in a 1997 paper by Solomon, Prkachin, & Farewell. The data from a subset of the experiment was analyzed by Farewell and Herberg (2003) and by Jarrett, Farewell and Herzberg (2020) using analyses-of-variance. A further re-analysis using linear mixed models that has been described by Brien (2022) will be outlined. Suggested improvements for the design of future experiments, based on the results of the re-analysis, will also be presented.