The University of South Australia offers support to research degree students undertaking their Master of Research or PhD, as well as to staff undertaking research studies. Advice and assistance may be provided in the following areas:
For HDR students, please speak with your supervisors first, as they may have a sufficient enough understanding of epidemiology and biostatistics to provide all the support you need. If further assistance is required, please contact the staff below.
If you require additional support beyond the initial consultation, a fee for service or negotiations around authorship should be discussed. Please contact the people below or your supervisors for support.
Professor Esterman has many years of experience working with clinicians from many different disciplines. He is comfortable working with the most senior researchers down to absolute beginners. He can provide support in all of the areas described above. However, epidemiology and biostatistics covers a massive area, and there are some specialist areas such as bioinformatics where further expertise might need to be brought in.
Professor Esterman primarily uses SPSS and Stata as general statistical packages, PASS for sample size calculation, and software such as Medcalc and XLSTAT for more specialised analyses.
E: adrian.esterman@unisa.edu.au
P: (08) 8302 2163
M: 0401 124 613
Skype: profesterman
Please note that Professor Esterman is based in James Cook University, Cairns from June to August. He is happy to provide support by email, phone or Skype over that period.
Dr Gentili can assist students with basic statistical tests such as the students T-test, repeated measures analysis and ANOVA's (one-, two-, three-way and nested).
Dr Gentili has experience in using Stata12 and SPSS.
E: sheridan.gentili@unisa.edu.au
P: (08) 8302 2452
Associate Professor Foster has many years experience working in biomedical research. His statistical knowledge is of a generalist nature.
His main expertise is in mathematical modelling, particularly Non-Linear Mixed Effects modelling in the context of population pharmacokinetic-pharmacodynamic analyses.
E: david.foster@unisa.edu.au
P (08) 8302 2055
Course title |
Dates and locations |
|
learnonline course in Research Methodology and Biostatistics |
TBA |
|
April 20-22, 2021 (Session 1) |
|
|
TBA |
Please email your relevant Academic Unit research team if you would like to advertise a biostatistical or epidemiological course for staff or students on this page:
UniSA Allied Health & Human Performance
UniSA Clinical & Health Sciences
Don't know what a MOOC is? Further information about MOOCs can be found here or browse from a wide list of MOOCs available here. The list below contains links to MOOC providers that offer courses in statistics.
A number of papers have been selected that staff and students may find useful when undertaking statistical analysis.
Foundation statistical knowledge |
|
|
Stages of development of a research project: putting the idea together |
This paper describes the process of developing a research idea into a research proposal, and focuses on key steps in the process that all students should consider. |
Statistical tips and tricks |
|
|
Increasing response rates to postal questionnaires: systematic review |
This paper will provide researchers with some tips and tricks that will enable them to optimise the response rate for postal questionnaires. |
|
Practical and statistical issues in missing data for longitudinal patient reported outcomes |
This paper provides strategies for handling missing data at each stage of research, and compares and contrasts various methods to assist researchers. Focus is placed on missing ‘patient reported outcomes’. This paper would very useful for any researcher using questionnaires as part of a clinical trial. |
|
Twenty statistical errors even YOU can find in biomedical research articles |
The author presents 20 common statistical reporting errors, which is a useful resource for students with limited statistical analysis knowledge. References are provided for those who would like to find out more. |
Statistical methods |
|
|
This paper goes beyond simple linear regression and focuses on the power of Generalized Estimating Equations (GEE) and Mixed Models, which incorporate random effects of repeated measures. This is particularly useful for using in longitudinal studies with a series of individuals. However, an advanced understanding of statistics is required for the reader to work through this paper. |
|
|
Ignoring the matching variables in cohort studies – when is it valid and why? |
This paper focuses on matched case-control and matched cohort studies and discusses the trade-off between bias and variance in deciding whether adjustment for matching factors is advisable. |
|
Reducing bias through directed acyclic graphs
|
This paper demonstrates a six-step process for determining whether a proposed set of covariates would reduce possible sources of bias when assessing the total causal effects of treatment on an outcome. This process can be used to help choose which covariates should be included in traditional statistical approaches in order to minimise the magnitude of bias in the estimate produced. |
|
This paper looks at the best methods for measuring continuous outcomes, e.g. pain, body weight, blood pressure, before and after treatment (or change from baseline) and would be useful to all researchers undertaking these types of studies. |
|
|
Statistical analyses in the physiology of exercise and kinanthropometry |
This paper reviews four statistical techniques that are relevant to researchers and students conducting physiological studies as well as providing a good overview of some of more common statistical methods. This paper is particularly useful as a starting guide for those that are less familiar with appropriate statistical models to use when designing and analysing data. |
|
What is an intracluster correlation coefficient? Crucial concepts for primary care researchers |
This paper discusses the concepts of intracluster correlation coefficient – or the impact of having subjects randomised at a group level but analysed on an individual level. This has important implications to sample size calculations. The paper also contains a useful case study that illustrates the concepts of clusters in a primary care setting. |