Consultancy
The Statistical Consultancy Unit (SCU) was established to provide practical statistical solutions to challenging problems arising in academia, industry, business, health and other organisations.
The aims of the group are:
- To increase the awareness and use of statistical methods.
- To promote correct usage of statistics for decision making.
- To provide probabilistic solutions to academic/industrial/business/health problems.
- To provide training in the use of statistical and related software.
Statistics and Mathematics play an important role in all fields of academic endeavour. Study in all disciplines involves firstly the use of some data collection mechanism and then the use of appropriate methods to analyse and interpret the data. If the data collection mechanism is not properly designed then no method can help you to analyse data and interpret the results correctly. Some common examples of easily overlooked flaws, which render the study outcomes inappropriate, are as follows:
- Sampling bias: Sample is not representative of the target group.
- Inadequate sample size or replications: Sample size is too small to make satisfactory inference about observed differences.
- Confounding: Experimental design makes it impossible to draw separate inferences about study factors.
- Missing essential information: Discarding original data and recording only a few summary statistics, not recording an important variable required for study.
- Misinterpretation of the statistical tools: e.g., interpreting correlation for causation, applying methods for discrete data on continuous data.
The list of such flaws is endless. Hence it is important that one has clear research hypotheses and well-defined data collection and analysis processes before beginning an experiment. A well-designed study can save both time and resources by avoiding common flaws and help one in extracting maximum information.
Proper design of any statistical study requires expertise in the field. A statistician can:
- Help to minimise the amount of experimental effort (and cost) required.
- Help design the study in order to avoid mistakes that may render obtained results inappropriate.
- Help interpret the results of a statistical analysis.
- Provide assistance to clients in order that they may use modern statistical and mathematical software, intelligently.
- Recommend sophisticated methods through which high precision results can be obtained using minimal resources.
- Develop and/or recommend statistical procedures tailored to meet the study’s requirement under special subject scenarios.
- Save time and resources, generally.
Clients are strongly encouraged to contact us at an early stage of their study. A visit to the unit can save considerable time, resources and avoid some unforeseen problems and classical mistakes.
For enquiries, please contact Professor Nicola Armstrong nicola.armstrong@curtin.edu.au
How to prepare
Before meeting a statistician it is advisable to prepare a description of the research project and intended methods of analysis if any, check on the availability of statistical software in your department, clarify the objectives of the study and have some idea about the intended uses of the results.
Clients who do not have sufficient background in statistics, are encouraged to attend training courses run by the Unit throughout the year. For Curtin staff and research students, fees are minimal.
Data
It is advisable that data are stored in some easily accessible electronic format e.g., ASCII code file, CSV or text files, spreadsheets. Sometimes it is better to consult a statistician even before preparing the data files for analysis.
Statistical Advice
We offer consulting services on:
- Design and analysis of experiments and surveys.
- Statistical and probabilistic methods in petroleum engineering and geology.
- Statistical modelling and general data analysis.
- Multivariate statistical analysis.
- Time series analysis and forecasting.
- Study of stochastic processes.
- Financial methods and risk analysis.
- Actuarial methods.
- Methodological research.
- Understanding statistical methods and presenting results in technical papers.
- Help in use of statistical software like SPLUS, SAS, MINITAB, SPSS, R.
The above list is indicative and not exhaustive.
Short Courses
The following is a list of past short courses:
- Basic Statistics for Research Workers
- Statistical Computing
- Applied Multivariate Techniques
- Design and Analysis of Experiments
- Statistics for Petroleum Engineers and Geoscientists.
Customised Software Solutions
It is often the case that client’s requirements are not met by commercially available software. The SCU has expertise in designing customised software to meet the specific requirements of the client. This could take the form of either an Excel add-in or a stand-alone software. The add-in is often a popular choice as the software can be run within a familiar environment and hence training requirements are minimal.
We provide consultation to Curtin researchers on statistical methods, presentation of Statistical methods for technical papers/research grants, formal supervision for masters and PhD students and conduct workshops and training sessions on special statistical methods.
Free initial consultation
The SCU will provide free initial consultation to Curtin researchers on statistical methods used in research. Consultation will be in terms of advice on statistical methods, provision of appropriate references, and sometimes by doing actual analysis. The focus of the consulting will be to promote statistical thinking and correct use of statistical methods. Statistical support requiring significant resources like longer consultations, computing and data management will be done on a fee paying basis or through collaborative arrangement between the Unit and the Department concerned.
Training courses in statistics
The SCU also runs a range of short courses on basic and special statistical methods throughout the year. The courses are interactive with lots examples and hands-on sessions. Special courses of interest can be organised for research groups.
Supervision of research projects
The SCU will provide formal supervision on statistical aspects of masters and PhD projects in different disciplines. Currently SCU is involved in many such projects. This service can be requested via a collaborative arrangement between the SCU and the Department concerned.
Research grants and technical papers
If you are preparing an application for competitive research grant (like ARC, NHMRC) SCU can offer you advise of the statistical aspect of the application. This could be vital for the success of your application.
We can also help you present statistical results in technical papers.
The SCU will provide formal supervision on statistical aspects of masters and PhD projects in different disciplines. Currently SCU is involved in many such projects. This service can be requested via a collaborative arrangement between the SCU and the Department concerned.