STATISTICS FOR NON-STATISTICIANS COURSE 2018
WESTIR Limited attended a short course over two evenings called ‘Statistics for Non-Statisticians’ in August and September 2018. The course was held at the University of Sydney’s Centre of Continuing Education. The course, which was facilitated by John Le Mesurier, aimed to refresh the skills of participants in statistical terminology and interpreting data through applying tests of statistical significance, interpreting p-values, determining confidence intervals, and considering Bayesian methodology. More information about this specific course can be found here.
Some takeaway points from the two-night course include:
- Statistics are a way of dealing with samples. The aim of statistics is to determine whether the characteristics of a sample comes from an underlying known population.
- We test using the Central Limit Theorem based on means and standard deviations. The Central Limit Theorem states that is you select samples greater than 30 from any population, the sample means will be normally distributed.
- You use a z-test when the population that you are sampling is greater than 120. For populations less than 120, you use a t-test.
- Non-parametric statistical testing is used when there is no underlying assumption that the distribution is normal. The most sensitive testing is parametric testing (which uses the Central Limit Theorem) but non-parametric testing can be used to study phenomenon like patterns. Some examples of non-parametric testing are ANOVA (Analysis of Variance) and chi-square tests.
- Many disciplines are now also recommending researchers shift from the reliance on Null Hypothesis Significance to other techniques such as confidence intervals. Confidence intervals are useful because they quantify the size of effects or differences.
- Bayesian statistics is emerging to challenge classical statistics and explores how the data can reduce uncertainty.
- Correlation and regression can also be used for determining the impact of independent variables on a dependent variable. Please remember that correlation does not equal causation.
The University of Sydney’s Centre for Continuing Education offers a range of courses on a variety of topics for those that are interested. More information can be found at their website: www.cce.sydney.edu.au