Classical Methods in Data Analysis

In order to analyze medical research data, it is essential to understand the basic biostatistical applications upon which it is based. This course provides you with this indispensable knowledge, and ensures you are confident in using it.

You will become familiar with interpreting and creating data based upon a whole range of applications that you are likely to come across in your professional life. For instance, we will discuss and utilize location and variability measures, samples and populations, distributions, confidence intervals, hypothesis testing, comparisons between two or more means or proportions (parametric and non-parametric methods), and relationships between two variables (correlation and simple linear regression). The course also includes an extensive discussion of the multiple linear regression model.

By the end of this course, you should be able to:

• Understand the principles and results of statistical analysis techniques including student t-tests, analysis of variance, simple and multiple linear regression analysis, and 1-sample, 2-sample and paired proportion tests

In particular, you should be able to:

• Understand the ‘√n’ law and its consequences for sample size
• Understand the general principles of decision procedures (‘testing’)
• Know in which situations different techniques can be applied and the conditions that should be met to obtain reliable results
• Understand the Kolmogorov Smirnov test and the Fisher test for equality of variances
• Understand the terms ‘explained variance’ and ‘multi-collinearity’
• Understand the principles of model reduction in regression analysis, and the basic principles of logistic regression analysis
• Select the appropriate non-parametric technique to be applied in case of non-normally distributed data
• Understand the results obtained with all the above techniques, and be able to apply all these tests in practice (e.g. to answer a research question) using common statistical packages (such as SPSS or R)