Research Methods 2

Week 1: Why Statistics?

Aims

To appreciate why statistics has a role in medical research

To use 'commonsense'  to identify some of the areas of medical research where statistical methods are needed

This module introduces some of the basic ideas of statistics.  Ways of summarising data - descriptive statistics - are explained.  Most of the course is concerned with inferential statistics, a more challenging branch of the subject.  It is of the utmost importance, as it allows you to use data from a group of patients you have studied to draw conclusions about a wider group of patients you have not studied.

The course does not assume any prior statistical or mathematical knowledge.  Despite this the first assignment is to read a paper and to write a brief commentary (no more than 1 A4 side), in which you identify those issues/areas where statistical aspects of the research arise.

The paper is by Hajek, Taylor and Mills, 'Brief intervention during hospital admission to help patients to give up smoking after myocardial infarction and bypass surgery: randomised controlled trial', British Medical Journal, 2002, 324, 87-89 (available in print and, in a longer version, at www.bmj.com).

In reading the paper and writing your commentary, you should bear in mind the following.

• It is not expected that you know any statistics - you are being asked to use your commonsense to identify the areas where you think statistical methods and statistical thinking have been used.
• In doing this the parts where numerical methods have been used will be the most obvious areas.  In commenting, try to say something more than that data were analysed.  For example, try to identify what comparisons were being made and why - how did the analyses performed add to the aims of the study?
• Statistics, especially inferential statistics, is about drawing conclusions in the face of uncertainty.  Trying to limit the amount of uncertainty is one device that could be used.  Try to think of areas where no calculations have been made in which statistical thinking has been deployed.
• The aim of this exercise is to demonstrate that this study, which is typical of many in the medical literature, could not be conducted properly without using statistical methods and thinking.
• The exercise is not assessed.