How do we solve a problem like choosing statistical tests?
Posted on 1st May 2013 by David Carroll
QMP Medical Statistics Tutorials: Tutorial 7: Choice of test and Confidence Intervals [1]
Description of what is being highlighted:
This statistics tutorial covers a few of the most essential areas of medical statistics: What test is appropriate? And what are confidence intervals and how do you use them? This is important because there are an overwhelming number of statistical tests that can be used but only a small number are appropriate and justifiable.
Who is it for?
I think it’s useful for anyone that is interested statistics but without a strong background in it. The text might be daunting for those new to the field but the glossary [2] is a really useful tool provided.
How long does it take and is it interesting?
The tutorial took me longer than other tutorials; I found it quite text heavy and hard to get through at times. Despite this, I plodded on and found it a really interesting resource, further reinforcing my knowledge of statistics, which can fall short at times.
What do I think?
Generally, I found it useful but it has strengths and weaknesses, like the other tutorials.
The flow charts are excellent; provide a really visual method to understand the complexities of choosing a statistical test. These flowcharts are really useful to download and keep/print out to use whenever you need to. The visual representations of central limit theorem are again a quite nice way to teach statistics.
However like other tutorials, this tutorial tends to be quite text driven, which could be incredibly daunting for the first time reader. Like the other tutorials however, there is a glossary provided.
The links to the additional articles are really useful, especially for the more experienced reader or those whose interest in statistics has been piqued.
The exercise questions are great but they would be much better if they had a text input or a MCQ type answer where appropriate. It’s incredibly tempting to peek at the answers. This being said the answers go into good detail and in themselves a really useful teaching tool.
Overall, it is a good tutorial but it is long and hard to get through.
So, put the kettle on.
References:
[1] Rachel Thompson, Mike Bennett, Ed Loughman. QMP Medical Statistics Tutorial: Choice of test and confidence intervals [Internet]. University of New South Wales, Faculty of Medicine; 2008 [updated 15 March 2013; cited 1 May 2013]. Available from: http://web.med.unsw.edu.au/QMP/QMPTut7_2012/Tut7_Intro.htm
[2] Rachel Thompson, Mike Bennett, Ed Loughman. QMP Medical Statistics Tutorial: Glossary [Internet]. University of New South Wales, Faculty of Medicine; 2008 [updated 15 March 2013; cited 1 May 2013]. Available from: http://web.med.unsw.edu.au/QMP/Glossary/Glossary.htm
No Comments on How do we solve a problem like choosing statistical tests?