Making sense of statistics
Making sense of statistical methods takes time to learn, especially choosin what statistical methods go with your research.
These resources present information that can help you learn to think about and choose the right statistics.
Workshop: statistics for reading and creating research (2018)
This workshop was given for Medical students in 2018, and covers the main ideas in understanding and choosing the right statistics.
Statistics lecture series (2016)
These lectures were given for Medical students in 2016, and they cover many specific ideas involved in various statistical procedures.
Lecture 1: the information you need to choose the right stats
In this lecture I listed the questions you should ask about your research and your data in order to choose what statistical methods you need. In particular, you should know about what the aim of your question is, and about the types of variables involved.
- The right questions about statistics (Prezi slides)
- VIDEO: Med 3rd year stats lecture 1 2016 - the information you need about stats (YouTube)
- VIDEO: Med 3rd year stats lecture 1 2016 - the information you need about stats (Echo360)
- The right questions about statistics handout (PDF)
- Question types activity handout
- Variable types activity handout
Lecture 2: how to think about your data using graphs
In this lecture I discussed how to think about data, including drawing several different kinds of graphs. Thinking about data in this way will help you understand how statisticians begin to make decisions about what stats they can do.
Lecture 3: how p-values and confidence intervals work
In this lecture I talked about how hypothesis tests and confidence intervals work, and how they help you answer yes-or-no and estimation questions. The aim is to help you have power over statistics by knowing more about it.
- VIDEO: Med 3rd year stats lecture 3 2016: p-values and confidence intervals (YouTube)
- VIDEO: Med 3rd year stats lecture 3 2016: p-values and confidence intervals (Echo360)
- Hypothesis test activity handout
- Confidence intervals activity handout
- How hypothesis testing and confidence intervals work handout (PDF)
Lecture 4: understanding how regression works
In this lecture I discussed how the process of regression works, and how it can help to answer questions about relationships between variables. The aim is to help you have power over statistics by knowing more about it. We also drew together some of what we have learned by discussing the statistics presented in a research paper (which happens to be about bed bugs).
Lecture 5: how to choose the right stats for your research
In this lecture I talked about how to choose what statistical procedures go with your research, based on the type of question, the type of variables, and the way the data is organised. We particularly focus on choosing between hypothesis tests such as the t-test, chi-squared test, ANOVA and regression.
Lecture 6: how to understand the t-test,ANOVA, chi-squared test
In this lecture I talked about how the most common hypothesis tests work, and how logistic regression works, giving more details so that you can understand what they are trying to acheive. Note the videos here are from the 2015 lectures.
- VIDEO: Med 3rd year stats lecture 6 2016 - specific stats procedures - part 1 (YouTube)
- VIDEO: Med 3rd year stats lecture 6 2016 - specific stats procedures - part 1 (Echo360)
- VIDEO: Med 3rd year stats lecture 6 2016 - specific stats procedures - part 2 (YouTube)
- VIDEO: Med 3rd year stats lecture 6 2016 - specific stats procedures - part 2 (Echo360)
- Online ANOVA graph activity
- Statistical tests handout (PDF)
Lecture 7: how to calculate sample size
In this lecture I talked about the information you need in order to calculate sample sizes, as well as how to calculate them in the most common situations of the t-test and the chi-square test.
Lecture 8: requests for specific stats
In this lecture I discussed specific stats procedures that were requested via email. I discussed. I discussed the Mann-Whitney U-test/Wilcoxon rank sum test, other nonparametric tests (such as the Kruskall-Wallis test, Friedman's test and Spearman's correlation), Fisher's exact test, confidence intervals for proportions and the difference between odds, risk and hazard ratios. In the second requests lecture, I discussed survival analysis (including Kaplan-Meier, log-rank and cox proportional hazards regression) and also "adjusting for covariates".
Lecture 9: how to write about statistics
In this lecture I talked about how to write about statistics in a research proposal.
David has written several questions, which are similar in style to the hypothesis test exercises in Lecture 4: Choosing the Right Stats. Below you will find a document containing similar questions to those David has put in past exams for medical students.