Teaching inferential statistics for AS and A Level Psychology

Teaching inferential statistics for AS and A Level Psychology 06 December 2022

Nichola Heath Nicola Heath, Psychology Subject Advisor

In this blog I am going to focus on teaching the criteria for, and use of, inferential statistical tests as this is a topic some find challenging. I will refer to four of our inferential statistics resources that I’ve found particularly useful:

What do students need to know about parametric and non-parametric tests?

AS/A Level Psychology students need to know the following about parametric and non-parametric tests:

the criteria for using a parametric test
the criteria for using a specific non-parametric inferential test (Mann Whitney U test, Wilcoxon Signed Ranks test, Chi-square, Binomial Sign test and Spearman’s Rho)
how to select an appropriate statistical test
understand the application of specific non-parametric inferential tests (Mann Whitney U test, Wilcoxon Signed Ranks test, Chi-square, Binomial Sign test and Spearman’s Rho).

They are required to calculate the non-parametric tests. Although this can seem daunting to some, it isn’t as difficult as it looks. After some practice, students can feel really positive when they get that eureka moment!

Setting the scene

Statistics can appear to students as a very abstract concept, so it is useful to give them a grounding in what probability and significance means in real life. Our probability workbook and answers resource provide an excellent introduction to these topic areas.

I would also introduce normal and skewed distribution curves at this point as this will be one of the criteria used to determine which statistical test to choose. We have a distribution curves workbook that would be an ideal resource to use.

Both workbooks introduce the new topics and provide opportunities for recapping previous mathematical skills. They are ready to use with your students but are also editable if you want to create a more personalised version for use in the classroom.

Teaching the criteria for using parametric and non-parametric tests

Now that your students have the underpinning knowledge and context as to why we use statistical tests and how they can be useful, we can look at them in more detail. Firstly, we need to teach the criteria for using parametric and non-parametric tests.

I recommend using the parametric and non-parametric workbook as page 2 gives a clear outline of the criteria for a parametric test, along with a neat diagram showing how they are separate to non-parametric tests. Page 4 provides a clear visual flowchart for choosing non-parametric tests that includes the set criteria.

In the paper 1 exam, a common multiple choice or short answer question is:

State criteria for using a parametric test.
State the criteria for choosing a particular non-parametric test.
Provide the name of the non-parametric test to use based on the information given in a source.

To help your students remember the criteria you could ask them to create a mnemonic and share these with the class. A good way to test application of this knowledge is to use a set of short scenarios and get students to give the correct statistical test. You can print cards with each test written on, use mini whiteboards or create a multiple-choice quiz. To add an element of competition, why not get students to work in pairs or teams?

Teaching how to calculate the non-parametric tests

Your students should now be familiar with the types of non-parametric test and when to choose them. The next step is to learn how to calculate them. Some students may be lacking in confidence when they start – it’s important to keep a growth mindset when teaching this topic. You could use this as an opportunity to discuss Dweck’s theory or the idea of the Learning Pit. The students who struggle the most will get the biggest boost once they have mastered it!

There are five tests that they need to know how to calculate: Mann Whitney U test, Wilcoxon Signed Ranks test, Chi-square, Binomial Sign test and Spearman’s Rho. Some of these use a formula and if needed, this will be given in the exam.

Three of the tests require the ranking of data (Mann Whitney U, Wilcoxon Signed Ranks and Spearman’s Rho). Although this is relatively straightforward to do, many small mistakes can occur at this stage, so I would advise investing some time on this and giving your students plenty of practice (especially with tied ranks/double values).

The parametric and non-parametric workbook has a clear worked-through example for each non-parametric test plus copies of the critical values tables needed to determine significance. This is a useful guide for students to refer to when they begin completing these calculations on their own. Page 29 provides a nice one-page summary of the steps needed to calculate each test which acts as a good reminder once students are more confident with what to do.

To avoid confusion, you may choose to take each test individually at first and use an “I do, we do, you do” technique:

Go through the example given in the workbook while the class follows along.
Set a new table of data to work on as a class, asking students at each step what to do next and calculating it together.
Set another table of data but this time students complete it themselves.

The data could be invented or from a practical activity they have carried out previously, or something they could do in class at the time. Where it’s possible to use data they collected, it can be helpful for students to see the context and application of statistical tests.

Students may also find it beneficial to write up their own “How to...” guide; a short summary of the steps involved or an annotated example in their own words.

Teaching statistical tests is likely to need some careful differentiation as there will be a range of confidence and ability. Those taking maths qualifications will find this easier than others and it is worth being aware that biology and geography students also cover some statistical tests on those courses.

The use of the workbook examples and summary will be a helpful scaffolding tool for those who need it. Students who complete the task first can have extension tasks – there are some review tasks at the back of the workbook. Depending on your class, you may be able to use your stronger students to support those who are struggling as this will benefit both students.

In conclusion

I hope you have found this blog helpful, and you feel more confident about teaching statistics. The best piece of advice I can give is practice. The individual steps of the calculations are not complicated, but confidence comes with repeated practice. I would expect your students to appear to be working harder than you in these lessons!

If you have any suggestions for teaching this topic or have any questions, please comment below, or get in touch.

Stay connected

If you have any questions, you can email us at psychology@ocr.org.uk, call us on 01223 553998 or tweet us @OCR_Psychology. You can also sign up to receive subject updates and information about resources and support.

About the author

Nicola joined OCR in 2022 as the subject advisor for psychology. Prior to joining OCR, she taught psychology for over 10 years and has had various other responsibilities in that time, including being Head of Year and Subject Leader. Outside of work, Nicola enjoys reading, baking and spending time outdoors.