Teaching Research Methods - Statistics

How to choose statistical tests

Statistics. The most commonly named area of the course when Psychology teachers are asked which part of the A-Level they are most nervous about teaching. When this is combined with how it is often named by students as the part of the course that they are most worried about learning, you've got a recipe for a whole lot of anxiety. As psychologists, we know that some anxiety is good, but too much can have many detrimental effects, including a negative impact on our students' abilities to learn and remember information.

As mentioned previously, I quite like teaching Research Methods, and I include statistics in that. The great part about statistics is that the exam questions are fairly predictable and there are so many examples out there that students are able to practise each of the required skills repeatedly. To help to quell the anxiety, I design lessons that build their confidence through repetition. This may not result in the most creative lessons ever, but my foci on teaching this part of the topic are instilling confidence, and simplifying the content as much as possible so that students look forward to the prospect of 'easy marks' when they see statistics questions on a paper.

To achieve this, I break down the teaching of statistics into three separate parts - choosing statistical tests, interpreting critical values tables, and errors and random rules to be aware of (e.g. how to draw a contingency table for Chi-squared and calculate degrees of freedom from it - just in case!). I then run an entirely separate lesson on the sign test after this so that students already have confidence on the topic before a lesson on calculating a test from beginning to end. Below is a step-by-step guide to the first of these lessons.

A couple of things to note before we get into the strategy:

1. My school uses a bronze, silver, gold medal system for differentiation. This can be used to give students choice (so they can complete either the bronze, silver or gold task) or they can work their way up the medals to earn gold and go from the 'easiest' task to the most 'difficult.' We then use challenge tasks to really extend the thinking of the top end.
2. This strategy is designed for the AQA A-Level specification.

Step 1 - introducing inferential testing

I start with a straightforward introduction of what inferential testing is, the purpose of it and what students have to know for this part of the topic (see the sections listed above). I deliberately stay nice and simple here. There is no mention of p values etc. at this point as it's not needed for students to be able to select an appropriate statistical test.

Step 2 - Levels of measurement recap

For my sequence of Research Methods lessons, the first statistics lesson follows on from a lesson on quantitative data analysis. As such, students already have some familiarity with levels of measurement as they used these to determine when it was appropriate to use different descriptive statistics. So for this step, I bolster confidence through a recap activity. In pairs, students match the names of the different levels of measurement to their definitions (see right) and an example, and then consider which level of measurement is 'best.' The gold task is particularly useful here as it can be referred back to when introducing parametric tests as the most powerful and robust tests in step 6.

Step 3 - Practising the levels of measurement

The entire lesson is built around one double-sided sheet where students are given a number of scenarios that they are tasked with determining the appropriate statistical tests for (see left). I introduce this sheet in step 3 and students practise identifying the level of measurement in each scenario. To help students to do this correctly, and to model the thought process for choosing the correct test, I initially get them to identify the dependent variable in the study (or co-variables in the case of correlations and associations). This is the bronze level task. From there, they can determine how this would be measured and therefore which level of measurement is appropriate (silver task). From there, I get students to justify their decisions using the scenarios because on some mark schemes justification marks specify that they must be contextualised in the example (gold task).


Step 4 - Think-pair-share and whiteboard practice

The stumbling block for many students when choosing the correct statistical test is determining whether the test is one of difference, a correlation or an association. Students complete a think-pair-share on the definitions of these concepts which we then discuss as a class. I also give my students some rules of thumb which potentially oversimplify things a little (see right), but really help the lower ability in distinguishing these ideas. I then display a couple of examples of past exam questions on the board for them to identify the type of test on a whiteboard. When I'm satisfied that students are confident distinguishing these concepts, we proceed to step 5.

Step 5 - Practising the types of test

For this step, students return to the sheet introduced in step 3. The bronze task is to determine whether the test is one of difference, correlation or association, and the silver task is to justify their decision using information from the scenario. For the gold task, I introduce that the final decision for choosing a test is whether the tests of difference have a related or unrelated design. I use two examples from the sheet (one with an unrelated design and one with a related) and get them to predict the difference between a related and unrelated design.

After discussing the answers as a class, students then quickly go through the tests of difference to complete the unrelated or related design column. To reinforce that there are only two decisions (levels of measurement and type of test) for correlations and associations, I get students to cross out the box in the unrelated or related design column for these tests.

Step 6 - Introducing the table of statistical test choices

Now that students are able to make all of the decisions needed to choose statistical tests appropriately, I display the table found in textbooks about when to choose each statistical test for students to note down, and explain the difference between parametric and non-parametric tests. I also introduce a couple of mnemonics for remembering the order of the tests in the table that I have come across: 'Carrots Should Come Mashed With Swede Under Roast Potatoes' and 'Can Simon Cowell Make Winners Sing Under Real Pressure.' With this information, students complete the table introduced in step 3 by deciding on which statistical test should be used in each case and justifying why a parametric or non-parametric test must be used.

Step 7 - Exam question practice

As mentioned earlier, there are a lot of past exam questions on statistics. Consequently, students can practise choosing the appropriate test and justifying their choices repeatedly, and this is what they do in the final part of the lesson. I ensure that they practise contextualising their justifications just to be safe, and introduce the challenge task of doing all of it without their notes.


I have found this method to be very successful in bolstering both mine and my students' confidence on statistics. Give it a go and let me know how it works out.