Section A: Methods

Paper 3 Section A examines your knowledge of Research Methods. There are 24 marks available in total distributed between one or more short extracts. You must use the extract carefully to contextualise every point – if you don’t contextualise, marks can be lost very easily despite evidence of detailed knowledge.

Here I have provided two question similar to those you might find in section A along with model answers to help you get the hang of the expected style.

  1. Arachnophobia is an extreme fear of spiders. This phobia can be treated using virtual reality (VR). In one study people with arachnophobia were asked to estimate the size of a real 7.5 cm wide tarantula. These estimates were compared with estimates made by a non-phobic control group. All participants also rated their fear on a scale of 1-10 as the spider was brought closer. Later all participants were confronted with four ‘digital spiders’ for five minutes using a virtual reality (VR) headset. This was repeated four times. Two weeks later, participants were shown the real tarantula again and asked to estimate its size and rate their fear level. Not all of the original participants came back for the second stage of the study.

Table 1: Tarantula size estimates before and after VR exposure treatment for participants with and without arachnophobia.

Estimate of tarantula’s size
Participants with arachnophobia. Participants without arachnophobia (control group)
Time 1: Before VR treatment 13.5 cm 8.3 cm
Time 2: After VR treatment 8.5 cm 8.1 cm

(Source: Adapted from Shiban et al. 2016)

(a) Describe one conclusion that can be determined from the figures in Table 1. (2)
Include actual figures from the extract but be sure to also state what can be inferred from them.

Full-mark answer: It could be concluded that arachnophobia is linked to exaggerated perception of spider size as before therapy, the average size estimate was 80% bigger than the real size (13.5 cm compared with 7.5 cm). After therapy, arachnophobics’ estimates were much more accurate, just one centimetre wrong in fact, (8.5 cm compared with 7.5 cm).

(b) Describe one reason why the researcher used a control group in this study. (2)
An accurate reason is given and then thoroughly explained. Sound understanding of the study is displayed through reference to the independent and dependent variables. Full-mark answer  Using a control group allowed the researcher to see whether the estimated spider size and the fear ratings (dependent variables) were different in the participants with arachnophobia and those without arachnophobia (the independent variable). The comparison would be useful because it could explain the higher fear ratings of those with arachnophobia, i.e. they were more scared as they perceived the spiders as bigger.
(c) The researcher wishes to know whether the reduction in the estimate of the spider’s size before and after VR treatment is statistically significant for the participants with arachnophobia. State which statistical test should be used and why. (3)
The candidate concisely states the three necessary conditions for a Wilcoxon test, systematically linking each condition to the extract context. Full-mark answer  The researcher would use a Wilcoxon test because: (1) they are testing for a difference between two conditions (before and after VR treatment), (2) the experimental design is repeated measures as they are looking at the same groups of participants (people with arachnophobia) before and after VR treatment, (3) the level of measurement for estimated spider size is interval (centimetres are standardised units with a true zero) and this can be collapsed to ordinal data.
(d) Explain two possible problems with this study which mean that VR treatment is not as effective as it may seem. (4)
The secret to a good answer is to state the problem and then explain it with specific reference to details of the study.

It is important to use the correct terminology wherever possible to demonstrate your detailed knowledge.

Full-mark answer First, internal validity may be low if participants gave smaller spider size estimates the second time due to demand characteristics. They knew they were expected to feel better after treatment and may have said what they thought the researcher expected to hear, maybe to avoid further VR sessions.

Second, the study arguably lacks generalisability because some participants did not attend stage two. The non-attenders may have found the VR terrifying, so their fear ratings and size estimates would have been much higher. Their lack of attendance means the treatments may have seemed more effective.

2. Sabrina has come across an interesting study showing that the more countries a person has visited, the more likely they were to cheat on a computerised quiz, if given the opportunity to do so.

She decided to investigate this for herself and asked nine of her friends how many countries they had visited and to complete a questionnaire including ten moral dilemmas. Her friends had to select one of four possible answers to describe what they would do in each situation. Sabrina calculated morality scores out of 40 for each friend, the higher the score, the higher their level of morality.

(Source: Adapted from Lu et al. 2017)

scatterplot sabrina.jpeg
(a) State a fully operationalised directional hypothesis for Sabrina’s study. (2)
Full-mark answer There is a negative correlation between the number of countries visited and morality scores based on moral dilemmas (maximum score of 40, where a high score equals greater morality). Both co-variables are operationalised using information from the extract.
(b) Analyse the data from the scattergram, with reference to original study upon which Sabrina based her investigation. (3)
Full-mark answer The scattergram shows a strong-moderate, negative correlation. This supports the original study as the more countries a person had visited, the lower they scored on morality questionnaire (i.e. less moral).

There is one outlier – the person who visited the fewest countries (5) scored 15 on the questionnaire. Removing this score would increase the correlation coefficient proving solid support for Lu et al.

The answer is analytical as it considers not only the trend shown by the majority of data points but also the possible impact of the outlier. Both points are contextualised well and the data used effectively in support.
(c) Measuring morality using a self-report may not be valid. Suggest one way that Sabrina could measure morality that has greater validity. (2)
Full-mark answer She could leave some money on the bar in a pub, making it look like someone had forgotten their change and then observe people’s reactions.

A scoring system could be used to rank the morality of different actions, e.g. one for take the money, two for give it to the bartender and three for put it in a charity donation box.

A sensible proposal is made showing understanding of validity – testing what you mean to test as opposed to what people say they would do. This is elaborated successfully for the second mark.
(d) Sabrina calculated the correlation coefficient using the Spearman’s test and got a value of -0.583. Using the appropriate critical values table, explain whether she can accept her directional hypothesis or not. (2)
Full-mark answer  The critical value for 9 participants (5% level of significance and a one-tailed test) is 0.600

The observed (calculated) value of rho (.583, ignoring the sign) is less than the critical value (0.600). Therefore Sabrina must accept the null hypothesis and conclude that there is not a significant negative correlation between the number of countries visited and morality scores based on moral dilemmas (maximum score of 40, where a high score equals greater morality) (p > 0.05).

The correct decision has been made based on a comparison of the observed and critical values, both of which are quoted. A full explanation of why 0.600 was chosen has been given. The final sentence states the conclusion.
(e) Explain one strength and one weakness of the use of quantitative data in Sabrina’s study. (4)
Full-mark answer One strength is that the questionnaire scores are an objective and reliable measure of morality. This is because participants choose one of four set answers and the researcher is therefore unable to bias the results in any way.

A weakness is that the data for morality may not represent a person’s actual feelings because there are only four possible answers for each moral dilemma. This means that participants may be forced to choose an answer which does not really match with what they would do in that situation meaning their answer is not valid.

Appropriate strengths and weaknesses are selected and well contextualised. Each is well elaborated to match the mark allocation. There is effective but also well explained use of terminology, e.g. objectivity.