OxfordAQA International A-level Psychology (9685): 3.3.3 Research methods 2 – Practice Questions

These are original questions written for Marlbridge, for revision and practice on this content. They are not reproduced past-paper questions, and they do not replicate the exam’s exact structure, question count or mark tariffs – examination boards hold copyright in their own papers. Use these alongside the official past papers from your board or school.

These questions cover section 3.3.3 Research methods 2 in the OxfordAQA International AS and A-level Psychology (9685) specification, Version 3.1 (International A-level exams May/June 2020 onwards). The specification assesses this content in Section C of Unit 3: Advanced topics and research methods 2, a 1 hour 30 minute written exam of 90 marks, with 30 marks for Research methods 2. You will not be asked to calculate a statistical test, so the test results below are given for you to interpret. All people, places and data are fictional.

Revise first with the Research methods 2 study guide and the revision notes for Research methods 2. See the 9685 course hub and the 9685 checklist.

Questions

1. Explain what is meant by a Type II error. [2]

2. At a fictional play centre, Elmstead, a researcher records three things about each child:

• (i) whether the child shares a toy (yes or no);
• (ii) the child’s finishing position in a puzzle race (1st to 8th);
• (iii) the time in seconds the child takes to finish the puzzle.

Identify the level of measurement of each. [3]

3. A researcher codes 180 anonymous posts from one term on the wellbeing forum of a fictional school, Ashgrove Academy.

Coding category	Number of posts
Exam worry	66
Sleep problems	39
Friendships	51
Family	24

(a) Calculate the percentage of posts coded as exam worry. Give your answer to one decimal place and show your working. [2] (b) A second coder independently codes a random 30 of the posts. The two coders agree on 27. Explain how this checks reliability and what the result suggests. [3] (c) Explain one limitation of using content analysis in this study. [2]

4. Distinguish between predictive validity and ecological validity. [4]

5. In a fictional town, Harwick, 15 shoppers visit a shop playing calm music and a different 15 visit the same shop with no music. Each rates their satisfaction from 1 (very poor) to 7 (excellent). The researcher wants to know if satisfaction differs between conditions.

Identify the statistical test the researcher should use. Justify your choice. [4]

6. Ten students at a fictional college, Coldmere, rate their confidence about giving presentations from 1 to 10, before and after a workshop. The hypothesis is that confidence will be higher after the workshop.

Student	A	B	C	D	E	F	G	H	I	J
Before	4	5	3	6	5	4	7	4	5	6
After	6	5	6	7	8	6	5	7	7	7

A Wilcoxon test gave T = 4.5, with N = 9.

Critical values of T, one-tailed	p ≤ 0.05	p ≤ 0.01
N = 9	8	3

The calculated T must be equal to or less than the critical value to be significant.

(a) Calculate the median rating after the workshop. Show your working. [2] (b) Explain why the Wilcoxon test was suitable. [3] (c) Explain why N is 9, not 10. [1] (d) Using the table, explain whether the result is significant at p ≤ 0.05 and at p ≤ 0.01. [3]

7. A researcher writes a new 10-item exam stress scale. Twelve students complete it and also complete an established questionnaire that measures exam stress. The correlation between the two sets of scores is r = +0.74. For N = 12, the two-tailed critical value of Pearson’s r at p ≤ 0.05 is 0.576. The calculated value must be equal to or greater than this.

(a) Explain how this procedure assesses concurrent validity. [2] (b) State whether the correlation is significant. Give a reason. [2] (c) Explain how the researcher could check the face validity of the scale. [2]

8. Explain how the researcher in Question 7 could assess the test-retest reliability of the exam stress scale, and what they should do if it is low. [4]

9. Identify the section of a scientific report in which each statement would appear.

• (i) “Thirty-two volunteers were tested one at a time in a quiet room using standardised instructions.”
• (ii) “A Mann-Whitney test showed that the difference between the conditions was significant at p ≤ 0.05.”
• (iii) “This fits earlier research, although the small sample limits how far the findings generalise.”
• (iv) A short summary at the start giving the aim, method, main result and conclusion.

[4]

10. A researcher surveys 100 students at a fictional college, Brightwater. The hypothesis is non-directional.

	Uses a revision app	Does not use one
Year 12	33	17
Year 13	22	28

A Chi-squared test gave χ² = 4.89 (df = 1). The two-tailed critical values are 3.84 at p ≤ 0.05 and 6.63 at p ≤ 0.01. The calculated value must be equal to or greater than the critical value.

(a) Explain why Chi-squared was suitable. [2] (b) Interpret the result at both significance levels. [3] (c) Explain why a Type I error is possible in this study. [2]

11. Discuss the use of content analysis in psychological research. Refer to the Ashgrove forum study in Question 3 in your answer. [8]

12. A fictional charity in Harwick wants to know whether a ten-minute daily walk improves mood. Design a study to test this. In your answer, refer to:

• the experimental design and sampling;
• how mood will be measured;
• the statistical test you would use, with justification;
• one way to check reliability or validity.

[12]

Answers

1. Accepting (retaining) the null hypothesis [1] when it is actually false, so a real effect or correlation is missed: a false negative [1]. [2] Examiner insight: the second mark needs the idea that the null was false; “a mistake in the results” scores nothing.

2. (i) Nominal [1]. (ii) Ordinal [1]. (iii) Interval [1]. [3] Examiner insight: finishing positions are ordinal even though they are numbers, because the gaps between places are not equal.

3. (a) 66 ÷ 180 = 0.3667 [1]; × 100 = 36.7% [1]. [2] (b) The two coders coded the same posts separately, using the same categories [1]. Agreement is 27 ÷ 30 × 100 = 90% [1]. This high agreement suggests the categories are clear and the coding is consistent, so inter-observer reliability is good [1]. [3] (c) Posts are read out of context, so a coder may judge a joke about exams as exam worry [1]. The tallies would then overstate how worried Ashgrove students are, which lowers validity [1]. [2] Examiner insight: in (b), show the 27 ÷ 30 working and then say what the figure means; a figure on its own earns one mark.

4. Predictive validity is whether scores on a measure predict a later behaviour or outcome [1], for example an aptitude test score predicting job ratings a year later [1]. Ecological validity is whether findings generalise from the research setting to real-life settings [1], for example whether recall of word lists in a lab reflects remembering in everyday life [1]. [4] Examiner insight: “distinguish” needs both types defined clearly, with each example matched to the right type.

5. Mann-Whitney [1]. The aim is to test for a difference between conditions [1]. The design is independent groups, as different shoppers are in each condition, so the data are unrelated [1]. Ratings from 1 to 7 are ordinal, as the gaps between points are not equal [1]. [4] Examiner insight: the three justification marks come from aim, design and level of measurement, each linked to this study.

6. (a) In order: 5, 5, 6, 6, 6, 7, 7, 7, 7, 8 [1]. The middle two values are 6 and 7, so the median is 6.5 [1]. [2] (b) The aim is a difference between before and after [1]. It is a repeated measures design, as the same students give both ratings, so the data are related [1]. Confidence ratings are ordinal [1]. [3] (c) Student B’s rating did not change, and zero differences are left out, so N is 9 [1]. [1] (d) The hypothesis is directional, so one-tailed values are used [1]. T = 4.5 is less than 8, so the result is significant at p ≤ 0.05 and the null hypothesis is rejected [1]. T = 4.5 is greater than 3, so it is not significant at p ≤ 0.01 [1]. [3] Examiner insight: in (d), state the rule direction for Wilcoxon and quote both numbers; “4.5 is significant” alone earns little.

7. (a) Scores on the new scale are compared with scores on an established measure of exam stress taken at the same time [1]. A strong positive correlation shows the new scale agrees with the established one, which supports its validity [1]. [2] (b) It is significant [1], because 0.74 is greater than the critical value of 0.576 [1]. [2] (c) Ask experts, such as clinical psychologists, to inspect the 10 items [1] and judge whether each item looks as if it measures exam stress, rewriting any that do not [1]. [2] Examiner insight: face validity is judged by inspecting items, so an answer that correlates scores here is describing a different type.

8. Give the scale to the same students on two occasions [1], with a gap of about two weeks, long enough that answers are not remembered but short enough that stress has not changed, for example not across an exam period [1]. Correlate the two sets of scores; a strong positive correlation shows good test-retest reliability [1]. If it is low, identify the items with the least consistent answers and rewrite or remove them, then retest [1]. [4] Examiner insight: many answers stop at “do it twice”; the marks come from the gap, the correlation and the fix.

9. (i) Method [1]. (ii) Results [1]. (iii) Discussion [1]. (iv) Abstract [1]. [4] Examiner insight: the test outcome belongs in the results; comparing it with past research belongs in the discussion.

10. (a) The data are nominal: each student is counted in one category [1]. Year 12 and Year 13 are different students, so the data are unrelated [1]. [2] (b) 4.89 is greater than 3.84, so the result is significant at p ≤ 0.05 and the null hypothesis is rejected [1]. Year 12 students were more likely to use an app: 33 of 50 (66%) against 22 of 50 (44%) [1]. 4.89 is less than 6.63, so it is not significant at p ≤ 0.01 [1]. [3] (c) The null hypothesis was rejected at p ≤ 0.05, so there is up to a 5% probability the association is due to chance [1]. If there is in fact no association, rejecting the null would be a Type I error; the result was not significant at the stricter 0.01 level, so replication would be needed [1]. [2] Examiner insight: (b) asks you to interpret, so add what the significant result means for the two year groups.

11. The real exam marks extended answers with levels of response, not point by point. These are indicative creditworthy points.

Content analysis is the indirect study of behaviour through material people produce, coded into categories and tallied [1]. In the Ashgrove study, posts were coded into four categories, such as exam worry, giving quantitative data [1].

A strength is that the posts already exist, so the data are cheap and quick to collect [1]. They are anonymous and written without a researcher present, so there are fewer demand characteristics than in an interview [1]. Reliability can be checked, as shown by the second coder’s 90% agreement [1].

A limitation is that coding needs judgement, and posts are read without context [1]. Tallies show that 36.7% of posts were about exam worry, but not why students posted [1]. A thematic analysis of the same posts would add depth, so combining both approaches gives a fuller picture than either alone [1]. [8] Examiner insight: top answers use the question’s own data, such as the 90% agreement, rather than general points that could fit any study.

12. The real exam marks extended answers with levels of response, not point by point. These are indicative creditworthy points.

Design and sampling. Use a repeated measures design: each participant has one week with the daily walk and one week without [1]. Counterbalance the order, so half walk in the first week, to control order effects [1]. Recruit 20 adults by random sampling from the charity’s members list, to reduce bias [1]. Gain informed consent and give the right to withdraw [1].

Measuring mood. Each evening, participants rate their mood from 1 (very low) to 10 (very high) [1]. Use the mean rating for each week, so one bad day has less effect [1]. Keep the walk standard: ten minutes at the same time of day [1].

Statistical test. Use the Wilcoxon test [1]. The aim is a difference, the design is repeated measures, so the data are related [1], and ratings are ordinal [1].

Reliability or validity. Check concurrent validity by also giving an established mood questionnaire at the end of each week and correlating the scores [1]. A strong positive correlation would show the daily rating measures mood [1]. [12] Examiner insight: each design choice needs a reason; a list of features with no justification stays in the lower bands.

Where marks are usually lost

• Naming a test without linking aim, design and level of measurement to the scenario.
• Using the “greater than” rule for Wilcoxon, Mann-Whitney or the Sign test.
• Forgetting to drop zero differences when finding N.
• Using two-tailed values for a directional hypothesis.
• Treating rating scales as interval data.
• Describing face validity as a correlation, or concurrent validity without one.
• Saying a significant result “proves” the hypothesis.
• Writing a design answer with no justification for each choice.
• In data questions, not quoting figures from the table.

Next steps

• Check gaps with the revision notes for Research methods 2, the Unit 2 practice questions and the 9685 exam preparation guide.
• Find your weaker topics with our diagnostics.
• Want a teacher to mark your design answers? Book a free trial class.

Official syllabus

OxfordAQA International AS and A-level Psychology (9685) specification, Version 3.1 (International A-level exams May/June 2020 onwards), section 3.3.3 Research methods 2; Unit 3 assessment details in section 2.3.