OxfordAQA International A-level Psychology (9685): 3.3.3 Research methods 2 – Revision Notes

These notes condense the Research methods 2 study guide, which has the full explanations and worked paragraphs.

They 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. It also expects the knowledge from Research methods 1, which you can revise in the revision notes for Unit 2.

Test yourself afterwards with the Research methods 2 practice questions. The 9685 course hub and 9685 checklist show where this section fits.

What the specification asks

• Methods: content analysis; case studies.
• Reliability across all methods; test-retest and inter-observer.
• Validity across all methods: face, concurrent, predictive, ecological; the use of correlation in assessing validity.
• Designing and reporting investigations; report sections: abstract, introduction, method, results, discussion, referencing.
• The purpose of statistical testing; probability and significance; statistical tables and critical values; Type I and Type II errors.
• Levels of measurement: nominal, ordinal, interval.
• Factors in choosing a test: aim, level of measurement, design.
• When to use Spearman’s rho, Pearson’s r, Wilcoxon, Mann-Whitney, related t-test, unrelated t-test, Sign test and Chi-squared.

You will not calculate tests in the exam, but you must comment on their use and interpret results.

Key terms

Term	Definition
Content analysis	Indirect study of behaviour through material people produce, by coding it into categories
Coding unit	What is counted in a content analysis, such as a word, image or post
Thematic analysis	Finding recurring themes in material; produces qualitative data
Case study	In-depth study of one person, group, institution or event, often over time
Reliability	Consistency of a measure or finding
Test-retest	Same test, same people, two occasions; scores correlated
Inter-observer reliability	Two observers record the same behaviour separately; tallies correlated
Validity	Whether a measure measures what it claims, and whether findings generalise
Face validity	The measure looks right on the surface
Concurrent validity	Agreement with an established measure taken at the same time
Predictive validity	Scores predict a later outcome
Ecological validity	Findings generalise to real-life settings
Null hypothesis	States there is no difference or no correlation
Significance	The result is unlikely to be due to chance at the chosen level
Critical value	The value in a table the calculated value is compared with
Type I error	Null rejected when true (false positive)
Type II error	Null accepted when false (false negative)
Nominal	Separate categories; frequencies
Ordinal	Ordered data with unequal gaps, such as ratings
Interval	Equal standard units, such as seconds

Studies at a glance

Who	What	Finding	One evaluation point
Scoville and Milner (1957), HM	Case study of a man whose medial temporal lobes were partly removed on both sides in 1953	Severe amnesia for events after the operation; later testing found immediate memory span normal	Rare, precise case that challenged single-store memory views, but one person with pre-existing epilepsy may not generalise
Harlow, Phineas Gage	Case of a man who survived an iron rod passing through his skull in 1848	Reported personality change after frontal damage	Retrospective, limited records; later historians argue many popular details were exaggerated

The test-choice grid

Aim and design	Nominal	Ordinal	Interval
Difference, unrelated	Chi-squared	Mann-Whitney	Unrelated t-test
Difference, related	Sign test	Wilcoxon	Related t-test
Correlation or association	Chi-squared	Spearman’s rho	Pearson’s r

Matched pairs count as related. Rating scales are ordinal.

Reading a critical value table

1. Find the right table for the test.
2. Choose one-tailed (directional hypothesis) or two-tailed (non-directional).
3. Choose the level, usually p ≤ 0.05.
4. Find N, or df, then read off the critical value.
5. Apply the rule: Mann-Whitney, Wilcoxon and Sign test need calculated ≤ critical; the tests with an “r” in their name need calculated ≥ critical.

Mini example (fictional). Twelve people try a breathing exercise before a public talk. Nine feel calmer, two feel less calm, one feels no change. The person with no change is dropped, so N = 11. S is the less frequent sign, so S = 2. The one-tailed critical value at p ≤ 0.05 for N = 11 is 2. S ≤ 2, so the result is significant. The two-tailed critical value is 1, so with a non-directional hypothesis it would not be significant.

Improving reliability and validity

Problem	Fix
Low inter-observer agreement	Operationalise behavioural or coding categories more clearly; train observers together; pilot the categories
Low test-retest correlation	Rewrite ambiguous items; use closed questions; check the gap between tests is suitable
Low face validity	Ask experts to inspect the items and rewrite those that do not fit
Low concurrent validity	Remove items that do not relate to the established measure, then re-correlate
Low ecological validity	Use a more natural task or setting, or follow up with a field study

Using correlation. Test-retest, inter-observer, concurrent and predictive checks all end in a correlation. A strong positive correlation supports reliability or validity. Many textbooks use +0.80 as a guide for reliability; your teacher will confirm what is expected.

Must-know distinctions

• Reliability vs validity: a bathroom scale that always reads 2 kg heavy is reliable but not valid.
• Concurrent vs predictive validity: same time, established measure vs later outcome.
• Test-retest vs inter-observer: same people twice vs two observers once.
• Type I vs Type II: false positive (too lenient a level) vs false negative (too strict a level).
• One-tailed vs two-tailed: directional vs non-directional hypothesis; changes the critical value.
• Quantitative vs qualitative content analysis: tallies of categories vs themes.
• Abstract vs discussion: short summary at the start vs interpretation, limitations and applications at the end.
• Significant vs important: a significant result is unlikely to be chance; it may still be a small effect.

Mini essay plan: “Discuss the use of case studies in psychology” (8 marks)

1. Define (AO1): in-depth study of one case, several methods, often longitudinal.
2. Example (AO1): HM, 1953 surgery, Scoville and Milner (1957), amnesia for new events.
3. Strength with analysis (AO3): allows study of rare cases that could not be created ethically; HM’s case helped separate short-term from long-term memory.
4. Strength (AO3): rich qualitative and quantitative detail over time.
5. Limitation (AO3): generalisation from one atypical case; HM had epilepsy before surgery.
6. Limitation (AO3): researcher bias and retrospective data; Gage’s records.
7. Judgement: most useful when combined with group studies or brain imaging that test the same idea.

Quick self-test

1. What is a coding unit?
2. How is inter-observer reliability assessed?
3. Name the type of validity checked by comparing a new test with an established one.
4. What does p ≤ 0.05 mean?
5. Define a Type II error.
6. Which test is used for a difference, related design, ordinal data?
7. Which test is used for a correlation with interval data?
8. Name the three factors that affect the choice of test, as the specification lists them.
9. For which three tests must the calculated value be equal to or less than the critical value?
10. In which section of a report would you find the statistical test result?
11. Why are rating scales usually treated as ordinal?
12. Which statistical test would you use for a difference between two unrelated groups with nominal data?

Answers

1. The thing counted, such as a word, image or post.
2. Two observers record the same behaviour separately and their tallies are correlated; a strong positive correlation shows reliability.
3. Concurrent validity.
4. There is a 5% or less probability that the result is due to chance.
5. Accepting the null hypothesis when it is false: a false negative.
6. Wilcoxon.
7. Pearson’s r.
8. Research aim, level of measurement and experimental design.
9. Mann-Whitney, Wilcoxon and the Sign test.
10. Results.
11. The gaps between points are not equal, since one person’s 6 may not equal another’s.
12. Chi-squared.

Where marks are usually lost

• Naming a test with no justification; you need aim, design and level of measurement.
• Saying “significant” without quoting the calculated and critical values and the rule.
• Using the two-tailed column for a directional hypothesis, or the reverse.
• Forgetting to drop zero differences before finding N in a Sign test or Wilcoxon.
• Describing face validity as “the study looks valid” with no idea of who judges it.
• Explaining test-retest reliability without saying the scores are correlated.
• Mixing up Type I and Type II errors, or linking them to the wrong significance level.
• Listing case study limitations with no reference to the case in the question.
• Writing content analysis steps without stating clear coding categories.

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.