Component 2: Principles of research (knowledge, understanding and evaluation of) — Eduqas A-Level Psychology
Test yourself on Component 2: Principles of research (knowledge, understanding and evaluation of) with EDUQAS A-Level practice questions.
Start free
7 days Premium · Then the free plan continues · No card, no charge
Your focus
- Social Psychology: Milgram, S. (1963). Behavioural study of Obedience. Journal of Abnormal and Social Psychology, 67, 371-8
- Developmental Psychology: Kohlberg, L. (1968). The child as a moral philosopher. Psychology Today, 2, 25-30.
Component 2: Principles of research (knowledge, understanding and evaluation of) exam tips
Quick Revision Summary (Key Takeaway)
Eduqas A-Level Psychology Component 2 (Principles of Research) assesses student mastery of scientific methodology, research design, descriptive and inferential statistics, and ethical considerations. Mastering this section requires students to evaluate methodologies critically, justify statistical test choices, and apply research principles to novel novel academic scenarios.
Topic Overview
Component 2 Section A focuses comprehensively on the principles of psychological inquiry, scientific methodology, and data analysis. Students must develop an in-depth understanding of how research is constructed, evaluated, and ethically managed across experimental, correlational, observational, and self-report designs.
This foundation is essential not only for achieving top marks in Section A novel scenarios and Component 2 personal investigations, but also for critically appraising key classical and contemporary studies across the entire A-Level specification. High performance requires precise use of technical terminology, statistical competence, and contextual evaluation.
Key Concepts
- →Scientific processes and validity: Internal validity, external validity (ecological, population, temporal), concurrent validity, face validity, and threats like demand characteristics and investigator effects.
- →Experimental and non-experimental designs: Independent groups, repeated measures, matched pairs, naturalistic vs controlled observations, structured/unstructured interviews, and correlational studies.
- →Sampling techniques: Random, stratified, systematic, opportunity, and volunteer sampling, including representativeness, sample bias, and generalisability.
- →Inferential statistics: Parametric tests (unrelated t-test, related t-test, Pearson r) versus non-parametric tests (Mann-Whitney, Wilcoxon, Chi-Square, Spearman rho), levels of measurement, Type I/Type II errors, and significance levels (p <= 0.05).
Examiner Tips
- 💡Always operationalise your variables fully when writing hypotheses or describing procedures: specify exact measurements, units, or scale boundaries rather than general concepts.
- 💡Memorise the decision pathway for inferential statistics: test of difference or correlation, experimental design (related or unrelated), and level of measurement (nominal, ordinal, interval/ratio).
- 💡When asked to improve a study, provide detailed, actionable modifications (e.g., precise counterbalancing sequence or double-blind protocol) rather than vague advice like 'use more participants'.
Common Mistakes
- Conflating reliability with validity: Students often claim a measure lacks reliability when discussing bias or confounding variables, whereas reliability specifically concerns consistency over time (test-retest) or across raters (inter-rater).
- Misidentifying levels of measurement: Treating subjective rating scales (e.g., Likert scales 1-10) as interval data rather than ordinal data, which leads to incorrect inferential test selection.
- Confusing directional (one-tailed) and non-directional (two-tailed) hypotheses: Stating a directional hypothesis without prior research evidence or failing to operationalise both variables precisely within the hypothesis statement.
Revision Plan
- 1Week 1 (Days 1-3): Review all core methodologies (experimental, observational, self-report, correlational) and create comparison tables contrasting strengths, weaknesses, and validity issues.
- 2Week 1 (Days 4-5): Master levels of measurement (nominal, ordinal, interval) and practice the statistical test decision tree until selection criteria can be stated from memory.
- 3Week 2 (Days 1-3): Work through past Eduqas Component 2 Section A novel scenario questions, writing fully contextualised evaluations and operationalised hypotheses.
- 4Week 2 (Days 4-5): Practise descriptive and inferential statistics calculations, significance tables reading, Type I and II error explanations, and peer-review marking with official mark schemes.
Exam Question Types
- 📋Novel Scenario Application: Unseen research summaries requiring identification of IV/DV, hypothesis writing, sampling critiques, and identification of confounding variables.
- 📋Statistical Justification and Interpretation: Selecting inferential tests, reading critical value tables, identifying significance at p <= 0.05, and calculating descriptive statistics.
- 📋Methodological Design / Modification (6-10 marks): Developing or improving an existing investigation, detailing sampling, control measures, ethical compliance, and data collection.
Command Word Expectations (EDUQAS)
Evaluate
Provide balanced commentary showing strengths and limitations of the research method, sampling, or design, explicitly contextualised to the scenario provided.
Justify
Provide clear, technical reasons based on criteria (e.g., design, data level, test type) explaining why a particular decision, test, or methodology was chosen.
Operationalise
State measurable, explicit details showing exactly how variables are manipulated, recorded, or quantified within the specified investigation.
How Students Lose Marks (Examiner Pitfalls)
Pitfall: Generic evaluation of experimental designs without contextual application to the novel scenario provided in the stem.
❌ Weak Answer (Loses Marks):A repeated measures design was used. A weakness is order effects, which means participants get tired or practice makes them better.
Example improved answer:A repeated measures design was used as all 20 participants completed both the memory recall test with background noise and without background noise. A major limitation is order effects; participants may have performed better in condition two purely due to practice on the word lists, confounding the effect of background noise. To control this, the researcher should have counterbalanced conditions using an ABBA design.
Examiner Tip: Always link methodological evaluation points directly back to the operationalised independent and dependent variables described in the prompt scenario.
Pitfall: Confusing standard deviation with range or treating standard deviation merely as a measure of central tendency.
❌ Weak Answer (Loses Marks):The standard deviation for Condition A is 4.2 and Condition B is 1.1, showing that Condition A did much better on the test.
Example improved answer:The standard deviation measures the dispersion of scores around the mean. The higher standard deviation in Condition A (SD = 4.2) indicates a greater spread of recall scores around the mean compared to Condition B (SD = 1.1), demonstrating that individual performance was far more varied and less consistent when background noise was present.
Examiner Tip: Interpret standard deviation explicitly in terms of dispersion, consistency, and variation around the mean rather than overall performance.
Step-by-Step Worked Solutions
Question: A psychologist investigated whether caffeine affects reaction time. Group 1 consumed caffeine; Group 2 consumed a decaffeinated placebo. Reaction time in seconds was recorded. Identify the most appropriate inferential statistical test and justify your choice using three distinct criteria. [4 marks]
- 1.Step 1: Determine the experimental design. Participants are in two distinct groups (caffeine vs placebo), meaning an independent groups design was utilised (unrelated data).
- 2.Step 2: Identify the level of measurement for the dependent variable. Reaction time is recorded in seconds on a continuous, meaningful scale with a true zero, making it ratio/interval data.
- 3.Step 3: Establish the purpose of the test. The researcher is testing for a difference in reaction times between two conditions, not a correlation.
- 4.Step 4: Check parametric assumptions. If data is assumed normally distributed with homogeneity of variance at interval level, an unrelated t-test is used; if non-parametric assumptions apply, the Mann-Whitney U test is selected. State both clearly based on Eduqas requirements.
Final Answer: The most appropriate parametric test is an Unrelated t-test (or Mann-Whitney U if non-parametric). Justification: 1) The study tests for a difference between conditions; 2) An independent groups design was used (unrelated data); 3) The data collected (reaction time in seconds) is interval/ratio level.
Question: Calculate the percentage increase in memory recall score from Baseline (Mean = 14) to Post-training (Mean = 21). Show your workings. [2 marks]
- 1.Step 1: Calculate the absolute difference between the two values: 21 - 14 = 7.
- 2.Step 2: Divide the difference by the original baseline value: 7 / 14 = 0.5.
- 3.Step 3: Multiply the resulting quotient by 100 to convert to a percentage: 0.5 * 100 = 50%.
Final Answer: Percentage increase = 50%.
Active Recall Memory Test
What are the three criteria required to use a Wilcoxon Signed Ranks test?
Key Fact: Test of difference, repeated measures or matched pairs design (related data), and ordinal (or non-parametric interval) level data.
What constitutes a Type I error and when is it most likely to occur?
Key Fact: A Type I error occurs when the researcher falsely rejects the null hypothesis (false positive), typically when the significance level is set too leniently (e.g., p <= 0.10).
What is the difference between concurrent validity and predictive validity?
Key Fact: Concurrent validity compares a new test with an established valid test administered at the same time; predictive validity assesses how well test scores predict future behaviour or performance.
How does stratified sampling differ from quota sampling?
Key Fact: Stratified sampling selects participants randomly from identified sub-groups (strata), whereas quota sampling selects participants non-randomly (often via opportunity) until the quota is filled.
Frequently Asked Questions
How do I know whether data is nominal, ordinal, or interval?
Nominal data consists of categorical headcounts or tallies (e.g., pass/fail, tallies of red vs blue). Ordinal data is ranked data where intervals between points are subjective and unequal (e.g., 1-10 rating scales or race positions). Interval data uses standardized, fixed objective units of measurement where intervals are equal, such as time in seconds, temperature in degrees Celsius, or weight in kilograms.
Do I have to calculate inferential statistical tests by hand in Eduqas exams?
No, Eduqas does not require full hand calculation of complex inferential tests like the Mann-Whitney U or Unrelated t-test in the exam. However, you are expected to calculate basic descriptive statistics (mean, median, mode, percentages), calculate degrees of freedom, and compare calculated/observed values with critical values in statistical tables to determine significance.
When is a result statistically significant when reading critical value tables?
For tests with the letter 't' or 'r' in their name (t-test, Chi-Square, Pearson r), the calculated/observed value must be equal to or greater than the critical value. For tests without those letters (Mann-Whitney U, Wilcoxon, Sign test), the calculated/observed value must be equal to or less than the critical value at the specified significance level (normally p <= 0.05).
What is the British Psychological Society (BPS) Code of Ethics structure tested by Eduqas?
The BPS Code of Ethics and Conduct is organised around four core ethical principles: Respect (privacy, confidentiality, informed consent), Competence (maintaining professional standards), Responsibility (protection from harm, debriefing), and Integrity (honesty, managing deception). In exam questions, you should explicitly name the core principle and explain how it applies to the scenario.
Why is p <= 0.05 used as the conventional level of significance in psychology?
A significance level of p <= 0.05 strikes the best balance between avoiding a Type I error (optimism bias/false positive) and avoiding a Type II error (false negative). It indicates that there is a 5% or lower probability that the observed results occurred due to chance factors rather than the manipulation of the independent variable.