QCE Psychology - Unit 3 - Memory

Psychological research, data and ethics

Design and evaluate psychological studies, select statistics, interpret uncertainty and significance, and apply ethics to Unit 3 investigations.

Part of the free QCE Psychology notes library for Unit 3: Memory.

Updated 2026-08-08 - 5 min read

Psychology evidence is only as strong as the operational definitions, design, measurement and reasoning that connect a question to a conclusion. This lesson is the practical bridge to the Unit 3 data test and student experiment.

Turn an idea into a testable question

A research question names the population, manipulated or grouping variable, measured variable and relevant conditions. An operational definition states exactly how a construct is manipulated or measured.

For an experiment on context and memory:

• construct: memory performance
• operational measure: number of correct responses on a 20-item test
• independent variable: study and test sound condition
• dependent variable: test score
• controlled variables: material, timing, sound level and instructions

A hypothesis predicts an observable relationship. The null hypothesis states that there is no effect or association in the population under the specified comparison.

Choose the correct research design

| Design | Defining feature | Strength | Key risk | | --- | --- | --- | --- | | Independent groups | different participants in each condition | avoids order effects | participant differences | | Repeated measures | same participants complete conditions | controls many participant variables | order, practice and fatigue | | Matched participants | different participants matched on relevant traits | reduces selected participant differences | imperfect matching | | Correlational | measures association without manipulation | studies naturally varying variables | cannot establish causation | | Quasi-experiment | compares a pre-existing grouping variable | feasible when manipulation is impossible | weaker control of group differences | | Observation | systematically records behaviour | can capture natural behaviour | observer effects and coding reliability |

Random sampling supports representativeness; random allocation supports comparable experimental groups. They solve different problems.

Measurement scales and descriptive statistics

Nominal data are categories, ordinal data have rank order, interval data have equal intervals without a true zero, and ratio data have equal intervals with a meaningful zero.

Use the mean for roughly symmetric quantitative data without extreme outliers; use the median for skewed data or ordinal scores. The range is sensitive to extremes. The interquartile range describes the middle half. Standard deviation describes spread around the mean.

Standard error estimates the precision of a sample mean. Confidence intervals express a plausible range for a population parameter under model assumptions. Overlap of confidence intervals is not a universal significance test.

Correlation and inferential tests

Pearson's (r) describes the direction and strength of a linear association from (-1) to (+1). It does not measure causation and can miss nonlinear relationships.

A paired two-sample (t)-test fits related scores such as repeated measures or matched pairs. An unpaired two-sample (t)-test fits independent groups. Both require attention to assumptions, including suitable quantitative data, independence at the relevant level and adequate sample conditions.

A (p)-value is the probability, assuming the null model and its assumptions, of obtaining results at least as incompatible with the null as those observed. It is not the probability that the null hypothesis is true.

• Type I error: reject a true null hypothesis
• Type II error: fail to reject a false null hypothesis

Worked example: interpret, do not overclaim

Validity, reliability and error

Internal validity asks whether the design supports the claimed causal explanation. External validity asks whether findings generalise across people, settings and tasks. Test validity asks whether a measure captures the intended construct.

Reliability concerns consistency. Test–retest reliability, observer agreement and internal consistency answer different consistency questions.

A confounding variable varies systematically with the independent variable and offers an alternative explanation. An extraneous variable can affect scores but is not necessarily distributed systematically. Standardisation, blinding, counterbalancing and careful allocation reduce threats; they do not make a study perfect.

Ethics are part of validity

Psychological research requires informed consent, voluntary participation, right to withdraw, confidentiality, secure information handling, justified and minimised deception, and timely debriefing. Researchers must manage distress, cultural safety and power imbalance. Experimenter expectations should be controlled where possible.

Ethics are not a final checklist after the method. A participant who feels unable to withdraw may behave differently, undermining both welfare and data validity.

IA1 and IA2 response logic

For a data-test response:

1. calculate or identify the requested feature
2. quote the relevant values and units
3. state the trend, relationship, limitation or uncertainty
4. interpret with the correct psychological concept
5. match the conclusion to the evidence

For a student experiment:

1. justify the modification from background science
2. write a focused research question
3. explain method and risk/ethics controls
4. present processed data clearly
5. analyse patterns and uncertainty
6. conclude against the question
7. evaluate validity and reliability
8. propose a specific improvement or extension

Try it yourself

Common exam traps

• confusing random sampling with random allocation
• describing a correlation as an experiment
• choosing an unpaired test for repeated scores
• saying (p>0.05) proves no effect
• treating a large correlation as causal
• naming “small sample” without explaining its effect on uncertainty or generalisability
• proposing “use more participants” without stating what it improves

Sources

• QCAA Psychology 2025 v1.3 syllabus
• QCAA Psychology IA1 sample assessment instrument
• QCAA Psychology IA2 sample assessment instrument
• American Psychological Association: Research ethics

Finished reading? Practise this topic free

Open Psychology past questions with this Unit 3 topic carried into the question bank, then save your progress for the next review.

Practise this topic free. Free to start. No payment details are required. Exact question coverage depends on the available past-paper syllabus mapping.