CAMBRIDGE INTERNATIONAL AS & A LEVEL PSYCHOLOGY · 9990

Research
Methods
Chapter 1 — Memorisation Notes

Every definition, table and evaluation point from the chapter —

condensed into bullet-point flashcards you can actually revise
from. Colour-coded, quick to scan, ready to memorise.
CONTENTS

Chapter 1 · Research Methods

1.1 Experiments

1.2 Self-Reports (Questionnaires & Interviews)

1.3 Case Studies

1.4 Observations

1.5 Correlations

1.6 Longitudinal Studies

1.7 Variables: Definition, Manipulation, Control

1.8 Sampling of Participants

1.9 Data and Data Analysis

1.10Ethical Considerations

1.11Evaluating Research: Methodological Issues

★ Whole-Chapter Summary
 1.1 Experiments

THE 10-SECOND VERSION

Researcher deliberately changes ONE thing (the IV) and measures the effect on another thing (the DV) — everything else is meant to stay the same.
Most scientific/controlled method psychology has → the only method that can show cause and effect.
Can happen in a lab (high control, low realism) or "in the field" (participants' natural setting, high realism, low control).

Aims & Hypotheses

Aim Hypothesis
The purpose of the study — what the researcher wants to find out or the A precise, testable, falsifiable prediction based on the aim — it must be
question they want answered. possible to prove it wrong.

Alternative hypothesis Null hypothesis

Predicts there WILL be a difference between conditions (or a relationship, in a States that any difference found is simply due to chance — there is no real
correlation). effect of the IV.

Directional (one-tailed) Non-directional (two-tailed)

Predicts the direction of the effect (e.g. "X will be higher than Y"). Used when Predicts a difference/effect exists but does NOT say which direction. Used when
previous research already points that way. there's no relevant previous research.

ELABORATE — FOR FULL-MARK EXAM ANSWERS

Every hypothesis must operationalise its variables — turn vague ideas into something precisely measurable (e.g. not "revision apps help" but
"students using the Gojimo app will score higher marks on a 20-item recall test than students using mind maps").
A hypothesis that can never be disproven isn't scientific — Freud's theories about unconscious motives are the classic example, because any outcome
could be claimed to "fit," so they cannot be falsified.
Statistical testing is really asking: "is this result too big to have happened by chance?" If yes → reject the null hypothesis, accept the alternative (result
is "significant"). If no → accept the null hypothesis.
Choosing directional vs non-directional isn't random — you justify it using existing research. Getting this wrong (e.g. writing directional with no
supporting research) loses marks even if the rest of the study is well designed.

IV, DV & Types of Experiment

Independent Variable (IV) Dependent Variable (DV)

What the researcher deliberately changes or manipulates to create different What the researcher measures — expected to change purely because of the IV.
conditions to compare.

Type Strengths Weaknesses

Laboratory experiment Good control of variables → raises validity. Causal relationships Artificial situation → participants may behave unnaturally → lower ecological validity.
Artificial, controlled can be established, since only the IV should affect the DV. Participants often notice the set-up and guess the aim → demand characteristics.
setting; participants Standardised procedures → high reliability and easy to replicate Easier to gain consent, but knowing they're being tested is itself an ethical
usually know they're being exactly. consideration.
tested

Field experiment Natural setting → participants behave naturally → higher Much harder to control extraneous variables → lower reliability and harder to
Participants' everyday, ecological validity. If participants are unaware they're in a study, replicate exactly. Less certain the IV (rather than something else) caused the change
natural environment; IV is demand characteristics are reduced. in the DV. Often impossible to gain informed consent beforehand.
still manipulated by the
researcher

ELABORATE — HOW TO WRITE THIS UP

When asked to "evaluate" an experiment, always link the strength/weakness back to a named concept: validity, reliability, generalisability, or ethics —
don't just say "it's more realistic," say "it has higher ecological validity because…".
A good exam answer explains the trade-off: labs sacrifice realism for control; field experiments sacrifice control for realism. You rarely get both at once.
Remember a "control condition" is simply the baseline level of the IV that the experimental condition gets compared against — useful vocabulary for
describing procedures accurately.

Experimental Designs

Design Definition Strength Weakness

Independent Different participants are used in each condition No order effects (each person only does the task Participant variables — pre-existing differences between
measures (level of the IV) once). Fewer demand characteristics, since each the two groups of people (age, ability, mood) might explain
participant only sees one condition. the result instead of the IV.

Repeated The SAME participants take part in every Individual differences can't bias the comparison, Order effects (practice/fatigue) from doing the task more
measures condition since each person acts as their own baseline. than once. Participants see every condition, so more likely
Needs fewer participants overall. to work out the aim (demand characteristics).
Matched Different participants, but each person in one Reduces participant variables without causing Time-consuming and difficult to find good matches;
pairs condition is matched with someone similar in the order effects, since each person still only does matching is rarely perfect, so unmeasured differences can
other condition (identical twins are ideal) one condition. still exist between pairs.

MUST KNOW: FIXING ORDER EFFECTS (REPEATED MEASURES)

✓ Practice effect — performance improves purely from doing the task before (familiarity/learning), not because of the IV.
✓ Fatigue effect — performance gets worse from doing the task before (tiredness or boredom), not because of the IV.
✓ Randomisation — randomly decide the order each participant does the conditions in, spreading order effects out unpredictably.
✓ Counterbalancing — split the sample in half: an "ABBA" design where half do condition A then B, and half do B then A — this cancels order effects
out systematically (more reliable than simple randomisation).
✓ Random allocation — the independent-measures equivalent fix: randomly assign participants to conditions (e.g. drawing numbers from a hat) so
individual differences are spread evenly across both groups.

Threats to Validity

Confounding variable Participant variable

An uncontrolled variable that acts on ONE level of the IV only, confusing the A type of confounding variable caused by differences between people — age,
results — it can hide or exaggerate the true effect of the IV. gender, personality, intelligence.

Situational variable Demand characteristics

A type of confounding variable caused by the environment — lighting, noise, Clues in the study that give away its aim, causing participants to (often
temperature, time of day. unconsciously) change their natural behaviour.

WORKED EXAMPLE
A researcher compares people working on soft chairs vs hard chairs to see which group works harder. If, by accident, all the "soft chair" participants
happen to also be arts students (who might naturally work differently to science students), "subject studied" becomes a confounding participant variable —
we can no longer be sure it was the chairs, rather than the subject group, that caused any difference in work rate. The fix: use random allocation so both
subject groups are evenly spread across conditions.

Ethics in Experiments

✓ Lab experiments: consent is usually easy to obtain, but deception may be necessary so participants don't guess the aim and create demand
characteristics.
✓ Field experiments: consent is often impossible to gain in advance, since participants may not even know a study is happening — this also makes their
right to withdraw unclear, since you can't withdraw from something you're unaware of.
✓ Privacy is easier to protect in a lab, because the tasks are pre-planned; a field setting carries more risk of accidentally invading someone's personal
space.
✓ Confidentiality matters in both, but is especially at risk in field studies if participants could be identified by, say, their workplace or school.

1.1 RECAP
Hypotheses: alternative (directional / non-directional) vs null — all variables operationalised.
IV = what's changed; DV = what's measured. Lab = high control, low realism. Field = high realism, low control.
3 designs: independent measures (order effects avoided, participant variables risk), repeated measures (opposite trade-off), matched pairs (middle
ground, but hard to arrange).
Confounding variables = participant variables (about people) + situational variables (about environment) — both threaten validity if uncontrolled.
 1.2 Self-Reports

THE 10-SECOND VERSION

Asking people directly to tell you about themselves — their opinions, feelings, or behaviour — rather than watching or testing them.
Two formats: questionnaires (written) and interviews (spoken).

Questionnaires

Closed questions Open questions

Offer a fixed set of possible answers (yes/no, rating scale, Likert scale from Ask participants to answer in their own words, with no fixed options → produce
"strongly agree" to "strongly disagree") → produce quantitative data. qualitative data.

Evaluating closed questions Evaluating open questions

Easy to analyse — you can total scores and calculate averages. But limited in Rich, detailed, valid data — participants aren't forced into a box. But harder to score consistently:
depth: a participant's true feeling might not fit any option offered, which can lower different researchers may interpret the same answer differently, which is a lack of inter-rater
validity. reliability.

MUST KNOW: QUESTIONNAIRE PITFALLS

✓ Low response rate — people can easily ignore a questionnaire, so those who DO reply may share certain traits (e.g. more free time), harming
generalisability.
✓ Social desirability bias — participants answer in a way that makes them look good rather than answering completely honestly.
✓ Filler questions — extra irrelevant questions inserted to disguise the real aim, so participants are less likely to guess it and change their answers.

ELABORATE — EXAM DEPTH

When evaluating a questionnaire, always specify WHICH question type caused the problem — e.g. "because Question 4 was closed, participants may
not have been able to express their true opinion, reducing validity" is far stronger than "questionnaires aren't valid."
A good improvement to suggest: pilot the questionnaire first, or mix open and closed questions so you get reliable, quick-to-analyse data AND some
richer detail.

Interviews

Structured interview Unstructured interview

Same fixed questions in the same fixed order for every participant — even Questions depend on what the participant says previously — flexible, but very
tone/posture may be standardised. Easy to compare between participants. hard to compare answers between participants.

Semi-structured interview
A mix: some fixed questions (so answers CAN be compared) plus the freedom to ask extra follow-up questions specific to that person — generally seen as the best
balance.

MUST KNOW: INTERVIEW EVALUATION

✓ Participants may lie — from social desirability bias, or because they've guessed the aim and want to "help" (or disrupt) the study.
✓ Time-consuming, which may put off certain types of people from volunteering — reducing how representative the sample is.
✓ Subjectivity risk when interpreting answers (the goal is objectivity — an unbiased, external viewpoint, improved by having other researchers help
interpret the data).
✓ Closed Qs in an interview → quantitative data, easier to analyse, generally more reliable. Open Qs → qualitative data, more in-depth/valid but less
reliably scored.

1.2 RECAP
Self-reports = questionnaires (written) or interviews (spoken).
Closed → quantitative, easy to analyse but can miss true feelings. Open → qualitative, rich but harder to score reliably.
3 interview types: structured, unstructured, semi-structured.
Watch for: social desirability bias, response rate, subjectivity/inter-rater reliability.
 1.3 Case Studies

THE 10-SECOND VERSION

ONE person (or family/institution) studied in huge detail, over time.
Combines several techniques + sources (interviews, tests, observations, records, relatives) — like a detective building a complete picture of one
individual.

Case study Triangulation

A research method studying a single instance — usually one person, but Using several different techniques to study the same thing; if they produce
possibly a family or institution — in great depth. similar findings, this suggests the results are valid.

MUST KNOW: WHEN CASE STUDIES ARE USED

✓ Rare cases where a detailed description is valuable (e.g. an unusual brain injury that couldn't ethically be created experimentally).
✓ Tracking developmental change over time, such as a child's progress or a patient's recovery.

Strengths Weaknesses

Very high validity — the person is studied in real depth, in a genuine real-life Subjectivity — researchers often build a close relationship with the participant, which can bias
context. Triangulation (multiple methods) supports validity further. how they interpret the data, lowering validity.

Ethical risk — such personal, detailed questions can feel intrusive; hard to keep the person's
identity confidential — even initials may not be enough if they're well known.

Low reliability — usually only one researcher and one participant, so it's hard to be sure the
interpretation is objective; a different researcher might interpret it differently.

Very low generalisability — findings are specific to that one person and may not apply to
anyone else at all.

ELABORATE — EXAM DEPTH

The strength and weakness of case studies pull in opposite directions: the SAME feature (extreme depth, one case) is what makes them so valid AND
so hard to generalise from — a great point to make explicitly in an evaluation answer.
Famous real case studies (e.g. of amnesia patients) are often cited because ethically you could never experimentally create that condition in a healthy
person — this justifies why a case study, rather than an experiment, was the right method.

1.3 RECAP
Case study = deep dive on ONE case, multiple methods/sources (triangulation).
Very valid, very detailed — but low reliability & generalisability, and real ethical risk.
 1.4 Observations

THE 10-SECOND VERSION

Watching people (or animals) and recording what they do — no questions asked, no tasks set.
4 key design choices to make, each with its own trade-off between validity, reliability, ethics and practicality.

Choice 1 — Setting

Naturalistic observation Controlled observation

Watching participants in their normal, everyday environment, with zero The social or physical environment has been deliberately manipulated by the
interference (social or physical) from the researcher. researcher — can happen in a natural setting OR an artificial one like a lab.

Choice 2 — Structure

Unstructured observation Structured observation

The observer records the WHOLE range of behaviours they see. Usually only The observer records only a limited, pre-decided set of behaviours, called
used at a "pilot" stage, to help decide what specifically matters. behavioural categories.

Behavioural categories
The specific, operationalised actions being recorded — they break a continuous stream of behaviour into distinct, observable events (never an "inferred" mental state
you can't actually see, like "feeling happy").

Choice 3 — Observer Role

Participant observer Non-participant observer

The researcher becomes part of the social group being studied (e.g. joining in The researcher stays separate/apart from the group (e.g. watching through one-
play or conversation). way glass).

Overt observer Covert observer

Participants know the researcher is observing them — role is obvious. Participants do NOT know they're being observed — role is hidden or disguised.

Inter-observer reliability
How consistently two or more observers record the same event — checked by comparing their independent recordings of the same footage/session.

Choice Strength Weakness

Naturalistic Behaviour is true-to-life → high ecological validity No guarantee the target behaviour will even happen

Controlled Ensures the behaviour of interest actually occurs Less natural → may reduce ecological validity

Unstructured Captures unexpected/important behaviours not predicted in advance Hard to record everything accurately → can lower reliability

Structured More reliable — observer only focuses on a small set of categories Might miss behaviours that aren't on the list

Covert Higher validity — no demand characteristics, since participants don't know they're Ethical issue (no informed consent); practically harder to arrange
watched

Overt Easier and more ethical to arrange Participants may change behaviour, knowing they're watched → lowers
validity

ELABORATE — EXAM DEPTH

Observation isn't only a stand-alone method — it's also used as a technique inside other methods (e.g. measuring the DV in an experiment, or as one
of several techniques in a case study). Recognising this cross-over is a common source of exam marks.
When justifying a choice, link it explicitly to the concept it improves: "covert observation increases validity because participants cannot produce demand
characteristics they're unaware of being watched" is a much stronger sentence than "covert is better because it's sneaky."

1.4 RECAP
4 choices: naturalistic/controlled · structured/unstructured · participant/non-participant · overt/covert.
Covert + naturalistic = most valid, but least ethical/practical.
Structured = most reliable (fewer categories to track, clearer definitions).
 1.5 Correlations

THE 10-SECOND VERSION

Looks for a relationship between two variables that are simply MEASURED — nothing is manipulated by the researcher.
Used when manipulating a variable would be unethical or impractical.
Never proves cause and effect — only a properly controlled experiment can do that.

Correlation Co-variables
A research method that looks for a relationship between two measured variables The two measured variables being compared in a correlation.
(co-variables) — a change in one is related to a change in the other.

Causal relationship
When a change in one variable is directly RESPONSIBLE for (causes) a change in another — establishing this always requires an experiment, never a correlation
alone.

MUST KNOW: 3 TYPES OF CORRELATION

✓ Positive correlation — as one variable increases, the other increases too (they rise together).
✓ Negative correlation — as one variable increases, the other decreases.
✓ No correlation — no consistent relationship; points scattered randomly on a graph.

CLASSIC WARNING EXAMPLE

Ice cream sales and murder rates show a positive correlation in some cities — but eating ice cream obviously doesn't cause murder. Both are linked to a
hidden third factor: hot weather (more people outside, tempers shorter, ice cream sells more). This is exactly why correlation ≠ causation, and why exam
answers should never claim a correlation "proves" one thing causes another.

MUST KNOW: EVALUATION

✓ Validity depends on both co-variables being clearly and effectively measured.
✓ Reliability depends on the consistency of the measurements — scientific scales (e.g. time in seconds) are highly reliable; self-report or observation-
based measures tend to be less reliable.
✓ The single most important limitation: you cannot conclude cause and effect — a hidden "third variable" could be responsible for both changes.
✓ Correlations are excellent as a first step, to justify running a full experiment later, or where an experiment isn't ethical/practical at all.

1.5 RECAP
Correlation = relationship between measured co-variables (positive/negative/none).
Big rule to repeat in every answer: correlation ≠ causation — a third variable might explain both.
 1.6 Longitudinal Studies

THE 10-SECOND VERSION

The SAME people are tested repeatedly over a long period — months, years, even decades.
Opposite of a "cross-sectional" study, which compares different age groups all at once.

Longitudinal design Cohort

Same participants tested on 2 or more occasions over a long period of time (e.g. A group of participants selected at the same age/stage of life, then tracked over
before/after a 6-month intervention, or repeatedly across years). time.

Cross-sectional study
Compares people at DIFFERENT ages/stages by testing different groups of people at ONE single point in time — the contrasting method to longitudinal.

REAL EXAMPLE
Terman's "Life Cycle Study of Children with High Ability" began in 1922 and followed highly intelligent children for decades. It found that a high IQ doesn't
guarantee success in life — but did produce famous graduates, including psychologist Lee Cronbach. Participants were nicknamed "Termites," and
knowing they'd been labelled "gifted" may itself have shaped their behaviour and choices.

MUST KNOW: KEY STRENGTH

✓ Because the SAME people are re-tested, researchers can be confident that any change is due to time/development — not due to differences between
separate groups of people, as could happen with a cross-sectional design.
✓ It avoids confounding situational variables that a cross-sectional design might introduce (e.g. one generation experiencing a completely different
education system to another).

MUST KNOW: KEY WEAKNESSES

✓ Sample attrition — the loss of participants over time (withdrawal, boredom with repeated testing, moving away, illness, death) → the sample shrinks
and becomes biased towards stable, healthy, cooperative people.
✓ Being part of a long-running study can make participants feel "special," which may itself change their behaviour — a validity problem.
✓ Reliability risk — measuring tools may need updating over such a long period, and the researchers/staff running the study may change.
✓ Ethical issues — consent must be re-confirmed at every time point (harder with children); keeping large banks of contact details for years raises
confidentiality concerns; harder to judge if a child genuinely wants to withdraw.

1.6 RECAP
Longitudinal = same people, tested repeatedly over time. Cross-sectional = different ages, tested once.
Key strength: rules out cohort/participant differences as an explanation for change.
Key weakness: sample attrition shrinks and biases the sample over time.
 1.7 Variables: Definition, Manipulation & Control

THE 10-SECOND VERSION

This section is all about being PRECISE: define variables so they're actually measurable, and stop unwanted variables sneaking in and ruining the
result.

Operationalisation
Turning a vague idea into something clearly defined and measurable (e.g. "young" becomes "under 20 years old"). Applies to the IV/DV in experiments, the co-
variables in correlations, and the behavioural categories in observations.

Confounding variable Participant variable

An uncontrolled variable that acts systematically on ONE level of the IV only, so
it can hide or exaggerate the true effect of the IV on the DV. A confounding variable caused by differences BETWEEN people — their natural
ability, personality, or mood.

Situational variable Pilot study

A confounding variable caused by an aspect of the environment — lighting, A small-scale trial run of the procedures BEFORE the real study, used to spot
noise, temperature. and fix problems (like uncontrolled variables) early.

Standardisation
Making sure every participant has exactly the same experience — same
instructions, procedure, equipment — no matter which condition they're in.

EXAMPLE: OPERATIONALISING
"Hard vs soft chairs" isn't precise enough → better: "chairs with wooden/plastic seats" vs "chairs with padded seats." "Students working better" isn't
precise → better: "number of homework pieces handed in on time" or "minutes spent on extra work."

ELABORATE — EXAM DEPTH

Controls exist to raise validity: they make sure the different levels of the IV really do represent what they're supposed to, so any change in the DV can
be confidently attributed to the IV.
Standardisation exists to raise reliability: it makes the study replicable, because a different researcher following the same standardised instructions
should get a similar procedure and similar results.
A strong exam answer names the SPECIFIC confound (e.g. "time of day" as a situational variable) rather than vaguely saying "other variables might
affect it."

1.7 RECAP
Operationalise everything: IV, DV, co-variables, behavioural categories.
Confounds = participant variables (about people) + situational variables (about environment).
Pilot study catches problems early; standardisation keeps every participant's experience identical, raising reliability.
 1.8 Sampling of Participants

THE 10-SECOND VERSION

You can't test every single person, so you pick a smaller group (a sample) to represent the population.
HOW you pick that group changes how representative — and therefore how generalisable — your findings will be.

Population Sample
The whole group of people who share a certain characteristic, from which a The smaller group actually selected to take part, ideally representative of the
sample is drawn. population.

Technique Definition Strength Weakness

Opportunity Choosing whoever is conveniently available Quick and easy — larger samples can be gathered fast Likely unrepresentative — available people
sampling tend to be alike (e.g. all similar
age/background)

Volunteer An advert/announcement is put out; those who Easy to arrange; volunteers tend to be committed (e.g. willing Volunteers often share traits (e.g. more free
sampling (self- respond become the sample to return for retesting); good for finding rare/unusual time, more curious) → unrepresentative
selected) participants

Random Every member of the population has an EQUAL Most likely to be genuinely representative Time-consuming to arrange properly; still
sampling chance of selection (e.g. numbers drawn from a biased if the original population list is
hat) incomplete

ELABORATE — EXAM DEPTH

Generalisability — how widely a study's findings apply beyond the sample studied — depends directly on how representative the sampling technique
was, and on sample size.
Smaller samples are less likely to capture the full range of variation that exists in the population, making findings less reliable and generalisable —
always worth mentioning sample size alongside sampling technique in an evaluation.
When asked to suggest an improvement, naming WHICH technique to switch to (and why it fixes the specific weakness) scores more marks than just
saying "use a bigger sample."

1.8 RECAP
3 techniques: opportunity (convenient), volunteer (self-selected), random (equal chance for everyone).
Random sampling tends to be most representative but is hardest to arrange in practice.
 1.9 Data and Data Analysis

THE 10-SECOND VERSION

Raw data (a big pile of numbers/descriptions) is hard to interpret on its own — summarise it using averages, spread, and graphs so the pattern
becomes clear.

Types of Data

Quantitative (numbers about amount) Qualitative (descriptions about quality)

Strengths Usually objective; reliable scales; unusual but important responses aren't Often more valid — participants can express themselves fully instead of being
hidden by averaging; easy to compare using averages/spread. squeezed into fixed choices.

Weaknesses Collection method may limit what participants can express, lowering validity if More subjective — recording/interpretation may be biased by the researcher's own
their true view doesn't fit the options given. views; results from a few individuals may not generalise widely.

Averages (central tendency)

Measure What it is Used with Pros / Cons

Mode The most frequently occurring score (can be more than one if tied) Any data, including categories Only average usable with categories, but ignores the actual
(e.g. favourite subject) values of scores → least informative

Median The middle value once scores are ranked smallest → largest (average Numerical/linear scale data only Unaffected by extreme outliers, but ignores the exact
the middle two if an even number of scores) values of most scores

Mean Sum of all scores ÷ number of scores Numerical/linear scale data only Most informative (uses every value), but CAN be
distorted/skewed by one or two extreme outliers

Spread (dispersion)

Range Standard deviation

(highest score − lowest score) + 1. The "+1" accounts for the real limits of the The average distance between each individual score and the mean. A bigger SD
measurement scale. Quick to calculate, but easily distorted by one extreme = more spread/variation; a smaller SD = scores clustered close to the mean.
outlier. Uses every score, so it's more accurate and less distorted than the range.

Graphs

Graph Used for Key feature

Bar chart Data in separate/discrete categories (e.g. totals, means or modes for different Bars have GAPS between them — the categories aren't part of one continuous
groups) scale

Histogram Continuous data (e.g. a whole distribution of test scores) Bars TOUCH — the x-axis is one continuous scale (DV on x-axis, frequency on
y-axis)

Scatter Correlational data (two co-variables) Each dot = one participant's score on BOTH variables; a line of best fit may be
graph added

MUST KNOW: R VALUE

✓ The r value ranges from −1 to +1 and describes the strength of a correlation. Closer to +1 or −1 = a stronger correlation (points close to the line of best
fit). Closer to 0 = a weaker or non-existent correlation.
✓ A scatter graph can show that a relationship EXISTS — it can never show that one variable is causing the change in the other.

ELABORATE — EXAM DEPTH

When choosing an average, always justify it by data type: "the mean was used because the data was on a numerical scale and there were no extreme
outliers" is a complete, exam-ready sentence.
Standard deviation is nearly always the stronger choice to discuss over range in an evaluation, because it uses every score rather than just the two
most extreme ones.

1.9 RECAP
Data: quantitative (numbers) vs qualitative (descriptions).
Averages: mode (most common), median (middle), mean (sum ÷ count).
Spread: range (quick, crude) vs standard deviation (uses every score, more accurate).
Graphs: bar chart (categories, gaps) · histogram (continuous, no gaps) · scatter graph (correlations).
 1.10 Ethical Considerations

THE 10-SECOND VERSION

Rules that protect participant welfare AND psychology's public reputation and trustworthiness.
Human guidelines (BPS Code of Ethics and Conduct, 2018) + separate animal guidelines.

Human Participants

Informed consent Presumptive consent

Participants are given enough information to decide, freely and while fully Used when true informed consent isn't possible (e.g. field experiments) — a
understanding, whether to take part. similar group is asked in advance whether they'd find it acceptable.

Right to withdraw Protection from harm

Participants can leave (and have data removed) at any time; rewards can't be No greater physical/psychological risk than everyday life; screen participants;
taken back; no pressure from researchers to stay. stop the study if unexpected risks appear.

Deception Confidentiality
Avoid where possible; if truly necessary, plan to minimise distress and debrief Data stored securely and never released; names replaced with numbers/initials;
fully afterwards. institutions anonymised too.

Privacy Debriefing
Never invade a participant's physical/emotional space; they can decline to
answer; only observe where they'd expect to be seen anyway. Full explanation given after the study so participants leave in at least as positive
a state as when they arrived.

MUST KNOW
✓ Extra care is needed with children, people with mental health conditions/learning difficulties, non-native speakers, or groups who might feel pressured
(e.g. prisoners) — their capacity to give truly informed consent may be limited.
✓ Debriefing does NOT replace designing an ethical study in the first place — it can reduce harm afterwards, but it can't undo real damage that's already
been done.

Animal Research

Bateson's cube (1986)

A decision-making model for whether animal research is justified, weighing: how CERTAIN the benefit is, how HIGH the QUALITY of the research is, and how much
SUFFERING the animals experience. Research is justified when certainty and quality are high, and suffering is low.

MUST KNOW: WELFARE PRINCIPLES

✓ Replacement — consider alternatives before using live animals (e.g. videos of animals in the wild, computer simulations).
✓ Species & numbers — use the species least likely to suffer, and only the minimum number needed for valid/reliable results.
✓ Procedures & housing — minimise pain, suffering and distress; respect each species' natural social behaviour (avoid isolating social animals, avoid
overcrowding); provide enough space, food and water.
✓ Reward, deprivation & aversive stimuli — consider normal feeding patterns; prefer a rewarding alternative (e.g. a favourite food) over depriving
animals or using unpleasant stimuli.

ELABORATE — EXAM DEPTH

Animals are used in psychology because they're convenient models for studying processes (like learning), allow procedures that couldn't ethically be
done on humans (e.g. brain lesion studies), or are interesting in their own right.
When evaluating an ethical issue in an exam, structure your point as: (1) name the guideline, (2) explain specifically how the study did/didn't meet it, (3)
say what effect this has on participants or the research. This 3-step structure is what separates a top-band answer from a vague one.

1.10 RECAP
Humans: informed consent, right to withdraw, protection from harm, avoiding deception, confidentiality, privacy, debriefing.
Animals: Bateson's cube (benefit + quality vs suffering); replacement, appropriate species/numbers, humane procedures.
 1.11 Evaluating Research: Methodological Issues

THE 10-SECOND VERSION

Pulls together ideas from the whole chapter — reliability, validity, generalisability, replicability — into a toolkit for judging whether ANY study is "good
science."

Reliability

Reliability Test-retest reliability

How CONSISTENT a measure or procedure is — would it give the same result Giving the same test to the same participants on two separate occasions; high
again? reliability = the two sets of scores correlate strongly.

Inter-rater reliability Inter-observer reliability

How consistently different researchers interpret the same qualitative data (e.g. How consistently different observers record the same event in an observation.
open-question answers).

ELABORATE
Reliability can be improved through: standardisation (same instructions/procedures/materials for everyone), clear operational definitions, and
researchers discussing/training together to interpret data consistently.

Validity

Validity Face validity

Whether a test, task or study actually measures what it CLAIMS to measure. Whether a test simply LOOKS, on the surface, like it measures what it's
supposed to.

Ecological validity Mundane realism

Whether findings from one situation (e.g. a lab) would also apply to more How true-to-everyday-life a task feels — higher mundane realism generally
realistic, real-world situations. raises ecological validity.

MUST KNOW: THREATS TO VALIDITY

✓ Subjectivity — a researcher's personal bias affecting how they interpret data (opposite: objectivity).
✓ Demand characteristics — clues that let participants guess the aim, causing them to change their natural behaviour.

Replicability & Generalisability

Replicability Generalisability
Whether a study's exact procedure can be repeated (by the same or different How widely a study's findings apply beyond the specific sample studied —
researchers) to check the results — requires very detailed, clear reporting of the depends heavily on how representative/large the sample was, and on ecological
original method. validity.

MUST KNOW: THE 4-QUESTION CHECKLIST (USE FOR ANY STUDY, INCLUDING UNFAMILIAR ONES)
✓ Is it valid? Does it test what it claims to? Think ecological validity, demand characteristics, subjectivity.
✓ Is it reliable? Are the measures/tools consistent? Could interpretation of the data be subjective?
✓ Is it generalisable? Would the findings apply to other people/places/times? Depends on the sample.
✓ Could it be replicated? Is there enough procedural detail reported to repeat it exactly?

MUST KNOW: WAYS TO IMPROVE A STUDY'S METHODOLOGY

✓ Method — e.g. switch field ↔ laboratory experiment, or questionnaire ↔ interview.
✓ Design — independent measures avoids order effects; repeated measures avoids individual-difference problems.
✓ Sample — opportunity for a bigger sample; volunteer to find rare participants; random for better generalisability.
✓ Tool — check and improve inter-rater or test-retest reliability.
✓ Procedure — reduce demand characteristics; increase mundane realism/ecological validity.

1.11 RECAP
Reliability = consistency (test-retest, inter-rater, inter-observer).
Validity = measures what it claims (face validity, ecological validity).
Generalisability = applies elsewhere. Replicability = the study can be repeated exactly.
Use the 4-question checklist to evaluate ANY study — including unseen exam scenarios.
★ Whole-Chapter Summary

RESEARCH METHODS EXPERIMENTAL DESIGNS

Experiments (lab/field) Independent measures
Self-reports (questionnaire/interview) Repeated measures
Case studies Matched pairs
Observations
Correlations
Longitudinal studies

HYPOTHESES VARIABLES
Alternative: directional / non-directional IV / DV (experiments)
Null (chance explains any difference) Co-variables (correlations)
Confounds: participant + situational
All must be operationalised

SAMPLING DATA & ANALYSIS

Opportunity — convenient Quantitative vs qualitative
Volunteer — self-selected Mode / median / mean
Random — equal chance Range / standard deviation
Bar chart / histogram / scatter graph

ETHICS — HUMANS ETHICS — ANIMALS

Consent, right to withdraw Bateson's cube
Protection from harm, deception Replacement, species/numbers
Confidentiality, privacy, debrief Humane procedures

METHODOLOGICAL EVALUATION
Reliability — test-retest, inter-rater, inter-observer
Validity — face, ecological
Generalisability & replicability

FINAL TIP
Cover each flashcard and try to recall it before checking the answer.
If you can explain WHY a rule exists (not just state it), you can apply it to unfamiliar exam scenarios too.