Component 1 — Research Methods & Statistics

Laboratory Experiment

High control over IV and extraneous variables in an artificial setting. High internal validity, easy to replicate, but low ecological validity and high demand characteristics.

Field Experiment

IV manipulated in a natural, real-world setting. High ecological validity, reduced demand characteristics, but low control over extraneous variables and ethical issues with consent.

Natural Experiment

IV changes naturally without researcher intervention (e.g., introduction of TV to an island). High ecological validity, allows study of unethical/impossible IVs, but no random allocation and low control over extraneous variables.

Quasi Experiment

IV is an inherent participant characteristic (e.g., age, gender, autism). High practical application, but cannot randomly allocate participants, meaning participant variables cannot be controlled.

Independent Measures Design

Different participants used in each condition. No order effects and lower demand characteristics, but individual differences (participant variables) can confound results and requires double the sample size.

Repeated Measures Design

Same participants take part in all conditions. Controls for participant variables and uses fewer participants, but susceptible to order effects (practice, fatigue) and higher demand characteristics.

Matched Pairs Design

Participants paired on key traits (e.g., IQ, age); one assigned to Condition A, one to B. Reduces participant variables and eliminates order effects, but time-consuming, expensive, and difficult to match perfectly.

Random Sampling

Every population member has an equal chance of selection (e.g., random name generator). Unbiased and representative, but difficult, time-consuming, and selected individuals may decline.

Opportunity Sampling

Selecting whoever is available at the time and place. Quick, cheap, and convenient, but highly unrepresentative and biased toward specific locations/times.

Volunteer (Self-Selected) Sampling

Participants opt-in response to an advert. Reaches specific target audiences and reduces drop-out, but sample is biased toward highly motivated or helpful individuals.

Stratified Sampling

Subgroups (strata) identified in population; participants randomly selected proportional to sub-group size. Highly representative and generalisable, but complex and time-consuming to construct.

Systematic Sampling

Every Nth member of a target population is selected from a list. Objective and avoids researcher bias, but can accidentally align with a periodic pattern in the list.

Operationalisation

Defining variables into measurable, concrete terms (e.g., defining "aggression" as "number of physical hits on a Bobo doll in 20 minutes").

Extraneous Variable

Any variable other than the IV that might affect the DV if not controlled (e.g., room temperature, noise).

Confounding Variable

A variable that varies systematically with the IV, making it impossible to determine if changes in the DV were caused by the IV alone.

Internal Validity

Extent to which a study measures what it claims to measure without interference from confounding or extraneous variables.

External Validity

Extent to which findings can be generalised beyond the study setting. Includes Ecological Validity (settings) and Population Validity (people).

Inter-Rater Reliability

Agreement level between two or more observers recording the same behavior. Checked using a correlation coefficient (r ≥ +0.8 indicates high reliability).

Test-Retest Reliability

Assessing consistency by administering the exact same test to the same participants on two separate occasions and correlating the scores.

BPS Ethical Principles

Respect (Consent, Confidentiality, Anonymity), Competence (Maintaining professional standards), Responsibility (Protection from harm, Debriefing), and Integrity (Honesty, Avoiding deception).

Covert vs Overt Observation

Covert: Participants unaware they are observed (high validity, ethical issues). Overt: Participants know they are observed (ethical, risk of demand characteristics).

Structured vs Unstructured Observation

Structured: Coding frame/categories used to tally behaviors. Unstructured: Researcher records all noteworthy behavior in qualitative detail.

Time Sampling vs Event Sampling

Time sampling: Behavior recorded at set time intervals (e.g., every 30 seconds). Event sampling: Behavior recorded every time a specific event occurs.

Nominal Data

Categorical data placed into discrete groups (e.g., Yes/No, Pass/Fail, tally counts).

Ordinal Data

Data placed in order/rank, where intervals between ranks are not equal or standardized (e.g., ranking happiness from 1 to 10).

Interval Data

Continuous data measured on a fixed, standardized scale with equal units (e.g., temperature in °C, time in seconds).

Type I Error

False Positive: Rejecting the null hypothesis when it is actually true (optimistic error, often caused by lenient significance level like p < 0.10).

Type II Error

False Negative: Retaining the null hypothesis when it is actually false (pessimistic error, often caused by overly strict significance level like p < 0.01).

Sign Test Rules

Used for Nominal data, Repeated measures design, Testing for Difference. S-statistic is the less frequent sign count (+ or -). Calculated S ≤ Critical Value for significance.

Mann-Whitney U Test Rules

Used for Ordinal/Interval data, Independent measures design, Testing for Difference. Calculated U ≤ Critical Value for significance.

Wilcoxon Signed-Ranks Test Rules

Used for Ordinal/Interval data, Repeated measures design, Testing for Difference. Calculated T ≤ Critical Value for significance.

Spearman’s Rho Test Rules

Used for Ordinal/Interval data, Correlational design, Testing for Relationship. Calculated r_s ≥ Critical Value for significance.

Chi-Square Test Rules

Used for Nominal data, Independent measures design, Testing for Difference/Association. Calculated Chi-Square ≥ Critical Value for significance.

Rule of 'U' & 'T' in Stats Tests

If the test name contains the letter 'U' or 'T' (Mann-Whitney U, Wilcoxon T, Sign Test S), Calculated Value must be LESS THAN OR EQUAL TO Critical Value (≤). Otherwise, Calculated ≥ Critical (≥).