• Objectives, contents, schedule, evaluation system, testing, common sense
• The role of common sense; how to think about ourselves and others

Sets up the course logistics and its central tension: our intuitions about behaviour feel obvious yet are often wrong or contradictory ("opposites attract" vs. "birds of a feather flock together"). Science exists precisely because common sense cannot adjudicate between such claims.

Key idea: "common sense" frequently produces mutually contradictory predictions, so we need systematic, empirical methods to decide which is actually true.

• Cognitive biases, confirmation bias, self-fulfilling prophecy
• Belief perseverance error, overconfidence effect

Confirmation bias is the tendency to seek and weight evidence that supports what we already believe. A self-fulfilling prophecy occurs when an expectation changes behaviour so that the expectation comes true. Belief perseverance is clinging to a belief even after its evidence is discredited; the overconfidence effect is systematically overrating the accuracy of our own judgments.

Why it matters: these biases are exactly what an experimental design — blinding, control groups, pre-registered hypotheses — is built to neutralise.

• Judgment under uncertainty — Tversky & Kahneman's classic on the heuristics (representativeness, availability, anchoring) behind systematic error. (Session 2)

• Fundamental attribution error, actor–observer effect, self-serving bias

The fundamental attribution error (FAE) is over-attributing others' behaviour to disposition while underrating the situation. The actor–observer effect is the flip: we explain our own behaviour situationally but others' dispositionally. The self-serving bias credits our successes to ourselves and blames failures on circumstances.

Key idea: attribution biases distort how we interpret data about people — a direct threat to the validity of observational and survey research.

• From the fundamental attribution error to the truly FAE — revisits how strongly observers discount situational forces. (Session 3)

• Hindsight bias, false consensus / false uniqueness effect

Hindsight bias ("I knew it all along") is the tendency, once an outcome is known, to see it as having been predictable. The false-consensus effect overestimates how many others share our views; the false-uniqueness effect underestimates how many share our abilities or desirable traits.

Key idea: hindsight bias is why predictions must be recorded before the data are seen — the logic behind pre-registration.

• Self-esteem & self-serving bias in reactions to positive and negative events — how self-esteem moderates the self-serving attribution pattern. (Session 4)
• Cross-cultural examination of the false-consensus effect — tests whether the bias generalises across cultures. (Session 4)

• What is science? Science vs. common sense; the scientific method
• Goals/types of research, features of empirical research, hypotheses & theories

Science is a systematic, empirical, self-correcting way of knowing with four goals: describe, predict, explain and (where possible) control. A theory is a broad explanatory framework; a hypothesis is a specific, falsifiable prediction derived from it. An operational definition specifies a construct in terms of the exact operations used to measure or manipulate it.

Key idea: a good hypothesis must be falsifiable — stated so that some conceivable observation could prove it wrong.

• Ch.1 Introduction to Scientific Thinking (pg. 3–19) — what makes knowledge scientific; the empirical, self-correcting nature of inquiry. (Session 5)
• Ch.2 Generating Testable Ideas (pg. 27–34) — moving from a vague question to a falsifiable, operationalised hypothesis. (Session 5)

• Science vs. pseudoscience, types of validity, replication, Type I & II error

Pseudoscience mimics science but resists falsification and ignores disconfirmation. Validity comes in kinds — construct, internal, external, statistical-conclusion. Replication is repeating a study to check that a result is real. A Type I error is a false positive (rejecting a true null); a Type II error is a false negative (missing a real effect).

Decision table: with significance level α = P(Type I) and β = P(Type II), statistical power = 1 − β is the chance of detecting a real effect.

• Ch.1 Distinguishing Science from Pseudoscience (pg. 20–27) — the demarcation criteria (falsifiability, openness to revision). (Session 6)
• Ch.4 Reliability & Validity of a Measurement (pg. 93–98) — consistency vs. accuracy of measures. (Session 6)
• Ch.4 Ethics in Focus: Replication as a Gauge for Fraud? (pg. 103) — replication as a safeguard against fabricated results. (Session 6)
• Ch.14 Types of Error and Power (pg. 401–402) — formal treatment of Type I/II error and statistical power. (Session 6)

• Features of descriptive research designs, representative samples, case studies

Descriptive designs aim to characterise a phenomenon as it naturally occurs. A representative sample mirrors the population on relevant characteristics so findings generalise; a case study is a deep, qualitative examination of a single person, group or event — rich but hard to generalise. Operationalisation turns abstract constructs into measurable variables.

Key idea: random sampling protects external validity (generalisability); random assignment, introduced later, protects internal validity (causal inference).

• Ch.4 Identifying Scientific Variables (pg. 83–88) — constructs, variables and how to operationalise them. (Session 7)
• Ch.5 Sampling from Populations (pg. 113–129) — probability vs. non-probability sampling and representativeness. (Session 7)

• Surveys, naturalistic observation; strengths & weaknesses of each
• Descriptive vs. correlational designs, causation, third variables, issues

Surveys efficiently capture self-reported attitudes at scale but are vulnerable to wording, order and social-desirability effects. Naturalistic observation records behaviour in its real setting with high ecological validity, but risks observer bias and reactivity. Descriptive statistics summarise data via measures of central tendency (mean, median, mode) and spread (range, variance, SD).

Key idea: the mean is x̄ = Σxᵢ / n and the standard deviation s = √(Σ(xᵢ − x̄)² / (n−1)) — together they describe "typical value" and "typical distance from typical".

• Ch.6 Choosing a Research Design (pg. 139–146) — matching the design to the research question and its constraints. (Session 8)
• Ch.13 Descriptive Statistics: Why Summarize Data? (pg. 368–378) — central tendency, variability and when each summary is appropriate. (Session 8)

• Correlational designs, causation, third variables, issues
• Scatterplots, directionality & third-variable problems, correlation vs. causation
• Spurious correlations; procedures to improve understanding of correlations

A correlation quantifies how two variables move together, from −1 (perfect inverse) through 0 (none) to +1 (perfect positive). Causation cannot be inferred because of the directionality problem (does A→B or B→A?) and the third-variable problem (some C drives both). Spurious correlations are strong-but-meaningless associations driven by chance or a lurking variable.

Key idea: Pearson's r = cov(X,Y) / (sₓ · s_y); and r² is the proportion of variance in one variable explained by the other.

• Ch.8 Correlational Designs (pg. 217–227) — computing and interpreting correlations and the limits on causal claims. (Session 9)

• Features of experimental studies: manipulation, randomization, control groups
• Confounding variables, placebo effect

An experiment manipulates an independent variable (IV) and measures a dependent variable (DV) while holding everything else constant. Random assignment distributes unknown differences evenly across conditions, so the control group provides a valid baseline. A confound is any variable that varies systematically with the IV and offers a rival explanation; the placebo effect is change caused by expectation rather than treatment.

Key idea: manipulation + random assignment + a control group together = internal validity, the licence to say "X caused Y".

• Ch.9 Single-Case Experimental Designs (pg. 256–272) — establishing causality with one or few participants via baseline/treatment phases. (Session 10)

• Participant/experimenter effects, single/double blind, applications, meta-analysis

Participant effects (e.g. demand characteristics) and experimenter effects (unintentional cueing) can bias results. Single-blind designs keep participants unaware of condition; double-blind designs keep the experimenter unaware too. A meta-analysis statistically combines many studies' effect sizes to estimate the true effect more precisely than any single study.

Key idea: blinding removes expectancy as a confound — neither side can unconsciously nudge the outcome toward the hypothesis.

• Ch.9 Single-Case Experimental Designs (pg. 256–272) — continued: reversal and multiple-baseline logic. (Session 11)

• Factorial designs, main effects & 2-way/3-way interactions, statistical tests

A factorial design crosses two or more IVs (factors) so every combination of levels is tested — a 2×2 has four cells. This is efficient and, crucially, reveals interactions: cases where the effect of one factor depends on the level of another. These designs are typically analysed with a factorial ANOVA.

Key idea: a 2×3 factorial = 2 levels of factor A × 3 of factor B = 6 conditions; it yields a main effect of A, a main effect of B, and an A×B interaction.

• Ch.12 Factorial Experimental Designs (pg. 335–346) — structure and notation of multi-factor experiments. (Session 12)
• Ch.12 Main Effects and Interactions (pg. 342–355) — defining and reading main effects vs. interactions. (Sessions 12–13)

• Main effects, interactions, simple main effects, floor/ceiling effects

A main effect is the overall effect of one factor averaged across the other. A simple main effect is the effect of one factor at a single level of another — what you probe when an interaction is significant. Floor/ceiling effects occur when a measure bottoms out or tops out, masking real differences.

Key idea: on an interaction plot, non-parallel lines signal an interaction; parallel lines signal main effects only.

• Ch.12 Main Effects and Interactions (pg. 342–355) — interpreting interaction plots and simple main effects. (Session 13)

• Identifying main effects, interactions, control groups & confounds in studies

Applied practice: working through published studies to spot the design's factors, diagnose confounds, and decide which pattern of means reflects a main effect versus an interaction.

• Participate in a between-participants design experiment (individual basis)

In a between-participants (between-subjects) design each person experiences only one condition, so comparisons are across different groups. This avoids carryover effects but requires more participants and relies on randomization to equate the groups. You take part as a participant, then analyse the resulting data next session.

• Ch.10 Between-Subjects Experimental Designs (pg. 273–282) — independent-groups logic and when to prefer it. (Session 15)

• Discussion of between-participants experimental data from Workshop 1

The data you generated are analysed with a factorial ANOVA, which partitions variance to test each main effect and interaction at once. The session connects the abstract design to a concrete output table and what its F-tests mean.

Key idea: ANOVA compares between-group to within-group variance — F = MS_between / MS_within; large F ⇒ groups differ more than chance.

• Ch.10 General Instructions for Conducting a Factorial ANOVA (pg. 356–364) — step-by-step running and reporting of a factorial ANOVA. (Session 16)

• Participate in a mixed factorial design experiment (individual basis)

A mixed factorial design combines at least one between-participants factor with at least one within-participants factor (where every person sees all levels). It blends the efficiency of repeated measures with the cleanliness of independent groups. As before, you participate now and analyse the data next session.

• Ch.10 Mixed Participants Experimental Designs (pg. 273–282) — combining within- and between-subjects factors in one study. (Session 17)

• Discussion of mixed-design experimental data from Workshop 2; midterm preparation

Analysing the mixed-design data and weighing the trade-offs: within-subjects designs gain power by removing individual differences but risk order/carryover effects (countered by counterbalancing); between-subjects designs avoid those but need larger samples. Doubles as midterm review.

• Ch.11 Comparing Between-Subjects / Within-Subjects Designs (pg. 328–334) — the power-vs-carryover trade-off and counterbalancing. (Session 18)

Covers material from the lecture slides only (Modules 1–3). Format: multiple-choice plus one long-answer question requiring you to explain a theory and apply it through two real-world examples.

• Quasi-experimental research, pre/post-test designs, interrupted time-series, issues

A quasi-experiment manipulates or compares conditions but without random assignment — often using pre-existing groups. A pre/post-test design measures before and after an intervention; an interrupted time-series takes many measurements before and after a change to see whether the trend shifts. Without randomization, selection bias and history are standing threats to internal validity.

Key idea: quasi-experiments trade some causal certainty for feasibility — strong on external validity, weaker on internal validity.

• Ch.11 Quasi-Experimental Designs (pg. 240–250) — nonequivalent-groups and time-series designs and their validity threats. (Session 20)

• Applied research: history, real-world examples, pre/post-tests, ethical issues

Applied research targets a concrete, practical problem (does this policy, therapy or product work?), in contrast to basic research, which seeks knowledge for its own sake. It typically prioritises external validity and real-world impact, and raises distinctive cost–benefit and ethical questions when interventions touch people's lives.

Key idea: applied vs. basic is about the goal (solve a problem vs. understand a phenomenon), not the method — either can use experiments or surveys.

• Quasi-Experimental Designs & Applied Research (Blackboard) — bridges quasi-experimental method with real-world applied problems. (Session 21)

• Indirect vs. direct self-report; behavioral/physiological measures
• Strengths/weaknesses of measurement tools, reliability of measures
• Ethical research, human/animal participants, research ethics boards

Measures vary in directness and reactivity: direct self-report (asking people) is easy but fakeable; indirect measures sidestep self-presentation; behavioural and physiological measures (e.g. reaction time, heart rate) are harder to fake but costlier. A measure must be both reliable (consistent) and valid (measures what it claims). Ethics — informed consent, minimising harm, and oversight by a research ethics board / IRB — constrains every choice.

Key idea: reliability is necessary but not sufficient for validity — a scale can be perfectly consistent yet consistently wrong.

• Should we trust web-based studies? (Blackboard) — data quality and validity of online data collection. (Session 22)
• Ch.3 Research Ethics (pg. 53–80) — consent, deception, debriefing, IRBs and animal welfare. (Session 22)
• Ethics of CIA & military contracting by psychiatrists/psychologists (Blackboard) — a hard case on dual-use research ethics. (Session 22)
• Scientific rewards and conflicts of ethical choices in human research (Blackboard) — how career incentives can collide with ethics. (Session 22)

• Qualitative vs. quantitative research, types of sampling, types of interviews
• Interview process; creating & conducting an interview; strengths/weaknesses

Qualitative research seeks depth and meaning (words, themes); quantitative seeks breadth and measurement (numbers, statistics). Interviews range from structured (fixed questions, comparable answers) through semi-structured to unstructured (open, exploratory). Qualitative work typically uses purposive rather than random sampling.

Key idea: more structure = more comparability but less depth; the interview type should follow the research question.

• Writing interview protocol and conducting interviews: tips for students new to qualitative research (Blackboard) — a practical how-to for first-time interviewers. (Session 23)

• Focus groups: processes & protocol, data analyses, issues

A focus group is a moderated discussion (typically 6–8 people) whose value lies in the interaction between participants, surfacing shared and divergent views. Challenges include groupthink, dominant speakers, and the difficulty of coding open-ended talk into themes. Focus groups and surveys are often combined — qualitative depth informs quantitative scale.

Key idea: the moderator's job is to elicit interaction without leading it — poor moderation is itself a confound.

• Focus groups and surveys as complementary research methods: a case example (Blackboard) — shows how the two methods triangulate. (Session 24)

• Focus groups continued…

Continued practice running and analysing focus-group data: moving from raw transcript to coded themes, and judging the trustworthiness of qualitative conclusions (credibility, transferability) — the qualitative analogues of validity.

• Applied qualitative methods: conducting field research / gathering data

The capstone of the qualitative module: in groups you conduct a real observational field study, gathering data in a natural setting and recording systematic field notes. This is where sampling, observation, ethics and analysis come together on data you collected yourselves — feeding directly into the group research report and presentation.

• Revisiting field experimentation: field notes for the future (Blackboard) — practical guidance on observing and note-taking in the field. (Session 26)

• Structured writing, formatting, title page, references, citing, plagiarism
• Group project (in-class / take-home if necessary)

APA format gives research a shared structure — title page, abstract, introduction, method, results, discussion and references — so readers can find and evaluate each part quickly. Correct citation credits prior work and lets claims be traced; failing to do so is plagiarism. The session prepares the written backbone of the group project.

Key idea: the IMRaD structure (Introduction, Method, Results, Discussion) maps onto the logic of inquiry itself — question, how, what, so-what.

• Ch.15 Communicating Research: Preparing Manuscripts, Posters, and Talks (pg. 425–433, 447–454) — how to write up and present results. (Session 27)
• APA-Style Writing, Sample Manuscript (pg. 455–503) — an annotated model paper to imitate. (Session 27)

• Discussion, evaluation and feedback · submit via Turnitin on presentation date

In groups of 4–5, present two studies from a single empirical article in the format of a conference talk, followed by class discussion and feedback. Graded against the rubric posted on Blackboard.

• Discussion, evaluation and feedback · submit via Turnitin on presentation date

Remaining groups present, with the same discussion-and-feedback format. Acts as a cumulative review of design concepts as the class critiques each study's methodology.

• Cumulative; multiple-choice plus long-answer. Minimum 3.5 required to pass.

Cumulative across all six modules (slides only). The long-answer asks you to weigh the advantages and disadvantages of methodological techniques and to identify every methodological flaw in a given study — the course's core skill in one question. You must score at least 3.5/10 here to pass, regardless of your weighted average.