EDUF3040 · Psychological Perspectives in Education
Foundations of Educational Psychology
Week 1 frames University of Sydney EDUF3040 as an applied, empirical science: theory generates testable hypotheses, which require valid and reliable measures and are judged against a hierarchy of evidence that runs from anecdote up to the true experiment. It hammers home that correlation is not causation, introduces reason-based practice and the Principle of Connectivity, and sets up APA's Top 20 Principles that organise the whole unit. In assessment this surfaces as multiple-choice questions on evidence levels and effect sizes, and as short-answer questions asking you to judge what causal claim a study design can support.
What this chapter covers
- 01Educational psychology as an empirical science: systematic empiricism → theory → testable hypothesis → valid/reliable measures
- 02Principles (few, general, evidence-based) vs methods (many, specific); the scientific teacher (Stanovich & Stanovich)
- 03Levels of evidence, weakest → strongest: anecdote → case study → correlational → pre-post → quasi → true experiment → time-series
- 04Correlation ≠ causation; the correlation coefficient r (−1 to +1: sign = direction, magnitude = strength)
- 05Experiments: manipulate an independent variable, control others; the 'CARE' criteria (Compare, Again, Relationship, Eliminate alternatives)
- 06Reviewing bodies of evidence: narrative review, meta-analysis, best-evidence synthesis; effect size (Cohen's d)
- 07Reason-based practice & the Principle of Connectivity: theory-linked methods can be justified even without direct evidence
- 08Law of Large Numbers: larger samples → tighter margin of error; why replication, meta-analysis and large studies (PISA, TIMSS) are trusted
Classify the evidence and judge the causal claim
- +1Identify the design. Two intact classes are compared after the fact, with no random assignment and (as described) a single post-measurement. That is a correlational / non-equivalent-groups comparison, not an experiment.
- +1Place it on the hierarchy. It sits above anecdote and case study (there is a quantitative comparison), but below quasi-experiment and true experiment, which are the designs that best rule out threats to a causal inference.
- +1Judge the causal claim. Correlation is not causation: a difference between two pre-existing groups cannot establish that the app caused the higher scores.
- +1Name a confound. The two classes may differ at baseline - prior reading ability, a stronger teacher, or socio-economic differences - any of which is a third variable that could explain the gap.
- +1Describe the stronger design. Run a true (randomised) experiment: randomly assign students to app vs control, add a pre-test to check baseline equivalence, and replicate. Random assignment plus a comparison condition satisfies the CARE criteria and lets you eliminate alternative explanations.
Key terms
- Levels of evidence
- A hierarchy from weakest to strongest: anecdote → case study → correlational study → pre-post comparison → quasi-experiment → true (randomised) experiment → time-series experiment. The last three best rule out threats to the validity of a causal inference; case studies are good for generating hypotheses but cannot support causal claims.
- Correlation ≠ causation
- A statistical association between two variables does not establish that one causes the other; a third (confounding) variable, reverse causation, or chance may explain the link. The correlation coefficient r runs from −1 to +1 (sign = direction, magnitude = strength).
- Effect size (Cohen's d)
- A standardised mean difference used as the common currency for 'what works'. Rough benchmarks used in the unit: d ≈ 0.20 is small (~8% improvement in learning rate), d ≈ 0.54 moderate (~21%), d = 1.0 large (~34%). Meta-analysis combines many studies' effect sizes, weighting by sample size.
- Best-evidence synthesis
- Slavin's approach to reviewing a body of research that combines effect-size quantification with systematic study selection and narrative attention to the strongest studies - a middle path between a pure narrative review and a purely statistical meta-analysis.
- Principle of Connectivity
- Stanovich & Stanovich's idea behind reason-based practice: when a pedagogy has no direct empirical evidence, do not automatically reject it - ask how the theory behind it connects to the wider research consensus. A theoretical link to established findings can support a method even without direct evidence.
- Law of Large Numbers
- As sample size increases, the sample relative frequency converges on the true probability and the margin of error shrinks (e.g. n ≈ 96 → ±10%, n ≈ 384 → ±5%, n ≈ 1,067 → ±3%). Small samples give imprecise results; precision improves with more data, which is why replication and very large studies (PISA, TIMSS) build confidence.
Foundations of Educational Psychology FAQ
Why does EDUF3040 start with research methods rather than teaching?
Because the unit's whole claim is that classroom decisions should rest on evidence, not opinion or fashion. If you cannot tell a correlational study from a controlled experiment, or judge whether a 'miracle method' has been replicated, you cannot evaluate the psychological efficacy of a teaching practice - which is exactly what the learning outcomes ask you to do. Week 1 gives you the toolkit (levels of evidence, effect sizes, meta-analysis, the Law of Large Numbers) that every later week's evidence is read through.
What is the difference between a principle and a method?
Educational psychology teaches principles - few, general, evidence-based statements about how learning works - rather than a long list of specific methods. A principle may have several practices attached to it, backed by strong, moderate, weak or no evidence. As the unit puts it (after Emerson), 'the teacher who grasps principles can successfully select their own methods.' This is why APA's Top 20 Principles, not a recipe book, organise the unit.
How is Week 1 assessed in EDUF3040?
Expect multiple-choice questions that ask you to rank evidence types, read a correlation coefficient, or identify what a design can prove, and short-answer questions asking you to classify a study and state its causal limits. The select-and-fill-in concept-map format is also introduced this week (the first map is due Friday of Week 2). Confirm exact dates on Canvas.
Can AI help me with the foundations material?
Yes. Sia can quiz you on the levels-of-evidence ladder, explain why correlation is not causation with fresh examples, walk through effect-size benchmarks, and check your reasoning on mock study vignettes. It is built to mirror how EDUF3040 is taught and assessed, explains each step, and does not complete graded work for you - the University of Sydney academic-integrity policy applies.
Exam move
Make the levels-of-evidence hierarchy automatic - you should be able to write it out weakest-to-strongest without hesitating - and pair it with the one-line reason each design can or cannot support a causal claim. Practise reading study vignettes and answering three questions fast: what design is this, what can it prove, and what confound threatens it. Memorise the effect-size benchmarks (d ≈ 0.2 / 0.54 / 1.0) and the margin-of-error intuition from the Law of Large Numbers so multiple-choice items on evidence strength are quick marks. Keep a running note of the APA Top 20 Principles as they appear, since this week sets up the frame the exam threads through. When a distinction blurs (correlational vs quasi-experiment, meta-analysis vs best-evidence synthesis), ask Sia to contrast them with fresh examples and test you.
Working through Foundations of Educational Psychology in EDUF3040? Sia is AskSia’s AI Educational Psychology tutor — ask any EDUF3040 Foundations of Educational Psychology question and get a clear, step-by-step explanation grounded in how EDUF3040 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
