Psychology Interview Questions: Experiment Design, Statistics Drilling, and Theory-Comparison Chains
A question bank for psychology interviews: the framework for live experiment-design questions, the expected depth of statistics follow-ups, and a matrix for theory-comparison items.
What this page helps you do first
- The killer move: design an experiment on the spot — five-step frame
- Statistics drilling in three layers: what — when — how to interpret
- Theory comparisons via a human-nature-method-application-critique matrix
The signature move: design an experiment live
“Design an experiment to test hypothesis XX” is a psychology-interview staple and the biggest differentiator. The frame: **state the hypothesis** (IV, DV, expected relation, one sentence) → **participants and grouping** (source, assignment, sample-size rationale) → **procedure** (materials, flow, controls) → **confound control** (within/between trade-offs, order effects, demand characteristics) → **statistics** (which test, why). Completing all five shows the panel complete scientific thinking.
The most common failure is stopping at step three — a design without variable-control awareness reads as uninitiated. Daily drill: narrate ten classic paradigms (Stroop, memory span, conformity) through the five steps.
Statistics drilling in three layers
| Entry question | Second layer | Third layer | | :--- | :--- | :--- | | t-test vs ANOVA? | Interpreting two-way interactions? | When do simple effects run? | | What is the significance level? | What p<0.05 does and does not mean? | Effect size vs significance? | | Correlation vs causation? | Which designs license causal claims? | Cross-lagged logic? | | What is reliability? | Reliability-validity relation? | Is a 0.7 scale usable? |
Note the second column’s “does not mean”: articulating “not a 5% probability of being true, not a large effect, not practical importance” instantly separates you from formula-memorizers.
A matrix for theory-comparison items
“Behaviorism vs psychoanalysis”, “compare the three memory models” — answer comparisons in a fixed matrix: **view of human nature** (determined by environment / driven by instinct / actively constructing) → **methodology** (experiment / clinical case / information-processing paradigm) → **application** (behavior modification / therapy / education and engineering) → **mutual critiques** (each school’s softest flank). The four-row matrix secures the structure score. Preparation: pair up the major schools from your general-psychology and history texts, one matrix card each.
Narrative capital for cross-major candidates
Psychology admits many cross-major candidates, and panels have an explicit lens: **what your origin discipline brings to psychological research**. STEM backgrounds sell statistics and programming (modeling, large-data handling); humanities sell theoretical reading and qualitative sensitivity; medicine sells the clinical and neuroscience interface. Avoid “psychology is interesting and easy” motivations — panels have heard them ten thousand times. The high-scoring shape: a concrete observation + an attempted psychological explanation + the discovery of that explanation’s limits.
Sprint checklist
- **Ten paradigm narrations**: two per day through the five-step frame;
- **Statistics three-layer cards**: what-conditions-interpretation per method;
- **School matrices**: pairwise four-row comparison cards;
- **One closely read paper**: hypothesis, method, limits retold;
- **Motivation story**: observation-explanation-boundary structure.
Frequently asked questions
- What if the panel pokes a hole in my live design?
- It will almost certainly happen, usually as stress-testing rather than rejection. Three steps: restate to confirm (“you mean the design fails to control XX?”) → acknowledge and localize (“yes, a between-subjects individual-difference problem”) → repair (“going within-subjects with counterbalancing mitigates it, at the cost of carryover”). Turning the poke into a display of design-cost literacy is exactly what panels watch for.
- How deep do statistics questions go — derivations?
- No derivations — “consumer-grade” depth suffices: what question each method answers, its preconditions, and a plain-language reading of results. Being asked to derive formulas live is rare. The real deep water is conceptual discrimination (significance vs effect size, reliability vs validity, main effects vs interactions) — being able to compare and exemplify those three pairs covers eighty percent of the drilling space.
- Do applied-psychology professional tracks still test basic theory?
- Yes, at application-anchored depth: panels prefer “how does this theory play in counseling / UX research / education”. Add one “application anchor” line per theory card: reinforcement paired with behavioral intervention, attachment with the therapeutic alliance, cognitive load with instructional design. Candidates who voluntarily slide theory toward application hold a natural edge on professional panels.
Where to go after this question bank
Question banks rehearse the follow-up chains; your own materials decide whether the answers hold. Use the thesis workflow to strengthen the draft behind your answers.