NZ Scholarship Psychology cheat sheet

Threats to internal validity: history, maturation, testing, instrumentation, regression to the mean, selection, attrition and confounds. Random assignment and a control condition address selection and most confounds, but they do not fix maturation or history in a one-group pre-test/post-test design.

Validity types: internal (causal inference), external (generalisability), construct (the measure captures the intended construct), ecological (real-world setting) and statistical conclusion validity. A study can be internally valid yet low in ecological validity; the trade-off is a design decision that must be justified.

Reliability types: test-retest (stability over time), inter-rater (agreement between observers) and internal consistency (Cronbach's alpha). Reliability is necessary but not sufficient for validity: a scale can measure the wrong construct consistently.

Sampling: probability methods (simple random, stratified, cluster) support statistical generalisation; non-probability methods (opportunity, snowball, purposive) dominate qualitative work. A small, biased sample limits external validity, not internal validity.

Designs: independent measures (between-subjects) avoids order effects but requires more participants; repeated measures (within-subjects) controls individual differences but risks order and practice effects, managed by counterbalancing or randomisation of order.

Correlation: r ranges from -1 to +1, and r-squared is the proportion of variance shared. Correlation does not establish causation; directionality and third variables must be ruled out by design, not by statistical control alone.

Inference: p < .05 is the probability of obtaining data at least as extreme as observed if the null hypothesis is true, not the probability that the null is true. Report effect size (Cohen's d: 0.2 small, 0.5 medium, 0.8 large) and confidence intervals alongside p.

Power equals 1 minus beta and rises with larger samples, larger effects and less stringent alpha. Underpowered studies yield unstable results, and publication bias inflates effect sizes in the published literature.

Qualitative quality criteria: credibility (member checking, triangulation), transferability (thick description), dependability (audit trail) and confirmability (reflexivity). Braun and Clarke's thematic analysis moves through familiarisation, coding, theme generation, review, definition and reporting.

Classical conditioning (Pavlov): a neutral stimulus repeatedly paired with an unconditioned stimulus becomes a conditioned stimulus eliciting a conditioned response. Extinction, spontaneous recovery and stimulus generalisation underpin exposure-based therapies such as systematic desensitisation.

Operant conditioning (Skinner): positive and negative reinforcement increase behaviour; positive and negative punishment decrease it. Variable ratio schedules produce high, extinction-resistant response rates, fixed interval schedules produce scalloped responding, and extinction typically produces a temporary burst.

Social learning theory (Bandura): observational learning requires attention, retention, motor reproduction and motivation. Vicarious reinforcement and self-efficacy beliefs determine whether a modelled behaviour is actually performed.

CBT (Beck, Ellis): the ABC model links activating events, beliefs and emotional consequences; cognitive restructuring and behavioural experiments test catastrophic predictions. Structured CBT is a guideline-recommended first-line treatment for mild to moderate anxiety and depression.

Attachment (Bowlby, Ainsworth): the Strange Situation classifies secure, insecure-avoidant and insecure-resistant patterns, with disorganised added by Main and Solomon. Internal working models predict later relationship styles, but continuity is moderate rather than deterministic.

Developmental theory: Piaget's stages (sensorimotor, preoperational, concrete operational, formal operational) use object permanence and conservation as markers; Vygotsky's zone of proximal development and scaffolding emphasise social guidance and language in learning.

Memory applications: Baddeley and Hitch's working memory model; Loftus's misinformation and weapon-focus research on eyewitness accuracy; retrieval practice and spaced testing consistently outperform rereading for long-term retention.

Clinical frameworks: the biopsychosocial model and diathesis-stress model explain disorder as an interaction between vulnerability and stress. Categorical DSM-5-TR diagnosis is criticised for comorbidity, thresholds and cultural bias, with dimensional and cultural-formulation alternatives proposed.

NZ practice frameworks: Te Whare Tapa Whā (taha wairua, taha hinengaro, taha tinana, taha whānau) and whānau-centred service design; culturally responsive practice requires more than translating materials, since engagement, assessment and outcome measures are all shaped by culture.

Four principles (Beauchamp and Childress): respect for autonomy, beneficence, non-maleficence and justice; APA adds fidelity and integrity. Principles frequently conflict, so an ethical judgement must justify the balance struck rather than merely name a principle.

Informed consent requires capacity, disclosure, understanding, voluntariness and documentation. Participants may withdraw at any time without penalty, including withdrawal of their data where feasible; minors provide assent alongside guardian consent.

Deception is permissible only when no non-deceptive alternative can answer the question, the deception carries no more than minimal risk, and full debriefing follows. Debriefing must explain the true aims and offer withdrawal of data.

Confidentiality limits should be disclosed in advance: serious risk of harm to self or others, court order and child protection concerns. In NZ the Privacy Act 2020 and Health Information Privacy Code govern collection, storage, access and disposal of participant data.

Power and dual relationships: psychologists avoid multiple relationships that impair objectivity or risk exploitation. Research with students, employees, prisoners or clients requires explicit safeguards, because consent obtained within a dependent relationship may be nominal.

Animal research follows the 3Rs: replacement with alternatives where possible, reduction to the minimum numbers needed, and refinement of procedures to minimise pain and distress, under the Animal Welfare Act 1999 and animal ethics committee approval.

Research integrity: fabrication (inventing data), falsification (altering or omitting data) and plagiarism are the three cardinal offences. Questionable practices include p-hacking, HARKing and selective outcome reporting; preregistration and open data reduce their prevalence.

Treaty of Waitangi obligations in research: partnership, participation and protection. This includes consulting mana whenua, honouring Māori data sovereignty (Te Mana Raraunga) and using Kaupapa Māori frameworks when research involves Māori.

Practice ethics: working within competence, supervision, accurate record-keeping, informed consent to treatment and cultural safety (Ramsden). A Tarasoff-style duty to warn permits limited disclosure where a client poses a serious threat to an identifiable third party.