GRE Psychology Subject Test common mistakes
GRE Psychology Subject Test · Understand the traps and how to avoid them.
When a question asks what a lesion or drug does, you answer from the name of the structure or drug class instead of from the described behaviour, choosing the answer that sounds most like the label (e.g., 'amygdala = fear') rather than the one matching the specific symptom pattern in the stem.
The exam rewards functional reasoning over vocabulary recall. Many candidates memorised structure-function pairings as one-to-one slogans, so they stop reading the stem once they recognise a term and pick the association that feels familiar. Distractor options are deliberately built from these popular but oversimplified pairings.
What to do instead
Before answering, restate the stem's described deficit or effect in your own words, then ask which structure or transmitter is necessary for that specific function. Practise with double dissociations (e.g., hippocampus vs. striatum in declarative vs. procedural memory; amygdala vs. hippocampus in fear conditioning vs. explicit recall) so you can discriminate rather than merely recognise.
You treat the hypothalamus, pituitary, and adrenal axis as a single unit and pick any answer mentioning 'stress hormone' or 'HPA axis' for questions about negative feedback, circadian rhythm, or sexual differentiation, without tracking which level of the axis the stem describes.
Endocrine content is taught briefly in most courses, so candidates learn a loose narrative ('stress raises cortisol') and never build a level-by-level model. Under time pressure, any option with the right-sounding hormone feels safe.
What to do instead
Draw the axis once from memory: hypothalamic CRH to anterior pituitary ACTH to adrenal cortex cortisol, with cortisol feeding back on hippocampus and hypothalamus. Then answer by locating the stem at a specific level. Do the same for the gonadal axis and for melatonin and the suprachiasmatic nucleus.
On psychophysics and perception items you compute or reason with the wrong threshold concept, for example treating the just-noticeable difference as constant across intensities, or confusing the absolute threshold with the difference threshold and with signal-detection d-prime.
Weber's law, Fechner's law, absolute threshold, difference threshold, and signal-detection measures are taught together and sound interchangeable. Candidates match on the phrase 'threshold' rather than on the quantitative relationship the stem describes.
What to do instead
Write one line for each: absolute threshold is detection of a stimulus present at all; difference threshold is the smallest detectable change; Weber's fraction is the constant ratio of change to background; d-prime is sensitivity independent of bias, while criterion reflects bias. When a stem gives numbers, substitute them into the definition before looking at the options.
You answer memory questions using the strength of your own familiarity with a term, so you pick 'encoding' for a retrieval failure or 'retroactive interference' for a proactive-interference scenario, without checking the direction of the effect described in the stem.
Encoding, storage, and retrieval, and proactive versus retroactive interference, are mirror-image pairs. The exam exploits this by describing a concrete situation (old material disrupts new learning, or new material disrupts old) and offering both labels. Candidates skim for the keyword 'interference' and choose the more frequently seen term.
What to do instead
For every interference item, label the stem explicitly: which material is old, which is new, and which one is being disrupted. For memory-failure items, ask whether the information was ever stored; if it was stored and cannot be accessed, the problem is retrieval. Practise by writing the direction arrow next to each option before choosing.
You read a social psychology vignette and choose the phenomenon whose label you recognise rather than testing each option against the described pattern, so you select diffusion of responsibility for a case of social loafing, or fundamental attribution error for a self-serving bias.
Social psychology has many closely related labels (conformity, compliance, obedience; bystander effect, diffusion of responsibility, pluralistic ignorance; ingroup bias, outgroup homogeneity). Candidates rely on surface resemblance because the vignettes are short and time is tight.
What to do instead
Convert the vignette into a one-sentence structural description (who is present, what is observable, whose behaviour is being explained) and eliminate options that require a feature the vignette lacks. Build a comparison table of the confusable triples with one distinguishing feature each, and rehearse it until the distinctions are automatic.
On developmental items you answer from the age in the stem alone, picking the stage label that matches the number of years rather than the described behaviour, and you treat Piagetian stages as fixed ages rather than as described competencies.
Developmental textbooks present stages with age ranges, so candidates memorise the ranges and use them as a lookup table. The exam instead describes a behaviour (e.g., conservation failure, egocentric speech, object permanence) and expects you to infer the stage from the competence, not from the age.
What to do instead
Learn each Piagetian, Eriksonian, and attachment stage by its defining behaviour and its characteristic error, not by its age band. For each item, name the behaviour in the stem first, then map it to the stage. Note that ages overlap across stages and that the exam tests the behaviour.
You confuse the direction of a clinical distinction, for example treating a positive symptom as a deficit or a negative symptom as an excess, or mixing up incidence and prevalence, sensitivity and specificity, in clinical and measurement items.
These are paired terms whose everyday meanings pull in the opposite direction from their technical meanings (positive symptoms are added experiences; sensitivity is the true-positive rate). Candidates answer from the everyday connotation rather than the technical definition.
What to do instead
Write the definitions as formulas before the exam: incidence is new cases per time; prevalence is all cases at a point; sensitivity is true positives divided by all who have the condition; specificity is true negatives divided by all who do not. For symptoms, classify each as an addition or a loss before choosing.
On research-methods items you identify the design by its label in the stem and then answer as if that design controlled everything, ignoring the specific confound, demand characteristic, or validity threat the stem describes (e.g., calling a study 'experimental' because it has two groups, despite random assignment being absent).
Methods questions are often answered by pattern-matching the design name. The exam tests whether you can detect a missing control or a threat to validity inside a described procedure, and the distractors are plausible design labels.
What to do instead
For every methods item, check four things explicitly: was assignment random, was there a control or comparison condition, was the measure manipulated or observed, and what alternative explanation remains? Then choose the option that names the actual flaw. Practise identifying internal, external, construct, and statistical conclusion validity threats in short scenarios.
You compute or interpret a statistic without checking its assumptions, so you use a parametric test on ordinal data, interpret a correlation as causation, or read a p-value as the probability the null hypothesis is true.
Statistics is often learned as a menu of tests keyed to design names, and the interpretation of p-values and effect sizes is memorised as a slogan. The exam probes the meaning of the numbers, not the arithmetic.
What to do instead
For each test, memorise the scale of measurement and distributional assumption it requires, then check the stem against them. Rehearse the correct sentence for a p-value (the probability of data this extreme or more, given the null) and distinguish it from effect size and from the probability that the null is true.
You spend too long on a small number of hard items and then rush the rest, leaving easy items unanswered or misread, because the exam is scored on correct answers and there is no penalty for guessing.
Candidates carry over habits from exams with penalty scoring or from coursework where partial reasoning matters, and they treat a hard item as a challenge to be solved rather than a cost to be managed. The result is a cluster of errors on items that would have been easy with normal pacing.
What to do instead
Adopt a two-pass strategy: answer every item you can resolve quickly, mark and skip the rest, then return. Never leave a blank, since there is no guessing penalty. Time yourself on full-length practice sets so the pacing decision is rehearsed rather than improvised on exam day.