Module 05
Personality and Individual Differences
Tools for trait structure and psychological measurement: where factors come from, what reliability and validity actually buy you, and why an individual profile is harder to interpret than a bar chart makes it look.
18 tools published
Tools in this module
-
Person–Situation Interaction Theatre
Four fictional characters are put through five situations. Students rank their behaviour twice — once where traits show through and once where a different affordance reorders everyone — then take control of situation strength and assigned role to find the setting where the four become behaviourally indistinguishable.
-
State versus Trait Tracker
Simulated experience-sampling data for four fictional people over a fortnight. Students rank them by typical level from a single moment, discover how badly that works, then control within-person variability, measurement error and life events — and watch two people with the same mean turn out to be living quite different fortnights.
-
Factor Rotation Playground
Twelve fictional adjective markers in a two-dimensional factor space. Students rotate the axes and watch every loading change while the configuration of points, each communality and the total variance explained stay exactly where they were — then meet a genuine cross-loading and a pair of clusters that no orthogonal rotation can fit.
-
Facet-Level Detective
Two fictional people share a broad domain score almost exactly and behave quite differently. Students assign behavioural evidence before the facet scores are shown, build their own matched pair, then watch the "fixed" broad score reverse its ranking when the questionnaire samples facets differently.
-
The Alpha Trap
Students maximise Cronbach's alpha by choosing items from a simulated bank, discover that the cheapest route is to ask the same question five ways, and then see the narrow scale they have built fail against a simulated outcome.
-
Reverse-Item Disaster
A six-item fictional scale with two reverse-keyed items, scored four ways: correctly, with no recoding, with the wording reversed mentally but not in the data, and with the wrong scale maximum. Totals, item-total correlations, alpha and the respondent ranking all update, and a step-by-step repair mode has students do the arithmetic themselves.
-
Response-Style Simulator
Seven fictional respondents with the same true standing answer the same twenty items in seven different ways. Students compare raw grids, distributions, totals and split-half correlations, watch balanced keying neutralise acquiescence and unbalanced keying disguise it, then identify a mystery pattern and are asked what it cannot tell them.
-
Measurement-Invariance Translator
Two fictional groups, four original items, and full control of loadings, intercepts and latent means. Students manufacture an observed group difference while the true latent difference is exactly zero, see configural, metric and scalar comparisons as pictures before meeting the vocabulary, and find out which comparisons remain defensible.
-
“Explain This Person” Courtroom
One fictional behaviour and eight explanations that all fit it. Students rate each, request evidence, and discover that fitting a case is cheap while predicting something distinctive is not. The verdict rewards evidence that discriminates and proportionate confidence, and scores down anyone who converges on a single cause.
-
Intelligence-Test Battery Builder
Seven generic cognitive task families, one testing session, and three different purposes. Students assemble a battery under time and burden constraints, watch breadth, reliability, exposure-dependence and burden trade against each other, then see the same battery judged against all three purposes without a single task changing.
-
Positive Manifold Visualiser
Six fictional cognitive tasks and a correlation matrix students control by deciding how much of each task's performance is shared with all tasks, shared within a group, or specific and error. Scatterplots, a loading diagram and a genuinely computed first factor follow — and the same positive first factor appears whether the structure underneath is one general ability or two groups.
-
Culture-Fair Test Challenge
Students design a non-verbal reasoning task through six decisions and watch six construct-irrelevant demands respond. Removing language removes one of them. Four fictional participants, described only by prior experience of testing formats, show how unevenly a design distributes demands that have nothing to do with reasoning.
-
Speed–Accuracy Trade-Off
An untimed arrow-discrimination task students can actually perform, followed by four fictional respondents — three with identical ability and different response caution, one with genuinely lower ability. Speed and accuracy are plotted jointly throughout, and a caution slider moves both at once while ability is held constant.
-
Twin-Study Simulator
Students set the true variance components, generate fictional twin and sibling pairs from them, and recover ACE estimates with Falconer's formulae — then switch on unequal environments, assortative mating and gene-environment correlation and watch the estimates fail in opposite directions while the truth stays put.
-
Gene × Environment Interaction Visualiser
Students name the pattern in a study that sampled only the adverse half of an environmental range, then watch the range widen and the same two groups turn a diathesis-stress finding into differential susceptibility. An explorer then lets them move slopes, crossover and sampled window and see which theory the visible data appear to support.
-
Emotional-Intelligence Claims Laboratory
Six fictional measures - a self-report questionnaire, two tasks with scored answers, colleague ratings of social effectiveness, and the ordinary personality and reasoning measures any claim has to beat. Students read the correlation matrix, test how much of the pattern is method rather than content, run incremental validity properly, and judge five pieces of marketing against the evidence they have in front of them.
-
Self-Esteem Stability Tracker
Four fictional people share a self-esteem baseline and differ on everything else. Students deliver praise, criticism, success and rejection across three life domains and watch level, day-to-day volatility, domain contingency and recovery speed come apart — four characteristics that a single score collapses into one.
-
Personality Disorder Continuum
Six dimensions of an entirely fictional profile - trait extremity, rigidity, cross-situational persistence, subjective distress, interpersonal impact and functional impairment - and two ways of drawing a category line across them. Students hold trait extremity constant while everything else changes, switch between a trait-counting rule and a rule requiring impairment, and move the threshold to see how many fictional profiles move with it. The tool never returns a diagnosis and asks nothing about the person using it.
What this module covers
These are the topics the module was planned around. The tools above cover them; the list is kept so the intended scope stays visible and so a contributor proposing a new tool can see what is already here.
- Trait structure and the Big Five How correlations among a large set of items resolve into a small number of broad dimensions, and what is lost in the reduction.
- Factor analysis, made visible Watching a correlation matrix reduce to factors, and seeing how the decision about how many to retain changes the story that gets told.
- Reliability How internal consistency and test–retest reliability respond to scale length and to item quality, rather than being a fixed property of a questionnaire.
- Validity and measurement error Why an unreliable measure puts a ceiling on the correlation it can show with anything else — including a perfectly measured outcome.
- Reading a trait profile Norms, percentiles and the standard error of measurement, and how wide the band around a single reported score really is.
A note on personality questionnaires
Tools in this module are for teaching measurement, not for assessing anyone. Where a tool asks students to answer items about themselves, it will use openly licensed public-domain item pools, say plainly that the result is a classroom illustration rather than an assessment, and keep every response inside the browser tab — nothing is stored or transmitted.
Traits are treated as probabilistic, dimensional and context-sensitive. No tool will imply that a score determines behaviour, and simulated values will be labelled as simulated rather than presented as norms or validated cut-offs.
Copyrighted commercial inventories will not be reproduced here, in whole or in part.
Building one of these
All five topics above are open. If you teach individual differences and have a demonstration you already run by hand, turning it into a page here makes it reusable by everyone else. The contributing guide covers the folder layout, the shared interactive shell and the accessibility checks a tool needs to pass before it is merged.