Cattell Personality Factor Questionnaire

Abstract

The Cattell Personality Factor Questionnaire, better known as the Sixteen Personality Factor Questionnaire or 16PF, is a type of personality inventory: a self-report instrument measuring normal adult personality along sixteen primary trait dimensions. Raymond Cattell built it by applying factor analysis to the thousands of trait words in ordinary language, seeking the smallest set of source traits that could account for the surface variety of human description. This article treats the 16PF as a case study in the factor-analytic approach to personality: how observed behaviour reduces to latent dimensions, why the count of those dimensions has been disputed for decades, and how the sixteen factors relate to the five-factor model that later came to dominate. Three interactive demonstrations let a reader convert raw scores to stens, build a factor profile, and roll primary factors into global ones.

Keywords: factor analysis, source traits, personality structure, sten scores, five-factor model

The 16PF holds a distinctive place in the history of personality measurement: it was among the first instruments to claim that the whole of normal personality could be captured by a fixed, factor-analytically derived set of dimensions, and it staked that claim decades before the field settled on any particular number. Reading it well means reading the argument it started rather than the score sheet it produces.

That argument is about reduction. A person can be described in thousands of ways, yet the descriptions are massively redundant, and the question Cattell posed was how few independent dimensions remain once the redundancy is removed. The 16PF is one answer — sixteen — and the century-long debate over whether the true number is sixteen, five, or something else is the subject this instrument is best used to teach.

What the Cattell Personality Factor Questionnaire Is

The Cattell Personality Factor Questionnaire is a self-report inventory that scores an adult respondent on sixteen dimensions of normal personality, each reported as a position between two opposite poles — reserved versus warm, deferential versus dominant, trusting versus vigilant, and so on (Cattell et al., 1970). The respondent answers a few hundred multiple-choice items about typical behaviour and preference, and the answers are aggregated into a score on each of the sixteen factors. The instrument is deliberately a measure of the normal range rather than of psychopathology: its poles are two ordinary ways of being, not a healthy state and a disordered one.

What sets the 16PF apart from a questionnaire simply written to cover topics an author thought important is the claim behind its structure. Cattell held that the sixteen factors were not chosen but discovered — that they were the latent dimensions a factor analysis of human trait description keeps recovering, the irreducible axes of the space in which people differ. The instrument is therefore inseparable from a theory of what personality is: a small number of continuous source traits whose combinations produce the endless surface variety of everyday character. The questionnaire is the measuring device; the sixteen-factor model is the claim it was built to operationalise.

Because the factors are continuous and bipolar, an individual is described not by a type but by a profile — a position on each of the sixteen dimensions at once. Two people can share the same score on one factor and differ sharply on the rest, so the number of distinguishable profiles is enormous even though the factors are few. This profile logic, rather than any single score, is the unit of interpretation the instrument was designed around, and it is what the demonstrations below make manipulable.

From the Lexical Hypothesis to Source Traits

The 16PF begins from the lexical hypothesis: the idea that the personality differences important enough to matter in human affairs will have become encoded as single words in the natural language. Working from an exhaustive dictionary compilation of nearly eighteen thousand trait-related terms (Allport & Odbert, 1936), Cattell reduced the list first by grouping obvious synonyms and then by empirical clustering of ratings, arriving at a manageable set of trait clusters he could submit to factor analysis (Cattell, 1943). The reduction is the whole method in miniature: begin with the redundancy of language, and strip it away until only independent dimensions remain.

Factor analysis is the statistical engine of that stripping. Given a large matrix of correlations among observed variables — here, ratings on many trait words — it identifies a smaller number of latent factors that can reproduce the pattern of correlations. Cattell drew a sharp distinction between the surface traits, the clusters of behaviours that visibly co-occur, and the source traits, the underlying factors that cause them to co-occur; the sixteen primary factors are his candidate source traits. The distinction matters because two surface traits can correlate not because either causes the other but because a common source trait drives both, and only the factor analysis can tell the difference.

The interpretive freedom in that procedure is also its central vulnerability. How many factors to retain, and how to rotate them into an interpretable form, are decisions the mathematics constrains but does not determine, and different defensible choices yield different numbers of factors. Cattell's sixteen was one resolution of that freedom; the five-factor tradition that followed was another, reached from much the same lexical starting point. The instrument thus embodies a methodological lesson as much as a personality theory: a factor solution is a considered judgment about where to cut a continuous reduction, not a bare readout of nature.

The 16PF and the Five-Factor Model

The sixteen primary factors do not stand entirely apart from one another. Factor-analysing the correlations among the primaries themselves yields a second, higher layer of five broad global factors — Extraversion, Anxiety, Tough-Mindedness, Independence, and Self-Control — each built from several primaries acting together (Cattell & Cattell, 1995). This two-level structure is the instrument's quiet anticipation of what became the dominant framework in personality psychology: the five global factors line up closely with the dimensions later formalised as the five-factor model, or Big Five (Digman, 1990).

That convergence was not a coincidence but a rediscovery of the same lexical territory from a different pass. The earliest recurrent-factor analyses of trait ratings had already found roughly five reproducible dimensions in the 1960s (Tupes & Christal, 1992), and the modern Big Five program confirmed a stable five-factor structure across independent lexical studies (Goldberg, 1990). Goldberg's synthesis argued that five phenotypic dimensions recur so reliably across languages and methods that they constitute a natural coordinate system for trait description (Goldberg, 1993), while the five-factor model was validated across quite different instruments and observer types (McCrae & Costa, 1987). Against that backdrop the 16PF's own five global factors read as an early, independent arrival at the same place.

The relationship has been examined directly. Joint factor analyses of the 16PF alongside explicit Big Five measures found that the instruments largely map onto one another, with the 16PF primaries reconstructing the five-factor space when treated as multiple indicators of broader dimensions (Gerbing & Tuley, 1991). But the mapping is not clean enough to settle the count. Re-examinations of the higher-order structure of Cattell's, Comrey's, and Eysenck's scales showed that the number of second-order factors one recovers depends on the extraction and rotation choices, so that neither five nor any other figure is forced by the data (Boyle, 1989). At the opposite pole from Cattell's sixteen stands Hans Eysenck's contention that three broad dimensions — extraversion, neuroticism, and psychoticism — suffice to capture the temperament that matters, and that Cattell's primaries splinter a smaller number of genuine factors into unstable pieces; Eysenck put the whole dispute as an explicit choice among sixteen, five, or three, to be settled by criteria that stand outside any single factor analysis (Eysenck, 1991). The 16PF thus sits at the centre of the field's longest-running structural question: whether personality is best described by a handful of broad factors or by a larger set of narrower ones, a question that turns on the purpose of the description as much as on the correlations.

The Instrument in Motion

The three demonstrations below make manipulable the parts of the 16PF that the prose can only describe. The first converts a raw factor score into a sten against a normative population. The second builds a full sixteen-factor profile and draws it as the instrument's characteristic zigzag. The third rolls a set of primary factors up into one of the five global factors, showing the hierarchy that links the 16PF to the Big Five.

Convert a raw factor score to a sten

A raw score is meaningless until placed against a norm group. The sten transformation re-expresses it on a fixed 1-to-10 scale with a mean of 5.5 and a standard deviation of 2.

The same raw score is high against a low-mean norm and unremarkable against a high-mean one, which is why factor scores are never read raw. Computed locally, not stored.

The sten demonstration makes the scoring transformation concrete. A raw score means nothing until it is placed against the distribution of a reference group; the sten conversion re-expresses it on a fixed one-to-ten scale centred at 5.5, so that a sten of 8 always denotes the same standing relative to the norm no matter which factor it describes. Dragging the raw score shows how the same number can be high on a factor with a low normative mean and unremarkable on one with a high mean.

Build a 16PF profile

Set each factor from 1 (low pole) to 10 (high pole). The connecting line is the profile — the shape an interpreter reads, not any single factor.

A profile that zigzags between the poles describes a very different person from a flat one near the mean, even at the same average sten. Computed locally, not stored.

The profile demonstration is the iconic output of the instrument. Adjusting the sixteen stens redraws the connecting profile line, and the shape of that line — not any single factor — is what an interpreter reads. A profile that zigzags between extremes describes a very different person from a flat one near the mean, even when the two share the same average sten.

Roll primary factors into a global factor

The five global factors are obtained by factor-analysing the primaries. Choose a global factor and set its contributing primaries — some enter reversed — to see the broad sten form from the narrow ones.

This is the same upward reduction that took the trait lexicon to sixteen factors, applied a second time to reach five. Composition weights are illustrative equal weights; published scoring uses regression weights. Computed locally, not stored.

The global-factor demonstration exposes the hierarchy. Setting the stens of the primaries that feed a global factor — some contributing with a positive sign, some reversed — updates the broad global sten in real time, making visible how five broad dimensions are composed from the sixteen narrow ones. It is the same upward reduction, applied a second time, that took the trait lexicon down to sixteen factors in the first place.

Worked Example

Consider a respondent whose raw score on Factor A (Warmth) is 25, in a normative sample where that factor has a mean of 20 and a standard deviation of 4. The sten conversion first expresses the raw score as a standard score: z = (25 − 20) / 4 = 1.25. It then rescales to the sten metric, which has a mean of 5.5 and a standard deviation of 2: sten = 5.5 + 2 × 1.25 = 5.5 + 2.5 = 8.0. A sten of 8 falls in the high band, so this respondent is described as warm and outgoing on Factor A relative to the norm group.

Now roll several primaries into a global factor. Suppose the respondent's stens on the five primaries that define the Extraversion global factor are A = 8, F = 7, H = 8, N = 4, and Q2 = 3. Two of these — Privateness (N) and Self-Reliance (Q2) — contribute with a reversed sign, because being forthright and group-oriented is the extraverted direction, so their contributions are reflected as 11 − 4 = 7 and 11 − 3 = 8. The global Extraversion sten is the average of the five signed contributions: (8 + 7 + 8 + 7 + 8) / 5 = 38 / 5 = 7.6, which rounds to a global sten of 8.

Both steps are the same operation at different levels. The first standardises a raw score against a population; the second standardises a person against the set of primaries that compose a broad trait. A careful reader can recompute every step by hand, from the raw item totals to the global profile — the transparency that makes the instrument a good vehicle for teaching what a factor score is and is not.

Discussion

The Cattell Personality Factor Questionnaire earns its place in cognitive psychology less for the sixteen numbers it reports than for the method it made concrete. It was the most sustained attempt to treat personality structure as an empirical question answerable by factor analysis, and in doing so it exposed both the power and the indeterminacy of that approach. The power is real: from the redundancy of ordinary language, a repeatable procedure extracts a small set of dimensions that reconstruct much of how people differ. The indeterminacy is equally real: the same procedure, run with different defensible settings, yields sixteen factors or five, and no purely statistical criterion adjudicates between them.

Whatever the disputed number of factors, the instrument's scores are held to the ordinary psychometric standards of any test. Its reliability — the consistency of a factor score across the items that compose it — is indexed by internal-consistency coefficients of the kind Cronbach formalised (Cronbach, 1951), and its validity is argued not from any single criterion but from the pattern of convergent and discriminant relationships its scores hold with other measures, the logic of the multitrait-multimethod matrix (Campbell & Fiske, 1959). These are the standards against which the sixteen-versus-five debate is ultimately adjudicated: a factor is worth measuring only if it can be measured reliably and shown to behave as its interpretation predicts.

That tension is the instrument's lasting contribution. The 16PF and the five-factor model are not rivals so much as two cuts through one continuous reduction, chosen for different purposes — the sixteen primaries for fine-grained prediction and clinical description, the five globals for parsimony and cross-study comparison. A field that understands this reads a factor solution as a decision about grain, not a discovery of the true number, and treats the question of how many traits there are as inseparable from the question of what the description is for.

Current Directions

Contemporary personality research has largely moved past the contest over a single correct number of factors toward questions the count alone cannot answer. One line argues that the field's long focus on a few broad traits — whether five or sixteen — has come at the expense of the narrower, more predictive units beneath them, and that descriptive, predictive, and explanatory goals each call for a different level of the hierarchy rather than one privileged layer (Mõttus et al., 2020). The hierarchical instruments that now dominate self-report, such as the facet-level Big Five inventories, are in part a response to that argument, reintroducing the fine grain the 16PF always claimed while keeping the broad factors for comparability (Soto & John, 2017).

The 16PF itself remains a live research instrument rather than a museum piece. Recent work has used it to test structural hypotheses about personality that require its full sixteen-factor resolution — for instance, whether the covariance among traits changes with intelligence, the differentiation-of-personality hypothesis, which needs many narrow factors to detect the effect at all (Schermer et al., 2020). A broad review of the field places both the instrument and the debate it started within a discipline now more concerned with how traits develop, predict outcomes, and change than with fixing their number once and for all (Roberts & Yoon, 2022).

Common Misconceptions

The 16PF sorts people into sixteen personality types.

It does not. Each of the sixteen factors is a continuous bipolar dimension, and a person receives a score on every one at once; the description is a profile across all sixteen, not membership in a category (Cattell et al., 1970).

Factor analysis proved there are exactly sixteen personality traits.

Factor analysis constrains but does not fix the number of factors; how many to retain and how to rotate them are judgment calls, and the same lexical data have been resolved into five factors as defensibly as into sixteen (Boyle, 1989).

The 16PF and the Big Five are incompatible theories.

They are better seen as two levels of one hierarchy: the sixteen primaries cluster into five global factors that align closely with the five-factor model, so the instruments largely map onto each other (Gerbing & Tuley, 1991).

A sten score is a percentage or a raw count.

A sten is a standardised score with a mean of 5.5 and a standard deviation of 2, expressing standing relative to a norm group; the same raw score yields different stens against different normative populations (Cattell et al., 1970).

Glossary

Bipolar factor.

A trait dimension defined by two opposite descriptors at its ends, so that both a low and a high score are substantive descriptions rather than merely more or less of one quality.

Construct validity.

The degree to which a test measures the theoretical construct it claims to, established by the pattern of relationships its scores hold with other measures rather than by any single criterion.

Factor analysis.

A statistical method that reduces a large set of correlated variables to a smaller set of latent factors able to reproduce the observed correlations.

Five-factor model.

The framework holding that normal personality is well described by five broad dimensions, commonly labelled extraversion, agreeableness, conscientiousness, neuroticism, and openness; also called the Big Five.

Global factor.

One of the five second-order dimensions the 16PF obtains by factor-analysing the correlations among its sixteen primary factors; the broad layer that links the instrument to the Big Five.

Lexical hypothesis.

The premise that the personality differences most important in human life become encoded as single words in the natural language, so that the trait lexicon is a map of personality structure.

Primary factor.

One of the sixteen first-order source traits the 16PF scores directly; the narrow, specific dimensions from which the broad global factors are composed.

Q-data.

Personality information gathered from a person's own answers to a questionnaire, as distinct from observer ratings or objective test performance; the data source the 16PF relies on.

Reliability.

The consistency of a measurement across the items, occasions, or raters that produce it; for a factor scale it is commonly indexed by an internal-consistency coefficient such as coefficient alpha.

Sixteen Personality Factor Questionnaire.

The full name of the 16PF; a self-report inventory scoring normal adult personality on sixteen primary factors, developed by Raymond Cattell.

Source trait.

An underlying factor that causes a cluster of behaviours to co-occur, as opposed to the visible cluster itself; the sixteen primary factors are Cattell's candidate source traits.

Sten score.

A standard-ten score with a mean of 5.5 and a standard deviation of 2, running from 1 to 10, onto which 16PF raw scores are converted for norm-referenced interpretation.

Surface trait.

A cluster of behaviours that visibly co-occur; distinguished from the source trait that causes the co-occurrence.

Trait.

A relatively stable disposition to behave, think, or feel in a consistent way across situations; the unit of individual difference the 16PF is built to quantify.

Key Researchers

Gregory J. Boyle (b. 1950). Personality and individual-differences researcher; re-examined the higher-order factor structure of the Cattell, Comrey, and Eysenck scales, showing that the number of second-order factors recovered depends on the analytic choices made. Google Scholar

Heather E. P. Cattell. Clinical psychologist and 16PF interpreter; co-author of the fifth edition of the instrument and of its restructured global-factor framework, whose work carried Cattell's model into contemporary assessment practice. Fifth-edition paper

Raymond B. Cattell (1905-1998). Psychologist at the University of Illinois; originator of the 16PF and of the factor-analytic approach to personality structure, who distilled the trait lexicon into sixteen source traits and built the instrument to measure them, defining the research program this article surveys. Wikipedia

Robert R. McCrae. Personality psychologist; co-developer of the five-factor model whose cross-instrument validation work framed the debate over how the 16PF primaries relate to the broader Big Five dimensions. ORCID

Brent W. Roberts. Professor of psychology at the University of Illinois at Urbana-Champaign; a leading contemporary personality researcher whose syntheses locate the trait-structure debate within a field now focused on development, prediction, and change. ORCID

Julie Aitken Schermer. Professor at the University of Western Ontario; individual-differences researcher who has used the 16PF to test structural hypotheses about personality, including whether trait covariance differs by level of intelligence. ORCID - Google Scholar - Faculty Page

Frequently Asked Questions

What does the 16PF measure?
The 16PF measures normal adult personality on sixteen primary factors, each a continuous bipolar dimension such as reserved versus warm or trusting versus vigilant, reported together as a profile rather than as a single type (Cattell et al., 1970).

Where did the sixteen factors come from?
Cattell derived them by factor-analysing the trait vocabulary of ordinary language, starting from an exhaustive dictionary list of nearly eighteen thousand trait words and reducing it to a set of independent underlying dimensions (Allport & Odbert, 1936).

What is a source trait?
A source trait is an underlying factor that causes a group of behaviours to co-occur, as opposed to the visible cluster of behaviours itself, which Cattell called a surface trait; the sixteen primary factors are his candidate source traits (Cattell, 1943).

How is the 16PF related to the Big Five?
The sixteen primaries cluster into five broad global factors that align closely with the five-factor model, so the 16PF is best seen as a finer-grained relative of the Big Five rather than a competitor (Gerbing & Tuley, 1991).

Is personality really made of sixteen factors, or five?
The number depends on how deep into the hierarchy one looks and on the factor-analytic choices made; the same lexical data support five broad factors or sixteen narrow ones, and neither figure is forced by the statistics alone (Boyle, 1989).

What is a sten score?
A sten, or standard-ten score, re-expresses a raw factor score on a one-to-ten scale with a mean of 5.5 and a standard deviation of 2, so a person's standing can be read against a normative population regardless of the factor's raw metric.

Is the 16PF still used in research today?
Yes. It remains a live instrument for testing structural hypotheses that require its full sixteen-factor resolution, such as whether the covariance among personality traits changes with intelligence (Schermer et al., 2020).

How does the 16PF establish that its scores are valid?
Its validity rests on construct validation, the pattern of relationships its scores hold with other measures and criteria, a standard the field formalised in the mid-twentieth century and applied to trait instruments across methods and observers (Cronbach & Meehl, 1955).

References