← All insights
Science Personality tests explained

16PF Personality Test: How Reliable Is Cattell's Test?

The CQ Team · 8 min read ·

The 16PF measures 15 traits and reasoning. Its scales score .68 to .87 on reliability in 10,261 adults. What it measures, and where evidence runs thin.

The 16PF is a reliable enough personality questionnaire, with gaps you should know about. In its 1993 US standardization sample of 10,261 adults, the 15 personality scales had internal consistency from .68 to .87, according to Del Giudice, Booth and Irwing (2012). That is solid for describing a person, and a little below the NEO PI-R, the standard Big Five inventory, in one direct comparison. Its retest figures sit in the publisher's 1994 technical manual, which we could not open, so this post does not quote them.

Who developed the 16PF?

Raymond Cattell built it from the words people use to describe each other. In his 1946 book Description and Measurement of Personality, Cattell started from the list of "some 4500 trait terms" compiled by Allport and Odbert. Grouping synonyms cut it to about 160, and he added technical terms to reach 171. Factor analysis then gave him the primary traits, which the publisher still describes as "identified through factor analysis".

Where do you land? The free MBTI test is 12 questions, about 3 minutes, no sign-up.

Start the test →

The questionnaire was first published in 1949 and revised in 1956, 1962, 1968 and 1993, as set out by Cattell and Cattell (1995). The 1993 revision is the fifth edition, the version the studies below use. The current publisher, Talogy, now names a sixth edition that it delivers as a computer-adaptive test.

What does the 16PF measure?

Fifteen personality scales and one ability scale. The fifth edition has 185 items: the 15 personality scales, a 15-item Reasoning scale and a 12-item Impression Management scale that flags answers polished to look good (Del Giudice and colleagues, 2012). So the sixteenth factor is a short reasoning test, a detail most summaries miss.

Each scale runs between two poles. The alpha column shows how consistently the items in each scale agreed with each other in the 10,261-person sample, where 1.0 would be perfect agreement.

ScaleLow pole vs high poleAlpha
WarmthReserved vs warm.69
Emotional StabilityReactive vs emotionally stable.79
DominanceDeferential vs dominant.68
LivelinessSerious vs lively.73
Rule-ConsciousnessExpedient vs rule-conscious.77
Social BoldnessShy vs socially bold.87
SensitivityUtilitarian vs sensitive.79
VigilanceTrusting vs vigilant.73
AbstractednessGrounded vs abstracted.78
PrivatenessForthright vs private.77
ApprehensionSelf-assured vs apprehensive.80
Openness to ChangeTraditional vs open to change.68
Self-RelianceGroup-oriented vs self-reliant.79
PerfectionismTolerates disorder vs perfectionistic.74
TensionRelaxed vs tense.79

The 15 scales group into five global factors. Some scales feed more than one global, which is why Warmth and Social Boldness each appear twice below.

Global factorBuilt fromClosest Big Five trait
ExtraversionWarmth, Liveliness, Social Boldness, Privateness, Self-RelianceExtraversion
AnxietyEmotional Stability, Vigilance, Apprehension, TensionNeuroticism
Tough-MindednessWarmth, Sensitivity, Abstractedness, Openness to ChangeOpenness, reversed
IndependenceDominance, Social Boldness, Vigilance, Openness to ChangeNone clear
Self-ControlLiveliness, Rule-Consciousness, PerfectionismConscientiousness

The groupings and the Big Five matches come from the same paper. Its authors write that Independence "has no clear-cut analogue" in the Big Five.

How reliable is the 16PF?

On internal consistency, adequate. The range of .68 to .87 comes from a large sample that was spread across US regions in proportion to population, though more educated than the US average. Two scales, Dominance and Openness to Change, sit at the bottom of the range at .68. Kaiser, Del Giudice and Booth (2020) put the average for the original scales at .76, against .83 for free public-domain versions built from the International Personality Item Pool.

Translations vary more. In 636 Spanish students, Aluja and Blanch (2003) found alphas from .54 to .84 on the Spanish fifth edition.

Retest reliability, whether you get the same scores weeks later, is the number most people want, and it is the hardest to check. The figures live in Conn and Rieke's 1994 technical manual, which is not openly available. Figures that circulate online disagree with each other, so we leave them out. The only retest study we could open covers Form E, a different low-literacy version: across one week in 100 rehabilitation clients, its scales ranged from .43 to .78, with a median of .68 (Bolton, 1993). That figure does not transfer to the fifth edition.

Gregory Boyle adds a caution in his 2008 chapter in the SAGE Handbook of Personality Theory and Assessment: a high alpha can also mean the items repeat each other, so it says little about whether a scale is stable over time.

Do the 16 factors hold up?

The answer depends on which study you read, and the newest one is the strongest. Noller, Law and Comrey (1987) gave the 16PF and two other inventories to 669 Australians and found the 16PF scales "highly overlapping", which suggested fewer than 16 separate traits. A year earlier, Cattell and Krug (1986) reviewed more than 50 studies and concluded there were no fewer than 16.

Both predate the fifth edition. Irwing, Booth and Batey (2014) tested the fifth edition on the full US sample split in two, 5,130 people to build the model and 5,131 to check it. The primary factors fit well. The five global factors showed slight misfit. Of the three, this study uses the current edition, the largest sample and confirmatory methods, so it carries the most weight: the 15 scales hold together, and the five-factor summary on top of them is looser.

Whether the extra detail earns its place is a separate question. Across 16 data sets, Mershon and Gorsuch (1988) found that moving from six factors to 16 raised prediction in 88% of them.

How does the 16PF compare with the Big Five?

They overlap on four broad traits and part ways on a fifth. Rossier and colleagues (2004) gave 386 people both the 16PF fifth edition and the NEO PI-R, the standard Big Five inventory. The two "roughly measure the same aspects of personality", but Agreeableness is not represented in the 16PF, and the 16PF data pointed to four factors rather than five. They judged the NEO PI-R the more reliable of the two. Boyle answered that their study reported no retest data, so it could not settle reliability.

The practical reading: if you want a broad Big Five profile, a Big Five test covers Agreeableness and the 16PF does not. If you want narrow detail, such as the difference between being dominant and being socially bold, the 16PF splits traits finer. For how these models stack up on accuracy more generally, see how accurate personality tests are and our HEXACO test guide.

Is the 16PF the same as the MBTI?

No. The 16PF gives you a score on every scale. The MBTI sorts you into one of 16 types from four either-or choices.

They do overlap on one trait. In 44 family-medicine residents who took both, the MBTI and 16PF measures of extraversion correlated at r = -.71, which the authors call highly correlated (Eicke and colleagues, 1993). On the type question, McCrae and Costa (1989) studied 468 adults and found no support for the MBTI's types as distinct categories, though its four scales did track four of the Big Five. Our post on whether the MBTI is accurate covers that evidence.

Can the 16PF predict job performance?

The publisher says its report scores 20 to 25 job competencies "that can be used to predict on-the-job success". Its product page cites no study for that claim. We searched for a peer-reviewed meta-analysis of 16PF scales against job performance and found none.

If an employer asks you to take it, one question tells you a lot: which study links this scale to this job? Our guide to pre-employment personality tests covers what employers may and may not do with the results.

Where can you take a 16PF-style test?

The official 16PF takes 30 to 40 minutes, and Talogy lists "accreditation required" for anyone giving it. For a free option, the International Personality Item Pool's 16PF-style scales are public domain, with alphas from .73 to .86.

Our free 16 personality factors test is a 16-statement test in the style of Cattell's model. It scores the five global factors in about three minutes. It is our own short measure, separate from the official 16PF questionnaire, so the reliability figures in this post do not apply to it. Use it as a first look at where you might sit, and browse the rest of our personality tests hub for the evidence behind other models.

Frequently asked questions

Is the 16PF a projective test?

No. It is a self-report questionnaire: you answer items about yourself, and each scale is scored from your answers. Its 185 items include a 12-item Impression Management scale that flags answers polished to look good.

How long does the 16PF take?

About 30 to 40 minutes for the current official version, according to the publisher. The fifth edition has 185 items.

Do men and women score differently on the 16PF?

On some scales, by a lot. In the 10,261-person US sample, Sensitivity showed the largest gap, with women scoring higher by d = 1.34, more than one standard deviation (Del Giudice and colleagues, 2012). Combining all 15 scales, the observed profiles of men and women overlapped by about 29%. The paper's model-based estimate, which corrects for measurement error, was larger still. A replication with 21,567 US adults on public-domain 16PF-style items found a combined gap its authors call comparable, if somewhat smaller (Kaiser and colleagues, 2020). These are group averages and say nothing about any one person.

Is there a free version of the 16PF?

The official questionnaire is sold through accredited users. The IPIP's public-domain 16PF-style scales are free to use.

What should you remember about the 16PF?

The 16PF gives a finer picture than a Big Five test, and its 15 scales hold together in the largest study of the current edition. Its public evidence is thinner than its reputation: the retest figures are locked in a 1994 manual, and no meta-analysis backs its job predictions.

How we checked: we read the full text of Del Giudice 2012, Kaiser 2020, Aluja and Blanch 2003, Boyle 2008, Cattell's 1946 book, the Talogy product page and the IPIP 16PF key. For Cattell and Cattell 1995, Irwing 2014, Rossier 2004, Noller 1987, Cattell and Krug 1986, Mershon and Gorsuch 1988, Eicke 1993, McCrae and Costa 1989 and Bolton 1993 we used the abstracts.

Curious where your needle sits? Take any assessment free and find out in minutes.

Take a test →

Keep reading

All Personality tests explained articles →
Guides

Pre-Employment Personality Tests: How Hiring Uses Them

7 min read →
Science

HEXACO personality test: what it measures and misses

8 min read →
Psychology 101

Enneagram Types: The 9 Types Explained

8 min read →