Comparison · with sources
The five-factor model, Big Five, is the personality model the research stands behind. The Enneagram is the model people use to understand themselves. Those are not the same question – and the nine studies that have measured them against each other show exactly why.
This page is written by teachers who make a living from the Enneagram. We have still let the five-factor model win where it wins – because anything else would be dishonest to anyone who has studied psychology.
The five-factor model was not devised. It was found. The starting point was the lexical hypothesis: that the most important personality traits are encoded in language. Researchers took the words a language uses about people, gave them to thousands of respondents, and let factor analysis show how many independent dimensions there were. The answer was five. Tupes and Christal found them in studies of US Air Force officers around 1961, and Lewis Goldberg named them the Big Five in 1981.
The Enneagram, by contrast, was devised. The nine types come from Oscar Ichazo and were translated into Western psychology by the Chilean psychiatrist Claudio Naranjo from 1970. The model is top-down: it states in advance that there are nine patterns, defined by a basic fear and a basic desire, distributed across the head centre, the heart centre and the body centre.
That difference explains most of the rest. Big Five can document its structure because the structure came out of the data. The Enneagram cannot, because the structure came first.
| The Enneagram | The five-factor model | |
|---|---|---|
| Form of the result | One of nine types | A position on five continuous scales |
| Built | Top-down, from a theory of nine patterns | Bottom-up, by factor analysis of the words of a language |
| Ownership | Nobody owns the model | Nobody owns the model. Some instruments such as NEO-PI-3 are commercial, others such as IPIP are freely available |
| Typical reliability | Mixed. Good instruments above 0.70, but down to 0.35 on one RHETI scale and 0.17 on one OEPS scale | NEO-PI-3: alpha 0.87–0.94 at domain level |
| Stability over time | One peer-reviewed instrument measured 0.29–0.51 after four weeks | NEO PI-R: 0.63–0.83 after six years |
| Describes development? | Yes – direction under pressure and in security is part of the model | No. It describes level, not movement |
| Accepted for selection? | No | Yes, with moderate expectations |
The 2021 systematic review of Enneagram research pooled nine studies that had measured Enneagram scales against the five-factor model. The pattern was strikingly consistent:
| Type | Consistent association |
|---|---|
| 1 | Higher conscientiousness (9 of 9 studies) |
| 2 | Higher extraversion (9 of 9), higher agreeableness (7 of 9) |
| 3 | Higher extraversion (6 of 9), higher conscientiousness (5 of 9), lower agreeableness (5 of 9) |
| 4 | Higher neuroticism (8 of 9), higher openness (6 of 9), lower conscientiousness (6 of 9) |
| 5 | Lower extraversion (9 of 9) |
| 6 | Higher neuroticism (8 of 9), lower extraversion (6 of 9), lower openness (5 of 9) |
| 7 | Higher extraversion (9 of 9), higher openness (9 of 9), lower conscientiousness (6 of 9) |
| 8 | Lower agreeableness (9 of 9), higher extraversion (8 of 9), lower neuroticism (6 of 9) |
| 9 | Higher agreeableness (9 of 9) |
On one point this is a strong result for the Enneagram: the scales measure something real, and what they measure sits where the theory says it should. A type 5 really is more introverted. A type 8 really is less accommodating.
And it is at the same time the hardest objection: if the nine types are simply combinations of the five traits, does the Enneagram add anything? That question is not settled. Nobody has shown that Enneagram types predict something the five-factor model does not already predict.
That is the honest question, and the answer is not comfortable for any typology.
In 2012 Haslam, Holland and Kuppens reviewed all published taxometric research – the method designed to test whether a property is categorical or continuous. They covered 177 articles, 311 findings and 533,377 participants. The conclusion reads: «most latent variables of interest to psychiatrists and personality and clinical psychologists are dimensional». Once methodological weaknesses were controlled for, the share of findings pointing to genuine categories fell to 14 per cent – and normal personality was not among them.
That hits the Enneagram, the MBTI and DISC equally hard. Personality appears to be continuous, not sorted into boxes. Enneagram research points the same way: factor analyses typically find fewer than nine factors, and nobody has used cluster analysis to derive the nine types from data.
We still think the nine types are useful. But they are a model, not a discovery. That is an important difference, and anyone who claims otherwise has not read the research.
The five-factor model is strongest when someone has to be measured. Selection, research, screening, comparison across large groups. NEO-PI-3 exists in more than 50 languages, and freely available versions exist through IPIP, including in Danish. The price is that the result rarely changes anything: knowing you score high on conscientiousness is not a basis for action.
The Enneagram is strongest when someone has to understand themselves. It gives a language for why you react as you do, and what happens to you under pressure. The 2021 review did find that Enneagram-based programmes were associated in several studies with positive changes in leadership versatility, communication and team effectiveness.
Note the wording: the programmes worked. That is not the same as the typology being true. A good conversation about yourself with a competent teacher does good, whichever model it hangs on.
If you want to try it yourself, take a free Enneagram test or see our Enneagram courses. We also compare the Enneagram with the MBTI and with DISC, and we have gathered the evidence on the page The Enneagram and the research.
Sources: Hook, Hall, Davis, Van Tongeren & Conner (2021), «The Enneagram: A systematic review of the literature and directions for future research», Journal of Clinical Psychology 77(4) – the tally of the nine studies and the findings on factor structure and interventions. · Haslam, Holland & Kuppens (2012), «Categories versus dimensions in personality and psychopathology», Psychological Medicine 42, 903–920. · Sutton, Allinson & Williams (2013), European Management Journal 31(3), n = 416. · Yanartaş et al. (2022), the ETASI validation, Anatolian Clinic 27(3), n = 3,531. · NEO-PI-3 and NEO PI-R reliability figures from PAR Inc. and the University of Oslo compilation. · Sackett, Zhang, Berry & Lievens (2022), Journal of Applied Psychology. · Goldberg et al. (2005) on IPIP.
Tell us whether you want an open course, an in-house course or a talk, and we'll find the right solution and send a no-obligation quote.