Big Five Test Scores and Percentiles: What High, Low, and Average Results Mean

7 days ago

16 min read

Author: Ukrainian Psychological Hub · Published: October 9, 2026 · Editorial Policy

Ukrainian Psychological Hub · October 9, 2026

A Big Five personality test score summarizes responses to a specific questionnaire designed to estimate a personality trait. The result may be shown as a raw score, item average, standardized score, transformed 0–100 number, or percentile rank. Each format answers a different question. A high, low, or average interpretation is justified only after the instrument, scoring direction, metric, and relevant comparison data have been identified.

The essential distinction is that a percentile describes where a score ranks in a reference distribution. An Extraversion score at the 80th percentile, for example, is approximately above 80 percent of scores in the stated norm group, subject to the method for handling ties. It does not mean the person is 80 percent extraverted or will seek company 80 percent of the time. A Big Five test estimates recurring tendencies; it cannot reduce a person's capacities, mental health, or future to five numbers.

Quick answer: What do high, low, and average Big Five scores mean?

A higher score usually means more endorsement of questionnaire responses keyed toward the trait. A lower score suggests a position toward the other pole of the continuous dimension. An average score may refer to the mathematical mean, a median rank, or an instrument-defined middle band. These meanings differ. The Big Five model itself has no universal percentile cutoffs for labeling a result high, average, low, or very low.

A report that calls the 70th percentile high may be using a particular convenience band, while another may classify the same percentile differently. Such labels should be traced to the validated instrument and its normative sample. The jointly developed Standards for Educational and Psychological Testing emphasize evidence for the specific interpretations and uses of test scores.

What is the Big Five measuring?

The broad Big Five domains are Openness to Experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism, often remembered as OCEAN. The corresponding psychological constructs are continuous dimensions of individual variation, not categories of people. Many established questionnaires assess these domains, but their exact items, scales, facet structures, and scores differ.

The 60-item Big Five Inventory–2, or BFI-2, uses the domain labels Extraversion, Agreeableness, Conscientiousness, Negative Emotionality, and Open-Mindedness. The latter two correspond broadly to traditional Neuroticism and Openness, respectively, while retaining instrument-specific measurement definitions. See the original BFI-2 validation study and the Berkeley Personality Lab instrument documentation. This distinction prevents a report from silently treating two differently named and constructed scales as interchangeable.

Five score formats that people often confuse

1. Raw sum

Suppose a fictional ten-item Conscientiousness scale uses responses numbered 1 through 5. After correctly reversing negatively keyed items, its total might be 38 out of a possible 50. This is a sum of keyed answers. It is only meaningful in the context of the scale's scoring key, item count, and response options. It cannot be compared directly with a total of 38 from a different instrument.

2. Item mean

If those ten keyed responses sum to 38, their mean is 3.8 on the 1–5 response scale. The item mean removes the effect of item count when all items have compatible coding, but its meaning still depends on what the items say and whether they measure the intended trait. A score of 3.8 is not the 76th percentile simply because 3.8 is 76 percent of five.

3. Percentage of maximum possible or a 0–100 transformation

Some reports map a raw-score range onto 0–100. A hypothetical transformation for a response mean ranging from 1 to 5 is (mean − 1) / 4 × 100. A mean of 3.8 becomes 70 under that formula. This number means 70 on the transformed possible-response range; it supplies no information about how many people in a real population scored lower.

A published study cited in the Berkeley BFI documentation displays Big Five means in a percentage-of-maximum-possible metric, often abbreviated POMP. Whether an online website uses that exact metric or a different transform is a question for the website's scoring documentation, not a property of the Big Five model.

4. Standard scores: z and T

A z score compares a raw score with a reference distribution by subtracting that distribution's mean and dividing by its standard deviation. A conventional T score is defined as 50 + 10 times the z score. If a reported T score equals 60, the position is one reference-sample standard deviation above its mean. It does not mean 60 percent of the trait or the 60th percentile.

Under an approximately normal reference distribution, a z score of +1 is near the 84th percentile and a z score of −1 is near the 16th. These values are illustrative properties of the normal distribution. Empirical norms may depart from normality, use other transformations, or account differently for tied scores; exact percentile conversion requires the relevant tables or method.

5. Percentile rank

A percentile rank expresses relative standing within a defined group. At approximately the 25th percentile, the result sits lower than about three-quarters of that group's scores. At the 50th percentile, it lies near the distribution's median. At the 95th percentile, it is higher than most reference-group scores. None of these figures describes a percentage of personality possessed, a percentage of test answers correct, or a probability of a particular future action.

Precise percentile definitions vary in their treatment of ties and interpolation. If several respondents have an identical raw score, one method may count scores strictly below it and another may count a proportion of ties. For that reason, a responsibly prepared test report describes its procedure rather than pretending that all percentiles are exact universal facts.

Examples: the same-looking number can convey completely different information

Example A: Conscientiousness 70/100

A score chart showing 70/100 could represent a transformed item average, a proprietary index, or a percentile. Without a label specifying the metric, it is impossible to tell whether the respondent scored relatively higher than average in a population. The number alone is insufficient for a high-score classification.

Example B: Conscientiousness, 70th percentile

When the normative method is documented, the 70th percentile indicates a higher rank than approximately 70 percent of the reference group. That group may be national, local, age-specific, organizational, or a self-selected online sample. It does not guarantee a fixed degree of diligence in every setting. Whether the report calls the result above average or high depends on its own evidence-based terminology.

Example C: Extraversion, T = 60

If a test uses the usual norm-based T scale, a 60 indicates one standard deviation above the comparison mean. Assuming an approximately normal norm distribution, this may correspond to roughly the 84th percentile. The relationship is a mathematical illustration, not an instrument-independent prediction of sociability or a justification for describing someone as an extrovert type.

Example D: Negative Emotionality, response average 4.0

For a correctly keyed BFI-2 scale, an average of 4.0 on a five-point response format leans toward the high end of that instrument's allowable response scores. It does not automatically equal the 80th percentile. An Emotional Stability inventory may use the opposite score direction; reading the labels alone without the scoring key risks reversing the intended interpretation.

What percentile counts as high, low, or average?

No scientific authority has established one set of universal Big Five percentile bands across all questionnaires. Instruments sometimes provide interpretation ranges for their particular scoring system. A researcher may also define operational categories for a specific study, such as the upper and lower thirds of that study's sample. Those definitions should be described as study-specific or instrument-specific, not as natural personality types.

The Berkeley Personality Lab's BFI documentation explicitly explains that there is no official BFI-2 manual with published norms. A website may build percentile tables from some dataset, but it must identify that sample and its calculation rather than present the tables as universal official BFI-2 norms.

The International Personality Item Pool's norms guidance likewise warns against generic canned norms that are poorly matched to a target population and favors appropriately developed local comparisons. IPIP is a public-domain item and scale ecosystem with different implementations, as explained in Goldberg and colleagues' foundational paper. There is consequently no single authoritative 'IPIP Big Five 70th percentile' without a named scale and comparison dataset.

What does an average score mean in a normed personality test?

The arithmetic mean is obtained by averaging reference-group raw scores. The median is the point around which observations are divided into two equal-ranked portions. In a symmetric distribution these may be close; in a skewed distribution they can differ. A 50th-percentile standing is a median position, while an 'average range' may cover a band defined by a scoring manual or feedback system.

A person can be close to the sample mean on every Big Five domain and still have a distinctive personality, values, interests, abilities, and life history. An average score is not an absence of the trait. Similarly, a median rank in one reference group could be above or below the median of another group using the same instrument.

Why the norm group matters more than the adjective

The meaning of a percentile changes when the reference group changes. Consider one hypothetical respondent who answers a questionnaire in exactly the same way on two occasions. A comparison with graduate students and a comparison with the general adult public can place the same raw result at different ranks if their score distributions differ. Neither percentile is a universal property of the respondent; both are comparisons conditioned on the data.

A good normative report therefore describes whom the sample represents, how participants were recruited, its approximate size, age and language coverage, time period if relevant, and whether test administration matches the intended use. It may also discuss the limitations of online volunteer samples and the uncertainty attached to unusually extreme percentile estimates.

In occupational personality testing, Tett and colleagues found that using different reference groups could materially change standardized results in the samples they studied. That result cannot be generalized as a fixed size of error for every Big Five inventory, but it demonstrates the practical consequences of norm relevance.

Age, language, and culture

Norm choices may reasonably take age and language into account when research supports those decisions. They also require caution: subgroup norms can obscure or exaggerate differences depending on purpose and statistical assumptions. Cultural adaptation requires careful item translation, construct interpretation, and examination of comparability. Evidence supporting Big Five factors across many settings does not mean every questionnaire or lexical study reproduces identical dimensions and response patterns in all cultures.

Online test-taking samples

A sample with tens of thousands of online volunteers may provide stable estimates for people resembling those volunteers. It is not automatically a representative national sample. Individuals drawn to personality websites can differ from nonparticipants in age, education, motivation, technology access, or interest in self-description. Statistical precision within a biased sample does not eliminate selection bias.

What if the report names no norm sample?

Its percentile should be treated as an unverified ranking claim. Readers may still find the questions useful for self-reflection, but the site has not provided enough evidence to interpret the percentile as population standing. One should seek a technical manual, peer-reviewed validation, published normative method, or clearly documented dataset.

How to interpret each Big Five domain

The general descriptions below explain what relatively higher and lower endorsement can mean. The precise emphasis and lower-level facet composition depend on the questionnaire. All five domains are continuous, and position on a domain is neither a moral grade nor a diagnosis.

Openness to Experience

Relatively higher scores often correspond to curiosity, receptiveness to ideas, imagination, or aesthetic interest. Relatively lower scores may correspond to preferences for the familiar, concrete, or established. There is no direct conversion from Openness percentile to intelligence, creativity, educational attainment, or cultural value. Instruments differ in how strongly they emphasize intellectual versus experiential forms of openness.

Conscientiousness

Higher scores often reflect organization, persistence, self-control, deliberation, or dutifulness. Lower scores may reflect more flexible, spontaneous, or less systematically organized behavioral preferences. The result neither determines work performance nor diagnoses attention difficulties; effective behavior also depends on skill, incentives, support, context, and resources.

Extraversion

Higher Extraversion is commonly associated with sociability, assertiveness, engagement, and positive energy. Lower Extraversion often reflects quieter or more reserved preferences and may be described as introversion within this dimensional model. Both ends allow for relationships, accomplishment, and psychological well-being. A percentile says little about behavior in an isolated social event.

Agreeableness

Higher Agreeableness typically reflects patterns of cooperation, compassion, respectfulness, or interpersonal trust, while lower scores may reflect greater skepticism, directness, competitiveness, or willingness to challenge. These tendencies can have context-dependent advantages and costs. A score is not a judgment about whether someone is kind, ethical, or a good partner.

Neuroticism / Negative Emotionality

A relatively higher score on a traditionally keyed Neuroticism measure suggests more frequent or intense vulnerability to negative emotional experiences in the questionnaire's intended sense. A relatively lower score indicates the opposite tendency. Some instruments label the pole Emotional Stability; the BFI-2 uses Negative Emotionality. Specific Neuroticism instruments, norm interpretation, and related questions are treated in the English Hub's published Big Five Neuroticism Test guide. This Big Five score article does not claim that either end is a clinical diagnosis.

Domains, aspects, and facets are different levels of measurement

A domain is one of the five broad factors. An aspect or facet is a more specific component assessed by a particular instrument or hierarchical model. The exact components and their names are not fixed universally across all Big Five assessments. The BFI-2 assesses 15 specific facets within five domains according to Soto and John's original research. The Big Five Aspect Scales study identifies two aspects within each domain.

The NEO family uses its own detailed facet system; the statement 'the Big Five have exactly 30 facets' only makes sense when referring to an operationalization with that structure. A brief questionnaire cannot inherit the facet-resolution of a longer instrument merely because both refer to the five domains.

A broad domain average can conceal differences among its subcomponents. For example, someone might endorse items about organization but less often endorse sustained persistence. That pattern may be informative if the relevant instrument measures those subcomponents with sufficient precision. The BFI-2 short-forms study investigates how reductions in item count affect domain-level and facet-level measurement.

Can you compare percentiles across the five traits?

A five-domain profile usually displays five separate ranks, each relative to a particular distribution of scores for that domain. The numbers do not add up to 100, because they are not parts of a single whole. It is perfectly possible to rank high on several, all, or none of the five scales in a particular normative reference.

Suppose a person receives Openness at the 75th percentile, Conscientiousness at the 40th, Extraversion at the 20th, Agreeableness at the 65th, and Neuroticism at the 55th. This describes five separate relative positions. Calling Openness that person's 'strongest skill' would be a leap: a percentile does not measure skill. Calling the combination a discrete personality type would also go beyond the Big Five's continuous structure.

Numerical differences between percentile ranks are not themselves equal-interval measures of trait magnitude. A 20-point gap in percentile ranks on two different domains is not directly comparable with a 20-point raw-score gap. Use plain-language descriptions of relative standing rather than false precision.

Why do two Big Five tests give different results?

Two instruments may have different items, reverse-scoring rules, domain definitions, response formats, facet coverage, time frames, and norms. They may also differ in reliability and validity evidence. The Big Five model names broad trait territory but does not force all tests to return the same numeric score.

The NEO inventory, Big Five Inventory, BFI-2, various IPIP scales, Big Five Aspect Scales, and Ten-Item Personality Inventory are separate tools. The TIPI's original validation explicitly examines an extremely brief measure suitable when questionnaire length is a major constraint, with tradeoffs compared with longer scales.

A formal conversion between two tests would require specific research linking their scales in a relevant population. Copying a raw score from one instrument into another test's norm table is methodologically unsound. The same applies to comparing '72' on a transformed 0–100 website report with the 72nd percentile on a different service.

Can Big Five scores and percentiles change over time?

Yes. A person's observed score may change because of short-term psychological states, variations in the circumstances of testing, response error, new information about oneself, or genuine longer-term personality development. Two scores at different times are estimates based on responses, not direct recordings of an unchanging inner quantity.

Longitudinal evidence shows personality traits have meaningful rank-order continuity alongside change across life. Roberts and DelVecchio's quantitative review addresses rank-order consistency across ages; Srivastava and colleagues' large-sample research describes adult age patterns in the Big Five. Neither finding permits a confident inference of lasting change from a small difference between two online quiz results.

Even a completely unchanged raw score can produce a different percentile if the platform updates its comparison dataset. If a report labels movement from 65th to 73rd percentile as improvement, it should first distinguish altered norms, measurement uncertainty, and genuine change in the underlying trait.

Measurement error and confidence in the result

A questionnaire samples a limited set of behaviors and self-descriptions. All such scores have uncertainty. Precision depends on item quality, number of items, reliability, interpretation validity, administration, and population. A score with a confidence interval should be interpreted as an estimate with uncertainty, rather than a perfectly precise location.

An interval expressed in raw-score units is not automatically an equally wide interval in percentiles. Percentile ranks are nonlinear, and a score difference near the center of a distribution may translate into a different percentile movement than the same raw-score difference near a tail. An unusually extreme percentile may also be less stable when reference groups are small or poorly matched.

The joint testing Standards make clear the importance of score precision and evidence for a particular use. A convenient online chart can be adequate for reflection yet insufficient for high-stakes selection, clinical assessment, or decisive predictions about an individual.

Self-report bias, social desirability, and observer ratings

Self-report responses can be shaped by self-awareness, memory, motivation, response style, misunderstanding of an item, context, or deliberate impression management. Someone taking a private research survey may answer differently from a job applicant who perceives a desirable response. Observer ratings can provide additional evidence but introduce their own perspectives and limitations.

Acquiescence and extreme-response patterns can also vary across individuals and cultural settings, creating score differences partly attributable to response style rather than the intended construct. Instrument-specific reverse-keying and scoring methods can address some issues but do not automatically make every score equally comparable across languages, populations, or purposes.

What Big Five scores can and cannot predict

Personality studies document group-level associations between traits and multiple outcomes. Such associations can contribute to statistical prediction when appropriately estimated and validated. They are not causal proof, and they do not guarantee how a particular person will perform at work, respond in a relationship, succeed academically, remain healthy, create art, or experience psychological distress.

A relatively high Conscientiousness rank might align with a group-level tendency toward organized behavior, but professional achievement depends on opportunity, ability, role demands, motivation, and many other factors. Extraversion percentile cannot determine leadership potential. Agreeableness percentile cannot rate compatibility with a partner. Trait profiles are one potentially informative input in a larger context.

Clinical boundaries matter particularly. High Neuroticism or Negative Emotionality is not by itself evidence that someone has an anxiety disorder or depression. Lower scores cannot rule those conditions out. Big Five measurement is trait assessment rather than diagnosis. A diagnostic conclusion requires appropriately conducted clinical assessment.

How to evaluate the quality of your Big Five score report

Start by finding the exact questionnaire name and version. Next check the wording and direction of its five scales. Identify whether each displayed number is a raw sum, item mean, transformed percentage, standard score, or percentile rank. If it is a percentile, inspect the reference sample and calculation method.

Then locate reliability and validity studies relevant to the age group, language, administration mode, and intended interpretation. Examine whether the report indicates uncertainty and whether the website discusses missing data, retesting, response styles, and limitations. Finally, look for exaggerated conclusions such as 'your score proves you will be a successful manager' or 'this percentile means you have a disorder.' Such claims require evidence beyond a personality profile.

A report whose instrument is unclear or whose norms cannot be identified may still spark useful reflection. Its precise numerical rankings should receive correspondingly limited evidential weight. Transparency about methods is more valuable than an elaborate color chart or a label such as 'scientifically proven personality type.'

Frequently asked questions

Does the 50th percentile mean an average Big Five score?

It represents a middle rank in the specified norm distribution. The arithmetic mean can differ from the median, and an instrument may define a wider average band. The precise meaning depends on the report's conventions.

Is the 90th percentile an exceptionally good personality score?

No. It is a relatively high rank on that dimension within the stated reference group, not a quality grade or moral assessment. Whether a tendency is useful depends on circumstances and the trait being considered.

What does a 70/100 Big Five score mean?

The answer depends on the format. It could be a normalized raw score, a proprietary display number, or the 70th percentile. Never assume a percentile unless the report explicitly defines the number that way.

What percentile is officially high on the Big Five?

There is no single official cutoff shared by all Big Five instruments. A specific measure or norm table may define ranges for its intended use; those definitions cannot be transferred automatically to another test.

Can I have high scores on all five traits?

Yes. The five domain scores are independent measures in the sense that their percentile ranks are not constrained to total 100. Correlations between domains may affect the frequencies of some combinations, but there is no fixed personality-score budget.

Does a low score mean I lack that trait completely?

No. It indicates a lower standing on a continuous measured dimension or a lower rate of endorsement of relevant questionnaire items. Most people can display behaviors associated with either end in appropriate situations.

Does a high Neuroticism percentile mean I have an anxiety disorder?

No. It describes a position on a personality dimension assessed by a particular instrument. A mental-health diagnosis requires a separate clinical evaluation and cannot be inferred from that percentile.

Why did two websites give me different percentiles?

They may use different item sets, scoring rules, norms, language versions, or assessment lengths. Some sites publish transformed raw scores as 0–100 displays. Check the technical documentation before comparing the results.

Does BFI-2 have official normative percentiles?

According to the Berkeley Personality Lab, there is no official BFI-2 manual with published norms. An independently developed percentile report should state its exact dataset and conversion method.

Should I trust a very short ten-item test?

A ten-item Big Five measure may be useful when brevity is essential, especially for group research. Its limited item count constrains precision and nuanced interpretation. Read the intended uses and evidence in the TIPI validation study.

Can my score change without my personality changing?

Yes. Changes can arise from random measurement variation, temporary circumstances, different instructions, or updated normative data. Meaningful trait change requires stronger evidence than one fluctuating number.

Can employers or clinicians make decisions from a percentile?

A high-stakes interpretation requires fit-for-purpose validity, suitable norms, fairness analysis, and relevant professional or legal safeguards. A generic personal-report percentile cannot by itself determine suitability for a job or establish a clinical diagnosis.

References

American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (2014). Standards for Educational and Psychological Testing. Official resource.

DeYoung, C. G., Quilty, L. C., & Peterson, J. B. (2007). Between facets and domains: 10 aspects of the Big Five. Journal of Personality and Social Psychology, 93(5), 880–896. DOI.

Goldberg, L. R., Johnson, J. A., Eber, H. W., Hogan, R., Ashton, M. C., Cloninger, C. R., & Gough, H. G. (2006). The international personality item pool and the future of public-domain personality measures. Journal of Research in Personality, 40(1), 84–96. DOI.

Gosling, S. D., Rentfrow, P. J., & Swann, W. B., Jr. (2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37(6), 504–528. DOI.

International Personality Item Pool. (n.d.). Norms and data sets. Oregon Research Institute. Official site.

John, O. P., & Soto, C. J. (n.d.). Big Five Inventory (BFI-2): Measures, scoring, and norms. Berkeley Personality Lab. Official site.

Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126(1), 3–25. DOI.

Soto, C. J., & John, O. P. (2017a). The next Big Five Inventory (BFI-2): Developing and assessing a hierarchical model with 15 facets to enhance bandwidth, fidelity, and predictive power. Journal of Personality and Social Psychology, 113(1), 117–143. PubMed.

Soto, C. J., & John, O. P. (2017b). Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS. Journal of Research in Personality, 68, 69–81. DOI.

Tett, R. P., et al. (2009). The use of personality test norms in work settings: Effects of sample size and relevance. Journal of Occupational and Organizational Psychology. DOI.

Psychological Encyclopedia