Can You Compare a Personality Domain and Facet Score?

Can you compare the numbers directly?

If two personality reports seem to disagree before a coaching conversation, resist the urge to subtract one score from the other. A broad domain and a narrower facet usually summarize different amounts or kinds of response content. Reading the numbers side by side does not make a difference, rank, or conversion meaningful. Direct numerical comparison requires evidence linking these specific scales for the population and purpose at hand. Without that evidence, keep the scores separate and compare each written description.

There are three reasonable outcomes. A supported numerical comparison is possible when a documented method relates the exact scales and states its population and limits. A descriptive-only comparison is appropriate when the scales appear related but have not been shown to share a score meaning. An unresolved comparison is the honest result when the reports do not provide enough information to tell what their scores represent. The Standards for Educational and Psychological Testing, a joint publication of AERA, APA, and NCME, addresses valid score interpretation and testing uses; its general guidance cannot establish a link between an unknown pair of personality scales.

The ETS report Principles and Practices of Test Score Equating explains equating as a way to give scores from test editions the same meaning when that is the intended purpose. It also notes that conversion errors can affect validity and fairness. The next step is to check whether the publishers document a suitable link. Until then, pause the arithmetic: describe the apparent difference without treating it as a measured distance between the two results.

Sources: Principles and Practices of Test Score Equating; Standards for Educational and Psychological Testing

What does a domain-to-facet difference mean before you look at either number?

A domain score gives a broad summary of a set of responses; a facet score describes a narrower part of that territory. The difference is about scope before it is about size. A domain may combine several component facets, but a facet printed in a second report is not automatically one of those components. So when one report gives a broad domain and another gives a narrower facet, the first question is not which number is higher. It is what each number summarizes, and how each publisher built that scale. The NEO-PI-3 offers a concrete, bounded example of a documented hierarchy. Its sample interpretive report says that each of five domains comprises six facets and that its domain factor score is a weighted combination of all 30 facet scores. The same report cautions that wide variation among facets within a domain can make the factor interpretation more complex. This describes how that particular NEO report constructs and interprets its scores. It does not establish that a similarly named facet in another assessment is one of the NEO components, receives the same weight, or can explain the NEO domain result. A domain score is not necessarily a simple average: the NEO example itself specifies weighting, so a reader should not reverse-engineer a domain from displayed facet numbers unless the manual explains how. “Facet” is a useful description of level, not a universal blueprint. In “Domains and Facets: Hierarchical Personality Assessment Using the Revised NEO Personality Inventory,” the authors describe six facet scales within each of five broad domains for the NEO-PI-R. They also explain that the choice of six was a design decision, not a fixed rule that every instrument must follow. Their account further notes that facet scales sample aspects of their domains rather than exhaust every possible aspect. That is a reminder to check the specific instrument’s definitions: two publishers may divide related content differently, choose different boundaries, or use the same label for scales with different coverage. The word “facet” by itself does not tell you which items are included or what the score represents. Related constructs can overlap without being identical. The same 1995 discussion examines how other measures converge with NEO domains while differing in their conceptualizations, with facet-level relationships helping reveal those differences. In practical terms, two scores might concern neighboring aspects of a broad trait and still answer different questions. One scale may summarize several kinds of content; another may focus on one selected aspect. Even if their scores tend to move together in some evidence, that association would not make their meanings or units interchangeable. Similarity is a reason to compare definitions carefully, not permission to translate one number into the other. A useful first pass is to read each report’s scale description and scoring notes side by side. Record the exact instrument and version, the construct definition, the level of the score, and whether the publisher says the domain is calculated from facets, items, or another documented structure. Then ask whether the facet’s definition falls inside that domain’s stated scope, and whether either report describes a formal relationship to the other instrument. If those details are absent, mark the relationship as unknown instead of filling the gap with a familiar label. Keep each report’s wording attached to its own score. This distinction also protects against two opposite mistakes. Treating a facet as if it were the whole domain overstates its reach; assuming a domain fully describes every narrower tendency can erase differences that matter to the reader’s question. The NEO sample report and the NEO-PI-R discussion support that point within their named frameworks, where domains summarize and facets add detail. They do not license a cross-report conclusion about an unspecified pair. For that, the scale definitions and any evidence linking them would need to be examined directly.

Sources: NEO-PI-3 Self-Interpretive Sample Report; NEO-PI-3; Domains and Facets: Hierarchical Personality Assessment Using the Revised NEO Personality Inventory; The Facet Structure of Personality: A Review and Future Directions

What evidence would make a numerical comparison defensible?

A numerical comparison is defensible only when evidence relates these particular scores in a way that fits the question being asked. A report’s broad domain and another report’s narrower facet can both use numbers, but that shared format does not make subtraction meaningful. Ask whether the publisher or an independent study has documented a method for connecting the named scales and versions, who was studied, and what interpretation the method supports. If that evidence is missing, the careful answer is that the scores cannot be placed on a common numerical scale from the reports alone. Score linking is a method for relating scores from different test forms or instruments. Equating is a narrower form of linking used when scores are intended to have the same meaning, such as successive editions of one assessment. The distinction matters here: two personality reports may address related tendencies without measuring the same content at the same level. A linking study could support a bounded conversion or comparison, but it does not make every score from either report interchangeable. Its purpose and limits determine what the numbers can mean. A correlation is not itself a conversion. If scores move together across a sample, that can support an association between them; it does not by itself show that a given score on one scale corresponds to a particular score on the other. Nor does a shared framework or similar label establish that the scales have matching item content, score construction, or interpretation. To convert, a reader needs a reported linking procedure and enough information to judge its assumptions, not merely evidence that the measures overlap. The title Principles and Practices of Test Score Equating describes equating as appropriate when successive test editions are expected to preserve the same score meaning. Its abstract also notes that errors in an equating function or score conversion can affect examinees and raise validity and fairness concerns. This is methodological guidance from testing programs, not evidence that any two personality scores have already been linked. When reviewing a claimed link, look for the exact instrument names and versions, the sample and population, the design used to relate scores, the resulting conversion or relationship, and uncertainty around that result. It should say whether participants took both measures, whether shared items or another design supplied the connection, and why it fits the scales. It should also identify the purpose the evidence can support. A result from one population may not settle interpretation for another. These are questions to ask of the actual technical report; they do not presume that any particular study has answered them. The Standards for Educational and Psychological Testing, jointly produced by AERA, APA, and NCME, cover appropriate score interpretation, fairness, and workplace testing and credentialing. The overview places test interpretation and use within a wider account of evidence and context. The standards do not certify a particular domain-to-facet comparison, and the overview is not a substitute for the full standards or the relevant instrument documentation. It offers a reason to ask what interpretation is supported and for whom, rather than treating a conversion as a neutral formatting choice. The practical test is therefore specific: does the evidence connect these scales, in versions and a population relevant to the reader, for the interpretation at issue? If yes, state only the comparison the method warrants and keep its uncertainty visible. If the evidence establishes association but no conversion, describe the relationship without translating one score into the other. If no suitable linking evidence is available, leave the numbers separate and compare their written meanings only. A publisher’s documentation or an appropriate linking study could change that conclusion; a similar label, common framework, or matching score display alone cannot.

Sources: Principles and Practices of Test Score Equating; Standards for Educational and Psychological Testing

Why do matching score formats and percentiles still fail to settle the comparison?

A familiar number format can create a false sense that two reports express the same measurement. A raw score is a total or average produced from particular responses under a scoring rule. Its size may depend on item count, response options, reverse-scored items, and weighting. A raw total of thirty in one report cannot be compared with thirty in another merely because the digits match; the scales may contain different items or assign different values to responses. A standardized score is transformed and expressed on a defined scale. “T score” is a familiar label, but seeing it on both reports is not enough: each may summarize different content and use its own transformation and reference group. The NEO-PI-3 product page says that its domain and facet scores are converted to T scores based on the selected normative comparison group. That describes this named version, not all personality reports. A percentile is a position relative to a comparison group. It does not mean “percent of the trait,” and the same percentile in a different group does not establish the same standing. The opened explainer “Why Two Personality Tests Give You Different Results” notes that different items, scoring, and comparison samples can yield different Big Five percentiles. This is useful orientation, not evidence about a reader’s specific pair. Check each report’s own documentation for its score definition and reference group. A verbal band such as “low,” “average,” or “high” adds another layer. The provider chooses boundaries and may base them on norms, a criterion, or an interpretive convention. Unless the report explains the basis, the label is not a shared unit. Two “average” descriptions can refer to different distributions or scale content. Record the band definition if provided; matching labels do not establish equivalent scores. When results seem to conflict, identify what has been compared. The difference might reflect broad versus narrow content, different items or scoring weights, instructions asking about different time frames, a different norm group, an older or newer version, or a language adaptation. These are possibilities to investigate, not automatic explanations for a particular gap. A display alone cannot choose among them. A compact record helps keep the visible number from dominating: | Field | Why it matters | If missing | |---|---|---| | Instrument and version | Identifies the measure and applicable documentation | Do not assume a familiar label means the same form | | Scale definition and level | Shows whether the score is a domain or facet and what it covers | Do not treat the scales as equivalent | | Scoring, metric, and norm group | Distinguishes score type and reference population | Do not subtract or rank the numbers | | Language and response instructions | Establishes which version and administration were used | Ask the provider for details | | Evidence connecting the scales | Shows whether a numerical relation has been studied | Keep the comparison descriptive or unresolved | This is a record of missing information, not a scoring method. The National Council on Measurement in Education’s overview of the *Standards for Educational and Psychological Testing* describes standards concerned with valid score interpretation, fairness, and uses including workplace testing. That supports asking whether an interpretation fits its evidence and intended use; it does not validate a link between two personality scales. A shared framework label, raw-score range, T score, or percentile can prompt investigation, but supplies neither scale definitions nor linking evidence. The practical rule is to compare meaning before format. If both reports clearly identify the score, version, and reference group, describe each result within its own system, then ask whether evidence relates the scales. If details are incomplete, keep the numbers separate and compare only the written meanings the documentation supports. If the construct or score basis is unclear, leave the comparison unresolved and request the manual or technical explanation before interpreting a difference.

Sources: Why Two Personality Tests Give You Different Results; NEO-PI-3; Standards for Educational and Psychological Testing

When can a broad domain hide meaningful internal variation?

A broad score can summarize a profile whose component scores do not all point in the same direction. In that case, the components may qualify the broad description. The NEO-PI-3 sample report says that wide scatter among its facet scores makes interpretation of its factor score more complex, and that facet statements can occasionally conflict with domain statements. This is a warning about interpreting the domain and facets inside that NEO report. It does not show that a facet from another assessment explains, equals, or corrects the NEO domain. The distinction matters because aggregation answers a different question from a narrower scale. A domain gives a broad summary across related response content. Its component pattern can show whether that summary describes an even profile or compresses differences among narrower tendencies. A facet, by design, focuses attention on a more specific part of that content. It may add detail about one area while leaving the rest of the broader domain unmeasured or unreported. Reading the facet as a full account of the domain would therefore exceed what the narrower result says. The NEO-PI-3 sample report provides an illustration. It describes five domains, each with six facets, and says its factor scores are weighted combinations of all 30 facet scores. The report explains that broad factor scores describe a person globally, but that wide scatter across facets can make the factor interpretation more complicated. They do not establish that another instrument uses the same facets, weights, items, or score construction. There is a counterpoint to treating every visible facet difference as the best explanation. In “Domains and Facets: Hierarchical Personality Assessment Using the Revised NEO Personality Inventory,” the discussion recommends studying all facets within a domain together and in relation to the domain score. Selecting only a few high or low points, it argues, can produce a superficial reading. That caution is useful: a single striking facet may attract attention while the rest of the profile changes its meaning. The paper also presents domains as a useful starting point, with facets adding detail. Broad scores and narrow scores can each contribute, provided the reader keeps their different scope in view. Consider a schematic example, not a real report: one assessment summarizes several aspects of a broad domain, while a second reports one narrow tendency that might sit somewhere within related content. The narrow result could help frame a specific question, but it cannot reveal how the unreported aspects of the domain scored. Nor can it show that the first assessment's broad result is mistaken. The reports may be describing related territory at different levels, or they may define the territory differently. Without the scoring documentation or evidence linking the two scales, the apparent tension remains a prompt for inquiry rather than a resolved measurement difference. This distinction separates internal variation from cross-report mismatch. Internal variation is visible when a report provides a domain and its own facets, with documented relationships among them. A cross-report mismatch concerns two separately developed score scales; matching labels or a plausible conceptual overlap cannot tell the reader how their values relate. Each report must establish its own content, scoring, and interpretive basis before comparison. For the reader, the practical implication is to treat a facet as a specific clue only when its definition and relationship to its own domain are clear. Read the domain alongside the complete set of reported facets when available, rather than selecting one appealing detail. Then inspect the other report on its own terms. Ask what the narrower score covers, what the broader score combines, and whether the publisher documents a relationship between them. If those details are missing, the responsible conclusion is limited: the descriptions may suggest a question to explore, but the facet does not automatically explain the broad score elsewhere.

Sources: NEO-PI-3 Self-Interpretive Sample Report; Domains and Facets: Hierarchical Personality Assessment Using the Revised NEO Personality Inventory

How much precision can you expect from a narrower facet?

A narrower facet can answer a focused question, but narrower does not mean more precise. Its wording may be more relevant to a particular reflection; the score’s precision still depends on the items, their consistency, and evidence for the interpretation. A detailed label alone does not show that a result is stable or exact. Reliability is evidence about score consistency or precision under specified conditions. It does not by itself show that a scale measures the intended construct. Internal consistency, often summarized with Cronbach’s alpha, describes how much items in a scale tend to vary together in a sample. It is one kind of reliability evidence, affected by item count and content breadth. It is not an accuracy percentage, nor does a group estimate give an individual reader a personal margin of error. The study “Short and extra-short forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS” gives a concrete example. It examined a 30-item short form, the BFI-2-S, with two items for each of its 15 facets, alongside the 60-item BFI-2, whose facets use four items each. Across an internet validation sample of 2,000 people and a university sample of 423, full BFI-2 facet alpha averaged .74 or .75, depending on sample; the two-item BFI-2-S facets averaged .60 or .61, with estimates ranging from .39 to .79. The researchers attributed the lower internal consistency partly to the short scales’ two-item length and choice to preserve content breadth rather than maximize item similarity. That result does not mean every short facet is unusable. The same study found that BFI-2-S facets retained information in its validation work and may be useful for examining facet-level associations in reasonably large samples, around 400 or more observations. The authors advised against using the 15-item BFI-2-XS, with one item per facet, to assess facets. Group estimates cannot establish how uncertain one reader’s score is; they show why item count and intended interpretation matter. The publisher’s technical information for the NEO-PI-3 Normative Update provides a separate example. PAR reports that its 2024 update included 1,855 Self-Report respondents aged 12 and above and 1,200 Informant Report respondents. For self-report, reported alpha ranges from .87 to .94 across five domains and .54 to .84 across 30 facets, with a facet median of .75. Informant values range from .87 to .96 for domains and .60 to .89 for facets, with a facet median of .80. These estimates apply to the named update and samples, not another report. It would be misleading to compare these NEO figures directly with BFI-2-S figures and declare one instrument more precise. They concern different instruments, versions, samples, and reporting contexts. A small difference between two personal results is hard to interpret when scales differ in content or item count, especially if neither report provides uncertainty estimates. For a specific facet, check its item count and whether technical documentation reports reliability or measurement error for the relevant version and population. Match evidence to the conclusion: support for group-level associations does not automatically establish precise interpretation of one person’s score. If a report gives a confidence interval or standard error, follow its explanation. If it gives only a point score and label, note that it does not show the degree of uncertainty. Use a facet’s narrower wording to ask a more specific question, not to assume a more trustworthy answer. A brief facet can prompt reflection when its items and limits are documented. Treat small differences cautiously when uncertainty is not shown; detail alone does not make a score more accurate.

Sources: Short and Extra-Short Forms of the Big Five Inventory–2: The BFI-2-S and BFI-2-XS; NEO-PI-3

How should you classify the comparison you actually have?

Use three outcomes to classify the comparison: a bounded numerical comparison when suitable evidence links the exact scales; a descriptive comparison when their meanings are related but not equivalent; and no conclusion when essential definitions or score documentation are missing. Similar wording is a reason to investigate, not proof that two numbers share a scale. For an illustrative comparison, imagine one report presents a broad domain score and another a facet with a similar label. Record each instrument and version, construct definition, score level, scoring method, metric, norm population, reliability or uncertainty information, intended use, and evidence connecting the scales. The definition shows the intended content; score level and scoring method explain how the displayed result was built. The norm population identifies the comparison reference when the result is norm-referenced. Reliability evidence does not show that scales mean the same thing; intended use bounds the conclusion. The connection field is decisive. The Standards for Educational and Psychological Testing, jointly produced by AERA, APA, and NCME, addresses fairness and appropriate score interpretation, including workplace testing. It supports context-appropriate interpretation, not a particular personality-score pairing. Principles and Practices of Test Score Equating describes equating where successive test forms are expected to carry the same meaning, and warns that conversion errors can affect validity and fairness. Its examples concern testing programs, not these reports. A numerical crosswalk therefore needs a documented basis, assumptions, and limits. The reader should be able to identify the evidence source, the particular scores it connects, and the context in which the conversion was evaluated, rather than relying on a label or a general claim that the tests are similar. If a technical report identifies both instrument versions, the relevant population, linking design, uncertainty around converted values, and supported purpose, a bounded numerical statement may be justified. A link in one population or use does not automatically apply to every group, version, or decision. Correlation or a shared framework may show association, not interchangeability. If the manuals define related content but no suitable linking evidence appears, classify the result as descriptive. One report offers a broad summary; the other points to a narrower pattern. Keep each value attached to its own instrument and reference group; do not subtract, average, or rank them as though arithmetic settled the question. In an illustrative case, a broad domain could summarize several components while a second report describes only one. It may refine reflection on that tendency, but cannot explain the broad score by itself. The NEO-PI-3 sample interpretive report shows why this distinction matters within one instrument. It says its five domain factor scores are weighted combinations of 30 facet scores, six under each domain, and notes that wide scatter among facets can make the domain interpretation more complex. This documents the NEO-PI-3 hierarchy, not that a similarly named facet elsewhere belongs to it or can be converted to its domain. A within-report relationship does not create a cross-report link. If a report offers a broad marketing label but no technical definition, score basis, version, or norm information, mark the comparison unresolved. Ask the publisher for the scale definition and whether a linking study supports this exact comparison, including the population and intended interpretation the evidence covers. Until then, keep the scores separate. The record can state the outcome plainly: linked numerically, comparable descriptively, or not interpretable from the material provided.

Sources: Principles and Practices of Test Score Equating; Standards for Educational and Psychological Testing; NEO-PI-3 Self-Interpretive Sample Report

What should you do when the two reports still do not line up?

Ask each publisher a question: What content and scoring rule define this scale, and is there evidence linking this exact version to the other report for my population and purpose? Request technical documentation, not just a statement that both reports measure a similarly named trait. The ETS report *Principles and Practices of Test Score Equating* concerns tests whose editions are intended to have the same meaning; it explains that conversion errors can affect validity and fairness. This supports asking for evidence and limits, but does not show these personality scales have been linked. If no suitable link is available, leave the numbers separate. Translate the original question into something observable. For planning under a deadline, note when you begin, what information you seek, and whether your approach changes with the task or setting. Record occasions over time instead of treating either report as the true account. This does not establish a stable trait; it gives you material for reflection or coaching and clarifies what you still want to understand. For a work question, keep the inference descriptive and tied to context. Personality scores do not establish job fit or employment suitability. A prompt may help you reflect on decisions and collaboration: the live Work Pattern Report at /assessment is a low-stakes self-report. It is non-validated and provides no norm, cutoff, selection score, or career recommendation. Compare meanings first and numbers only when linking evidence supports that step. The conclusion could change if documentation relates these instruments and versions for a relevant population and purpose. Until then, a behavior question is more useful than forcing unlike scales onto one number.

Sources: Principles and Practices of Test Score Equating