What Hogan HPI Percentiles Mean Beside Raw Subscale Counts

What does an HPI percentile mean when subscales show raw counts?

The primary-scale percentile and raw subscale count answer different questions. A percentile rank describes where a score falls in a score distribution: it is the percentage of scores below that score. The NCME Assessment Glossary ties this comparison to a defined population, so a percentile is not the percentage of items endorsed, a probability, or a rating of merit. Its meaning depends on the reference group and scoring edition used for the report. In its 2013 Interpreting HPI Subscales guide, Hogan describes primary-scale results as often presented in percentiles relative to the general workforce, while subscales can be reported as raw scores that add detail to the main-scale interpretation. A raw count is the result of the subscale’s item-scoring rule; seeing it beside a percentile does not put both numbers on the same scale. Dividing that count by the maximum possible points gives a fraction, not a population comparison. Before translating a count into relative standing, identify the report’s form, scoring and norm edition, then consult the matching guide or request its documented conversion. Until that mapping is verified, keep the value labeled as a raw count. It cannot establish that the person is at any particular percentile.

Sources: Assessment Glossary; Interpreting HPI Subscales

What exactly is being compared when the report prints a percentile?

A percentile rank tells you where a score falls in a specified distribution. It does not tell you how many questionnaire items a person endorsed. The National Council on Measurement in Education (NCME) defines percentile rank as the percentage of scores in a specified distribution below a given score; a norm-referenced interpretation compares a test taker’s performance with the distribution for a specified reference population. For an HPI report, therefore, the percentile is meaningful only alongside the norm population and scoring edition used to produce it. A raw subscale count printed nearby is a different kind of result. That distinction is easy to miss because reports often place several numbers together and use familiar labels for both scales and subscales. But the numbers answer different questions. A primary-scale percentile asks, in effect, “How does this scale score compare with scores in this reference group?” A raw subscale count asks, “What total did this person receive under the scoring rule for these subscale items?” A percentile beside a subscale does not silently convert its count into a percentile. Hogan’s 2013 Interpreting HPI Subscales guide illustrates this distinction for a specific older form and report context. It describes an HPI with 206 True/False items, seven primary scales, and 42 subscales. In that guide, primary-scale results are often presented as percentiles relative to the general workforce, while raw subscale scores can add detail to interpretation beyond the main-scale scores. This is useful historical evidence for why a reader may see the two formats side by side. The guide is historical evidence, not documentation for every later form or current template. A percentile is not a percentage of answers. A person at a given percentile has not necessarily endorsed that share of the items, received that share of possible points, or answered that share “correctly.” Personality inventories do not use correct answers. Instead, the reported rank locates a scale score within the reference distribution. Nor is it the probability of a behavior or a prediction of what a person will do next. The reference group matters because “higher than” is incomplete unless the comparison group is named. Hogan’s participant FAQ says HPI and HDS norms are based on samples of more than 100,000 working adults and describes the data as representing a cross-section of the current workforce. This is the publisher’s description of its norm base, not evidence that it represents every population or purpose. Comparisons with a local group or another norm edition answer different questions. NCME’s glossary likewise defines norms as data summarizing score distributions for specified groups, and a reference population as the population represented by those norms. This also means that two identical percentile figures need not carry identical practical meaning if they came from different reference populations, editions, or report purposes. Even when the number is the same, the “standing” it expresses is conditional on the distribution used. Hogan’s older guide says “general workforce” for its described primary-scale percentiles; readers should not silently expand that label to all adults, all occupations, or an individual’s own workplace. A percentile is a comparison statement, not a universal rank attached to the person. Keep three questions separate when reading the page: Which construct or scale does the number refer to? Is the number a raw score or a norm-referenced percentile? Which reference group and scoring edition anchor the comparison? The label and report documentation should answer the first two; norm information should answer the third. If these details are absent, the reader cannot say who the person is being compared with. Before comparing percentile values across people or reports, record the report date and the norm named by the administrator or documentation. That small step preserves what the number actually says: relative standing on a specified scale, within a specified score distribution, under a specified scoring context. It prevents a primary-scale percentile from being mistaken for a raw count, a subscale count from being treated as a percentile, or either quantity from being read as a fixed description of the person.

Sources: Assessment Glossary; Interpreting HPI Subscales; Frequently Asked Questions

Why doesn't a raw count divided by its maximum become a percentile?

A raw count divided by its maximum tells you what fraction of the available item points the scoring rule awarded. It does not tell you the score’s percentile. A percentile rank describes where a score falls in a specified reference-group distribution. To move from a count to that relative position, you need information about how people in the relevant norm group scored, plus the scoring and conversion rules for the report. Arithmetic alone supplies neither. The distinction is between rescaling a number and comparing it with other people’s scores. Suppose a hypothetical subscale has a possible range from zero to ten, and a respondent’s raw total is five. The ratio, five divided by ten, is one half: half of the possible points under that scale’s scoring rule. That result remains the same whether very few, many, or nearly all people in a reference group also score five. Yet those different distributions would place five at different relative positions. This is a mathematical illustration, not an HPI result or conversion example. A percentage can look familiar while answering a different question. “Half of the possible points” describes one raw total in relation to its scale’s maximum. “At the 50th percentile” describes a score’s position in a specified distribution. The statements might happen to align in a particular scoring system, but one does not imply the other. The National Council on Measurement in Education’s Assessment Glossary defines percentile rank in relation to the percentage of scores below a given score in a specified distribution, and norm-referenced interpretation as comparison with a specified reference population. The reference distribution is part of the claim, not an optional detail. For the HPI, “FAQ Blog Series: The Hogan Personality Inventory (HPI)” says Hogan uses direct summative scoring for HPI Homogenous Item Composites (HICs), as well as primary and occupational scales, without item weights. The FAQ explains that item and subscale counts differ because the constructs represented vary in breadth and depth. This helps explain why a raw total depends on the items assigned to that particular subscale. It does not provide a population distribution for the total or a rule to convert it to a percentile. The older guide “Interpreting HPI Subscales” also distinguishes primary-scale percentiles from raw subscale scores in the report format it describes. These sources support keeping the score types separate; they do not authorize conversion for every report edition. The denominator still matters. It tells a reader what “five” is measured against when the maximum is known, and may help establish whether two printed values are even on the same raw scale. But equal fractions across two subscales do not establish equal standing among respondents. Each can have different items and a different distribution of raw totals. The fraction alone does not show how common a total was in the relevant norm group. That comparison requires a matching norm-based table or documented conversion, not simply the item count. Treat count divided by maximum as a descriptive fraction only. Do not label it a percentile, use it to rank someone against a population, or compare it across subscales as if it were a shared ruler. If the report supplies a separate norm-derived category or conversion, interpret it under documentation for the relevant HPI form and scoring edition. Without that link, the denominator identifies the raw scale; it does not license a population comparison.

Sources: FAQ Blog Series: The Hogan Personality Inventory (HPI); Assessment Glossary; Interpreting HPI Subscales

Why can two subscale counts with different denominators mean different things?

Two HPI subscale counts can differ in meaning because they do not necessarily count the same number or kind of keyed responses. Hogan's 2015 “FAQ Blog Series: The Hogan Personality Inventory (HPI)” says subscales, also called Homogeneous Item Composites (HICs), have uneven item counts because their themes vary in breadth and depth. It gives Adjustment and Sociability as examples of primary scales involving different ranges of concepts. The FAQ also says HIC scores are produced by direct summative scoring, without item weights. It does not put every HIC on a shared ruler. A count must therefore be read with its subscale label, assigned items, and scoring key. It records keyed outcomes, not a free-standing quantity of a personality characteristic. Where raw totals differ, the same number can occupy a different place on each scale. Even equal totals or equal fractions of the maximum do not mean a respondent gave equivalent answers: the subscales may concern different themes. The denominator identifies available item points, but does not supply a common interpretation. The older “Interpreting HPI Subscales” guide shows the range of these themes through labels such as Calmness, Likes Parties, Mastery, and Curiosity. It places them under broader primary scales and says subscales add detail to those scales. The guide cautions against overemphasizing a particular Inquisitive subscale unless the job or role context calls for it. A subscale may suggest a question about a broader result, but its number belongs to that construct and context. The scoring key matters because some subscale labels are phrased negatively. In “Interpreting HPI Subscales,” an asterisk identifies cases where a higher rate of endorsement means absence of the characteristic named. For example, the guide says high endorsement on Not Anxious indicates someone who seems calm and relaxed. Everyday interpretation of the label could reverse the direction. An integer is not self-explanatory; its definition, polarity note, and applicable scoring edition are part of its meaning. Take an explicitly hypothetical comparison, not an HPI result: imagine the same raw count beside two subscales, one with fewer possible keyed items and one with more. The first may represent a larger fraction of its available points. That arithmetic fact does not show that its tendency is more common, stronger in daily behavior, or more unusual relative to a reference group. Hogan's 2015 FAQ explains differing item counts through breadth of themes; it does not provide a rule for comparing raw HIC totals across themes. Dividing a raw count by its maximum should not be treated as a percentile or cross-subscale score. A percentile interpretation needs a score distribution for a specified reference group; a fraction is calculated from possible points. They answer different questions: what share of available keyed points is this, and where does this score fall relative to a norm group? Without a matched conversion, a fraction cannot establish relative standing. Hogan's “Subscale Brick Format Scoring Updates” FAQ provides a historical contrast. For named reports generated from May 2, 2016, it says raw subscale scores were replaced with bricks. For HPI and MVPI, these were four categories aligned to cumulative frequencies. The FAQ says Hogan computed each subscale's cumulative frequencies against a global working population, then aligned raw scores with one of four bricks. It also says subscales supplement, rather than replace, interpretation at the primary-scale level. A brick conveys a broad norm-derived category that a raw tally does not. It cannot be recreated by division, and this FAQ is not a universal conversion table. The practical rule is to avoid comparing counts with different denominators as if they used a common scale. Identify the subscale names, report form, and matching scoring guide. Check which items contribute, what endorsement direction means, and whether the value is a raw total or a documented norm-derived category. Treat the detail as a prompt to examine an observable example, attached to its primary scale and context. If no matching conversion is supplied, retain the label “raw count.” A larger number alone cannot establish higher population standing or greater importance than another subscale.

Sources: FAQ Blog Series: The Hogan Personality Inventory (HPI); Interpreting HPI Subscales; Subscale Brick Format Scoring Updates

How did Hogan's later subscale bricks relate to raw counts?

For the HPI reports named in Hogan’s 2016 materials, subscale bricks displayed how a score compared with a working population. They replaced earlier raw subscale scores on specified reports; they were not a linear rescaling, nor the primary-scale percentile printed elsewhere. A brick added a coarse population comparison that a bare count did not provide, while remaining tied to the report format and scoring edition.

The change had a defined scope and date. Hogan’s Subscale Brick Format Scoring Updates FAQ says the new display would appear on reports generated from May 2, 2016, and lists HPI Potential, Insight, and Flash among the affected reports. Raw subscale scores would no longer appear on those reports. The FAQ describes four HPI and Motives, Values, Preferences Inventory (MVPI) categories: Low, Below Average, Above Average, and High. It gives different categories for the Hogan Development Survey (HDS): No Risk through High Risk. These labels are not interchangeable. An HPI brick should not be read using HDS risk language, and the list does not establish what every report template displayed.

Hogan explains the HPI and MVPI mapping as a cumulative-frequency calculation: the scoring process considers how often raw subscale values occur in a reference population, then assigns a value to one of four categories. The FAQ gives the bands as 1–25 percent, 26–50 percent, 51–75 percent, and 76–100 percent. The Subscale Interpretive Guide says each brick represents approximately one quartile and that bars indicate the quartile in which a score falls relative to the global working population. A quartile is a broad quarter of a distribution, not a fine-grained percentile reading. Four categories cannot preserve the detail of every raw score or express the exact standing that a primary-scale percentile can.

The relationship between count and category can be nonlinear because raw values are not necessarily equally common. Hogan’s Subscale Interpretive Guide illustrates this with the No Guilt subscale: it says scores from 1 through 3 were relatively rare in its cited global working population, with fewer than 25 percent of respondents in that range. In the guide’s example, all three raw values fall in the same lowest category. Its Not Anxious example shows a more proportional mapping for a four-item subscale. The contrast explains why dividing by item count, rounding the fraction, or assuming equal steps between bricks cannot reproduce the norm-based conversion. These examples illustrate the method; they are not conversion tables for other subscales or report versions.

The documents also limit reconstruction. The guide says Hogan could not supply a full conversion chart, while offering sample conversions to show the general relationship between the older raw format and newer bars. A reader with an older report should not infer its brick from the count alone. Request the conversion or interpretive guidance matched to the report’s subscale and scoring edition. A later chart may look compatible, but that does not establish the same items, reference data, or mapping.

Hogan presents bricks as a quicker way to interpret subscales and says they supplement interpretation of the primary scale rather than replace it. That is a useful counterpoint to treating the format change as cosmetic: a norm-derived category does add relative context to a raw tally. But the added context is limited. It supports only the category and comparison described for that edition; it does not turn the subscale into a second primary-scale percentile or establish a person’s behavior in a particular situation. For an old count, request its matched version-specific mapping. Do not divide, round, or borrow a later HPI chart to assign a brick.

Sources: Subscale Brick Format Scoring Updates; Subscale Interpretive Guide

What changed in later HPI forms and norms, and why does that matter here?

Hogan documents three distinct changes: a 2016 subscale display transition, a response-format change beginning in 2018, and a later norm update. They concern different scoring stages, so a report date alone cannot settle how an older raw count should be interpreted. The first change was presentational. Hogan’s 2016 Subscale Brick Format FAQ says reports generated from May 2, 2016 would show subscale bricks instead of raw scores in named HPI Potential, Insight, and Flash reports. HPI bricks represented four cumulative-frequency categories, from Low through High, based on comparison with a global working population. The FAQ says primary-scale interpretation remained central and bricks added context. This was a shift from raw count to coarse norm-referenced display, not a universal conversion rule. A separate change affected response options. In the 2023 Norm and Scoring Upgrade FAQ, Hogan says HPI moved from True/False to four choices: Strongly Agree, Agree, Disagree, and Strongly Disagree. Hogan began collecting responses in this format in September 2018. The FAQ says the broader range required a new data set to provide corresponding percentile scores. The response values and raw scores belong to a scoring setup; equal numbers from different forms are not automatically equivalent. The norm update followed. Hogan says it finalized a new global norm in late 2022 and began using it for most off-the-shelf reports on June 1, 2023. A norm is the reference data used to locate a score among other scores; Hogan explains that raw scores are converted to percentiles through those data. The response form and norm are related but distinct: one governs recording and scoring, the other the comparison distribution. Percentiles may shift when that distribution changes. Hogan describes the newer norm as more representative of work-relevant demographic groups and reports coverage of 190 countries and territories, up from 179. These are publisher-reported properties, not proof that every individual score changed or that the norm is more accurate for every purpose. Hogan also says most scores tend to remain in similar interpretive ranges across norms, while recommending that comparisons between people use the same norm for accuracy and fairness. Hogan says most standard-report clients moved to the new norms automatically, while some custom scoring needed client approval. Users of web-service integrations were told to confirm the latest norm ID. Full technical documentation is available through Hogan representatives, but the public FAQ gives no universal conversion table for older raw counts. A report date is a clue, not a complete answer. For a fair comparison, establish both axes: which response form and scoring setup produced the result, and which norm edition supplied the reference distribution. The 2016 brick transition explains why some reports show categories while others show raw counts; it does not establish that they use the same response form or norm. If comparing people, ask whether both reports use the same norm and applicable scoring rules. If interpreting an older count, request the exact guide or conversion for that report rather than applying a current percentile table. These checks identify what score and comparison the report presents; they do not establish its reliability or suitability for a particular decision. Ask the report administrator which response format was used, what scoring and norm edition generated the percentile, and whether the subscale figure is a raw count or mapped brick category. Request the version-matched guide. Until those details are confirmed, keep the number labeled as raw and do not compare it directly with a percentile or another report’s count. The matching documentation could support a more specific interpretation; the number alone cannot.

Sources: FAQ: 2023 Norm and Scoring Upgrade; Subscale Brick Format Scoring Updates; Frequently Asked Questions

When does a score comparison stop being fair or interpretable?

A comparison is interpretable only when the numbers refer to sufficiently comparable scores. For an HPI report, ask whether the same response form and scoring rules were used, whether the scores share a norm or documented equivalent, and whether the reports serve the same purpose. If any answer is unknown, a numerical difference may reflect the measurement setup as well as a difference between people. A percentile describes relative standing within its specified norm group. A raw subscale count, without a matching conversion, cannot establish population rank, actual conduct, or job suitability.

The Standards for Educational and Psychological Testing frame validity as evidence supporting a particular interpretation and use of scores. Evidence supporting one use does not automatically support another. A percentile may describe where a primary-scale score falls relative to its reference distribution. By itself, it does not explain why someone answered as they did, what they will do in a particular setting, or whether they meet a role requirement. Without its scoring rule and any norm-based conversion, a raw subscale tally remains a count produced by that rule.

Think of the comparison as a chain: response form, item key and scoring, reference group, score interpretation, then proposed use. A mismatch early in the chain can weaken later conclusions. Reports from different periods may use different response formats or score keys even when a scale label looks unchanged. If forms match but norms differ, percentiles still refer to their respective distributions. Hogan's public scoring and norm FAQ recommends the same norm when comparing people, while noting that most scores tend to stay in similar interpretive ranges across norms. A norm update does not mean every individual changes category, but the group-level statement cannot tell whether a particular score moved.

Purpose adds a boundary. In private reflection, a person can treat a subscale label as a prompt and ask whether a recent decision fits. In coaching, the result may suggest a question to test against examples and context. In selection or another consequential employment decision, evidence must support the score interpretation for that defined use. A number useful for reflection does not become a defensible hiring criterion simply because it appears in a professional report.

Hogan's HPI FAQ says results should be considered in context, that there is no ideal HPI profile, and that interpretation should consider goals and situation. This is publisher guidance about its assessment, not independent proof that a subscale predicts an outcome. The 2016 Subscale FAQ describes subscales as supplementary to primary-scale interpretation. Detail can sharpen what to explore within a broader scale, but a count is not an observed event: it does not show that someone interrupted a colleague, avoided a decision, or handled a deadline in a particular manner.

The strongest counterpoint is that subscales can add useful nuance. Hogan's materials support using them to enrich primary-scale interpretation; reducing every result to one broad scale could discard distinctions the report is designed to show. Their meaning still depends on scoring and interpretive documentation. A norm-derived category may support a coarser comparison if its mapping matches the report edition. It is not a precise percentile, and a historical guide cannot automatically convert every later report. The conclusion would change if the administrator supplied version-matched documentation naming the form, scoring key, relevant norm, conversion method, and evidence for the intended interpretation and use.

In work feedback, use the score to frame a question rather than deliver a verdict. A coach might ask, “The report raises a question about how you approach disagreement; can we look at a recent example?” Then examine the exchange, task demands, constraints, and other appropriate evidence. That inquiry may show whether the interpretation fits or whether another explanation, such as unclear responsibilities, is more relevant. It applies the evidence boundary; it does not claim the score explains the event. If the report gives only a raw count, call it a raw count. Do not translate it into a percentile, infer a work habit, or attach a performance judgment without documented grounds. A fair comparison names its form, scoring, norm, and intended conclusion.

Sources: Standards for Educational and Psychological Testing; Frequently Asked Questions; Subscale Brick Format Scoring Updates

What should I ask before acting on this HPI report?

Before acting on the number, ask the report administrator for the report title and date, HPI response format, scoring and norm edition, and the matching subscale guide. Ask what the printed value represents: a raw item tally, a fraction of possible points, or a norm-derived category. Hogan’s “FAQ: 2023 Norm and Scoring Upgrade” says norms convert raw scores into percentiles and notes that custom reports or web integrations may need confirmation of the norm ID. “Subscale Brick Format Scoring Updates” describes a 2016 change to bricks, based on comparison with a working population. Neither history identifies the format. If no version-matched conversion is available, keep the value labeled as a raw count. Do not estimate its percentile from the count or treat a fraction of possible points as relative standing. Instead, use the subscale label as a cautious prompt: what observable example would support or complicate this interpretation in the situation you care about? The number alone does not establish what you did or will do. For low-stakes reflection on decision and collaboration patterns, the Work Pattern Report offers a prompt. It does not convert or validate an HPI score, and it is not a hiring score or job recommendation. Start with the matching documentation. If it is unavailable, keep the count raw and begin a conversation about an example.

Sources: FAQ: 2023 Norm and Scoring Upgrade; Subscale Brick Format Scoring Updates

Questions readers ask