How Should Missing Answers Be Handled in a Personality Report?

The claim under review: just average what was answered

This claim sounds reasonable because many personality scales combine several answers into one score. If a scale asks about planning, follow-through, and keeping commitments, a reader may assume that the average of the answered items is a fair substitute for the full scale. Sometimes an instrument manual permits exactly that. But the arithmetic is not the rule. The rule must come from the instrument's design, validation work, and intended use.

A missing answer is not the same thing as a response at the middle of the scale. It tells the scorer that the person's position on that item is unknown. The reason may be accidental, such as a lost click, or deliberate, such as finding the wording uncomfortable or unclear. Those reasons can matter. A report should not silently turn uncertainty into a number just to complete a page.

Why the shortcut is tempting

A scale score is often a sum or average of multiple items intended to represent a broader construct. When only one answer is absent, replacing it with the person's average on the other items can appear to preserve the scale's range. This is commonly called proration or person-mean imputation. It is not automatically careless. Some published scoring manuals set a threshold and allow it for a small amount of missing data.

The practical appeal is clear. A report can remain usable after a respondent skips one confusing sentence. A provider can avoid discarding every scale because of a single omission. Yet the same shortcut gives different people scores based on different subsets of items. Research on multi-item scales notes that this can affect bias, precision, and the meaning of the resulting score. A neat number is not proof that the underlying information was complete.

Before discussing scoring, separate three situations. An unanswered item is missing. A selected response such as “not applicable” may be a valid option with its own scoring rule. A technical failure may mean the answer was never recorded even though the person responded. These cases should not be collapsed into one code. A transparent report should identify whether the item was skipped, omitted by design, unavailable because of an accessibility issue, or lost during administration.

The instrument's rule matters more than a universal percentage

There is no responsible universal rule such as “up to 10 percent is always fine.” An assessment may set a minimum number of answered items for each facet, use a different threshold for a broad domain, or require every item because the scale is short. The rule may also change when scores are used for self-reflection, coaching, research, or a consequential decision.

A report should therefore name the relevant unit. Is the threshold calculated across the whole questionnaire, within each facet, or within each domain? A person could have enough answers overall but too few answers for one narrow facet. In that situation, the broad report might be partly usable while the facet should remain unscored.

This is why copying a rule from another questionnaire is unsafe. A threshold belongs to a specified instrument and scoring method. The joint testing standards from AERA, APA, and NCME are available as a reference point for sound testing practice, but they do not turn one instrument's rule into a rule for every personality report.

What proration can and cannot tell you

Suppose a five-item scale has four answered items and the manual explicitly permits an average when at least four items are present. A scorer may calculate the mean of those four responses and label the result as prorated. That number can be useful as an estimate under the instrument's stated conditions. It is not the same as observing the missing response.

The report should preserve that distinction. It might say that the scale was calculated from four of five items, identify the permitted rule, and caution that the result is less complete than a score based on all items. It should not present the result as though all five answers were supplied. If the missing item concerns a distinctive facet, the limitation may be more important than the percentage alone.

Research is especially cautious about informal person-mean replacement. A 2016 methodological paper describes how proration can rely on assumptions about item means and relationships, and reports that it can produce bias even under a missing-completely-at-random scenario in the study it examined. That does not prove every permitted prorated score is unusable. It does show why the rule needs evidence and disclosure.

When withholding the score is the clearer answer

A scale should generally be marked incomplete when the instrument's minimum has not been met, when a required item is absent, or when the provider cannot explain how partial responses are scored. “Not reported” is more honest than a precise-looking number whose basis is unknown.

Withholding one scale does not necessarily invalidate the entire assessment. A report can distinguish a completed domain from an incomplete facet and explain whether the missing information prevents interpretation of a higher-level score. It should avoid filling the gap with a whole-test average unless that procedure is part of the documented scoring model.

There is a trade-off. Omitting incomplete cases can reduce the amount of information available, especially in research. But retaining a score can also introduce bias or false precision. The right choice depends on the use and the evidence behind the method. The reader should be able to see that choice rather than discover it only after challenging the result.

Do not confuse scoring with statistical imputation

In research, analysts may use multiple imputation or full-information methods to estimate missing values while carrying uncertainty into later analyses. Multiple imputation creates several plausible completed data sets and combines the results, including variation between those estimates. That is a method for population or study inference, not a casual instruction to invent an individual's unanswered response.

A personal report has a different communication problem. If a provider estimates a score from missing information, the reader needs to know that an estimate was used, what information informed it, and how uncertainty was represented. A single filled-in answer or rounded percentile can hide more than it explains.

The distinction also helps when reading a research claim about a personality questionnaire. A study may handle missing data with a sophisticated model, while the consumer report uses a simple threshold rule. Evidence about the study's analysis does not automatically validate the consumer report's scoring practice.

Why the pattern of omissions may matter

The number of missing answers is only one part of the picture. Which items are missing, and whether omissions cluster in one section, may matter. A person who skips one item in several different facets presents a different scoring problem from a person who leaves an entire facet unanswered.

The reason for omission may matter as well, although a report should not guess the reason. If wording, language, disability access, privacy, or relevance made certain items difficult, the pattern may indicate a problem with administration or fit rather than a personality tendency. This is a reason to request an accommodation or clarification before treating the result as a stable description.

A 2013 study of simulated missing values in a personality test examined effects on reliability, factor structure, and the ordering of test scores. Its abstract reports that low levels of missingness could be handled adequately by some simple methods in its applied simulations, while other replacement procedures performed poorly. The study does not provide a universal cutoff. It supports examining the pattern and method rather than relying on a blank count alone.

A worked example without pretending to know the result

Imagine a report with three broad domains and several narrower facets. One facet contains a skipped item. The provider's documentation says that the facet may be scored when at least four of five items are answered, and the report labels the result as prorated. In that case, the reader can ask: Was the stated minimum met? Is the score clearly marked as partial? Does the domain score depend on that facet? What uncertainty or limitation is attached to the interpretation?

Now change one fact. The documentation says every item is required, but the report still displays a facet percentile. The problem is not that the percentile is low or high. The problem is that the report has not shown a valid basis for displaying it. The appropriate response is to request correction, completion, or an explanation, not to interpret the number more confidently.

This example deliberately gives no fictional score, respondent, norm group, or personality conclusion. The useful comparison is between a disclosed, instrument-specific partial-score rule and an unexplained number.

What to ask before using the report

Ask the provider or assessor for the missing-response policy before making a decision from the report. The answer should be concrete enough to reproduce the scoring choice, even if the full algorithm is proprietary.

Check whether the report tells you the number and location of missing items, the minimum required for each scored section, whether any items were prorated or imputed, and whether incomplete sections were withheld. Ask what population and purpose support the interpretation of the resulting score. A norm-referenced percentile based on a partial score needs a particularly clear explanation because the comparison depends on the scoring method.

For self-reflection, you may decide that a cautiously labeled partial result is enough to prompt a question or observation. For coaching, selection, or other consequential use, demand stronger documentation and avoid treating an incomplete score as a verdict. A general personality report should not be converted into a diagnosis merely because missing data make the output look authoritative.

The practical conclusion: transparency beats completion

The best handling of missing answers is not always to discard the report and not always to calculate every possible score. It is to follow the instrument's documented rule, preserve the distinction between observed and estimated information, and show the reader what uncertainty remains.

A trustworthy report makes incomplete sections visible. It explains whether a partial score was allowed, how much data supported it, and what uses are outside the evidence. If the provider cannot answer those questions, use the report as a prompt for reflection at most, not as a precise comparison or employment judgment.

Start with the checklist below, then ask one direct question: “Which of my scores were calculated from incomplete answers, what rule allowed that, and how should that limitation change the way I use this report?” The quality of the answer is part of the evidence about the report.

Questions readers ask