What Convergent Validity Adds to a Personality Report

What does convergence add to the meaning of a score?

Convergent validity adds evidence that a personality report’s interpretation relates to other measures in the way theory says it should. It supports a bounded interpretation, not universal accuracy. The APA PsycTests “Methodology Field Values” defines convergent validity through relationships between scores on conceptually similar tests. It defines test validity more broadly: evidence and theory must support specific score interpretations for a proposed use. One finding does not support every report sentence, population, or decision. A report may describe a score as reflecting a tendency, such as preferring advance planning. The score is an observed result; the tendency is the construct the report proposes it represents. A construct is an attribute inferred from response patterns and relationships, not directly observed like a marked answer. Convergence is useful when those relationships fit a prior account of the construct. If measures designed to capture related aspects tend to move together as expected, that pattern gives the interpretation support beyond familiar-sounding report wording. It does not show the tendency in every response or setting. The Standards for Educational and Psychological Testing, prepared by the American Educational Research Association, American Psychological Association, and National Council on Measurement in Education, place relations to other variables among several sources of validity evidence. They describe convergent and discriminant evidence as ways to examine whether scores relate to measures of similar and different constructs as expected. The comparison matters more than a broad label such as “validated.” Ask which interpretation is being tested, what relationship was expected, and what comparison was observed. A statement that a measure “correlates with personality” leaves the construct, comparison measure, population, and proposed use unspecified. No single correlation carries an interpretation by itself. A relationship between two scores is an observation; deciding what it means requires an account of what each measure captures and why the relationship should occur. Evidence is more informative when a clear expectation is credibly tested. Distinguish observing a relationship from explaining why it supports an interpretation. So “convergent validity” should lead to a narrower question than “Is this whole report valid?” Ask what claim the evidence supports: does it support interpreting this score as a tendency toward a named pattern, in a population and for a purpose relevant to the report? Both the Standards and APA definition tie validity to intended use. Convergence can give a reader reason to take a bounded interpretation seriously while leaving broader claims unestablished. Next, inspect what the second measure contributes.

Sources: Methodology Field Values; Standards for Educational and Psychological Testing

When does a second measure add independent information?

A second measure adds more information when it reaches the proposed construct through a meaningfully different route. If two questionnaires use similar wording, the same response scale, and ask respondents to describe themselves in much the same way, their agreement is relevant. But it leaves open whether both track the intended tendency or partly share the same way of asking. A stronger comparison changes features and checks whether the relationship is specific to the target construct. This is the practical logic of the multitrait-multimethod approach described in “Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix.” Rather than treating one correlation as a verdict, it asks how several relationships fit together. Do measures of the same trait relate as theory expects when their methods differ? Are measures of distinct traits less closely related, including when they share a method? This pattern can help separate evidence about the construct from evidence about the measuring process. The 2014 Standards for Educational and Psychological Testing likewise describe relations to other variables as validity evidence and note that comparisons across methods can sharpen score meaning. A method effect is variation associated with how information is collected, rather than solely with the intended trait. Two self-reports may share influences such as how respondents use rating scales or interpret questions. Their items may also cover nearly the same content. Their agreement is still evidence, but shared wording, response habits, or content remain plausible explanations. Suppose a report compares its self-description scale with another questionnaire that asks about similar tendencies in similar language. If the scores align as predicted, that is a piece of convergent evidence. Ask whether the second questionnaire adds a new test: does it use distinct items or another source, and were different traits checked to test whether the relationship is selective? Without such comparisons, one correlation cannot show how much reflects the intended construct and how much may reflect shared method or content. The matrix approach compares traits and methods instead of treating the largest isolated relationship as best. If the report also compares the self-report with ratings from an observer who has had enough opportunity to see relevant behavior, the evidence changes. A relationship in the predicted direction can reduce some shared-method explanations and add another viewpoint. But a different method is not automatically better, more objective, or a gold standard. An observer may not see private intentions or behavior in settings they do not share; self-report may capture experiences unavailable to that observer. Ask whether the pattern was predicted and what each source could observe. The comparison also needs a meaningful contrast. If the target relates equally to measures of several distinct traits, its apparent convergence may be less specific than claimed. Conversely, an imperfect relationship with a related measure is not automatically failure: theory may predict overlap without identity, and methods may capture different parts of a tendency. The Campbell and Fiske framework makes relationships interpretable as a pattern of convergence and discrimination, not a contest to maximize similarity. When a report gives one correlation, inspect the design alongside the number. What construct was expected to overlap, and why? How similar were items, response format, source, and setting? Were distinct constructs checked? Did the comparison measure have a clear interpretation of its own? A second measure adds independent information to the extent that its design tests the intended relationship while making plausible alternatives visible. Agreement across similar self-reports can support an interpretation; agreement across suitably different methods, considered with relevant contrasting measures, can make the case more discriminating. Neither pattern establishes every claim a reader might want to make from the report.

Sources: Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix; Standards for Educational and Psychological Testing

What can agreement and disagreement between self and observers tell you?

Agreement between a person's self-report and an observer's ratings can strengthen the case that both are registering something in common about a personality tendency. It does not establish that either account is complete, that the score is exact, or that one source is the final authority. Disagreement is equally limited: it raises questions about what each source can see and how the measure frames the tendency; it does not, by itself, invalidate the report. The useful evidence lies in the pattern and its conditions, not in a simple vote between the person and an observer. A 2007 meta-analysis, “The Convergent Validity between Self and Observer Ratings of Personality: A meta-analytic review,” pooled studies of self and observer ratings across the Big Five dimensions. The abstract reports corrected mean correlations ranging from .46 for agreeableness to .62 for extraversion, with values for conscientiousness, emotional stability, and openness between those figures. These are aggregated results across dozens of studies and thousands of ratings, not a target that an individual personality report should meet. The authors also report substantial unique variance in both self and observer ratings. In other words, the sources overlapped, yet each contained information not captured by the other. The analysis found that duration of acquaintance had a large moderating effect on convergence; observer type, comparing work peers with relatives, did not moderate it. That result makes familiarity relevant, but it does not show that every close observer knows every aspect of a person better. Self-reports can include private thoughts and intentions; observers can describe behavior they have seen. Neither view maps perfectly onto a broad trait. Limited exposure may leave an observer with an impression, while a self-rating can reflect how a person understands the questions or their current context. These are possibilities to investigate, not explanations established by disagreement alone. Consider an illustrative report that describes a tendency to plan ahead. The person may recognize careful preparation before familiar tasks, while a colleague sees last-minute changes when priorities shift. Those accounts may reflect different situations or definitions of “planning,” or an uneven pattern: preparation when goals are stable, improvisation when they are not. Ask what each source observed, over what period, and under which conditions. A concrete follow-up is: “When did you see me prepare early, and what was happening when I changed plans?” This can refine a tendency into a conditional description. This interpretation fits the logic of convergent and discriminant evidence: agreement can help show how an instrument operates, but it cannot settle every claim behind the instrument. The ETS research bulletin “Validation by the Multigroup-Multiscale Matrix” applied an adapted comparison to perceptions of college environments, using faculty, administrator, and student groups at 17 colleges and universities. Its abstract says convergent and discriminant information offered insight into the instrument's operation and possible improvement, while not confirming the theory underlying the measurement. That is a different subject from personality ratings, so it is not direct evidence about a personality report. It illustrates the narrower point that patterned agreement is informative without proving the whole explanatory account. A report is easier to interpret when it says whose perspective supports a statement and what the comparison involved. If a self-rating aligns with an observer rating, ask whether the observer had enough exposure to the relevant behavior and whether the evidence spans more than one setting. If they differ, look for access, context, timing, wording, and the possibility that the tendency varies by situation. Then compare the report's description with repeated, observable examples rather than treating either score as a verdict. The meta-analysis supports expecting overlap alongside distinct perspectives in its Big Five research base; it does not provide a universal coefficient, decide which source is right for one reader, or validate a different instrument or use. Convergence adds weight when the sources are relevant and their conditions are understood. Disagreement can identify what needs closer interpretation. Neither substitutes for that interpretation.

Sources: The Convergent Validity between Self and Observer Ratings of Personality: A meta-analytic review; Validation by the Multigroup-Multiscale Matrix; Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix

Why is convergence different from predicting an outcome?

A relationship with a similar personality measure supports an interpretation of what a score represents; it does not show that the score predicts a separate outcome such as work performance. A reader should ask what the report actually compared before carrying its conclusion into a decision. If the comparison was another measure of a related personality tendency, the result may support that interpretation. It does not establish what the person will do in a particular job or which role would suit them.

A criterion is an operationally distinct outcome that an assessment is proposed to relate to or predict. For example, a report might be evaluated against a defined work outcome, measured in a particular setting and population. That is a different question from whether its score aligns with another questionnaire intended to measure a related trait. The outcome must be described in assessable terms rather than left as a broad label such as “success.” It must also be relevant to the purpose: an outcome that does not represent the decision at hand cannot make the evidence answer that decision well.

The 2014 Standards for Educational and Psychological Testing distinguish evidence based on relations to other variables, including convergent and discriminant evidence, from test–criterion relationships. The latter concern prediction accuracy for a specified purpose, and the Standards emphasize the relevance and quality of the criterion. This framework clarifies what evidence a claim needs; it is not a finding that any particular personality report predicts an outcome. The distinction is plain: “This scale aligns with another measure of a related construct” describes convergence; “This score predicts a defined outcome in this setting” makes a criterion-related claim. The second statement needs evidence about that outcome and setting, not just a stronger version of the first comparison.

The APA PsycTests methodology definitions make the boundary useful for report readers: validity concerns evidence and theory supporting a particular score interpretation for a proposed use, while convergent validity concerns relations with conceptually similar tests. A report can offer a supported interpretation of a tendency and still provide no evidence for a proposed prediction. The use changes the question. A report for reflection asks whether its description is grounded enough to be a useful prompt. A report used to decide who should be hired asks whether the score supports that selection decision under relevant conditions, with an appropriate outcome and process.

A claim about convergence should not be stretched into career fit, selection value, diagnosis, or certainty about behavior. A general personality score does not become a job recommendation because it resembles another personality score, and agreement across measures does not establish a clinical diagnosis. Nor does a relationship with one outcome automatically transfer to another role, population, or decision. Keep the inference close to the comparison actually tested.

When a report makes a consequential claim, ask for evidence designed around that outcome and decision process. What outcome was measured? Does it match the decision? In which setting and population was the relation examined, and how was the outcome assessed? These questions locate convergence's contribution rather than dismissing it. A coherent pattern across related measures may strengthen confidence in a score's meaning. To support a claim about what that score predicts, the evidence must also address the criterion, context, and intended decision.

Sources: Standards for Educational and Psychological Testing; Methodology Field Values; MBTI Facts

How should you judge the comparison measures named in a report?

Check whether the comparison measure is a relevant yardstick for the stated construct, and whether its population, language, method, and evidence fit the claim. A report cannot borrow strength from a comparison measure whose meaning is unclear. When a report says one scale converges with another, ask why theory expects that relationship, what each instrument measures, and what it adds beyond a second label.

Start with the predicted relationship. Does the report explain which aspect of the target construct should overlap with the comparison measure, and which should remain distinct? A broad statement that both assessments measure personality is not enough. The Standards for Educational and Psychological Testing frame evidence from relations to other variables in relation to an interpretation and proposed use. A correlation without a reasoned expectation is difficult to interpret, even when it looks impressive.

Next, find out what the comparison measure assesses and how its scores are interpreted. A measure's name is not a description of its content. The report or technical material should identify the construct, relevant scale or facet, and basis for treating its scores as meaningful. The APA PsycTests methodology reference defines convergent validity through relationships with conceptually similar tests; ‘conceptually’ matters. Similarity needs a defensible account, not just matching labels. If the comparison instrument has uncertain meaning or weak documentation, its association cannot establish much about the target report.

The size of a relationship also needs context. A high association is not automatically stronger evidence. If two questionnaires repeat nearly identical wording, use the same response choices, and ask the same person to describe themselves at the same sitting, their scores may align partly because the measurement situations are alike. That can support overlap, but leaves shared item content and response habits as plausible explanations. A less-than-perfect association may fit a theory in which constructs overlap without being identical. The Campbell and Fiske article on the multitrait-multimethod approach recommends examining same-trait measures across methods alongside different-trait comparisons. The surrounding pattern helps separate construct from method.

Then inspect who was studied and under what conditions. A comparison result applies most directly to the people, language, version, and administration represented in its evidence. A report should make the sample and procedure clear enough to judge whether the comparison bears on the reader's situation. A result from one population does not automatically transfer to another language or context, or show that the same score interpretation works for every purpose. The Standards treat validity as evidence supporting specified score interpretations and uses, not as a permanent property detached from context.

Consider whether the measures offered independent information. Different methods can reduce some shared-method explanations, but a different method is not automatically a gold standard. A knowledgeable observer may report visible behavior but have limited access to private intentions; another questionnaire may capture a different perspective while sharing other biases. The Campbell and Fiske framework asks readers to compare patterns across traits and methods rather than treat one pair of scores as decisive.

A relation to a theoretically similar construct helps assess whether an interpretation behaves as expected. It does not establish prediction of an external outcome; that requires a separately justified outcome and evidence relevant to the proposed use. The MBTI Facts page from The Myers-Briggs Company describes comparisons with similar and different measures and points to manuals and supplements. This is a provider-authored account of that instrument, not independent confirmation or evidence that another assessment has the same properties. Trace claims to the relevant version's technical material rather than transfer a provider summary across instruments.

A report lets you follow the chain: why the comparison was expected, what the second measure means, who supplied the data, how methods relate, and what kind of result was observed. If details are missing, lower confidence in that particular convergence claim and ask for the technical report or manual. Missing detail creates uncertainty; it does not prove that either measure failed.

Sources: Standards for Educational and Psychological Testing; MBTI Facts; Methodology Field Values

What does convergent validity leave uncertain?

Even a well-designed convergence result leaves questions unanswered. It may support an interpretation that a score reflects a personality tendency, but it does not show how precisely this person’s score was measured, whether the assessment covers the full construct, whether it distinguishes that tendency from neighboring traits, or whether the interpretation fits the relevant population and intended use. The Standards for Educational and Psychological Testing frames validity as evidence for a proposed interpretation and use, not a permanent property stamped on a test. Reliability, or precision, concerns how consistently scores are produced under specified conditions. A scale may align with a related measure and still leave readers unsure how much its score might vary across occasions or forms of administration. The Standards says precision evidence should fit the population and intended use, and that reported indices need details about methods, samples, and testing conditions. Agreement between measures does not substitute for this evidence. The APA PsycTests Methodology Field Values defines reliability as consistency across test replications; consistency alone does not establish that the intended interpretation is supported. Convergence also cannot establish that an assessment represents the full range of what its label suggests. Content evidence asks whether items adequately represent the intended domain and population. A scale might relate to one recognizable facet while leaving other parts of a broader construct unexamined. The Standards treats the match between test content, construct, and intended population as a separate question. Thus, a relationship between two measures does not mean every sentence in a report has equal support. The scale’s definition and item coverage still matter. Nor does convergence alone show that nearby traits remain distinct. The Campbell and Fiske article, “Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix,” recommends examining same-trait relationships across methods alongside different-trait relationships. If a scale aligns with a related measure, readers still need to know whether it is less closely related to measures theory says should differ. Otherwise, apparent support may partly reflect broad overlap or weak separation between constructs. This comparison bounds interpretation; it does not require every related trait to be unrelated. Population fit matters too. Evidence from one language, setting, or group does not automatically establish that the same score meaning carries into another. The Standards connects content and interpretation to the intended population and discusses whether a construct has equivalent meaning across groups and cultural contexts. That guidance does not prove a particular report fails across populations. It means readers should look for evidence relevant to the people and conditions named in the report, especially when claims extend beyond the validation sample. Finally, convergence does not establish an outcome claim. If a report suggests that a tendency predicts a result, such as success in a role, that is a different claim requiring evidence about that outcome, setting, and proposed use. A relationship with another personality measure cannot supply it. Nor does convergence turn a general personality description into a diagnosis or fixed account of behavior. The interpretation should remain a description of tendencies, bounded by its evidence. When comparison context or intended use is missing, confidence should narrow accordingly. Readers can treat a result as a plausible prompt for reflection while withholding stronger conclusions about precision, coverage, distinction, population fit, or outcomes. The practical question is whether the report states those remaining questions clearly enough for its conclusion to stay within the evidence.

Sources: Standards for Educational and Psychological Testing; Methodology Field Values; Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix

How can you use the evidence without turning it into a verdict?

For reflection or coaching, convergent evidence is a reason to examine a reported tendency, not a verdict about who you are. When different measures relate as theory predicts, that pattern can make one interpretation more credible. The practical question is what it helps you notice in ordinary situations, and where its limits become visible. Consider an illustrative case, not a study result: a report says someone tends to plan ahead, and a second measure collected in a different way points in a similar direction. The person can look for examples. Do they sketch a plan before a complex task? Do colleagues see them identify dependencies early? Do they improvise when a task is familiar? These questions translate an abstract description into observable behavior while leaving room for a tendency to appear often without appearing everywhere. The comparison becomes useful when the person recalls specific episodes. They might ask whether planning happened before a deadline, after a request for structure, or only when several people depended on the work. These details do not prove that the report is right. They help test whether its wording fits recurring behavior. Now add a boundary to the illustration. When a task arrives with little notice, the same person may begin immediately and plan while working. That need not cancel the earlier pattern. Preparation may depend on time, stakes, familiarity, or coordination with others. If the report does not account for context, the reader can raise it as a question rather than assume context was measured. This is a proportionate use of convergence: let agreement guide attention, then let examples refine the interpretation. The 2014 Standards for Educational and Psychological Testing frame validity around evidence and theory supporting a particular interpretation for a proposed use. A relationship among measures can support a bounded reading of a score; it does not make that score a complete account of a person. The Standards support this scope rule, not the illustrative tendency or any particular report. A useful conversation stays close to behavior. The reader might ask a coach, “When have you seen me prepare early, and when do I figure things out as I go?” The other person can name a setting and action, or say they have not seen enough to judge. An observer sees visible actions; a person completing a self-report may also describe private intentions or experiences. Neither account has to be forced into agreement. Keep the purpose low stakes. Reflection can help someone choose a discussion question or notice a work habit. Convergence alone cannot diagnose a condition, determine suitability for a role, or predict what someone will do; those are different claims requiring evidence matched to those uses. Even in reflection, revise an interpretation when repeated examples do not fit. Turn one report sentence into an observation question: “In which situations does this tendency show up, and what changes when it does not?” Include occasions that complicate the description as well as those that seem to confirm it. If the pattern holds across relevant situations, it may be a useful working description. If it varies, the variation may be more useful. Convergence gives a reason to investigate; conversation and observation keep the interpretation connected to actual contexts.

Sources: Standards for Educational and Psychological Testing

What should you do next with a report's convergence claim?

Start with the report’s exact claim. Does the comparison support an interpretation, or invite a broader inference? The Standards for Educational and Psychological Testing frame validity evidence around a a proposed interpretation and use. Ask what this evidence permits you to say. Support for a narrow interpretation does not justify a different decision. Next, ask what the comparison adds. The Multitrait-Multimethod Matrix Reconsidered explains why relationships between measures should be read alongside their traits and methods. Two questionnaires with similar wording and formats may agree with shared measurement influences. Suitable evidence across methods adds information when theory predicts a relation. Check whether the report identifies what was compared, and under what conditions. Missing detail limits the conclusion; it does not show the claim is false. Finally, turn the supported tendency into an observable question. If the report describes planning ahead, ask when that pattern appears, what changes it, and whether someone else sees it there. Treat the description as a prompt, not a verdict about who you are. A clear theory, suitable comparisons, different methods, and a relevant population strengthen confidence. Shared-method agreement alone, or a use the evidence did not address, leaves more uncertain. Use the report to start a conversation about the pattern and its limits.

Sources: Standards for Educational and Psychological Testing; Convergent and Discriminant Validation by the Multitrait-Multimethod Matrix

Questions readers ask