What an Emotional Intelligence SJT Can Show About Work

What can an emotional-intelligence scenario score establish?

An emotional-intelligence situational judgment test (SJT) shows which response a person selected, or how they rated possible responses, in the situations its items describe. Evidence for a named instrument may support a careful claim about the kind of emotional judgment sampled. The score alone cannot establish how that person will act in a live workplace exchange. That matters if someone uses it to make a broad claim about listening, self-regulation, or repairing disagreement: these behaviors unfold with another person under particular conditions.

Think of the inference as steps. First is the answer recorded. Next is the emotional skill the item is meant to represent. Then comes any demonstrated link between the score and a relevant performance measure. The final step is the person’s conduct at work. Each needs its own evidence; reaching one does not complete the next. The U.S. Office of Personnel Management’s “Situational Judgment Tests” guidance describes an SJT as a response to a described problem and distinguishes that from performing behavior in a simulated setting. The score records a response to a scenario. Whether it says something useful about work depends on the instrument and evidence behind that further interpretation.

Sources: Situational Judgment Tests

What is the test actually sampling?

An emotional-intelligence situational judgment test (SJT) presents a described situation and asks the test taker to judge what is happening or choose an intended response. The score therefore begins with an answer to authored situations. What that answer represents depends on the instrument’s theory, the situations included, the response format, and the scoring rule. A test called an EQ test does not, by its name alone, cover every emotional skill or show how someone behaves when a real exchange becomes difficult. The U.S. Office of Personnel Management describes SJTs as low-fidelity simulations: people respond to a work problem but do not perform the task as they would in an assessment center or work sample. This samples judgment, not the live work of noticing uncertain cues, managing one’s reaction, speaking, and adapting to another person’s response. Two ability-based measures discussed in The Measurement of Emotional Intelligence: A Critical Review of the Literature and Recommendations for Researchers and Practitioners. The Situational Test of Emotional Understanding (STEU) targets emotional understanding: judging how an emotion relates to a situation. Its items draw on an emotional appraisal theory, and the review notes that several concern workplace behavior. The Situational Test of Emotion Management (STEM) targets emotion regulation, especially the management of emotions. The review says its items were developed from interviews in which 50 people described 290 emotional situations they had experienced during the previous two weeks. This also explains why a test’s score key matters. The U.S. Office of Personnel Management notes that SJT situations and response options are typically developed from critical incidents generated by job experts, and that subject-matter experts’ judgments commonly inform which alternatives count as best or worst. That guidance describes a general method, not the key for every EI SJT. To interpret a particular result, a reader needs to know how that particular instrument defines an effective answer. Consider a prompt in which a colleague says that a handoff was incomplete and a deadline is close. A scenario might ask which response is most effective. One option could clarify what is missing and agree on the next step; another could immediately defend the earlier work. Choosing the first may show that, in this framed problem, the respondent recognized a potentially constructive course. It does not show whether, during an actual tense exchange, the person would notice their defensiveness, ask a useful question, set a needed boundary, or follow through on the agreement. Those actions are related to the judgment in the prompt, but the test has not observed them. The instrument’s design narrows the claim further. A test focused on understanding emotions cannot automatically establish emotion recognition, regulation, empathy, or communication if those processes are not sampled and scored. A regulation scenario does not show whether someone can carry out the response under pressure, even when its setting is a workplace. As The Measurement of Emotional Intelligence review emphasizes, measures can cover different parts of emotional intelligence; using a particular instrument may leave other parts unmeasured. These examples describe research instruments, not the current EQ Test product. The private 32-item Emotional Skills Profile is an educational reflection tool. Its recent-behavior summaries and authored-scenario judgments are kept separate; it has no validated norms or established ability-test validity. It should not be treated as STEU, STEM, or another research measure, and findings about those instruments do not validate this profile. To read an SJT result, name the instrument, identify the skill sampled, and check how its response key was justified. The result can inform a question about emotional judgment in those situations. To say what a person does at work, look for examples of the relevant conduct in context.

Sources: Situational Judgment Tests; The Measurement of Emotional Intelligence: A Critical Review of the Literature and Recommendations for Researchers and Practitioners; Measurement of Ability Emotional Intelligence: Results for Two New Tests

What does validation of an EI SJT add?

Validation adds reasons to interpret a score as evidence about a defined ability, rather than merely a tally of answers. It asks whether items and scores behave as the test’s theory anticipates: do responses to emotion-understanding situations relate to other measures of understanding, and remain distinguishable from a questionnaire about perceived ability? Such comparisons support a particular interpretation in studied samples. They do not show that a respondent will enact the judgment in a pressured conversation, where cues are incomplete and others respond in real time.

The named instruments STEU and STEM illustrate how the theory shapes that claim. The Situational Test of Emotional Understanding (STEU) was built from Roseman’s emotion-appraisal theory. In “Measurement of Ability Emotional Intelligence: Results for Two New Tests,” Austin describes 42 multiple-choice situations, some abstract and some set in personal or work contexts; respondents choose which emotion is most likely. The theory supplies the scoring rationale: a response counts as correct when it fits the model’s account of how appraisals of a situation relate to emotion. The Situational Test of Emotion Management (STEM), by contrast, presents scenarios and asks which response would be most effective for managing the emotion. Austin reports that its situations were drawn from interviews with 50 people, with candidate responses generated by another group of 99 and then weighted by experts.

A favorable STEU result supports a claim about performance on its emotional-appraisal items, under its theory and scoring. It does not establish broad competence in recognizing emotion, regulating one’s reaction, or repairing a work relationship. STEM evidence concerns its own response choices and key. Neither finding automatically applies to another test with different situations or self-described habits. The instrument name is part of the claim.

Austin’s 2010 study tested STEU and STEM among 339 undergraduates at a UK university, with an average age of about 22. Scores on both tests were positively related to the MSCEIT emotional-intelligence measure and most of its branches; neither was significantly related to the trait-EI questionnaires included in the study. STEU, but not STEM, was significantly associated with the study’s general intelligence measures. This pattern gives some support to the idea that the tests sampled emotion-related abilities distinct from self-perceived typical behavior. It also cautions against describing the pair as one broad, settled capacity: their links to comparison measures differed by component. The student sample and cross-measure comparisons cannot establish how working adults would behave in ordinary workplace exchanges.

The 2019 critical review, “The Measurement of Emotional Intelligence,” describes STEU as emotional understanding and STEM as emotion regulation. It notes STEU’s workplace items, while warning that the tests cover only parts of the abilities grouped under emotional intelligence. Its review of earlier evidence records variable internal consistency estimates across studies, including lower estimates for STEU in some samples. Internal consistency concerns how closely a test’s items relate to one another; it is one kind of evidence about scores, not proof that the test captures all of a skill or predicts conduct. The review supports specific interpretations of these measures while flagging limits that matter when readers generalize.

Agreement with another measure, a related task, and a theory behind scoring answer different questions. A strong result on one comparison cannot substitute for evidence on another. Moving from “can select a response” to “will use it with a colleague” involves attention, emotion, incentives, authority, and the exchange itself. These studies do not observe that chain in everyday work.

Validation raises confidence that a specified score represents performance on a defined emotional problem for the version and populations examined. A broader claim needs evidence connecting that score to relevant behavior in the setting at issue. Until then, the result can identify a question to examine; it is not a record of how its holder handles work.

Sources: Measurement of Ability Emotional Intelligence: Results for Two New Tests; The Measurement of Emotional Intelligence: A Critical Review of the Literature and Recommendations for Researchers and Practitioners

Why does the scoring key change the meaning of a good answer?

A “good” answer on an emotional-intelligence situational judgment test (SJT) is good according to the test’s key and the assumptions behind it. The score therefore means more than which option a person selected: it also reflects how the test defines an appropriate response. In a typical SJT, a person reads a described situation and chooses, ranks, or rates possible responses. The U.S. Office of Personnel Management’s “Situational Judgment Tests” guidance says items are often built from critical incidents identified by job experts and scored using experts’ views of the best and worst alternatives. The score reflects agreement with judgments by people familiar with the work. Experts can judge the options presented, but cannot supply context the item left out or establish one universally correct move for every situation. A test may ask someone to infer what another person could be feeling, understand what may have prompted an emotion, select a way to regulate a reaction, or choose how to respond to another person. The critical review “The Measurement of Emotional Intelligence” describes the Situational Test of Emotional Understanding (STEU) as targeting understanding, and the Situational Test of Emotion Management (STEM) as targeting regulation. It cautions that one measure may leave other parts of emotional intelligence unmeasured. Thus, a high score on an emotion-understanding item should not be translated into a general claim about empathy, restraint, or conflict repair. The STEU shows how theory can shape what counts as a correct answer. “Measurement of Ability Emotional Intelligence: Results for Two New Tests” describes its items as built from an emotion-appraisal theory: a situation’s features are used to infer the emotion most likely to follow. The scoring key can then mark an answer as correct within that model. The answer records whether the respondent selected the emotion predicted by the model, not what they would actually feel or do. The model makes the standard explicit; different facts or appraisals could change the interpretation. STEM uses another route. In “Measurement of Ability Emotional Intelligence: Results for Two New Tests,” researchers describe developing STEM scenarios from interviews about recent emotional situations, then having experts assess response options to create a key. A score reflects how the selected response compares with that judgment. That gap is built into the format. OPM classifies SJTs as low-fidelity simulations because respondents describe or select what they would do rather than perform the behavior in a work sample or assessment-center exercise. On a prompt, someone can identify a measured, respectful way to respond to criticism. In the exchange itself, the person must first notice their defensiveness, understand what the criticism refers to, choose words, listen to the reply, and decide whether to follow up. The prompt records an answer, not those steps unfolding. Scoring keys should therefore be read as part of the evidence, not as a hidden technical detail. Ask what the item is meant to assess, whether the key follows a named theory or expert judgment, and whether the interpretation fits that target. Agreement with a key can support a claim about judgment on the test’s situations. A broader claim about someone’s interpersonal virtue or ability in every setting requires other evidence. The useful reading is local: this score reflects these responses, under this scoring rationale. It can point toward a behavior worth noticing, while leaving actual conduct to be judged from relevant examples. That remains a scoring choice, not an observation.

Sources: Situational Judgment Tests; The Measurement of Emotional Intelligence: A Critical Review of the Literature and Recommendations for Researchers and Practitioners; Measurement of Ability Emotional Intelligence: Results for Two New Tests

Does workplace-focused design bridge the gap to behavior?

A workplace setting makes an emotional intelligence test more relevant to work, but relevance is not the same as a record of what someone does at work. The Geneva Emotional Competence Test (GECo) is a useful case because its designers built the situations around workplace events and tested whether scores related to performance on computerized assessment centre tasks. That gives a stronger bridge from selected answers to behavior than a scenario test with no work context or performance comparison. The bridge still ends before ordinary conduct across meetings, deadlines, relationships, and organizations can be inferred.

The GECo research separates four competencies that are often blended under the label “emotional intelligence.” Emotion recognition is assessed through video clips of actors. Emotion understanding, regulation of one’s own feelings, and management of other people’s emotions are assessed through situational judgment items based on workplace scenarios. The distinctions matter: recognizing an expression, reasoning about why an emotion arose, and selecting a way to regulate it are related tasks, but success on one does not demonstrate all the others. The University of Geneva’s GECo page says the scenarios were informed by interviews with managers and professionals during the development phase. That makes the content recognizably work-focused; it does not mean the test observes a respondent in a real exchange.

Scoring also reflects design choices. For GECo’s scenario items, the page says correct and incorrect options were developed using established theories from emotion and organizational research. It describes the answers as determined through theoretical foundations, expert judgments, and general-population consensus. Those sources of judgment give the key a stated rationale. A keyed response therefore indicates how a person answered under the test’s defined alternatives and scoring rules. In a live conversation, the facts may be incomplete, the other person’s intent may be unclear, and several responses may be reasonable depending on authority, timing, or the harm at stake. A good score cannot settle which action a particular workplace required.

The GECo study reported five studies with 888 participants in total. According to the abstract reproduced by the University of Geneva research centre, the subtests showed high measurement precision using Item Response Theory and correlated in expected ways with other emotional intelligence tests, cognitive intelligence, personality, and demographic variables. The abstract also reports that GECo predicted performance on computerized assessment centre tasks in a sample of professionals. These findings support an instrument-specific claim: in the studied samples, the test’s components showed the reported measurement and relationship patterns, and scores were associated with a particular kind of simulated work performance. The abstract does not give the task-level effect size or details needed to judge the size of that association, so the word “predicted” should not be read as a claim of precise individual forecasting.

An assessment centre task is closer to behavior than choosing an option on a page because it asks a participant to do something in a structured simulation. It remains a simulation, however. The GECo result described in the abstract concerns computerized tasks, not a longitudinal record of how those professionals handled everyday disagreements or feedback. The page also reports an incremental association with academic performance in students beyond another ability emotional intelligence test. That is a different population and outcome; it adds context about the instrument’s research program, but it cannot strengthen a direct claim about routine conduct by employees. The source’s conclusion calls GECo promising for studying emotional competencies in organizational outcomes, which is appropriately narrower than saying it proves how a person will behave at work.

The practical lesson is to match the strength of the conclusion to the closeness of the evidence. GECo gives a more substantial bridge than workplace wording alone: its design uses work scenarios, and its research reports an association with computerized assessment centre performance among professionals. A reader can reasonably say that the result offers evidence about performance on the test’s emotion tasks and, for the studied samples, about its relationship to those simulated tasks. To claim what one person usually does in a real feedback conversation, the relevant evidence would need to come from that behavior in a relevant setting, described specifically enough to inspect. Workplace content makes a test pertinent; it does not make a scenario score a work history.

Sources: Geneva Emotional Competence Test (GECo)

What do broader SJT findings say about work performance?

Broader SJT research checks two tempting conclusions: that choosing an answer to a written situation says nothing about work, or that a test score tells us what one person will do in an exchange. The evidence supports neither extreme. Scores have been associated with performance criteria in some settings. Those associations describe patterns across samples and tests, across varied settings; they do not turn an emotional-intelligence scenario score into a forecast of an individual colleague’s conduct. The 2025 meta-analysis “The Validity of Single-Response Situational Judgment Tests: A Nomological Network Meta-Analysis” pooled results from 20 samples comprising 3,685 participants. For single-response SJTs, the authors report a correlation of .18 with job performance before correction and .20 after correction. A correlation describes how two measures varied together; it does not show causation or predict what one respondent will do. The review also found positive associations between SJT scores and emotional intelligence. This makes the broad claim that scenario responses are unrelated to work difficult to sustain. The same review marks the boundary. Its evidence base was still developing, and most studies used low-stakes laboratory settings. It examined single-response SJTs across content areas, rather than EI SJTs alone. The pooled relationship cannot be assigned to a particular EI test, scoring key, occupation, or person. The .20 corrected estimate is an average across studies, not a personal probability or percentage of work behavior explained. Extending it to “this person will handle feedback well” would require evidence the meta-analysis does not provide. A second synthesis looks more directly at interpersonal performance, but in a narrower domain. “Situational judgement test validity for selection: A systematic review and meta-analysis” examined SJTs used in medical selection. It included 30 studies, judged overall to be at moderate risk of bias; 26 reported validity correlations. Their pooled estimate was .32, with a 95% confidence interval from .26 to .39. These figures come from separate research bases and should not be ranked as if they compared tests in the same people. The medical review found very high heterogeneity, I² = 96.5%, meaning results varied substantially across studies. Its setting and selection use differ from an EI SJT used for personal reflection. The medical review’s relevant caution concerns the outcome. The authors called for more robust criteria that capture interpersonal aspects of typical workplace performance. A relationship with a selection criterion answers the reader’s question only as closely as that criterion does. A test might relate to selection scores while leaving open whether a person listens, regulates a reaction, or follows up during ordinary work. Heterogeneity and the call for better typical-performance measures limit how directly the review speaks to day-to-day conduct. Taken together, these reviews support a careful middle position. SJT scores can relate to performance evidence, so a scenario answer is not necessarily empty or arbitrary. But general SJT findings are bridge evidence, not direct evidence about the emotional skill captured by every instrument. The single-response meta-analysis covers a format across domains; the medical-selection review covers a profession and selection context. Neither validates a named EI test by association, and neither establishes what a particular person will do in a specific meeting. Instrument content and scoring, the studied population, and the criterion shape the inference. When a source cites “SJT validity,” ask what counted as performance. Was it a later job rating, another test, a simulated task, or an observed exchange? How close was the measure to the behavior you care about? These questions keep a pooled finding at the right scale. A sample-level association can justify investigating a test’s work relevance, but cannot substitute for evidence about an individual’s behavior. For someone considering an EI SJT, the practical conclusion is modest. Broader findings make it reasonable to treat scenario performance as potentially informative about work-related judgment when a named instrument has suitable evidence in those samples. They do not justify using a score alone to decide how someone handles disagreement or feedback at work. To answer that individual question, seek concrete examples or observations in the relevant setting, and tie the claim to the behavior actually seen. The distinction is between a test’s relationship with performance in studied groups and what this person does in the exchange that matters.

Sources: The Validity of Single-Response Situational Judgment Tests: A Nomological Network Meta-Analysis; Situational judgement test validity for selection: A systematic review and meta-analysis; Situational judgement test validity for selection: A systematic review and meta-analysis

Why is a scenario score different from work behavior or self-report?

A situational judgment test (SJT) records a response to a designed problem. A self-report records how someone describes usual behavior. A work account records what happened in one setting. Each is a different kind of evidence, at a different distance from conduct in the moment. Treating them as interchangeable turns a narrow result into a broader claim than the evidence can carry.

| Evidence | What it records | What it can help answer | |---|---|---| | EI SJT | A response to an authored emotional situation | How someone judges the presented problem under the test’s rules | | Self-report | A person’s description of typical feelings, habits, or competencies | How they see their usual pattern | | Work example | Actions and words in a specific exchange | What occurred in that setting, and what to examine next | The U.S. Office of Personnel Management’s “Situational Judgment Tests” guidance calls an SJT a low-fidelity simulation: it presents a work problem but does not place the respondent in a simulated setting to perform the behavior. Suppose a person selects a tactful response to a feedback prompt. The answer shows what they chose when the situation was described and options were available. It does not show whether they noticed defensiveness, heard criticism accurately, asked for clarification, or later closed the loop. These are possible observations, not a validated checklist or a claim about a particular person.

Self-report sits at a different point. It asks someone to summarize what they generally feel or do, so it concerns perceived typical behavior rather than performance on an authored emotion problem. “Measurement of ability emotional intelligence: Results for two new tests” distinguishes hypothetical EI items, which ask what someone should do, from items that directly assess a capability in action. Its instruments, STEU and STEM, are research measures; evidence about them cannot simply be transferred to another test. A person may understand a habit while still lacking evidence about how it unfolded in one exchange.

“Why does self-reported emotional intelligence predict job performance? A meta-analytic investigation of mixed EI” makes this boundary important. The analysis examined mixed EI measures, combining ability-related content with self-perceptions and other material. It reported substantial overlap with ability EI, self-efficacy, self-rated performance, several personality traits, and general mental ability (multiple R = .79). The updated association between mixed EI and supervisor-rated job performance was ρ = .29; after controlling for those overlapping constructs, it was β = −.02. These findings concern mixed EI and supervisor ratings across studies. They do not estimate an individual’s behavior after an ability-oriented SJT, or show that self-report and scenario performance mean the same thing.

A concrete work account also needs its conditions. Consider an illustrative feedback exchange: a colleague says a handoff omitted an important detail. An account could separate what was said, what the recipient did next, and whether the detail was later supplied. It might record a clarification or boundary. Urgency, authority, missing information, cultural expectations, and conversational safety can shape what response was possible and how it was understood. Those conditions keep an inference local. One tense reply cannot settle how someone usually handles feedback.

The forms answer related but distinct questions. An SJT can prompt reflection on judgment; a self-report can surface a pattern someone believes they show; a specific work example grounds discussion in what happened. To understand collaboration, use the test result to choose a behavior worth noticing, then seek a recent example or concrete feedback about it. Ask for an instance rather than a general label, and keep the description close to the behavior at issue. This helps make the next conversation about an action that can be described, rather than a trait inferred from a score. Keep conclusions close to the evidence collected. A score may suggest what to examine; what someone did in the relevant exchange requires evidence from that exchange.

Sources: Why does self-reported emotional intelligence predict job performance? A meta-analytic investigation of mixed EI; Situational Judgment Tests; Measurement of Ability Emotional Intelligence: Results for Two New Tests

How should you use the result in a real conversation?

Use a scenario result as a question to carry into a conversation. Identify the instrument and the kind of response it sampled: recognizing an emotion, interpreting its cause, or choosing how to respond. The EQ Test Emotional Skills Profile differs from research instruments including STEU, STEM, and GECo. It is a 32-item educational reflection tool. Its recent-behavior summaries are separate from judgments on authored scenarios; it has no validated norms and is not an ability score or work-performance forecast.

Choose one behavior to notice, such as whether you pause to clarify feedback before answering. Afterward, record what was said, what you did, and what followed. “I asked which part of the handoff needed changing” is more useful than “I handled criticism well.” If a trusted colleague was present, ask for an example of what they observed and one alternative response they think might help. Their account adds a perspective; one exchange cannot define a pattern.

The 2019 critical review of emotional-intelligence measures distinguishes tests of emotional understanding and regulation from other ways of assessing EI. Keep each result tied to its instrument. For a structured prompt for private reflection, explore the Emotional Skills Profile at /assessment. Broader claims about work behavior need relevant evidence; the scenario answer can help you decide what to look for.

Sources: Emotional Skills Profile; The Measurement of Emotional Intelligence: A Critical Review of the Literature and Recommendations for Researchers and Practitioners

Questions readers ask