Outgrace
OutgraceThe Leadership Magazine, home
LeadershipExplainer

Emotional intelligence scores and leadership: what they can predict

The sources we checked say nothing about whether emotional intelligence scores predict leader performance. Here is how to test an instrument before it shapes a hire.

This piece sets out what a leader should expect an emotional intelligence instrument to tell a selection panel, and where the evidence for that claim runs out. One caveat comes first. The sources compiled for this article say nothing about emotional intelligence instruments or leadership outcomes, so the predictive claims cannot be checked from them. What follows uses the one body of guidance we could check, the US Office of Personnel Management's material on assessment, and then sets out the tests a reader can apply to any instrument on the table.

How it works

The OPM defines personnel assessment as any method of collecting information on individuals for a selection decision. In its usage, selection decisions cover hiring, placement, promotion, referral, retention and program entry. That breadth matters. A leadership questionnaire used to pick a successor is an assessment in exactly the sense the OPM means, whatever the vendor chooses to call it.

An assessment tool is any test or procedure that evaluates job-related competencies, interests or fitness for employment. The phrase job-related is the hinge. A tool earns its place by measuring something the job requires, not by being interesting. On that definition an emotional intelligence score has no standing until someone shows that the job needs what the score measures.

Job-related and predictive are not the same thing. A tool can describe a candidate accurately and still forecast nothing about how that candidate will perform. The OPM names predictive validity as the most important characteristic of a tool, meaning the accuracy of its scores in forecasting on-job performance. The page attributes that definition to Schmidt and Hunter's 1998 paper in Psychological Bulletin, volume 124, pages 262 to 274, on the validity and utility of selection methods. We have not checked that paper, so treat the citation as the OPM's, not ours. The practical question is simple to state: do people who score high go on to do the job better? Answering it needs performance data, not a brochure.

Start with the job, not the instrument. The OPM says effective personnel assessment starts with a job analysis that identifies the duties and competencies needed for effective performance. It adds that basing an assessment on that analysis makes the link between job requirements and tools more transparent, and that this improves perceived fairness. Its examples of success factors are oral communication and problem solving. Emotional intelligence does not appear among them.

The same logic applies to how much effort an assessment deserves. The OPM notes that the choice of assessment strategy depends on hiring timetables, staff and financial resources, the number of positions, and the nature and complexity of the work. A search for three senior roles and a search for two hundred call for different tools. A board may not have the resources to validate a bespoke instrument for either, and that limit belongs in the decision, not in the footnotes.

Who it affects

The sources in this set concern US federal agencies and their hiring and promotion processes. The OPM page names no companies, sectors or regions in connection with emotional intelligence, and nothing in the set addresses leaders in private firms. Readers outside government should treat the framework as a method, not a rule. It transfers because the questions are the same: what the job needs, what the tool measures, and what a wrong decision costs.

The cost of a wrong decision is the part leaders most often underweight. The OPM lists agency-wide consequences of selection error: customer complaints, accidents and injuries, absenteeism, poor work quality, turnover and reputational damage. It also says that selection errors can require managers to train and counsel marginal employees, while coworkers absorb the extra workload. A leader chosen badly does not fail alone. The cost lands on the team that reports to them.

The same page links good assessment to employee organisational commitment and job satisfaction, through better job matching. That is a claim about assessment in general, not about emotional intelligence, and the sources in this set do not connect the two. A reader who wants the link for leaders will have to build the evidence from their own hiring and performance records.

What leaders get wrong

  • Treating a high score as a forecast. None of the sources gives a validity coefficient for any emotional intelligence instrument, and none tests an instrument against leader performance, promotion or retention. A score may be meaningful. Nothing in this set shows that it is.
  • Skipping the question of what is measured. No source says whether emotional intelligence instruments measure ability, self-report or ratings by others. In our judgement these are different measurements, and a report should say which one it delivers before anyone reads the number.
  • Using one tool for every role. The OPM states that not all assessment tools are appropriate for every job and organisational setting, and that agencies must weigh several factors. A tool that suits a sales director may say little about a plant manager.
  • Mistaking volume for quality. The OPM states that the quantity of applicants does not guarantee quality, and that assessments narrow large applicant pools. A long shortlist produced by an attractive instrument is still a shortlist.
  • Assuming academic sources have covered it. The NBER and Harvard pages checked for this set cover AI innovation, health spending, entrepreneurship training, dementia coverage, interoception and meditation. Neither addresses leader assessment or links emotional intelligence to leadership.

What to do about it

Before any emotional intelligence instrument enters a selection process, write down the job. List the duties and competencies the role needs, in the words the hiring manager would use. Then ask the vendor which of those competencies the instrument measures, and what evidence links its scores to performance in roles like this one. If the answer is a general norm sample or a sales deck, the instrument is not yet a selection tool.

Second, ask for validity evidence in the form the OPM frames it: scores set against on-job performance. A reasonable test is whether the vendor can show that people who scored higher were later rated higher by their managers, and whether the study population resembles your own roles. If no such study exists, the instrument can still inform a conversation. It should not carry the decision.

Third, treat a score as a prompt rather than a verdict. A result may suggest a structured question in an interview, or a probe into how a candidate handled a difficult team. Ask the question, record the answer, and judge it against the job analysis. The OPM's examples of success factors are communication and problem solving. A panel that replaces those with a single number has stopped doing job analysis.

The OPM puts the underlying point more plainly than most vendors do:

It is very simple — using effective assessment tools will reduce the degree of error in making hiring decisions.

Office of Personnel Management, assessment guidance, opm.gov

The ambition is right; the guarantee is not. An assessment reduces error only when the tool measures what the job needs, and the sources in this set do not show that for emotional intelligence. Keep a record of each score and what happened next, including how the hire performed at six and twelve months. Without that record, the predictive question cannot be checked later, and the instrument will keep being judged on its reputation.

Finally, have your own employment counsel review how any assessment is used in hiring and promotion, including any fairness or disclosure obligations where you operate. The sources here do not address employment law for private firms, and this article does not give that advice.

Newsletter

The Briefing

One email a week: the leaders we profiled, the ideas worth stealing, and nothing else.

Weekly, free, one click to leave. Read our privacy policy.