What EQ Training Gains Remain After a Course?

The headline result is about improvement, not duration

The 2024 review, “Training emotional competencies at the workplace: a systematic review and metaanalysis,” asks whether emotional-competency measures improve by the end of workplace training. Across 50 studies, its pre-to-post analysis found an average standardized mean difference (SMD) of 0.44 (95% CI 0.29–0.59). In 27 controlled trials, the training-versus-control difference was 0.47 (95% CI 0.31–0.62). Both averages favor training, but neither tells a worker how many points they should gain or how long change will last. An SMD expresses a difference in standard-deviation units, allowing results from different scales to be summarized. It is not an EQ score or a percentage improvement. Results varied substantially: for the pre-to-post estimate, I² was 88%. The studies did not point to one consistent-sized effect. The pooled figure summarizes evidence; it is not a forecast for every course or participant. The intervention pool matters as much as the headline. The authors use “emotional competencies” for training in emotional intelligence, empathy, and emotion regulation. Of the 50 studies, 41 focused on emotional intelligence, seven on empathy, and two on emotion regulation. These programs and measures do not represent one standard course or interchangeable skill. The positive result supports trainability across a mixed field of workplace interventions, rather than proving that a particular feedback, boundary-setting, or conflict-repair technique will improve. The two main estimates compare scores before and after training, or trained groups with control groups. They address improvement associated with training, not what remains after the course. The authors say effects persisted beyond three months, but that duration claim depends on a much smaller set of studies with follow-up measurements. To judge persistence, we need to examine which studies measured participants later, when, and against what comparison. The headline establishes a positive training signal; it does not settle retention. These broad comparisons establish the case for improvement; a separate follow-up analysis is required to assess whether measured gains remain after training ends.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis

What does the follow-up estimate measure?

The follow-up estimate asks a narrower question than “Did the course work?” In the 11 studies that measured participants again after the immediate post-course assessment, did their later emotional-competency scores differ from the scores recorded just after training? In “Training emotional competencies at the workplace: a systematic review and metaanalysis,” that post-to-follow-up comparison produced a standardized mean difference (SMD) of 0.37 (95% confidence interval [0.01, 0.74]; I² = 94.91%). An SMD expresses an average difference in standard-deviation units, allowing results from different scales to be combined. It is not a number of EQ points. The positive estimate points to a small average change between the two measurement occasions in the direction the review treats as improved scores. Its confidence interval is the range of values compatible with the analysis under its assumptions: here, the lower end sits just above zero, while the upper end allows a substantially larger difference. That interval conveys uncertainty about the pooled average. I² describes how much the study estimates varied beyond what sampling chance alone would explain. At 94.91%, the variation was very high. The average therefore does not describe one dependable course effect shared by all 11 studies. Programs, participants, measures and timing differed; a worker cannot read 0.37 as the gain they should expect to keep. A second comparison in the same subset helps prevent a common misreading. From before training to follow-up, the pooled SMD was 0.97 (95% CI [0.42, 1.52]; I² = 97.11%). That pre-to-follow-up contrast asks whether follow-up scores stood above the original baseline. The post-to-follow-up estimate instead compares later scores with the immediate post-course level. Those are different reference points. A positive pre-to-follow-up difference can coexist with some decline from the post-course score: participants may remain above where they began without retaining every measured gain seen just after training. The two averages cannot be subtracted to calculate an individual’s retained share, and the review does not offer a personal retention percentage. The distinction matters when someone reads “lasting effect” in the paper. The authors interpret their pooled follow-up findings as evidence that some training effect remained after the course. The data support a positive average post-to-follow-up contrast across this selected set of studies, but its wide interval and extreme between-study variation leave the size uncertain. It does not tell a person whether their own ability to pause before responding, for example, improved or endured; the pooled estimate combines emotional-competency measures rather than a single named behavior. Nor does it turn a self-reflection score into proof that a skill was retained. The careful conclusion is possibility at group level: later scores were, on average, higher than immediate post-course scores in this subset, with substantial uncertainty about how much and for whom.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis

Why the three-month phrase is easy to overread

The review says that training effects persisted more than three months after courses ended. That is the authors’ interpretation of their pooled evidence, and it deserves to be reported. The timing details, however, make the sentence easier to overread than its confident wording suggests. Across the 50 studies in the main review, only 11 included a follow-up assessment. The average interval from the end of training to the first follow-up was 26 days. About 20% of studies first measured participants later than four weeks. The figures do not describe a common three-month appointment for these workers. They describe a small set of studies whose measurement schedules differed. (Training emotional competencies at the workplace: a systematic review and metaanalysis.) A mean of 26 days is not a claim that every study checked at day 26. Some first assessments were later, and roughly one fifth were later than four weeks; subsequent assessments could also contribute where studies reported them. The mean compresses that uneven schedule into one summary. It cannot tell a reader how many people were measured at precisely three months, or whether a particular course’s participants were followed for that long. In the paper’s discussion, the authors explain that they included all available time points because some earlier research suggested emotional-competency changes might become detectable only after practice and time. The pooled analysis therefore answers a broader question across the observations available, rather than estimating change at one shared calendar point. This is why the authors’ wording and the timing summary should travel together. They report a positive average difference between immediate post-training and follow-up scores among the 11 studies, then interpret the pattern as evidence that some effects persisted beyond three months. But the reported average first-follow-up interval is much shorter than three months, and the included studies did not all use the same schedule. A short average interval does not prove that no later observations support the interpretation; equally, the authors’ interpretation does not turn the varied schedule into a standardized three-month retention test. The evidence supports a positive longer-term signal as the authors describe it, with sparse and uneven timing limiting how specifically its duration can be stated. A reader summarizing this review can say that some measured gains remained detectable at follow-up and that the authors interpret their findings as persistence beyond three months. The timing qualification matters if the claim is being used to judge a course, compare programs, or set expectations for personal progress. “Skills last three months” sounds like a typical duration established by repeated checks at that interval. This review does not provide that kind of common checkpoint or guarantee for an individual. Its follow-up results are better read as evidence that gains may remain measurable after training, while the exact time course remains uncertain. That distinction keeps the positive result intact without making the calendar claim more precise than the study design allows.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis

Does the timing analysis settle how long gains last?

The timing comparison does not tell us when a training effect fades. In “Training emotional competencies at the workplace: a systematic review and metaanalysis,” the researchers found no statistically significant difference between effect sizes from studies whose follow-up took place within four weeks and those with a longer interval. That result leaves the timing question unresolved. It does not show that early and later follow-ups produce equivalent results, and it does not identify a point at which gains begin to shrink. A non-significant difference can arise when two groups really are similar, but it can also reflect limited information or estimates too imprecise to distinguish. Here the comparison drew on a small follow-up pool, and the underlying studies varied in timing, training, and measures. The review reports that only 11 of its 50 studies had follow-up data; for the broader set, around one fifth first followed participants later than four weeks. A comparison across these uneven intervals cannot function like a controlled timeline in which the same course is measured at fixed points. Without more and better spaced observations, the study cannot rule out a timing effect simply because its subgroup test did not detect one. The authors also report smaller effect sizes at second follow-up than at first follow-up. This pattern is consistent with some decline in measured gains over time, but it is not a decay curve: the available studies do not establish one common starting point, interval, or rate of change. Nor does the pattern reveal why scores were smaller later. The authors suggest emotional competencies may need continuing practice or reinforcement to sustain improvement. That is a reasonable interpretation to investigate, not a tested booster schedule or proof that a particular follow-up session would prevent decline. A course provider should not turn it into a promise that refreshers maintain a known percentage of skill. So what can a reader say about duration? The review offers evidence that measured gains remained detectable at follow-up on average, while its timing analysis cannot locate a reliable fade point. It supports neither “the effect disappears after four weeks” nor “the effect lasts for three months” as a typical course-level rule. The paper’s broad persistence language should be read alongside the uneven follow-up dates and the authors’ own finding that the within-four-week and later groups did not differ significantly. Those observations describe limits in the evidence; they do not cancel the positive follow-up estimate. For a specific course, maintenance claims need repeated assessments at stated intervals and a measure tied to the skill being taught. If the claim concerns handling disagreement, for example, an evaluation could report what was measured after the course and again later, rather than attaching a duration to “emotional intelligence” in general. Until such evidence is available, a worker can treat continued practice as a sensible development choice, not as a proven requirement or guarantee. The study leaves the practical question open: whether the targeted behavior is still showing up when work creates another chance to use it.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis

The review cannot name the skill that lasted

A positive follow-up average cannot tell a reader whether a course taught empathy, emotion regulation, or a specific way to handle disagreement. In “Training emotional competencies at the workplace: a systematic review and metaanalysis,” the authors deliberately combined interventions aimed at emotional intelligence (EI), empathy, and emotion regulation. They found no statistically significant difference in effects by the stated training construct. That result may sound like these courses work equally well across skills. The authors offer a narrower explanation: the categories and the training methods overlap, so the analysis may not be comparing cleanly separated kinds of learning.

The overlap is visible in both the concepts and the exercises. Some mixed models of EI include empathy and emotion regulation as facets. In the studies reviewed, EI courses could include lectures about emotion or practice in perspective-taking, which also appeared in empathy training. A course label therefore did not guarantee a distinct lesson or outcome. The review says it remained unclear which aspects of emotional skills were actually trained and what outcome assessors measured. Its pooled result combines these different targets rather than showing that any one target improved by the same amount.

This matters when moving from a broad scale to a workplace action. A score may summarize several facets, and the review says that when studies reported subscales, the researchers sometimes averaged those effects into an overall score. They acknowledge that this choice could hide improvement in particular facets. The reported persistence estimate therefore cannot be assigned to receiving feedback without defensiveness, setting a boundary, noticing rising frustration, or repairing a strained exchange. Those are plausible training goals, but the meta-analysis does not isolate them as lasting outcomes.

The authors suggest self-awareness or self-reflection might be a shared element across programs, and they note that group learning, cohesion, and feedback could contribute to personal change. These are explanations they propose, not mechanisms identified by the analysis. Most reviewed trainings included reflection, discussion, role-play, or practice, but the paper does not show which ingredient accounts for the follow-up effect. Nor does a shared ingredient mean participants acquired the same skill: reflection could help someone name an emotion while leaving their response in a tense conversation unchanged.

For a course-specific claim, ask what behavior was practiced and whether the later measure sampled it. If a workshop teaches a pause before responding to criticism, a useful follow-up would name that action and check it again after the course, rather than rely only on a broad emotional-competency total. This is a way to make a claim testable, not a procedure validated by this review. The key distinction is between a positive average across varied programs and evidence that one defined skill remained available in ordinary work. The study supports the former; it leaves the latter unresolved.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis

A score can persist without proving a work behavior

A later questionnaire or ability-test score records performance on that measure. By itself, it cannot show that a coworker saw a lasting change in a meeting, handoff, or disagreement. Training research treats these as related but distinct outcomes: a person may retain learning on an assessment without demonstrating the skill at work. For a course claim, “the score remained higher” and “the behavior changed on the job” therefore require different evidence.

The 2024 review, “Training emotional competencies at the workplace: a systematic review and metaanalysis,” makes the distinction important. Of its 50 studies, 34 used self-report, nine peer or 360-degree reports, and seven ability tests. The authors say these approaches capture different aspects. These counts describe the full pool, not necessarily the 11 follow-up studies, so the review does not specify which measures produced that pooled estimate. They do show why “emotional skills score” is too broad a label for treating every result as the same thing.

A self-report asks people to describe their usual behavior or perceived capability. An ability test asks them to solve emotion-related problems under its scoring rules; a peer or 360-degree report gathers judgments from others. None is a direct window onto every work interaction. Someone may report using a calming strategy more often, while a colleague notices whether they interrupted during a disagreement. These measures concern connected parts of emotional competence, but a change in one does not establish a matching change in another.

Training transfer means applying what was learned in the work setting. In “A Meta-Analysis of the Relations Among Training Criteria,” the authors analyzed 34 studies and 115 correlations, distinguishing post-training learning, retention, and behavior or skill demonstration. Their analysis treats on-the-job performance as a transfer criterion separate from immediate or retained learning. This is a general training framework, not an estimate of emotional-intelligence course effects. Its relevance is conceptual: a follow-up test can show that measured learning remains, while a claim about changed work behavior needs evidence that samples behavior in that setting.

Suppose a course teaches one way to receive disagreement: pause, summarize the concern, then respond. A later questionnaire about confidence could indicate a reported change. To support a claim that the skill appeared at work, an evaluation could also examine later disagreement exchanges against that behavior, or ask people familiar with them what they observed. This pairing illustrates how evidence might match a claim; the 2024 review did not test it. A colleague’s report remains a perspective shaped by what they saw and how they interpreted it, not an automatically objective record.

A higher follow-up score can support a claim about a measured emotional competency, with the measure named. A lasting-behavior claim calls for a later observation or work-context report tied to the taught behavior. Neither alone establishes improved team outcomes; those need separate measurement. The review’s follow-up signal is meaningful, but readers should ask what the measure captured before translating a retained score into a claim about how people work together.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis; A Meta-Analysis of the Relations Among Training Criteria

Why the positive signal survives, but stays uncertain

The positive result does not rest only on the 18 studies that the review judged at critical risk of bias. In its sensitivity analysis, the authors removed those studies and recalculated the overall pre-to-post change across the remaining workplace interventions. The estimate shifted from SMD 0.44 to 0.40, with a 95% confidence interval of 0.26 to 0.54. A standardized mean difference (SMD) expresses the average change in standard-deviation units, allowing results from different scales to be combined. The smaller estimate still points toward improvement in emotional-competency measures by the end of training. That is a useful robustness check: the main positive signal remains when the most serious risk category is excluded. It is not a test of whether skills persist after training, because this calculation addresses the broader pre-to-post evidence. The quality review also shows why the size of that signal deserves care. Among 50 studies, 18 were rated at critical risk of bias, 28 at serious risk, three at moderate risk, and two at low risk. Critical ratings came from possible confounding in before-and-after studies without a comparison group; a change over time could not be cleanly separated from the intervention. Many studies rated serious risk had outcome-measurement concerns because participants and assessors knew about the training, and outcomes included self-reports or peer reports. The authors also found evidence consistent with publication bias in the broader pre-to-post analysis: smaller or less favorable studies may be underrepresented. These weaknesses can make an average look more convincing or larger than the underlying training effect. Yet throwing out the entire review would ignore the sensitivity result. The more proportionate reading is that workplace emotional-competency measures generally improved in the available studies, while confidence about how much they improved remains limited by study design and measurement. The estimate after excluding critical-risk studies still had substantial variation across studies, so it does not imply every course worked or that workers changed by the same amount. Nor does it establish that a change on a questionnaire or test was visible in an interaction at work. Most important for the question of what lasts, this sensitivity analysis belongs to the review’s broad pre-to-post estimate, not its smaller set of follow-up studies. Excluding critical-risk studies does not add missing later observations, align follow-up dates, or show that one named skill endured. It makes the case for trainability harder to dismiss; it does not make the duration claim precise. The evidence therefore supports cautious confidence that training can improve measured emotional competencies, alongside an open question about how much of that improvement remains later and appears in a specific work behavior.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis

A narrower healthcare comparison adds context, not confirmation

A second 2024 review points in a similar direction, but it does not answer the same question. “Emotional intelligence training among the healthcare workforce: a systematic review and meta-analysis” examined 17 longitudinal studies of EI training among healthcare workers aged 18 and older. It reports that all 17 studies found an increase in EI after intervention. The authors also warn that limitations in the underlying studies may have overstated effects, and call for further research on training duration and delivery. This is evidence of post-intervention improvement in a particular workforce, not a shared estimate of how much skill remained at a common later date.

The difference in scope matters. “Training emotional competencies at the workplace: a systematic review and metaanalysis” brings together workplace programs addressing emotional intelligence, empathy and emotion regulation, across occupations and measures. The healthcare review focuses on EI in one employment sector. It describes varied study designs and, in its full text, variation in intervention length, format and assessment methods. That makes it useful context for the focal review’s positive direction and measurement problem. It does not make the two results interchangeable: the healthcare review is not an independent replication of the workplace review’s post-to-follow-up estimate.

There is also a difference between an increase after a course and retention. The healthcare review reports EI measured after intervention; its accessible abstract gives no common follow-up interval or pooled comparison between immediate post-course scores and later scores. Even if some included studies took later measurements, the abstract’s summary does not establish that their timing and findings support a single duration claim. It cannot confirm that emotional skills typically last three months across workplaces. The focal workplace review is more directly relevant because it analyzes change from post-course assessment to follow-up, although only for a small, unevenly timed subset.

Together, the reviews make a measured case for possibility: training can be followed by higher emotional-intelligence or emotional-competency measures in the populations studied. The healthcare review’s cautions also show why a positive direction is not a maintenance estimate. Its findings should not be generalized automatically to every occupation. A rise in a measure does not show that a colleague later observed a lasting change in a disagreement, handoff or feedback conversation; that behavioral claim needs evidence gathered in work situations.

This comparison strengthens the case that emotional skills are trainable while leaving duration unresolved. It cannot tell an employee how long a specific course’s effect will last, which skill persisted, or whether a measured gain transferred into daily conduct. A provider making a retention claim needs evidence matched to it: the skill named, the follow-up interval reported and a suitable measure at that time. If the claim concerns work behavior, follow-up should examine that behavior in context. The healthcare review offers support for post-course improvement, not a substitute for such evidence.

Sources: Emotional intelligence training among the healthcare workforce: a systematic review and meta-analysis; Emotional intelligence training among the healthcare workforce: a systematic review and meta-analysis; A systematic review and meta-analysis of the effects of emotional intelligence training in adults

What stronger follow-up evidence would look like

A follow-up study should name the emotional skill, describe the course, and state when the skill is measured again. In *Training emotional competencies at the workplace*, Mehler and colleagues call for higher-quality studies, including randomized controlled trials, and better intervention descriptions. A comparison group can help distinguish course-related change from other explanations. Researchers should also report who completed later assessments, since missing responses affect whose results are represented.

The outcome should fit the claim. If a course teaches people to summarize a concern before replying in disagreement, researchers could measure that skill after training and at later intervals, then separately assess whether it appears in relevant work interactions. A scale score and a work-context report answer different questions. Ford, Baldwin and Prasad’s *Transfer of Training: The Known and the Unknown* recommends precise intervention and variable measurement, explicit transfer criteria, and trajectories over time. Applied here, that means tracking what was learned and whether it appears at work, rather than treating irregular assessment dates as one duration. A team-culture claim needs a defined measure.

This gives readers a way to distinguish immediate learning, later measured retention, and behavior at work when weighing course claims. It is an evidence standard to look for, not a protocol tested by the 2024 review.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis; Transfer of Training: The Known and the Unknown

The proportionate conclusion and next step

The workplace meta-analysis supports a measured conclusion: some emotional-competency scores remained higher at follow-up than immediately after training, on average. Its 11 follow-up studies were few, varied widely, and assessed participants at uneven times. The authors describe effects as persisting beyond three months, but average first follow-up was 26 days. This cannot promise a skill will last three months or that coworkers will notice change. It justifies practice and a closer look at what a course’s follow-up measured.

A personal check starts with one behavior the course addressed. If the practice was receiving disagreement, notice whether you can summarize the other person’s concern before replying in a later conversation. You might ask a colleague what they observed in that exchange. This is a practical application, not a retention method tested by the meta-analysis; one conversation cannot establish a pattern. It gives you a prompt for practice beyond asking, “Has my EQ improved?”

The private 32-item Emotional Skills Profile can help you reflect on recent emotional habits and judgments about authored scenarios. It is an educational reflection tool without validated norms or established ability-test validity, so it cannot verify training retention or workplace behavior. The research signal is positive; a personal conclusion requires a defined behavior in context. After your course, revisit one behavior and ask a colleague about that exchange rather than treating a follow-up score as proof of lasting change.

Sources: Training emotional competencies at the workplace: a systematic review and metaanalysis

Questions readers ask