The Executive Function Curriculum Problem

I’m busy trying to choose a school for our oldest. “Executive Function” courses seem to be all the rage right now. But.. why? On what basis? Are they any good?

A head of school told me they “follow the research” on school curriculum but what does the research actually say about “executive function”?

Beyond even this — curriculum shapes values. What values do “executive function” courses smuggle into your child?

The Machine That Shapes the Souls

Before examining what the research says about executive function training, it is worth asking a prior question that rarely gets asked: why this curriculum, and in service of what vision of human flourishing?

The logic that generates EF curricula follows a pattern that has repeated across decades of education reform. Schools observe the economy as it currently exists. They identify the traits that correlate with success within it. They reify those traits as intrinsic cognitive capacities — “executive function,” “grit,” “21st century skills.” They build curricula to install those capacities in children. And at no point in the process does anyone ask whether “success in the current economy” is a worthy aim for the formation of a human soul.

Consider the three core components of executive function as the research literature defines them: working memory (holding information in mind for task completion), inhibitory control (suppressing impulses that interfere with productivity), and cognitive flexibility (adapting to changing task demands). Read generously, these describe aspects of mature cognition. Read honestly, they are a spec sheet for a compliant knowledge worker. They describe what the contemporary economy needs from its inputs, not what a human being needs to flourish. And when schools package this as a curriculum — as a set of skills to be trained — they are smuggling in a set of values under the cover of neutral science. They are not saying “we want to produce adaptable workers.” They are saying “we want to develop executive function,” which sounds like cognitive science but encodes the same commitment.

William Morris saw this clearly in the industrial age: when education takes its marching orders from the economy, it produces souls shaped to serve the machine rather than to question whether this is the machine we want. Wendell Berry observed that rural schools systematically train children to leave — to acquire the skills valued by the urban knowledge economy, to see their home communities as places to escape from rather than places to build. C.S. Lewis warned in The Abolition of Man about education that conditions rather than initiates — that trains children to perform the right behaviors without ever forming the moral imagination that would let them evaluate whether those behaviors are right. An EF curriculum that trains inhibitory control without asking “in service of what?“ is producing what Lewis called “men without chests”: technically capable and morally vacant.

There is also a temporal absurdity. The correlations that underpin EF curricula were derived by studying people who grew up in the 1980s and 1990s, identifying which childhood traits predicted success in the 2000s and 2010s economy. Today’s kindergarteners will enter the workforce around 2044. The labor market that generated the data will not be the labor market they encounter. Every generation of education reformers fits a curve to historical data and extrapolates into a future that has no obligation to follow the trend. The “21st century skills” movement was designed for a world that is already being eaten by artificial intelligence. EF curricula are built on correlations from a world that may be unrecognizable by the time their students graduate.

The circularity is the deepest problem: the machine shapes what we value; what we value shapes education; education shapes the next generation; the next generation perpetuates the machine. At no point does anyone step outside the loop to ask whether the machine deserves perpetuation.

This essay does not resolve that philosophical question. But it proceeds from the suspicion that when a curriculum arrives pre-fitted to the needs of the existing economic order, dressed in the language of cognitive science, and promising to optimize children for “success,” the burden of proof should be extraordinarily high. As it happens, EF training curricula cannot meet even a modest evidentiary burden — let alone the extraordinary one their ambitions demand.

The Empirical Case

Executive function — the set of cognitive processes encompassing working memory, inhibitory control, and cognitive flexibility — is a real and well-studied construct in cognitive science. Children with stronger executive function do better in school, earn more as adults, and have better health outcomes. None of that is in dispute.

What is in dispute, and what the weight of evidence increasingly answers in the negative, is whether executive function can be trained through standalone curricula in ways that actually improve academic performance or life outcomes. Schools across the country are adopting packaged EF programs on the strength of a compelling but fundamentally flawed syllogism: EF predicts success; therefore, training EF causes success.

What follows traces six lines of evidence suggesting that EF training curricula, as currently marketed and implemented, do not deliver on their central promise.

1. Far Transfer Is Essentially Zero

The entire value proposition of standalone EF curricula rests on far transfer — the idea that training a general cognitive capacity (like working memory) in one context will improve performance in a different domain (like reading or math). If you train inhibitory control with a computer game, the promise is that your students will also become better at ignoring distractions during a math lesson.

The meta-analytic evidence on this question is unusually clear by social science standards, and the answer is not encouraging. A meta-analysis published in Psychological Bulletin examining children’s executive function training found:

“A significant near-transfer effect (g+ = 0.44, k = 43, p < .001) showing that the interventions in the primary studies were successful in training the targeted components. However, we found no convincing evidence of far-transfer (g+ = 0.11, k = 17, p = .11). That is, training a component did not have a significant effect on the untrained components.”

— Kassai, R., Futo, J., Demetrovics, Z., & Takacs, Z.K. (2019). Psychological Bulletin, 145(2), 165–188.

In plain language: EF training makes children better at the specific tasks they practiced. It does not make them better at other executive function tasks, let alone academic subjects. Children who drill working memory get better at working memory games. Their reading and math do not improve.

This finding has been replicated at the highest level of meta-analytic aggregation. A second-order meta-analysis (a meta-analysis of meta-analyses) across multiple types of cognitive training reached the same conclusion:

“Recent meta-analyses summarizing the extent empirical evidence have resolved the apparent lack of consensus in the field and led to a crystal-clear conclusion: The overall effect of far transfer is null, and there is little to no true variability between the types of cognitive training.”

— Gobet, F. & Sala, G. (2023). Cognitive Training: A Field in Search of a Phenomenon. Perspectives on Psychological Science, 18(1), 125–141.

This is not a marginal or contested finding. Across working memory training, video-game interventions, exergames, and music-based training, the pattern is the same: near transfer yes, far transfer no. Without far transfer, standalone EF curricula have no mechanism to improve academic outcomes. They are, at best, making children better at brain games.

2. Correlation Is Not Causation: EF May Be an Indicator, Not a Lever

The sales pitch for EF curricula leans heavily on correlational data — and the correlational data is genuinely impressive. Executive function predicts academic outcomes, sometimes more strongly than IQ. Children with better self-regulation at age 3 have better life outcomes at age 33. But prediction is not the same as intervention, and the history of one of psychology’s most famous experiments illustrates exactly why.

The Marshmallow Test: A Cautionary Tale

Walter Mischel’s marshmallow test, begun in the 1960s at Stanford’s Bing Nursery School, became foundational to the EF-as-lever framework. Children who could delay gratification at age 4 — waiting for two marshmallows instead of eating one immediately — had better SAT scores, better health, and better life outcomes decades later. The interpretation seemed obvious: self-control is an intrinsic cognitive trait, and it causes later success. Train the trait, improve the outcomes.

That interpretation has not survived scrutiny.

In 2013, Kidd, Palmeri, and Aslin ran an elegant experiment. Before the marshmallow task, they exposed children to either a reliable adult (who followed through on a prior promise of better art supplies) or an unreliable adult (who did not). The result was dramatic: children in the reliable condition waited four times longer than those in the unreliable condition. Same children, same marshmallows, same supposed “self-control capacity” — but wildly different behavior based on whether the environment had proven trustworthy. As the researchers concluded, children’s wait-times reflected not a deficit in self-control, but rather:

“An implicit, rational decision-making process that considers environmental reliability.”

— Kidd, C., Palmeri, H., & Aslin, R.N. (2013). Rational Snacking: Young children’s decision-making on the marshmallow task is moderated by beliefs about environmental reliability. Cognition, 126(1), 109–114.

The children who ate the marshmallow immediately were not exhibiting poor executive function. They were making a rational decision given their evidence about whether promises get kept. And children from chaotic, resource-scarce, or unstable homes carry a lifetime of such evidence.

Then came the large-scale replication. In 2018, Watts, Duncan, and Quan tested the marshmallow paradigm with a much larger and more diverse sample of approximately 900 children — far beyond Mischel’s small, homogeneous Stanford sample. They found that:

“This bivariate correlation was only half the size of those reported in the original studies, and was reduced by two-thirds in the presence of controls for family background, early cognitive ability, and the home environment.”

— Watts, T.W., Duncan, G.J., & Quan, H. (2018). Revisiting the Marshmallow Test: A Conceptual Replication Investigating Links Between Early Delay of Gratification and Later Outcomes. Psychological Science, 29(7), 1159–1177.

Most recently, a 2024 follow-up tracked marshmallow test participants to age 26, and the original narrative collapsed entirely:

“Marshmallow Test performance was not strongly predictive of adult achievement, health, or behavior. Although modest bivariate associations were detected with educational attainment, almost all regression-adjusted coefficients were nonsignificant. Results indicate that Marshmallow Test performance does not reliably predict adult outcomes.”

— Sperber, J., Vandell, D.L., Duncan, G.J., & Watts, T.W. (2024). Delay of gratification and adult outcomes: The Marshmallow Test does not reliably predict adult functioning. Child Development.

The Same Error, Scaled Up

The logical structure of the marshmallow test’s collapse maps precisely onto the EF curriculum problem:

1. Observation (real): Children who score well on EF tasks do better in school and life.

2. Causal inference (unwarranted): EF is an intrinsic cognitive capacity that causes good outcomes. Training it will improve outcomes.

3. What actually explains the correlation: Children from stable, resource-rich, trustworthy environments simultaneously develop better EF scores and better life outcomes — not because one causes the other, but because the same environmental factors produce both. The EF score and the life outcome are both downstream of the environment.

4. Why training fails: Training the downstream indicator (EF task performance) does not touch the upstream cause (environmental stability, resource availability, trust). You have moved the thermometer reading without curing the fever.

Researchers within the EF field have acknowledged this possibility directly:

“One explanation for the limited evidence for far transfer of executive function training to academic achievement might be that executive functions are simply not causally relevant to these domains.”

— Smid, C.R., Karbach, J., & Steinbeis, N. (2021). Training Executive Functions to Improve Academic Achievement: Tackling Avenues to Far Transfer. Frontiers in Psychology.

Executive function may be downstream of more fundamental factors — neural development, nutrition, household stability, genetic endowment — that independently drive both EF scores and academic performance. If EF is a symptom of developmental advantage rather than a cause of academic achievement, then training EF scores upward on laboratory tasks leaves the actual causal factors untouched. The very data that makes the sales pitch so compelling (”EF predicts success better than IQ!”) is equally consistent with this deflationary interpretation.

3. The Construct Itself Is Poorly Defined

“Executive function” has become a remarkably elastic term. A review commissioned by the Institute of Education Sciences acknowledged the problem:

“Historically, EF has been an ill-defined construct, often including broad and diverse processes relevant to many forms of self-regulation, from sustained attention to planning.”

— Zelazo, P.D. et al. Executive Function: Implications for Education. Institute of Education Sciences (IES/NCER).

Different researchers define EF differently. Some include emotional regulation; others treat it separately. Some include metacognition; others do not. The three-component model (working memory, inhibitory control, cognitive flexibility) has broad support, but commercial curricula routinely expand the umbrella to include planning, organization, time management, goal-setting, and self-monitoring.

This definitional looseness has practical consequences. A program that teaches children to use a planner for their homework, a program that runs computerized n-back memory drills, and a program that does mindfulness breathing exercises all get marketed as “executive function training.” When one approach shows positive results in a study, the others borrow the credibility. The construct is broad enough that nearly any pedagogical intervention can be reframed as EF training, making the category functionally unfalsifiable at the product level.

The distinction that matters most is between teaching domain-specific organizational strategies (using checklists, breaking tasks into steps, previewing assignments) and training abstract cognitive capacities (computerized working memory drills, inhibition tasks). The former is just good teaching. The latter is what the meta-analyses keep finding doesn’t transfer. EF curricula systematically blur this line.

4. Positive Results Come from Weaker Study Designs

A persistent methodological pattern runs through the EF training literature: the more rigorous the study design, the smaller the effects.

The critical variable is the control group. Studies that compare EF training to doing nothing (passive control) tend to find positive effects. Studies that compare EF training to a different engaging activity (active control) tend to find null effects. This pattern strongly suggests that the apparent benefits of EF training are driven by engagement, novelty, and adult attention — not by EF improvement per se.

Gobet and Sala’s meta-analyses documented this systematically, finding that across multiple forms of cognitive training, far-transfer effects vanish when measured against active controls. They further noted:

“Placebos always occur in WM training when it comes to far transfer... These placebos are around 0.15 to 0.20 standardized mean difference at best and often affected by publication bias.”

— Gobet, F. & Sala, G. (2023). Perspectives on Psychological Science, 18(1), 125–141.

Meanwhile, the most impressive claims for EF curricula tend to originate from studies conducted by the companies or researchers who developed the programs, rather than from independent replications. Axiom Learning cites a multi-year pilot it conducted itself. The SMARTS curriculum cites decades of research by its own creator. This is not necessarily indicative of fraud, but it is exactly the pattern one would expect in a field where effect sizes shrink toward zero as study quality and independence increase.

Schools evaluating these products should ask: Has this specific curriculum been evaluated by independent researchers, with an active control group, using academic outcome measures (not just EF task performance), and replicated at least once? For most commercially available EF curricula, the answer is no.

5. This Is a Recurring Historical Pattern in Education

Executive function training is not the first time education has been promised a domain-general cognitive upgrade that would lift all academic boats. A review of the transfer literature placed EF squarely in a long lineage of similar claims:

“The history of education has been littered with claims of resolving educational problems by improving underlying processes presumed to be important. For example, visual perception, vergence, strabismus, scotopic sensitivity, balance, primitive reflexes, auditory processing speed, executive function, perceptual motor skill, brain patterns, balance, working memory, and so on.”

— Hempenstall, K. Near and Far Transfer in Cognitive Training. National Institute for Direct Instruction (NIFDI).

Every generation of educators encounters a new version of “train the underlying process and everything improves.” In each case, the initial studies are promising, a commercial ecosystem develops, schools adopt programs at scale, and eventually the meta-analyses converge on the same verdict:

“Directly teaching the requisite academic skills has produced superior outcomes to process training.”

— Hempenstall, K. NIFDI.

The deeper theoretical point, drawn from decades of expertise research, is that domain-general cognitive transfer may be a fundamental limit of human cognition rather than a technical problem awaiting a better training protocol:

“The lack of training-induced far transfer is an invariant of human cognition.”

— Sala, G. & Gobet, F. (2019). Near and Far Transfer in Cognitive Training: A Second-Order Meta-Analysis. Collabra: Psychology, 5(1), 18.

This is not a claim that children cannot learn or improve. It is the claim, supported by over a century of research stretching back to Thorndike’s experiments on “formal discipline,” that learning is domain-specific. There is no general-purpose cognitive upgrade available. Children get better at what they practice, in the contexts where they practice it. Programs premised on the existence of a transferable cognitive engine are swimming against a very strong empirical current.

6. The Field’s Own Researchers Acknowledge the Problem

Perhaps the most telling sign is the gap between what researchers write in their academic papers and what gets packaged and sold to schools.

Even researchers sympathetic to EF-informed education acknowledge the evidentiary thinness. A 2025 pilot study of an SEL program targeting executive functions noted:

“Although SEL programs are widely endorsed for promoting student well-being, evidence regarding their effectiveness remains mixed... direct causal links [between SEL interventions and EF improvement] are not yet well established.”

— Rosas, R. et al. (2025). Pilot evaluation of a socio-emotional learning program on executive functions in elementary school students. Frontiers in Psychology.

A review of school-based SEL adoption observed:

“It is common for schools to adopt SEL programs without providing the essential initial training, ongoing coaching and mentoring, or technical support.”

— Jones, S. et al. (2023). Evidence for Social and Emotional Learning in Schools. Learning Policy Institute.

This suggests a pattern where the adoption of programs has outpaced the evidence for programs — a familiar dynamic in education, where institutional momentum, parental demand, and vendor marketing can drive purchasing decisions faster than research can evaluate them.

Gobet and Sala’s conclusion stands as the most direct assessment from within the research community:

“Given that the available empirical evidence on cognitive training and other fields of research suggests that the likelihood of finding reliable and robust far-transfer effects is low, research efforts should be redirected to near transfer or other methods for improving cognition.”

— Gobet, F. & Sala, G. (2023). Perspectives on Psychological Science, 18(1), 125–141.

What Schools Should Do Instead

None of this means that executive function is unimportant, or that schools should ignore self-regulation entirely. It means the specific intervention model — standalone EF curricula promising domain-general cognitive improvement — is not supported by the evidence.

What the evidence does support:

• Teach organizational strategies within academic content. Showing students how to break a writing assignment into steps, or how to check their work on a math problem, works — but it works because it’s domain-specific instruction, not because it’s upgrading a general cognitive engine.

• Invest in the upstream determinants of EF. Physical activity, adequate nutrition, reduced chaos and unpredictability, and strong adult relationships all predict better executive function. These are harder to package and sell than a curriculum, but they address root causes rather than downstream indicators.

• Demand rigorous evidence from vendors. Before adopting any EF or SEL curriculum, ask for independent (not developer-conducted) evaluations, with active control groups, measuring academic outcomes (not just EF task performance), replicated across multiple sites. Most products cannot meet this standard.

• Recognize that good teaching already develops EF. Structured classrooms with clear expectations, scaffolded instruction, and opportunities for increasing student autonomy already exercise the cognitive processes that EF curricula claim to train — without the overhead of a separate program.

The appeal of executive function training is understandable. It offers a tidy, packageable solution to the messy reality that academic achievement is shaped by dozens of interacting factors, many of them outside a school’s control. But the evidence consistently points to a less glamorous truth: children learn what they are taught, in the domains where they are taught it, and there are no cognitive shortcuts.

Sources cited in this essay draw from peer-reviewed meta-analyses, systematic reviews, and institutional research reports. Key references include Kassai et al. (2019) in Psychological Bulletin, Gobet & Sala (2023) in Perspectives on Psychological Science, Sala & Gobet (2019) in Collabra: Psychology, Smid et al. (2021) in Frontiers in Psychology, Zelazo et al. via IES/NCER, and reviews from the Learning Policy Institute and the National Institute for Direct Instruction.