Harold Pashler

Introduction

Harold Pashler (b. 1958) is an American experimental psychologist at the University of California, San Diego, who has spent his career measuring the limits of what the mind can do at once — and then working out how to arrange learning so that the limits matter less.

His central empirical claim is that dual-task interference is structured rather than general. When two tasks are attempted close together, the delay is not a diffuse competition for capacity but a discrete bottleneck at one specific stage: selecting a response. Everything else can proceed in parallel, but only one response can be selected at a time.

He then applied the same instinct — find the structure, then exploit it — to memory. His work on the spacing of study sessions produced a quantitative rule that has ended up inside the spaced-repetition software used by millions of students.

He is also one of the more effective methodological critics in the field, best known for a paper arguing that some of the most-cited correlations in social neuroscience were statistical artefacts.

Early Life and Education

Pashler was born in New York in 1958. He took his bachelor's degree at Brown University in 1980 in logic and philosophy of science — a training in formal reasoning that is visible in everything he has done since, and which partly explains why he later became a critic of statistical practice rather than only a producer of findings.

He completed his doctorate in psychology at the University of Pennsylvania in 1985 and joined the faculty at UC San Diego the same year, where he has remained for his entire career and is now Distinguished Professor of Psychology.

Academic and Professional Career

Pashler's first line of work, developed with colleagues through the 1980s, was the response selection bottleneck model of dual-task interference. Building on Alan Welford's earlier observations, the model made quantitative predictions about the pattern of reaction times in the psychological refractory period experiment, and those predictions have been confirmed both by mathematical analysis of behavioural data and by neuroimaging studies of people multitasking.

In 1988 he published what appears to be the first experimental demonstration of the phenomenon later named change blindness — using displays of letters that appeared, vanished, and reappeared, sometimes altered. Observers reported a subjective sense of having seen the entire display yet missed large changes in it. He drew the contrast explicitly: the feeling of rich visual awareness against the very limited ability to detect even substantial changes.

From the 1990s onward his attention shifted increasingly to learning and memory. In 1992, with Mark Carrier, he showed that actively retrieving information strengthens associative learning more than spending the same time re-studying it — an early and influential contribution to the literature on what is now called retrieval practice or the testing effect.

In 2007, with Liqiang Huang, he proposed the Boolean map theory of visual attention, which holds that the contents of visual awareness at any instant have the structure of a Boolean map: a representation that marks, for each feature, which locations in the visual field possess it. The claim is that this single data structure constrains what can be attended at once and explains a range of otherwise separate findings.

He was elected to the Society of Experimental Psychologists in 2000, is a fellow of the Association for Psychological Science and the Psychonomic Society, and received the Troland Research Award from the National Academy of Sciences in 1999 — the Academy citing his "experimental breakthroughs in the study of spatial attention and executive control" and his "insightful analysis of human cognitive architecture."

Major Contributions to Psychology

The response selection bottleneck

In the psychological refractory period paradigm, a person must respond to two stimuli presented in quick succession. The reliable finding is that the second response is delayed, and the delay grows as the interval between the stimuli shrinks.

The obvious explanation is that capacity is shared and dividing it slows everything. Pashler's argument was that the interference is localised: perceptual encoding of the second stimulus and the motor execution of the first response can proceed together, but the stage at which a response is selected is serial. Only one selection operation runs at a time, and everything downstream waits for it.

This matters beyond the paradigm. It implies that multitasking failure is not a general limit on how much the brain can do, but a specific architectural constraint with specific workarounds — tasks that can be completed without concurrent response selection interfere far less than ones that cannot.

Change blindness

His 1988 demonstration preceded the wave of research on the phenomenon in the 1990s. The finding — that people systematically overestimate what they have seen, and that large changes to a scene go undetected when they coincide with a brief interruption — has since become one of the standard illustrations of the gap between phenomenal consciousness and perceptual content.

The spacing of learning

The most practically consequential work is the series of studies, beginning with Cepeda and colleagues and extended in 2008 with Melody Wiseheart and others, that mapped the optimal interval between study sessions.

Holding total study time constant, the question is: given that you will be tested after a delay D, how long a gap G should you leave between two study sessions? The answer they found is a stable ratio rather than a fixed duration — performance is best when G is roughly 10 to 20 percent of D.

So if you want to retain something for a year, the two sessions should be about a month apart. If you want it for a week, a day apart is better. This "temporal ridgeline" is what spaced-repetition applications implement, and the reason their schedules adapt to how long you intend to remember.

The inner crowd

In 2008, with Edward Vul, he showed that the wisdom-of-crowds effect — averaging independent estimates improves accuracy — can be obtained within a single person. Averaging someone's first estimate with a second, later estimate from the same person produced about a third of the benefit that averaging two different people would. The Economist called it the inner crowd effect, and it opened a line of work on improving judgment without additional information.

Methodological criticism

Pashler has been as influential in catching bad inferences as in producing findings.

The "learning styles" critique (Pashler, McDaniel, Rohrer, and Bjork, 2008) examined the claim that instruction is more effective when matched to a student's supposed visual, auditory, or kinaesthetic style, and found the evidence absent: studies that claim to support it generally fail to establish that matching produces better outcomes than mismatching. The concept remains enormously popular in education and unsupported by data.

In 2009, with Vul, Harris, and Winkielman, he published "Voodoo Correlations in Social Neuroscience," arguing that a widespread practice in neuroimaging — selecting voxels by their correlation with a behavioural measure, then reporting that same correlation — inflates the reported relationship, sometimes drastically. The paper produced an intense public and academic controversy. Whether or not every example cited was fairly chosen, the target practice was real, and cross-validation subsequently became much more common in the field.

He has also written on the discriminating power of data: the argument that a good fit does not confirm a model unless the data would have distinguished it from plausible alternatives.

Major Publications

  • "Dual-Task Interference in Simple Tasks: Data and Theory" (1994, Psychological Bulletin, 116, 220–244)
  • The Psychology of Attention (1998) — a synthesis of the selective-attention literature
  • "Learning Styles: Concepts and Evidence" (2008, with Mark McDaniel, Doug Rohrer, and Robert Bjork, Psychological Science in the Public Interest)
  • "A Boolean Map Theory of Visual Attention" (2007, with Liqiang Huang, Psychological Review, 114, 599–631)
  • "Voodoo Correlations in Social Neuroscience" (2009, with Edward Vul, Christine Harris, and Piotr Winkielman, Perspectives on Psychological Science)
  • Organizing Instruction and Study to Improve Student Learning (2007, US Department of Education Practice Guide, with Patrice Bain and others)

Impact and Legacy

Pashler's bottleneck account is the standard explanation of dual-task interference, and it reframed a phenomenon that had been described vaguely as a limit on attention into a claim about architecture that can be tested and, to some extent, designed around.

His spacing work has had more direct practical effect than most cognitive research ever does: the ratio he and his colleagues established is implemented in spaced-repetition software used by students worldwide, and the Department of Education practice guide he chaired distributed the same findings to teachers at no cost.

Criticism and controversy. The "voodoo correlations" paper was the most contested. Its targets objected that the analysis mischaracterised their methods, that the examples were selectively chosen, and that the title itself was inflammatory; the resulting exchange was acrimonious and played out partly in the popular press. There is a reasonable complaint that the paper's rhetoric ran ahead of its evidence for individual cases, even where the methodological point was sound.

More broadly, the style of critique he represents — auditing the statistical practices of a neighbouring field — is valuable and unpopular in equal measure, and it has made him a controversial figure in areas he does not work in.

He has also published outside his area of expertise in ways that have drawn criticism. His 2013 paper on immigrants' attitudes toward libertarian values, which argued that lower support among foreign-born residents implies a dimmer political future for libertarianism in the United States, is a normative projection from survey data by a perception researcher, and it sits oddly beside the rest of his output.

His position on the replication crisis is also not the majority one. He has argued that the core fields of cognition, attention, and perception are probably in better shape than the crisis literature implies, because within-subject designs give them high statistical power — and that the problems are concentrated elsewhere. That is a defensible reading of the evidence, and it is also an argument that the fields he works in are the healthy ones.

Learn More

References

  • Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. (2008). Learning styles: Concepts and evidence. Psychological Science in the Public Interest, 9(3), 105-119.

Last updated: 2026-10-20

Menu
Loading