Can Mental Practice Actually Improve Real Performance?

Yes, and the size of the effect is the interesting part. Meta-analysis puts mental practice at a moderate benefit to performance, roughly half a standard deviation, which is real, replicable, and consistently smaller than the benefit of actually doing the thing. It also decays: the longer the gap between rehearsing and performing, the less it helps. Those two qualifiers - smaller than physical practice, and perishable - are what separate the research from the version of it that gets sold, where imagining an outcome vividly enough is supposed to produce the outcome.

The meta-analysis

Driskell, Copper and Moran published the definitive review in Journal of Applied Psychology in 1994. They pooled the mental practice literature and found a positive, statistically significant overall effect on performance of about d equals 0.53. Mental practice works. That is the first sentence of the honest summary and it should not be buried.

The second sentence is that the effect was moderated by three things, and the moderators are more informative than the average.

Task type. Cognitive tasks benefited more than physical ones. If the task has a substantial thinking component - a sequence to remember, a decision to make, a procedure to follow - rehearsing it mentally captures more of what the task requires. For tasks that are mostly about force, speed, or fine physical execution, the mental version captures less, because a large part of the skill lives in the body rather than in the plan.

Retention interval. The advantage shrank as the delay between mental practice and performance grew. This is the finding most often left out, and it turns mental practice from an achievement into a maintenance activity. The authors' practical recommendation followed directly: refresh the rehearsal periodically rather than treat one block of it as banked.

Length of practice. More was not linearly better. Very long mental practice sessions showed diminishing returns, which fits the observation that sustained, vivid motor imagery is genuinely effortful and degrades with fatigue.

The result has held. Toth and colleagues ran a replication and extension in Psychology of Sport and Exercise in 2020, twenty-four years on, and recovered a comparable overall effect. In a period when a lot of psychology did not survive re-examination, this one did, which is worth saying plainly.

Why it works at all

The mechanistic story is reasonably well grounded. Imagining a movement partially activates the circuits that execute it, which is why mental rehearsal of a specific, already-learned motor sequence can change corticospinal excitability, as covered in whether thinking alone can change your brain. Ranganathan and colleagues reported in Neuropsychologia in 2004 that imagined contractions produced measurable strength gains, and attributed them to increased central drive rather than to any change in the muscle itself.

That framing explains both the effect and its ceiling. Mental practice improves the instruction being sent. It cannot improve the thing receiving the instruction, and it cannot supply the physical feedback that only execution produces. A rehearsed movement is missing the error signal that comes from getting it slightly wrong and feeling it.

The distinction that decides whether visualization helps

Here is where the research and the marketing diverge, and it comes down to what you visualize.

Pham and Taylor tested this directly in Personality and Social Psychology Bulletin in 1999. Students facing a midterm were assigned to mentally simulate either the process of studying - sitting down, opening the book, working through the material - or the outcome, imagining themselves receiving a high grade and how good it would feel. Process simulation improved exam performance and increased actual study time. Outcome simulation did not.

Oettingen and Mayer found the same asymmetry looking forward in the Journal of Personality and Social Psychology in 2002. Positive expectations, judgments about how likely success is, predicted better outcomes. Positive fantasies, indulgent imagery of having already succeeded, predicted worse ones, including fewer job offers and lower salaries among graduating students.

The pattern is consistent enough to state as a rule. Rehearsing the process is a form of practice. Savoring the outcome is a form of consumption, and it appears to discharge some of the motivation that would otherwise go into the work. The full account of that second mechanism is in why visualizing success reduces effort.

Conditions under which mental practice actually helps, ranked

1. You already have the skill, at least partially. Every study in this literature rehearses something the participant has been taught. Imagery reheats an existing motor program; it does not create one from nothing.
2. You rehearse the process, in sequence, in detail. The steps, in order, at something like real tempo. This is the condition Pham and Taylor isolated, and it is the one that survives contact with an actual exam.
3. The task has a cognitive component. Procedures, sequences, and decisions gain more from rehearsal than pure strength or speed, per the Driskell moderator analysis.
4. You rehearse close to when you perform, and repeatedly. The effect decays with delay. Treat it as a warm-up, not a deposit.
5. You are still doing the physical practice. Mental practice is a supplement with a smaller effect size. Substituting it for the real thing gives up the larger effect to keep the smaller one.
6. You imagine the reward. This is the condition the evidence does not support, and the one most commonly sold.

What should temper all of this

A 1994 meta-analysis inherits whatever publication bias existed in the literature it pooled, and mental practice studies are small, easy to run, and were published in an era with no preregistration. The 2020 replication is reassuring on this point but does not eliminate the concern.

Blinding is also close to impossible. A participant told to spend twenty minutes vividly imagining a task knows they are in the interesting condition, and control groups are frequently given nothing to do rather than an equally engaging alternative. Some portion of a d of 0.53 is plausibly expectancy and attention rather than rehearsal.

And compliance is unverifiable. There is no instrument that confirms someone spent the session imagining the task instead of thinking about lunch. Researchers rely on self-report of imagery vividness, which is exactly the sort of measure that behaves poorly. None of this sinks the finding. It suggests the true effect is probably somewhat smaller than the headline number, which is the usual direction of travel and a habit of mind worth applying across the wider set of brain and behavior claims.