The observation is common enough to be a cliché: by the end of a demanding day, small decisions feel disproportionately difficult. What to eat, which email to answer, whether to go out. The explanation attached to it is equally familiar — you have used up your self-control, like a battery running down. That explanation has been tested more thoroughly than most, and it has not done well, which makes this a more interesting topic than it first appears.

The Phenomenon, as the Literature Describes It

A 2025 integrative review took the question seriously enough to search three databases and retrieve 1,027 articles on decision fatigue. After screening with PRISMA and appraisal using Joanna Briggs Institute methods, 23 remained, which were analysed through root cause analysis and thematic synthesis (Frontiers in Cognition, 2025).

Its starting definition is operational rather than mechanistic: decision fatigue leads to reduced efficiency in the rate and quality of decisions. Both, and they are separable — deciding more slowly and deciding worse are different failures.

The review identified ten causes and organised them into a framework intended to apply across professional domains. Note what that is and is not: a synthesis of what the literature attributes decision fatigue to, not an experimental demonstration that any of them produces it.

The Explanation Everybody Cites

The dominant account has been the strength or limited-resource model: self-control draws on a finite physiological or cognitive resource, so exercising it makes subsequent acts of self-control more likely to fail. That predicted effect has a name — the depletion effect — and it inspired hundreds of experiments.

It is an appealing model because it matches the felt experience precisely. Something does feel used up.

It is also, for exactly that reason, the kind of model worth testing hard. An explanation that matches intuition perfectly will be believed whether or not it is correct.

What Happened When It Was Tested

A 2014 paper reported four studies — two with community samples, two with young adults — using a frequently used depletion procedure (the crossing-out-letters protocol) and the two most frequently used outcome measures (handgrip perseverance and a modified Stroop task). Each participant completed a baseline measure, then a randomised depletion or control task, then the same measure again.

The result: no evidence for significant depletion effects in any of the four studies. The authors noted the null results may indicate depletion has more limited effects than prior publications implied, and called for further replication with full disclosure of positive and negative findings (PLoS ONE, 2014).

A 2015 paper went further, running a series of focused meta-analytic tests designed to address limitations in earlier appraisals of the evidence. Its conclusion, in the authors’ own words, was that there is very little evidence that the depletion effect is a real phenomenon, at least when assessed with the methods most frequently used in the laboratory — a result they described as strongly challenging the idea that self-control relies on a limited resource (Journal of Experimental Psychology: General, 2015).

That is about as direct as scientific disagreement gets, and it concerns the mechanism, not the experience.

Holding Both at Once

So there is a 2025 review treating decision fatigue as a phenomenon substantial enough to warrant a cross-domain framework, and there are meta-analytic tests concluding the classic laboratory demonstration of its mechanism does not hold up.

These are less contradictory than they look. The review is about decision-making in real professional settings — where the working day involves genuine time pressure, accumulating cognitive load, sleep debt, hunger, interruption and consequence. The depletion experiments tested something much narrower: whether one brief controlled act of self-control impairs the next one, measured minutes later.

A laboratory task failing to reproduce an effect does not establish that the effect never occurs outside the laboratory. It establishes that the specific paradigm used to demonstrate it does not reliably do so — which is a claim about the evidence, not about the world.

The honest position is therefore: the experience is widely reported and has been reviewed systematically; the mechanism most often used to explain it is not well supported; and what is actually happening at the end of a demanding day remains an open question.

Why the Distinction Is Worth Making

Because the limited-resource story carries advice with it. If self-control is a battery, then the implications are to make important decisions early, to minimise trivial choices, and to treat a difficult evening as an inevitable consequence of a demanding morning.

Some of that may be sensible for other reasons. None of it is supported by a mechanism that has survived meta-analytic testing, and it is worth knowing which recommendations rest on evidence and which rest on a metaphor.

It is also a useful example of how to read this kind of research generally. A phenomenon can be widely described, extensively studied and genuinely experienced while its standard explanation is unsettled — and that combination is far more common in psychology and neuroscience than popular coverage suggests.

The related question of why sustained concentration is itself effortful is covered in what happens to thinking under pressure.

What Nobody Can Tell You From This

Not whether you are experiencing decision fatigue. There is no measure of it, no threshold, and no diagnostic use.

Not how many decisions are too many. The review identifies causes across domains; it does not quantify a limit, and no such number exists.

And not whether the effect is real outside the laboratory. The negative findings concern specific paradigms. Absence of a reliable laboratory effect is genuinely informative and is not proof that nothing happens over a working day.

What it does support: reduced efficiency in the rate and quality of decisions is a documented concern serious enough for cross-domain review, and the explanation most people have been given for it is on considerably weaker ground than they have been told.