Mind & Commute • Statistics
Decision Fatigue Statistics, the real numbers
A research-backed compilation of decision fatigue statistics, with every number traced to its primary source and contested findings clearly flagged. We started this project because the most-cited statistic in decision fatigue content doesn’t actually exist.
49% → 86%
401(k) participation when default flips from opt-in to auto-enroll
Madrian & Shea, 2001
43%
Of daily behaviors performed in the same context, almost daily
Wood, Quinn & Kashy, 2002
66 days
Median time to form a habit, ranging 18 to 254 days
Lally et al., 2010
The Stat We Need to Talk About First
35,000?
The most-cited decision fatigue statistic has no peer-reviewed source. We tried to find one. Here’s what we discovered.
Search “decision fatigue” and you’ll find the same number repeated everywhere: the average person makes 35,000 decisions per day. It appears in productivity blogs, coaching sites, medical association content, and the occasional university press release. It’s been quoted by therapists, executives, and television personalities.
It also has no traceable primary source.
This matters because decision fatigue is real. The phenomenon is well-documented across decades of research: defaults change behavior dramatically, cognitive performance follows circadian rhythms, professional judgments degrade through extended sessions. We’ve covered why it hits hardest in the morning and why discipline isn’t the fix. But the most-cited number in popular content about it was never in a study. It became gospel through repetition.
We spent several hours trying to verify the claim. The number is commonly attributed to “a Cornell study” or “research from Georgia Tech.” Neither institution has published any such finding. There is one Cornell paper that produced an unrelated figure about food decisions (227 per day, the “mindless eating” stat), but that paper came from a research lab that was later found to have data integrity problems and had multiple papers retracted. It shouldn’t be cited as reliable evidence either.
Beyond that, the trail goes cold. There is no peer-reviewed paper that produces a daily decision count of 35,000. The number was never measured. It was repeated until it sounded official.
The propagation chain appears to start with a 2015 faculty blog post by Dr. Joel Hoomans at Roberts Wesleyan University, titled “35,000 Decisions: The Great Choices of Strategic Leaders.” The figure then traveled through a 2016 Wall Street Journal opinion piece by Jim Sollisch and went viral from there. Neither cites original research. The number is also commonly attributed to Sahakian and LaBuzetta’s 2013 book Bad Moves, but academics who searched the book for the figure have reported being unable to find it there. The citation is a ghost.
This isn’t a small problem. The 35,000 figure shapes how people think about cognitive load, how productivity advice gets framed, and how publications structure their decision fatigue content. Every page that repeats it without sourcing inherits the urban legend.
The decision fatigue phenomenon doesn’t need an inflated, unsourced number to be real. The actual research is more interesting than the inflated number. Here are the decision fatigue statistics that are real, the ones that are contested, and the ones that have been quietly debunked.
Key Takeaways
- Decision fatigue is real. Sustained decision-making degrades subsequent decisions through worse choices, increased impulsivity, status quo bias, and reversion to defaults. Different research traditions emphasize different effects, but the underlying phenomenon is consistent.
- The 35,000 figure has no source. It appears in countless articles but no peer-reviewed study supports it.
- Defaults are the most powerful behavior change tool ever measured. Auto-enrollment raises 401(k) participation from 49% to 86%. Organ donation consent goes from 12% (opt-in) to 99.98% (opt-out). These dramatic effects work because depleted decision-makers reliably accept whatever requires the least effort.
- Decision quality degrades in some professional contexts. Physicians prescribed 73% more inappropriate antibiotics by end of shift (Linder 2014). The effect is real but context-dependent.
- About 43% of daily behaviors are habitual, performed in the same context almost daily (Wood, Quinn, Kashy 2002).
- The famous parole judges study is methodologically contested. The 65% to 0% favorable ruling drop may be partly explained by case scheduling. But the broader phenomenon the study points at is supported by other research with cleaner methodology.
- The “paradox of choice” is also contested. The famous jam study found a 10x conversion gap between 6 and 24 options, but a 2010 meta-analysis of 50 experiments found no consistent effect (Scheibehenne et al.).
- The simple “willpower fuel tank” model failed replication (Hagger et al. 2016). The specific mechanism was wrong, but the broader phenomenon of decision fatigue is supported by experience, neuroscience, and multiple research traditions.
- Cognitive performance follows circadian rhythms, with a measurable 10-20% afternoon vigilance dip (Monk 2005).
- Habit formation takes a median of 66 days, ranging from 18 to 254 (Lally et al. 2010). The “21 days” figure is not supported.
Key Decision Fatigue Statistics (Verified)
- 43% of daily behaviors are habitual, performed in the same context daily (Wood, Quinn & Kashy, 2002)
- 49% → 86% 401(k) participation increase under auto-enrollment (Madrian & Shea, 2001)
- 12% vs 99.98% organ donation consent: opt-in vs opt-out countries (Johnson & Goldstein, 2003)
- 10-20% cognitive vigilance decline in early afternoon (Monk, 2005)
- 3x more effective: environmental nudges vs informational nudges (Cadario & Chandon, 2020)
- 66 days median to form a habit (range: 18-254 days) (Lally et al., 2010)
- 73% increase in physician inappropriate antibiotic prescribing by end of shift (Linder et al., 2014)
What Decision Fatigue Actually Is
Decision fatigue is the phenomenon where sustained decision-making degrades subsequent decisions. It manifests through several observable effects: people make worse choices later in the day, become more impulsive, increasingly default to whatever requires the least effort, and stick with the status quo even when better options are available. The pattern is supported by neuroscience, by behavioral economics, and by the universal experience of feeling mentally taxed at the end of a long day of choices.
The mechanism has been debated. The simple “willpower as fuel tank” model that dominated popular content in the 2000s, where willpower drains like a battery and is restored by glucose, has not held up cleanly to laboratory replication. But the underlying phenomenon, that effortful cognition is metabolically costly and that this cost compounds across decisions, is well-established across multiple research traditions and consistent with the lived experience of nearly everyone who has ever made decisions for a living.
What follows is a research-backed compilation of what we actually know. We’ve traced every figure to its primary source. Where studies are methodologically contested, we say so. Where the popular framing exaggerated specific mechanisms, we correct it. But we don’t editorialize the phenomenon itself out of existence. Decision fatigue is real, and the question is not whether it happens but how to design around it.
Decisions are metabolically expensive
Making conscious choices recruits the prefrontal cortex, which is among the most metabolically demanding regions of the brain. Decision-making activates these effortful processing systems and is experienced subjectively as cognitive work. The brain treats deliberation as a costly process, not a free background activity.
General neuroscience finding, summarized in Anderson (2003), Psychological Bulletin, 129(1), 139-167, and supported across decades of executive function research.
Decision fatigue produces multiple measurable effects
When cognitive load accumulates, depleted people show a cluster of changes: they make worse decisions in some professional contexts (Linder 2014 on antibiotic prescribing), become more impulsive and less deliberative (behavioral economics literature), are more likely to defer or accept defaults (Anderson 2003, choice architecture literature), and exhibit stronger status quo bias (Samuelson and Zeckhauser 1988). Different research traditions emphasize different effects, but the underlying pattern is consistent across all of them.
Multiple effects documented across the choice architecture, behavioral economics, and clinical decision-making literatures. See specific findings throughout this article.
The subjective experience of mental fatigue is universal
Almost everyone reports that mental effort feels increasingly costly through the day. This is consistent with neuroscience showing prefrontal cortex glucose consumption during effortful cognition, and with research on cognitive control, vigilance decrement, and executive function. The specific “ego depletion” laboratory paradigm failed replication (Hagger et al. 2016), meaning we should be cautious about strong claims about specific mechanisms. But the underlying phenomenon of subjective cognitive fatigue, supported by both first-person experience and converging research lines, is not in serious dispute.
Subjective experience is universally reported. Neuroscience and cognitive control literature support metabolic cost of effortful cognition. Hagger et al. (2016) challenged specific ego depletion paradigm, not the broader phenomenon.
This is the basic phenomenon. Now we can ask the harder questions: how many decisions are we actually making, when does cognitive performance peak and decline, and what interventions actually help.
The Volume of Daily Decisions
If we can’t count daily decisions reliably, what can we say? The answer is: much less than productivity content claims, and what we can say is more useful anyway.
35,000 decisions per day
Widely cited, completely unsourced. No peer-reviewed paper produces this figure. Commonly misattributed to Cornell or Georgia Tech research that does not exist.
Origin untraceable. Appears in popular press from approximately 2014 onward.
227 food decisions per day
From Wansink & Sobal (2007), “Mindless Eating.” This paper originates from a research lab whose work was later found to have significant data integrity problems, with multiple papers retracted from the same lab. While this specific paper was not formally retracted, the broader research record makes it unreliable. The figure should not be cited.
Wansink & Sobal (2007), Environment & Behavior, 39(1), 106-123.
~96 phone pickups per day
Average smartphone user picks up their device approximately 96 times per day, according to a 2019 Asurion survey-based estimate from 2,000 US adults. Each pickup creates a context-switching micro-decision. Apple Screen Time data corroborates this range, and more recent surveys suggest the number has continued to climb. Note that survey-based phone use estimates have known reliability issues compared to behavioral tracking data.
Asurion (2019), survey of 1,998 US smartphone users by Solidea Solutions, August 2019.
Corporate research has produced its own widely-circulated figures. Oracle’s 2023 “Decision Dilemma” study reported that 74% of respondents felt their daily decisions had increased tenfold over three years, and 86% felt overwhelmed by data when making decisions. The survey reached 14,250 respondents across 17 countries and gets cited frequently in productivity content and AI-generated overviews. But it has the same fundamental measurement problem as the 35,000 claim: it measures self-reported perception of decision overload, not actual decision volume. Self-reports are known to be unreliable for estimating cognitive activity. The Oracle figures tell us people feel overwhelmed by decisions. They do not tell us how many decisions people actually make.
Cognitive Performance by Time of Day
This is one of the strongest research areas in decision science. Cognitive performance is not constant across the day. It follows circadian rhythms with measurable peaks and dips.
10-20% afternoon vigilance decline
The post-lunch dip is real and measurable. Vigilance and reaction time decline 10-20% in early afternoon, independent of food intake. This is a circadian effect, not a glucose effect.
Monk (2005), Clinics in Sports Medicine, 24(2), e15-e23.
Executive function peaks mid-morning, dips 2-4 PM
Working memory, complex reasoning, and decision quality follow a predictable curve: ramp-up after waking, peak in mid-morning, gradual decline through afternoon, partial recovery in early evening for some chronotypes.
Schmidt et al. (2007), Cognitive Neuropsychology, 24(7), 755-789.
The morning morality effect
People behaved more ethically in morning sessions than afternoon sessions across four experiments. Moral self-regulation appears to deplete across the day. Effect sizes are modest, with mixed results in subsequent replications.
Kouchaki & Smith (2014), Psychological Science, 25(1), 95-102.
Cognitive Performance Through the Day
Cognitive performance through the day, based on Schmidt et al. (2007) and Monk (2005). Peak mid-morning, decline 2-4 PM, partial recovery early evening.
The Israeli Parole Judges Study
This is the most famous decision fatigue study, and the most contested. It deserves a longer treatment than usual.
65% after breaks → near 0% before next break
Israeli parole judges granted favorable rulings about 65% of the time after a food break, with the rate falling to near zero by the end of each session. The original interpretation: cognitive depletion through extended decision-making sessions.
Danziger, Levav & Avnaim-Pesso (2011), PNAS, 108(17), 6889-6892.
Glöckner (2016) proposed an alternative explanation. Cases were not heard in random order. Easy cases (likely to be granted) tended to be heard first after breaks. Complex cases (likely to be denied) accumulated toward the end. The pattern, in this view, is a scheduling artifact, not decision fatigue.
Weinshall-Margel and Shapard (2011) raised similar methodological concerns. The data does not include information about case difficulty or attorney representation, both of which correlate with case ordering.
The current consensus is that the pattern is real, but the causal mechanism in this specific study is genuinely disputed. The most-cited decision fatigue study is also the most contested one. Importantly, this does not mean decision fatigue is fake. It means this particular study cannot serve as the sole evidence for the phenomenon. The broader pattern, that sustained decision-making degrades subsequent decisions, is supported by other research with cleaner methodology, including the Linder 2014 antibiotic prescribing study covered in the next section.
What the Depleted Brain Actually Does
Here’s where the research gets interesting. When the brain is cognitively loaded, it doesn’t simply make worse decisions. It stops making them.
Decision avoidance under cognitive load
When cognitive load gets too high, people don’t just choose worse. They stop choosing entirely. They default. They defer. They take whatever requires the smallest action. The depleted brain isn’t lazy. It’s economizing effort.
Anderson (2003), Psychological Bulletin, 129(1), 139-167.
Status quo bias when cognitively loaded
People disproportionately stick with defaults when their cognitive resources are taxed. This is a foundational behavioral economics finding, replicated across decades and contexts.
Samuelson & Zeckhauser (1988), Journal of Risk and Uncertainty, 1, 7-59.
Physicians prescribed unnecessary antibiotics 73% more by end of shift
Inappropriate antibiotic prescribing rose from approximately 15% early in clinical sessions to 26% by session’s end. Decision fatigue in healthcare has measurable patient outcomes.
Linder et al. (2014), JAMA Internal Medicine, 174(12), 2029-2031.
3% vs 30% purchase rate: the jam study
In a famous 2000 field experiment at a California grocery store, shoppers offered 24 jam varieties stopped at the display more often (60%) but rarely bought (3%). Shoppers offered just 6 varieties stopped less (40%) but bought far more often (30%), a tenfold conversion gap. The finding launched the “paradox of choice” framing in popular discourse. However, a 2010 meta-analysis of 50 choice overload experiments (n=5,036) found a mean effect size of virtually zero (d=0.02), with no consistent conditions identified. The effect appears under specific circumstances but is not a universal law of consumer behavior. The “22 options” version of this stat circulating in popular content is a paraphrase artifact, not the original finding.
Original: Iyengar & Lepper (2000), Journal of Personality and Social Psychology, 79(6), 995-1006. Meta-analysis: Scheibehenne, Greifeneder & Todd (2010), Journal of Consumer Research, 37(3), 409-425.
Choice Architecture and Default Effects
This is where decision fatigue becomes most measurable. The default effect literature is the strongest evidence we have that decision-making is cognitively expensive, because it shows what happens when people stop deciding actively. They accept whatever the default is, and behavior shifts dramatically as a result. The size of these effects, in some cases approaching 90 percentage points, would be impossible if active decision-making were costless.
49% → 86% 401(k) participation
When a US company switched 401(k) enrollment from opt-in to auto-enroll, participation rose from 49% to 86% within three years. Same employees, same plan, same salaries. The form did the deciding.
Madrian & Shea (2001), Quarterly Journal of Economics, 116(4), 1149-1187.
12% vs 99.98% organ donation
Germany (opt-in system): 12% consent rate. Austria (opt-out system): 99.98% consent rate. Same language, same border, same human beings. The default was the decision.
Johnson & Goldstein (2003), Science, 302(5649), 1338-1339.
Environmental nudges 3x more effective than information
Behavioral nudges that change defaults and convenience outperformed informational nudges (calorie labels, reminders) by a factor of three for changing eating behavior. A 2020 meta-analysis of 96 studies.
Cadario & Chandon (2020), Marketing Science, 39(3), 465-486.
The Default Effect: Same Choice, Different Outcome
The two cleanest demonstrations of the default effect ever measured. The default isn’t a small nudge. For most people, it is the decision.
Habit Formation
How much of daily life runs on autopilot? More than we think, and habit formation takes longer than the popular myths suggest.
~43% of daily behaviors are habitual
Participants recorded their behaviors hourly. About 43% of daily actions were performed in the same location and context almost every day. Nearly half of daily life is context-triggered repetition, not deliberate decision.
Wood, Quinn & Kashy (2002), Journal of Personality and Social Psychology, 83(6), 1281-1297.
66 days median to form a habit (range: 18-254)
The widely-cited “21 days to form a habit” claim is wrong. Lally et al. tracked 96 participants forming new habits and found the median time to reach 95% behavioral automaticity was 66 days. Some habits formed in 18 days, others took 254. Variation depends on the behavior and the person.
Lally et al. (2010), European Journal of Social Psychology, 40(6), 998-1009.
The Ego Depletion Failure
One of the most-cited models in decision fatigue content is also the one with the weakest empirical support. The “willpower as fuel tank” idea, popularized by Roy Baumeister’s research in the 1990s and 2000s, did not survive large-scale replication.
Ego depletion: d = 0.04
A 23-laboratory pre-registered replication (n=2,141) found an effect size of essentially zero. The original studies showed exercising self-control depleted a finite “willpower” resource, leading to worse subsequent decisions. The replication found no such effect.
Hagger et al. (2016), Perspectives on Psychological Science, 11(4), 546-573.
No evidence for decision fatigue in large-scale healthcare data
A 2025 pre-registered study analyzed high-resolution data from a national telephone triage and medical advice service, where healthcare professionals make repeated medical judgments under varying levels of fatigue. The Bayesian analysis found no credible evidence for decision fatigue effects in this specific setting. The authors note this does not rule out weaker or more context-specific versions of the phenomenon. The finding suggests the effect may be stronger in some decision contexts than others, not that decision fatigue itself is fake.
Andersson, Lindberg, Tinghög & Persson (2025), Communications Psychology, 3, 33.
What does this mean? The popular “eat a snack to restore willpower” advice has weak empirical backing. The glucose-as-willpower-fuel model is not supported by laboratory replication. The specific “fuel tank emptying and refilling with sugar” mechanism failed.
But this does not mean decision fatigue isn’t real. It means we should be cautious about specific oversimplified mechanism claims while continuing to take the broader phenomenon seriously. The Linder 2014 antibiotic study, the morning-versus-evening cognitive performance research, the choice architecture literature, the universal subjective experience of mental fatigue, and decades of research on cognitive control and executive function all point to the same conclusion: sustained decision-making is costly, and the cost manifests in measurable ways.
The honest framing: decision fatigue produces a cluster of effects when cognitive resources are taxed. People make worse decisions in some professional contexts. They become more impulsive. They defer or accept defaults. They exhibit stronger status quo bias. They feel mentally tired. The mechanism may be more nuanced than the original “willpower fuel tank” model suggested, but the phenomenon is supported by experience, neuroscience, and multiple research traditions. The popular framing oversimplified the mechanism. It did not invent the phenomenon.
Recovery and Reset
What helps? The research on recovery from cognitive load is more reliable than the research on what causes it.
Sleep deprivation degrades complex decisions
Sleep-deprived subjects made significantly worse decisions on novel and complex tasks. Routine decisions held up better. The deficit is largest for decisions requiring integration of multiple variables.
Harrison & Horne (2000), Journal of Experimental Psychology: Applied, 6(3), 236-249.
50-minute nature walk → 20% attention restoration
A 50-minute walk in a natural environment improved directed-attention performance by approximately 20%, compared to a matched urban walk. Brief nature exposure meaningfully restores cognitive resources depleted by sustained decision-making.
Berman, Jonides & Kaplan (2008), Psychological Science, 19(12), 1207-1212.
Why This Matters for Commuters
The reason CarryCommute compiled these decision fatigue statistics is that the research has direct implications for the morning commute. The hours when our brain is sharpest are the hours we spend making logistical micro-decisions before reaching the desk.
Three pieces of the research are particularly relevant:
The morning peak. Cognitive performance peaks in the first one to three hours after waking. This is the most valuable cognitive real estate of the day. Most commuters spend it deciding what to wear, what to pack, which route to take, and whether to stop for gas.
The cluster of effects. When morning decisions accumulate, the consequences show up in multiple ways: we make worse choices about food and route, we become more impulsive about purchases or screen time, we defer to defaults like grabbing coffee instead of breakfast or taking the familiar congested route, and we stick with the status quo even when better options exist. The morning compresses enough decisions to push us into all of these states before we’ve reached the office.
Choice architecture as the practical fix. Default effects are the largest behavioral interventions ever measured. Engineering the morning environment to remove decisions works because it accepts the cognitive limit rather than trying to overcome it through discipline. We’ve covered this in our companion pieces on moving morning decisions to the night before and designing the environment to remove decisions entirely.
How we selected these statistics
We started with every decision fatigue statistic we could find in popular productivity content. We traced each one to its claimed primary source. Where the source didn’t exist, we said so. Where the source existed but was contested or had failed replication, we flagged it.
We prioritized peer-reviewed studies in major journals. Where industry data appears (such as the phone pickup figure), we noted the source and limitations. We excluded research from labs with documented data integrity problems, findings from the broader ego depletion paradigm that failed replication, and any uncited “studies” that couldn’t be traced.
This page will be updated as new research is published.
One friction point at a time.
Short emails covering overlooked commute problems, why they drain us, and the simplest fixes. No spam, no fluff.
Keep reading
Mind & Commute
We’ve Already Used Our Best Thinking Before We Leave the House
The morning version of decision fatigue and the night-before fix.
Mind & Commute
Most of Our Morning Was Already Decided
How to remove morning decisions instead of just simplifying them.
Mind & Commute
Commute Stress Is Ruining Your Day (Not Your Job)
The small frustrations we’ve stopped noticing are draining us.
Mind & Commute
Why Our Commute Feels Longer Than It Is
Time perception, anticipatory anxiety, and why the same drive feels different on different days.