What Showing the Work Actually Does

Answer to the question

What does the labour illusion actually do?

It raises what people think a service is worth when they can see work being done on their behalf. In Ryan Buell and Michael Norton's 2011 experiments, showing a travel search in progress lifted perceived value above a blind progress bar at every wait length tested, and the route ran through perceived effort and then reciprocity. The wait itself still lowered value.

Key takeaways

• In the original labour illusion experiments wait time had its own negative effect on perceived value; transparency raised the whole line without changing its direction.
• Showing the work beat describing it, 5.13 against 4.49 against 3.82 for a blind progress bar, even though describing it made people believe about 53 sites had been searched against about 28.
• The mediation runs through perceived effort and then reciprocity, not through reduced uncertainty, because a progress bar sat in every waiting condition.
• It reverses when the outcome is poor, 2.47 against 3.24 for an instant result, and across 4.68 million parcel deliveries customers punish visible idleness, late idleness hardest.

Two things are called effort, and conflating them has cost this publication an argument. The labour illusion is about the provider's effort. Checkout friction is about the customer's. Keep them apart and both bodies of evidence mostly stand.

The Cost of the Wait

The labour illusion is Ryan Buell's finding, with Michael Norton as second author, published in 2011.

Their own definition, from the paper's abstract, is that because of what they term the labour illusion, when websites signal that they are exerting effort, people can prefer websites with longer waits to those that return instantaneous results, even when those results are identical.

Read quickly, that sounds like a licence to slow things down.

It is not. In their first experiment, 266 people online saw a travel search that either returned fares instantly or made them wait between ten and sixty seconds. The waiting conditions were split. Transparent ones showed a changing list of the sites being searched with fares visibly compiling. Blind ones showed a progress bar filling at a uniform rate. Perceived value, measured on a seven-point scale, came out at 5.36 with transparency against 4.96 without, and the difference cleared significance comfortably.

Then the part the retellings drop. Wait time had an effect of its own and it was negative. Value showed what the authors call a general downward trend over time, and the test for that trend was significant too. Transparency scored higher at every single time point, and the two factors did not interact.

So transparency raised the whole line. It did not change the direction the line was travelling in.

Be precise about that last clause, because the paper's discussion also offers a rule of thumb, that the benefit of transparency starts to fade after about thirty seconds in the travel simulation and after about fifteen in the dating one. The direct evidence on whether the benefit varies with the length of the wait is the interaction, and the interaction came back null.

We report the test. The rule of thumb is the authors' reading of it, and we do not print it as a measured decay.

Either way, anyone selling a longer wait on the strength of this paper is selling something the paper does not contain, because the wait itself cost value at every length measured.

The Showing and the Telling

The third experiment is the most interesting single result in this literature, and it is the one that tells you what the labour illusion runs on. 143 people online were put in one of three positions. Some were shown the work. Some were merely told about it, with a flash of the message that the site was preparing to search a hundred sites and a list of about a hundred names. Some got a blind progress bar.

Condition	Perceived value	Perceived effort	Sites believed searched
The work shown	5.13	5.50	about 28
The work described	4.49	4.75	about 53
Blind progress bar	3.82	4.59	not asked

Telling made people believe in more work and buy less of it.

Telling made people believe in more work and buy less of it.

That is the finding. Being told about a hundred sites left participants estimating roughly 53 sites searched, nearly twice the estimate of the people who watched the search happen, and yet the describing condition landed almost two thirds of a scale point below the showing condition on value. The believed quantity of labour predicted neither perceived effort nor perceived value.

Volume of work is not the currency. Witnessed work is.

The Route Through Reciprocity

The mediation in the paper is explicit and it has three steps. Transparency raises perceived effort. Perceived effort raises reciprocity, meaning appreciation and gratitude directed at the provider. Reciprocity raises perceived value. Every link in that chain reached significance, the mediation was full, and nothing off the hypothesised path reached significance at all.

What the mediation is not is uncertainty reduction. A progress bar sat in every non-zero wait condition in the paper, transparent ones included, put there deliberately so that uncertainty could not do the explaining. Reported uncertainty did not differ across the conditions.

This is the point at which the adjacent literature becomes useful and also inconvenient. The closest relative is the effort heuristic, named by Justin Kruger and colleagues in 2004, which holds that information about how much effort went into something raises judgements of its quality and its worth. We did not read that paper in this pass, and what we know of it here comes from the description inside its replication.

That replication, by Ignazio Ziano and colleagues in 2023, is the sharpest limit available. Two preregistered attempts, 1,405 participants in total, produced a split result in the authors' own words: support for an effort heuristic effect on liking and quality, and no support for an effort heuristic on monetary value.

Effort information still moves what people think of a thing. It no longer moves what they will pay.

One absence belongs here, and it was searched before it was allowed on the page. Three lines of enquiry went after a replication of the 2011 labour illusion paper, over four passes in standard and extended modes: failed and preregistered replications, null results in operational transparency experiments, and recent transparency experiments. Nothing came back in either direction. The honest description is that the labour illusion has been extended and has not been independently replicated, which is a narrower claim than either of the two things people usually say about it.

The Reversal at the Outcome

The paper's own fifth experiment is where the effect turns over, and it is to the authors' credit that it is in the paper and was not left for somebody else to find.

280 people in a laboratory used a simulated dating site. The design crossed a fifteen or thirty second wait with transparency or blindness and with a favourable, average or unfavourable outcome, where the outcome varied only by the photograph shown. The compatibility score was held at 96.4 per cent in every condition. The interaction between transparency and outcome was significant.

With an unfavourable outcome, the transparent fifteen-second wait scored 2.47 against 3.24 for the instantaneous version. At thirty seconds the transparent condition fell to 2.10. The blind condition, meanwhile, was indistinguishable from instantaneous. Their reading, quoted:

> When a service demonstrates that it is trying hard and yet still fails to come up with anything but poor results, people blame the service for this failure and rate it accordingly.

Visible effort amplifies the verdict. It does not substitute for one.

Visible effort amplifies the verdict. It does not substitute for one.

A second limit sits in their fourth experiment, which varied how much work was actually done behind the display, three sites against thirty-six. The headline effect of transparency held across the design. The simple effect, though, appeared only in the cell where real work was being done, 5.16 against 4.47. In the low-work cell the figures were 4.65 against 4.44, a gap the study could not distinguish from nothing. The authors report no interaction and treat the quantity of work as irrelevant. Read strictly, the only cell where transparency measurably worked was the one with work in it.

The authors also draw the ethical line themselves, writing in the paper that the fact firms can induce the labour illusion does not mean they therefore should. Operational transparency, they say, requires honesty. Inducing the illusion moves closer to an ethical boundary.

The Evidence from the Field

Buell's later work with Tami Kim and Chia-Jung Tsay takes the question out of the browser. Only the working paper would load for us, so every study detail below belongs to that version and not to the published one.

In a university dining hall, iPads were used across two weeks to open or close the sight lines between customers and the chefs at a grill station, in a staggered design, with 299 customers surveyed. Reciprocal transparency, where each side could see the other, produced a 22.2 per cent increase in customer-reported quality, and throughput time fell to 80.8 per cent of its standard level.

The second field study needs care. It is 48 people.

At a sandwich counter, 48 participants, 23 against 25, were assigned either to a transparent collection with a wait or to a blind instantaneous one. Perceived value rose against each participant's own earlier rating, and held up with controls. The transparent group reported waiting more than four times as long, 3.16 minutes against 0.72. Because the sandwiches were pre-made, objective quality did not move at all, which is the cleanest separation of perceived from real improvement anywhere in this literature. A companion study with 91 participants recruited in isolated rural Kenya found the full-transparency contrast only marginal.

Then a city. Buell, Elizabeth Porter and Norton studied Boston's Citizen's Connect app, where closing a service request could include a photograph of the work performed. After exposure, residents submitted an average of 59.8 per cent more service requests in 37.7 per cent more categories, with the effect persisting across most of the following thirteen months, drawn from 371,992 monthly observations of 21,786 reporters. This is real behaviour at scale and it is not a randomised experiment: treatment arrived through staggered app diffusion and the chance that a works team posted a photograph.

The same programme carries its own backlog problem. With 125 Boston residents and three versions of a civic website, transparency about work recently opened and closed raised trust and support for government programmes, while a version that additionally displayed the substantial backlog of unresolved requests came out marginally less trusted than the functional version, though no worse than the blind baseline. The authors' summary is that showing what government is doing increases trust, and showing what government is not doing may not. In the app data, where residents had experienced government to be unresponsive, transparency was associated with reduced engagement.

Wait psychology has evidence of its own, and it points at progress, not at effort. Nira Munichor and Anat Rafaeli held 123 first-time callers to a university laboratory's sign-up line on a 108-second wait, filled with music, with apologies, or with queue-position information. These were people ringing a line the researchers controlled, in a call staged for the study, and not customers of an ordinary business. Abandonment was 69.4 per cent with music, 66.7 per cent with apologies and 35.9 per cent with position information. In their follow-up, the sense of progress carried the effect and perceived waiting time explained none of it.

The Idleness on Display

The strongest counter-evidence is not a laboratory result. Robert Bray, sole author, analysed 4.68 million parcel deliveries from the Cainiao Network, Alibaba's logistics arm, each carrying between four and ten customer-visible tracking events alongside a customer delivery score, with instrumental-variable regressions. His objection to the earlier work is about an assumption, not about anybody who made it:

> those authors implicitly assume that customers will perceive worker effort but not worker loafing. Yet true operational transparency will make both industry and idleness visible.

What he finds is that customers punish late idleness far more than early idleness. On a hundred-hour shipment, moving the average tracking action from hour 20 to hour 80 buys about as much delivery score as cutting the arrival from a hundred hours to seventy-three. He attributes it to [the peak-end rule](https://thewhyweproject.com/articles/why-you-remember-the-end-of-a-holiday), and to memory limits that make customers especially sensitive to how an operation ends.

David Maister had the practitioner version of this in 1985, and he was honest about what he had. His chapter on waiting lines offers testable propositions rather than findings, supported by anecdote, and he says so in the text. The proposition that matters here runs against the show-the-work case: fast-food chains instruct staff to take rest breaks out of sight of waiting customers, because the sight of serving personnel not actually serving customers reads as unexplained.

Maister had Bray's result as a hunch, thirty-eight years early.

The Case for Removing Friction

The case against everything above has more behavioural evidence behind it than the case for it, and it gets its full strength here. We have made it ourselves, in [What Cognitive Load Costs at Checkout](https://thewhyweproject.com/articles/what-cognitive-load-costs-at-checkout), which separates two things the trade bills at one rate. One is a price per form field, which rests on folklore. The other is effort reduction, which rests on more than 2.2 million people in field experiments plus the consent literature and research on dark patterns and where the blame sits. Only the second has anything behind it.

That piece reads an abandoned basket as a person declining to spend effort on a problem the interface invented. Its anchor is a meta-analytic framework by Yuxi Luo and colleagues covering 184 field-experiment papers and 2,245,373 participants, in which interventions targeting effort outperformed interventions targeting intrinsic motivation, with the authors cautioning about possible publication bias. We carry that record from our own earlier source file and did not re-verify it at source in this pass.

That is the largest pooled behavioural evidence base either side of this argument has. It is also about real behaviour and not rating scales, which is more than most of the transparency literature can say.

The remove-friction case is not weak. It is the strongest lever measured.

The Signal and the Cost

Both cases hold, and they may not be about the same thing.

In 1927 Stuart Chase and Frederick Schlink described a gap, not a mechanism. Mass production, they argued in Your Money's Worth, had severed the buyer from any standard by which to judge goods: the whole system for judging sound values went into decay, and trial and error alone remained. Their one measured result, Rosamond Cook's nine makes of cotton sheeting at Teachers College, found that a shopper had one chance in nine of judging relative quality from her own shopping experience, which is what a guess would give her, and that the only make she reliably recognised was the worst. Trained shop assistants did no better. Their remedy was published impartial testing.

Our inference, and ours alone, is this: visible effort is the cue people reach for exactly where they cannot judge the thing itself, so it fills the judgement gap Chase and Schlink described using the one instrument they distrusted, the seller's own display. They wrote nothing about waiting, effort or process transparency. [The Thing You Cannot Judge](https://thewhyweproject.com/articles/the-thing-you-cannot-judge) takes a different line out of the same book.

Which gives a rule, offered as our hypothesis and not as a finding, because no study here was built to test it. The customer's effort is a cost. The provider's effort is a signal. Take the first away, and put the second where it can be seen.

Bray's parcels look like a counter-example, and they are the reason the rule needs its second clause. A tracking gap is a provider's own process made visible, so on the loose version of the rule it should have helped, and instead it cost score. What those customers could see, though, was not work being done. It was a long hole in the middle of the job where work was not being done, which is the same thing the fast-food case Maister describes keeps off the floor when staff take their breaks out of sight.

So the signal is witnessed work, and witnessed absence of work is a signal too, pointing the other way.

That second clause is our refinement and it is in no paper here. A reader who thinks it rescues the rule instead of correcting it should treat the parcel result as a live objection and keep it on the table. We think the refinement is right. We cannot say it was tested.

Where we disagree with our own earlier piece is narrower and firmer. That article treats a wait as a cost to be minimised and nothing else, and on this evidence that is incomplete. A wait with visible work is not the same object as a wait without it. Buell and Norton's second experiment put that to a choice. 118 people in the laboratory, each paid, used two rival simulated travel sites, one of which returned results instantly. Offered identical fares, 62 per cent waited thirty seconds for the transparent site and 63 per cent waited sixty, against 42 and 23 per cent choosing the blind one.

Three checks, then, before anyone builds a progress display:

1. Is there real work happening behind the display, or only a display?
2. Is the outcome likely to be good, given that visible effort made a poor outcome worse?
3. Does the display have gaps in it, especially late ones?

Underneath all three sits the distinction this piece opened with, and it is the one most teams get wrong: ask whether the thing you are about to make visible is the customer's work or your own, because taking a field off a form and showing a kitchen are opposite moves that happen to share a word, and a team that confuses them will reliably remove the signal it needed and keep the cost it should have deleted.

Common questions

What does the labour illusion actually do?

It raises what people think a service is worth when they can see work being done for them. In the original experiments, showing a travel search in progress lifted perceived value above a blind progress bar at every wait length tested. The route runs through perceived effort and then reciprocity, not through reduced uncertainty, since a progress bar sat in every waiting condition.

Does the labour illusion mean longer waits are better?

No, and the paper says otherwise. Wait time had its own negative effect on perceived value, with a general downward trend over time. Transparency raised the whole line without changing its direction, and the interaction test was null, meaning the lift did not vary with how long people were kept waiting. Showing the work is the lever. The wait is still a cost.

Has the labour illusion been independently replicated?

Not as far as we can establish. We went looking down three lines, through failed and preregistered replications, nulls in operational transparency and recent transparency experiments, and came back with neither a replication nor a failure. What exists is extension into dining halls, a city app, hospitality and healthcare interfaces. Treat the effect as extended, not replicated.

When does showing the work backfire?

When the outcome is poor, when the visible work is idleness, and when the provider is unresponsive. With an unfavourable result, a transparent fifteen-second wait scored 2.47 against 3.24 for an instant one. Across 4.68 million parcel deliveries, shifting the average tracking event on a hundred-hour shipment from hour 20 to hour 80 gained about as much delivery score as cutting the arrival to seventy-three hours.

Sources

Buell, R. W., & Norton, M. I. (2011). 'The Labor Illusion: How Operational Transparency Increases Perceived Value.' Management Science, 57(9), 1564-1579. doi:10.1287/mnsc.1110.1376. [FULL TEXT READ, from the publisher PDF hosted at hbs.edu; the published page at pubsonline.informs.org returned an empty body and is recorded as not read. Author order confirmed at both the Crossref and the OpenAlex item records: BUELL FIRST, NORTON SECOND, both at Harvard Business School. The trade almost always credits Norton and should not.] Five experiments, online travel and online dating, with a uniform progress bar in every non-zero wait condition including the transparent ones, specifically to hold uncertainty constant. Experiment 1: 266 participants online, 13 conditions, namely instantaneous or 2 (transparent versus blind) by 6 (wait of 10 to 60 seconds); perceived value on a 7-point scale 5.36 with transparency against 4.96 without, significant, AND a significant main effect of wait time with value showing 'a general downward trend over time', transparency higher at every time point and no interaction. Experiment 2: 118 in the laboratory, within participants across two rival simulated sites one of which was instantaneous; preferring to wait rather than take identical instant results was 62 per cent at 30 seconds and 63 per cent at 60 seconds with transparency against 42 per cent and 23 per cent without. Experiment 3: 143 online, 17 excluded for implausible estimates; the third condition told rather than showed, flashing that the site was preparing to search 100 sites with a list of about 100 sites; perceived value transparent 5.13, told 4.49, blind 3.82, all contrasts significant; perceived effort 5.50, 4.75, 4.59; reciprocity 5.19, 4.43, 3.44; telling raised the believed quantity of work far more than showing, 53.28 sites against 28.13, yet quantity predicted neither perceived effort nor perceived value; uncertainty did not differ across conditions; path analysis gave transparency to perceived effort, effort to reciprocity and reciprocity to value, each significant, full mediation with nothing significant off the hypothesised path. Experiment 4: 116 participants, 16 outliers removed, 2 (transparent versus blind) by 2 (actual labour low, 3 sites and 15 results, versus high, 36 sites and 433 results), all waiting 30 seconds; main effect of transparency on value 4.89 against 4.46 and on satisfaction and repurchase intentions; BUT the simple effect held only in the high-labour cell, 5.16 against 4.47, while the low-labour cell was 4.65 against 4.44, not distinguishable from nothing. Their own managerial hedge: benefits of transparency 'began to decline after 30 seconds' in travel and 'even earlier, at 15 seconds' in dating. Their own ethics caveat, in the paper: 'the fact that firms can induce the labor illusion does not mean that they therefore should induce it', operational transparency being 'ethically unproblematic' because it 'requires honesty' while inducing the illusion 'moves closer to an ethical boundary'.

Buell, R. W., & Norton, M. I. (2011), Experiment 5, the authors' own boundary condition, same paper and DOI as above. [FULL TEXT READ.] 280 participants in the laboratory on a simulated dating site, design 2 (wait 15 or 30 seconds) by 2 (transparent or blind) by 3 (outcome favourable, average or unfavourable), plus three instantaneous conditions; outcome favourability was manipulated only by the photograph while the compatibility score was held at 96.4 per cent in every condition. Significant main effect of wait time, significant main effect of outcome, and the predicted significant interaction of transparency with outcome. With an unfavourable outcome, waiting 15 seconds WITH transparency scored 2.47 against 3.24 for instantaneous, while the blind condition at 3.23 was indistinguishable from instantaneous; at 30 seconds the transparent condition fell to 2.10. For average outcomes no contrast was significant. The authors' own reading: 'When a service demonstrates that it is trying hard and yet still fails to come up with anything but poor results, people blame the service for this failure and rate it accordingly.'

Buell, R. W., Kim, T., & Tsay, C.-J. (2017). 'Creating Reciprocal Value Through Operational Transparency.' Management Science, 63(6), 1673-1695. doi:10.1287/mnsc.2015.2411. [WORKING PAPER FULL TEXT READ, from the Harvard DASH bitstream; the published version at pubsonline.informs.org was not retrieved, so study detail is attributed to the working paper. Author order confirmed at OpenAlex: Buell first at Harvard Business School, Kim second, Tsay third at University College London.] Headline from the published abstract: reciprocal operational transparency 'contributed to a 22.2% increase in customer-reported quality'. Field experiment 1: a grill station in a university dining hall over two weeks, iPads used to open or close sight lines in a 2 by 2 staggered treatment design, 299 customers surveyed, with throughput time falling to 80.8 per cent of standard in the fully specified model. Field experiment 2, the small one: 48 participants, 23 against 25, assigned to a transparent-with-wait or blind-instantaneous sandwich collection; perceived value rose relative to each participant's own earlier rating and held with controls, EVEN THOUGH the transparent group reported waiting more than four times longer, 3.16 minutes against 0.72; because the sandwiches were pre-made, objective quality did not move, which separates perceived from real improvement cleanly. Mechanism studies: 272 recruited on Mechanical Turk with 254 retained; 91 participants recruited from isolated rural areas of Kenya, where the full-transparency against no-transparency contrast on perceived value was marginal; and 599 participants on Mechanical Turk for the employee side.

Buell, R. W., Porter, E., & Norton, M. I. (2021). 'Surfacing the Submerged State: Operational Transparency Increases Trust in and Engagement with Government.' Manufacturing & Service Operations Management, 23(4). doi:10.1287/msom.2020.0877. [MARCH 2017 WORKING PAPER FULL TEXT READ, from the copy hosted at utdallas.edu; the published version was not retrieved and study numbering differs between versions, so studies are described rather than numbered. Buell first at Harvard Business School, Porter second at George Washington University, Norton third.] The website study: 125 voting-age Boston residents randomly assigned to one of three versions of a map-based civic website, namely blind with tallies only, functional transparency showing recently opened and closed service requests, or dysfunctional transparency additionally showing the substantial backlog of unresolved requests; functional transparency raised trust and support for government programmes relative to blind, while the backlog condition was marginally LESS trusted than functional transparency though no less supported and not below the blind baseline. The authors' own summary: 'showing what the government is doing increases citizen trust and in turn, support for government; showing what government is not doing (e.g. reported potholes that go unfilled) may not'. The behavioural study: Boston's Citizen's Connect app, where from 25 September 2014 closing a request could include a photograph of the work performed; treatment spread through staggered app diffusion plus the chance that a public works team posted a photograph, so this is NOT randomised assignment; residents submitted an average of 59.8 per cent more service requests in 37.7 per cent more categories after exposure, persisting with significance in most of the following 13 months, from 371,992 monthly observations of 21,786 resident reporters, excluding the top percentile of submitters. In the same data, gains were largest where residents had experienced government to be responsive, and 'under relatively low levels of responsiveness, transparency was associated with diminished citizen engagement'.

Bray, R. L. (2023). 'Operational Transparency: Showing When Work Gets Done.' Manufacturing & Service Operations Management, 25(3), 812-826. doi:10.1287/msom.2020.0899. [FULL TEXT READ, from the copy hosted at kellogg.northwestern.edu. SOLE AUTHOR, Operations Management Department, Kellogg School of Management, Northwestern University.] The strongest when-it-backfires evidence in the literature and the only item running on a very large commercial dataset: 4.68 million parcel deliveries from the Cainiao Network, the logistics arm of Alibaba, each with between 4 and 10 customer-visible track-package activities and a customer-supplied delivery service score, analysed with five specifications including three sets of two-stage least-squares instrumental-variable regressions using calendar-based instruments. His objection to the earlier literature, quoted: those authors 'implicitly assume that customers will perceive worker effort but not worker loafing. Yet true operational transparency will make both industry and idleness visible.' Finding: customers punish early idleness less than late idleness and leave higher scores when tracking activity clusters towards the end of the shipping horizon, such that on a 100-hour shipment 'delaying the time of the average action from hour 20 to hour 80 increases the expected delivery score by approximately the same amount as expediting the arrival time from hour 100 to hour 73'. He attributes it to the peak-end rule and concludes that 'memory limitations make customers especially sensitive to how service operations end'.

Maister, D. H. (1985). 'The Psychology of Waiting Lines.' In J. A. Czepiel, M. R. Solomon & C. F. Surprenant (eds), The Service Encounter, Lexington Books, pp. 113-123. [FULL TEXT READ, from the author's own copy at davidmaister.com; citation cross-checked against the deposited reference list of Buell and Norton 2011.] IT IS A FRAMEWORK AND NOT A STUDY, AND MAISTER SAYS SO: he offers 'testable propositions offering the opportunity for future research' and his support is anecdote and practitioner report. His propositions include occupied time feeling shorter, pre-process waits feeling longer than in-process waits, anxiety lengthening waits, uncertain waits feeling longer than known finite ones, unexplained waits feeling longer than explained ones, unfair waits feeling longer, tolerance rising with the value of the service, and solo waits feeling longer. THE PROPOSITION THIS ARTICLE NEEDS MOST CUTS AGAINST THE SHOW-THE-WORK CASE: under unexplained waits he notes that fast-food chains instruct serving personnel to take rest breaks out of sight of waiting customers, because 'the sight of serving personnel not actually serving customers is "unexplained"'.

Munichor, N., & Rafaeli, A. (2007). 'Numbers or apologies? Customer reactions to telephone waiting time fillers.' Journal of Applied Psychology, 92(2), 511-518. doi:10.1037/0021-9010.92.2.511. [FULL TEXT READ, from the copy hosted at serveng.technion.ac.il. Munichor first, Rafaeli second, both at the Technion.] Study 1: 123 real first-time calls to a university laboratory's experiment sign-up line, randomly assigned to music, music plus three apologies, or music plus three position updates, every wait exactly 108 seconds. Abandonment: 69.4 per cent with music, 66.7 per cent with apologies, 35.9 per cent with queue-position information; music against apologies not significant. Satisfaction among the 48 callers who stayed, 2.87 music, 2.58 apologies, 3.82 position information. Study 2: 83 students on a simulated call, same conditions and duration, with the mechanism result that sense of progress mediated reactions while perceived waiting time did not and was non-significant in every model. Their own limitation: fixed wait duration and queue length, one stimulus per condition, and low position numbers chosen to signal proximity to the end.

Kruger, J., Wirtz, D., Van Boven, L., & Altermatt, T. W. (2004). 'The effort heuristic.' Journal of Experimental Social Psychology, 40(1), 91-98. doi:10.1016/S0022-1031(03)00065-9. [BIBLIOGRAPHIC RECORD CONFIRMED AT CROSSREF, Kruger first; the paper itself was NOT READ in this pass, and its content here comes from the replication paper's own description of it, which the article states plainly.] The adjacent literature closest to this article's subject and the one Buell and Norton cite. As the replication describes it, effort information affects judgements of quality: telling participants that a poem or a painting took 26 hours rather than 4 raised quality judgements and monetary valuations. It concerns judging somebody else's output from information about their effort, which is the mechanism this article is about.

Ziano, I., Yeung, S. K., Lee, C. S., Shi, J., & Feldman, G. (2023). '"The Effort Heuristic" Revisited: Mixed Results for Replications of Kruger et al. (2004)'s Experiments 1 and 2.' Collabra: Psychology, 9(1), 87489. doi:10.1525/collabra.87489. [ABSTRACT AND OPENING READ IN FULL at online.ucpress.edu; the full results sections were not read. Ziano first, University of Geneva, confirmed at OpenAlex.] Two preregistered replications, 1,405 participants in total, United States participants from MTurk and Prolific. In the authors' words the first replication 'found support for the original's findings regarding Experiment 2, yet failed to find support for the original's findings in Experiment 1', and the second revised attempt 'was mixed, with more nuanced findings, showing support for an effort heuristic effect for liking/quality and no support for an effort heuristic on monetary value'. The effort-to-quality link survived and the effort-to-money link did not, which is a direct limit on any claim that visible effort raises willingness to pay.

Luo, Y., Li, A., Soman, D., & Zhao, J. (2023). 'A meta-analytic cognitive framework of nudge and sludge.' Royal Society Open Science, 10(11), 230053. doi:10.1098/rsos.230053. [NOT RE-VERIFIED AT SOURCE IN THIS PASS, and the article says so. Carried verbatim from the vault's own verified source record for Articles/What Cognitive Load Costs at Checkout, where it is logged as abstract read in full: field experiments with real behavioural measures, 184 papers, 2,245,373 participants, 2008 to 2021, with interventions targeting effort outperforming interventions targeting intrinsic motivation, and the authors cautioning about possible publication bias. Luo first.] The remove-friction side's strongest pooled evidence and our own publication's anchor for that position. The detail that resolves the apparent conflict, as our own article reports it, is that interventions adding friction and interventions removing it came out similar in size, so the effort channel runs in both directions.

Chase, S., & Schlink, F. J. (1927). Your Money's Worth: A Study in the Waste of the Consumer's Dollar. The Macmillan Company, New York. [FULL TEXT READ by the vault as OCR of the Internet Archive scan, item yourmoneysworths00chas, 586,451 characters end to end, audited 20 September 2026; two tables at pp. 112 to 113 and pp. 160 to 161 are destroyed in the OCR and were not read. Nothing in the book's apparatus has been externally verified.] The seed, cited for its argument and never as a source of psychological evidence. The one citable psychological result belongs to Rosamond C. Cook at Teachers College, Columbia, not to Chase and Schlink, and carries her own limiting clause that consumers judge accurately only where quality varies widely. Both authors are long dead.

ABSENCE CHECK, licensed by the fact file and stated in the article. NO INDEPENDENT REPLICATION AND NO FAILED REPLICATION of Buell and Norton 2011 was found. Searched with four WebSearch passes in standard and extended modes: on failed and preregistered replications of the labour illusion; on nulls and non-significant results in operational transparency experiments; on recent labour-illusion work in chatbot and AI-interface settings; and on the published record via OpenAlex and Crossref citation metadata. What exists is conceptual extension, by the original author's group and by Hwang, Lee and Legendre 2024 and Anderson, Anderson and Nguyen 2025, both read as OpenAlex reconstructed abstracts only, plus Bray's field challenge to the premise. The article says 'not independently replicated' and does not say 'failed to replicate'. Separately NOT USED: the widely repeated '8 per cent higher value' figure, which appears nowhere in the paper's reported statistics and is NOT VERIFIED; and an unverified report of a 2026 chatbot delay study, for which no primary source was retrieved.

People

• Stuart Chase
• Michael Norton