A daycare centre had a problem with parents collecting their children late. It introduced a fine. The number of late collections went up, and when the fine was removed some months later, it did not go back down.

Key Takeaway

Whether extrinsic rewards undermine intrinsic motivation has been meta-analysed at least four times on overlapping literatures, producing opposite conclusions[1][5]. A comment on the dispute warned of significant costs to the unthinking or automatic use of meta-analysis on literatures of this kind[4]. Separately, field experiments found students paid roughly three cents per correct answer performed substantially worse than those paid nothing[9], and that a fine for late daycare pickup increased lateness[6]. A later replication attempt found the fine effect transitory and that fines deterred[8]. We set out both.

Our Grades For These Claims

Applying the scheme from the first article in this series, because this literature needs it more than most.

"Money does not motivate" is Grade D. No source we located supports it, and the field experiments below show incentives changing behaviour substantially, sometimes in the wrong direction.

"Tangible rewards undermine intrinsic motivation" is contested, and we grade it C. Not because the evidence is thin, but because competent researchers meta-analysing overlapping literatures reached opposite conclusions and the dispute is not resolved in anything we obtained.

"Small incentives can produce worse performance than no incentive" is Grade B. It rests on published field and laboratory experiments in economics[9][10].

"Fines can crowd out norms and the effect persists" is Grade C. The original is a randomised field experiment; a later replication attempt contradicts it on the persistence point[8].

Our own position, stated up front: this is the least settled literature we have covered, and any adviser presenting it as settled in either direction is overreaching.

A Note On Method

Everything here is verified to August 2026 against the primary literature or its published abstracts.

We did not obtain the full text of any of the four meta-analyses. We rely on abstracts, on the authors' own summaries of their work, and on a published comment that compares them.

We cite Jenkins and colleagues (1998), the principal workplace meta-analysis on financial incentives, by citation only. We could not obtain its findings and therefore state none, which is a material gap addressed in its own section.

The same applies to Rynes, Gerhart and Minette (2004), whose title we can report and whose findings we cannot.

Our sources give the page range of the daycare paper as both 1–17 and 1–18, and we do not resolve it.

Several supporting claims come from a commercial or practitioner publication and are flagged where they appear.

This article reviews research on motivation. It is not compensation, employment law or psychological advice, and pay design has legal and contractual dimensions this article does not address.

The version that reached business. This section is our own analysis.

Some form of the following proposition is now standard in management writing: money is a poor motivator, extrinsic rewards damage the intrinsic drive that produces good work, and organisations should therefore rely on autonomy, mastery and purpose instead.

Three things are worth separating in that claim, because they are not equally supported.

That rewards can reduce intrinsic motivation under some conditions is a real research finding with a large literature behind it.

That this makes money a poor motivator does not follow, and is contradicted by the field experiments below in which money changed behaviour dramatically.

And that organisations should therefore de-emphasise pay is a policy recommendation resting on a contested empirical base, offered to people who have an obvious interest in believing it.

That last point deserves saying plainly. A doctrine that concludes employers should worry less about pay is one that employers will find congenial, and this series has already noted that a finding supporting what you were inclined to do receives less scepticism than one that contradicts you.

Where It Started

The founding study.

The finding that extrinsic rewards can undermine intrinsic motivation first appeared in Deci (1971), and has been highly controversial since[1]. Research in that and subsequent papers revealed that tangible rewards, specifically money, could undermine college students' intrinsic motivation[3].

The theoretical apparatus that grew from it is cognitive evaluation theory, formulated by Deci and Ryan in 1980 to explain both positive and negative reward effects[2], later developed into self-determination theory.

Two observations, ours.

The original participants were college students in laboratory tasks. That is the population and setting from which a great deal of workplace advice has been derived, and the gap is worth holding in mind.

And the phrase "highly controversial since it first appeared" is the authors' own characterisation of their field[1]. This was never a settled finding that was later questioned. It was disputed from the start.

The Meta-Analysis War

What happened when people tried to settle it quantitatively.

Cameron and Pierce (1994), in the Review of Educational Research, concluded that the undermining effect was minimal and largely inconsequential for educational policy, and advocated abandoning cognitive evaluation theory altogether[1][2].

Tang and Hall (1995) published a third meta-analysis of the same broad literature[4].

Deci, Koestner and Ryan (1999), in Psychological Bulletin, meta-analysed 128 experiments, organised so as to test cognitive evaluation theory much as Cameron and Pierce had done, and concluded that tangible rewards do significantly and substantially undermine intrinsic motivation[2].

They further stated that the Cameron and Pierce meta-analysis was seriously flawed and that its conclusions were incorrect[1].

Two observations, ours.

These are not marginal differences of emphasis. One team concluded the effect was inconsequential and the governing theory should be discarded; the other concluded the effect was substantial and the theory strongly supported.

And both were published in leading peer-reviewed journals, by established researchers, on overlapping bodies of experiments.

Round Two

The exchange continued, and its titles convey the temperature.

Ryan and Deci published When paradigms clash: Comments on Cameron and Pierce's claim that rewards do not undermine intrinsic motivation in the Review of Educational Research in 1996[2].

Deci, Koestner and Ryan returned in 2001 with Extrinsic Rewards and Intrinsic Motivation in Education: Reconsidered Once Again[1], arguing that there is reason for teachers to exercise great care when using reward-based incentive systems[2].

And Cameron, Banko and Pierce replied with Pervasive Negative Effects of Rewards on Intrinsic Motivation: The Myth Continues, stating that their 1994 analysis had concluded negative effects were limited and could be easily prevented in applied settings, that Deci and colleagues showed pervasive negative effects, and that their purpose was to resolve differences in previous meta-analytic findings[5].

Our own observation. A paper title containing the phrase "The Myth Continues" is not a neutral summary of the evidence, and neither is a rejoinder describing the other side's work as seriously flawed. That does not make either wrong. It means a reader encountering one of these papers alone is receiving advocacy as well as analysis, and should know which side they have picked up.

The Comment That Refereed It

A third party's assessment, published alongside the dispute, and the most useful document in it.

A comment on the 1999 meta-analysis noted that three different meta-analytic reviews of the same literature had appeared and, despite their common focus, had offered dramatically opposed bottom-line conclusions about its meaning and implications[4].

Its verdict: the findings of this literature had been more accurately captured by the reviews of Deci and colleagues and of Tang and Hall (1995) than by that of Cameron and Pierce (1994)[4].

But it did not stop there, and the second half is the part worth carrying away. The authors suggested there may be significant short- and long-term costs to the unthinking or automatic use of meta-analysis with theoretically derived, procedurally diverse, and empirically complex literatures like this one[4].

Three observations, ours.

The comment sides with the undermining effect, two of three reviews against one, which is the closest thing to a referee's decision we located.

But it also makes a methodological point that outlives the dispute: meta-analysis is not a neutral machine that resolves disagreements. Applied to a literature where studies differ in procedure and derive from competing theories, it can encode the analyst's judgments about what counts as comparable.

And that qualifies the first article in this series. We described meta-analytic bias correction predicting a replication result in the ego depletion case, which it did. This case shows the other outcome: the same technique, applied by different competent teams, producing opposite answers.

What A Contested Meta-Analysis Means

How a practitioner should read a dispute of this kind. This section is our own analysis.

The instinct is to look for the winner. That instinct is usually wrong, for three reasons.

When meta-analyses disagree, the disagreement is typically about inclusion criteria and coding decisions rather than about the underlying data. Both teams are looking at broadly the same experiments.

Which means the honest inference is not that one side is right, but that the effect is sensitive to how you define it. An effect that appears or vanishes depending on which studies are judged comparable is a real effect operating under conditions nobody has yet specified.

And a practitioner's question is not who won. It is whether the conditions in my business resemble the conditions where the effect appears, which requires knowing what those are.

The choice overload article in this series described the same resolution arriving through moderators. This literature has not obviously had that moment, and until it does, confident application is not warranted.

Where We Come Out

Our own position, stated so a reader can disagree with it.

We think the undermining effect is real, conditional, and smaller in workplaces than in laboratories, for three reasons, none of which is decisive.

The referee comment favoured it, two reviews to one[4].

The economics field experiments below independently produce crowding-out results by entirely different methods, in real settings, which is convergent evidence of a kind meta-analytic disputes cannot supply.

And the laboratory paradigms concern activities people already enjoy, being puzzles and games, where there is intrinsic motivation available to undermine. A great deal of paid work does not have that property, which would attenuate the effect.

Against that, we hold it loosely because a competent team meta-analysing the same literature reached the opposite conclusion and defended it across a decade.

We state this as our reading rather than as a finding, and a reader is entitled to weigh the same evidence differently.

Pay Enough Or Do Not Pay At All

The economics literature, which approaches the question from a different direction and is where the practically useful results are.

Gneezy and Rustichini published Pay Enough or Don't Pay At All in the Quarterly Journal of Economics in 2000[10].

A description of the study reports that Israeli college students given a very small amount of money, the equivalent of three cents in US currency, for each IQ question answered correctly performed substantially worse than those offered no incentive at all[9].

We flag that this description comes from a practitioner publication[9] and that we did not obtain the paper.

Three observations, ours.

The comparison that matters is small payment against no payment, not against large payment. The finding is not that money fails; it is that a token amount can be worse than nothing.

The title states the practical rule as clearly as any paper we have encountered: pay enough, or do not pay at all. There is a range in the middle that is worse than either end.

And this is directly actionable in a way the meta-analysis dispute is not. Referral bonuses, small spot awards, nominal per-item rates and token performance payments all sit in exactly that middle range.

Why Small Can Be Worse Than Nothing

The mechanism, as the literature describes it and as we read it.

A practitioner source states the underlying idea: any addition of money into the equation, even a small amount, has the potential to change the nature of the interaction between the parties, so that what was before viewed as a favour could be transformed in the mind of the agent into a market interaction when even fractions of a dollar are on the table[9], an idea it attributes to Fiske (1991).

Our own reading of what that implies.

The token payment does not add a small amount of motivation to the existing stock. It replaces the frame. Before the payment, the question was how well should I do this. After it, the question is is this worth three cents.

Three consequences.

The relevant comparison for the person is now the payment against the effort, and a small payment loses that comparison.

The effect should be largest where the pre-existing motivation was social or personal rather than financial, because that is what gets displaced.

And it should be hard to reverse, since removing the payment does not obviously restore the earlier frame. That prediction is tested, in both directions, by the study discussed next.

A Fine Is A Price

The most cited demonstration of crowding out, and a genuine field experiment.

Gneezy and Rustichini published A Fine is a Price in the Journal of Legal Studies in 2000[6]. Our sources give the page range as both 1–17[11] and 1–18[12], and we do not resolve it.

The design was a randomised field experiment at Israeli daycare centres, testing whether, as standard assumptions in psychology, economics and legal studies hold, imposing negative consequences on a behaviour produces a reduction of that behaviour[8].

The paper reports four periods: before the fine (weeks 1–4), the first four weeks with the fine (weeks 5–8), the entire period with the fine (weeks 5–16), and the post-fine period[6].

Its first stated result: the effect of introducing the fine was a significant increase in late collections[6].

A practitioner source adds a useful defence against the obvious objection. One might wonder whether the fine was simply too small; but since more parents arrived late after the fine rather than the rate merely holding constant, the fine was not merely ineffective, it produced the opposite of the intended result[9].

The Part That Should Not Have Happened

The detail that made this study famous, and which the authors themselves found difficult.

When the fine was removed, lateness did not return to its previous level.

The authors' own account of this is unusually candid. They write that the behaviour in the third period is still difficult to explain; that a social norm account would predict a low level of delay; that one might suppose four weeks was insufficient for behaviour to return to the appropriate level; but that considering two or three weeks had been enough to double the level of delays after the fine was introduced, the learning rate of the parents is too fast to justify so much persistence[6].

Two observations, ours.

That is a paper reporting a result its own authors could not account for, and saying so in print. It is a model of how to handle an anomaly and the opposite of the pattern this series described in the ego depletion case.

And the asymmetry they identify is the interesting part. The norm collapsed in two or three weeks and did not rebuild in four. If that is right, it suggests social norms are easier to destroy than to restore, which would be a substantial claim about organisational behaviour generally.

We flag that this is an interpretation of a single field experiment, that the authors expressly declined to explain it, and that the next section describes a study reaching the opposite conclusion on precisely this point.

And The Replication That Contradicts It

The evidence on the other side, which is rarely mentioned alongside the original.

A study titled Is a fine still a price? Replication as robustness in empirical legal studies sought to replicate the results using experimental surveys administered on a crowdsourcing platform, a methodology the authors describe as increasingly common in empirical legal studies, psychology and economics[8].

Their findings run against the original on both key points.

The effects of fines on outcome behaviours and on respondents' reasons were transitory. Once the fines were removed, respondents returned to their baseline behaviours[8].

And the results were consistent with intuitive judgments and standard rational-choice theory that fines deter[8].

The authors state their own limitation plainly, and we quote it because it is the sort of caveat that usually gets stripped in retelling: a survey is not a field experiment; their results suggest more research is required to understand when and how any "fine is a price" effect may arise[8]. They also describe their work as not an exact replication, undertaken because it allowed them to control aspects of the design difficult to replicate in the field[8].

Weighing The Two

Our own assessment, offered as reasoning rather than as a finding.

The original is a randomised field experiment observing real parents making real decisions with real consequences over sixteen-plus weeks.

The replication is a survey experiment on a crowdsourcing platform, which its own authors distinguish from a field experiment.

Three consequences.

On ecological validity, the original is stronger. Stated behaviour in a survey and actual behaviour under a real fine are different things, and the gap between them is the subject of the section below.

On statistical power and control, the replication is likely stronger, and its authors chose the method for that reason.

And the honest summary is that the crowding-out effect is demonstrated in one field setting and not confirmed in a survey setting, which is a considerably weaker position than the popular retelling implies but is not a refutation.

Our practical reading: treat crowding out as a risk to be considered when introducing a penalty or a small payment into a relationship previously governed by norms, not as a mechanism you can count on operating.

Where Crowding Out Is Most Likely

The conditions under which the effect appears most consistently, from the practitioner literature.

A source notes that prosocial behaviours are an interesting subset of potentially incentivised behaviours to which people often react paradoxically, because motives for performing them are often intrinsic, focusing either on the opinion of others or on one's own self-image, so that offering a financial reward for such acts often undermines the agent's true motive[9].

It cites two supporting results: that schoolchildren collecting donations for a charitable organisation collected less money when given performance incentives, and that volunteers, when rewarded with pay, reduced the number of hours they volunteered[9].

We report these as that source describes them and did not obtain either study.

Our own translation into a business setting, which is where the exposure sits.

Behaviours in a firm that are currently unpaid and socially motivated are the candidates. Helping a colleague, mentoring a junior, tidying a shared process, flagging a problem outside one's remit, covering for someone.

Those are precisely the behaviours an organisation notices, values, and then tries to encourage by attaching a small reward to them.

And if the crowding-out literature is right about anything, that is the most dangerous thing you can do to them.

Large Stakes And Big Mistakes

The other end of the range, and a caution against the simple fix.

Ariely, Gneezy, Loewenstein and Mazar published Large Stakes and Big Mistakes in the Review of Economic Studies in 2009[7]. A working-paper version of the same research states its scope as concerning what happens when introducing incentives where there previously were none, or raising existing ones[11].

We did not obtain the paper's findings and state none. We include it because it is the standard citation for the proposition that very large incentives can impair performance, and because a reader following this literature will encounter it.

What we will say is why the title matters even without the results, and this is ours.

If the Pay Enough finding says small incentives can be worse than none, and this literature examines whether very large ones create their own problems, then the space of safe incentive sizes is bounded at both ends.

That is a considerably more demanding design problem than the popular framing admits, and it is not solved by either paying nothing or paying a lot.

Anyone designing a scheme where the sums are large relative to base pay should obtain this paper rather than relying on our description of its title.

The Workplace Evidence We Could Not Obtain

A gap we are declaring rather than papering over.

The principal meta-analysis on this question in an employment setting is Jenkins, Mitra, Gupta and Shaw (1998), Are financial incentives related to performance? A meta-analytic review of empirical research, Journal of Applied Psychology 83(5), 777[11].

We obtained the citation and not the findings. We therefore report nothing about what it concluded.

We flag this prominently for two reasons.

It is the most directly relevant study in this article's subject area, being about financial incentives and performance in workplaces rather than about children, students or laboratory puzzles.

And its absence means that everything above is evidence from adjacent settings. Education, laboratory tasks, daycare parents and charity collection are all informative, and none of them is an employment relationship with a salary, a contract and a career attached.

Anyone making a compensation decision should obtain that paper. Our not having it is a limitation of this article, not an indication that the workplace evidence is absent.

What People Say Versus What They Do

A second citation we can name and not report, whose title alone is instructive.

Rynes, Gerhart and Minette published The importance of pay in employee motivation: discrepancies between what people say and what they do in Human Resource Management in 2004[11].

We did not obtain its findings and state none.

What we can observe, and this is our own reasoning, is that the title identifies a methodological problem running through this entire subject.

Much of what is known about motivation comes from asking people what motivates them. That is a self-report about a matter on which people have both limited introspective access and a clear social incentive to answer in a particular way.

Three consequences.

Survey evidence that employees rank pay low among motivators is weak evidence that pay does not motivate them, because saying so is what a person is expected to say.

This is a specific instance of the pattern the second article in this series described: a measurement that depends on voluntary disclosure captures the willingness to disclose alongside the thing measured.

And it is a reason to weight the field experiments in this article more heavily than the survey literature, whichever direction they point, because behaviour under a real incentive is not subject to it.

Design Rules That Survive The Uncertainty

What can be said despite an unsettled literature. These are our own inferences, and we mark them as such.

Avoid the middle range. The clearest result we located is that small payments can produce worse performance than none. If an amount is not material to the recipient, it may be doing harm rather than nothing.

Be most careful with currently unpaid prosocial behaviour. That is where the crowding-out literature reports the most consistent paradoxical results, and it is what organisations most often try to incentivise.

Treat penalties as a pricing decision. The daycare finding, whatever its robustness, establishes that a fine can be read as a price for permission. Ask what you are implicitly selling.

Assume norms are asymmetric. On the original field data, the norm collapsed in weeks and had not rebuilt in a month. Even if that specific result is contested, the asymmetry is the prudent assumption when contemplating a change.

Do not rely on what people tell you motivates them. A published paper's title identifies exactly this discrepancy, and self-report on this subject carries an obvious social incentive.

Separate the measured from the unmeasured. Any scheme rewards a specific dimension of performance, and the effort has to come from somewhere.

Pilot rather than roll out. An incentive change is one of the few interventions a business can run as a genuine experiment, by team, by branch or by period.

Watch what happens when you remove it. The most informative moment in any incentive scheme is its withdrawal, and it is the moment nobody plans to observe.

The Honest Summary

Where a reader should land.

The popular proposition, that money does not motivate, is not supported by anything we located, and the field experiments show money changing behaviour substantially.

The academic proposition, that tangible rewards undermine intrinsic motivation, is supported by the weight of the meta-analytic dispute as assessed by a third-party comment, and is disputed by competent researchers who defended the contrary position for a decade.

The economics field experiments provide convergent evidence for crowding out by different methods, and the single most famous of them has a replication attempt contradicting it on the persistence question.

And the workplace meta-analysis we most needed, we could not obtain.

Two things follow.

The most defensible practical rule is about size: pay enough or do not pay at all, because the middle is where the demonstrated harm sits.

And the most defensible practical posture is caution about newly incentivising what people currently do for other reasons, since that is the setting where the paradoxical results concentrate.

That is less than a doctrine and more than nothing, which on this literature is the honest amount to claim.

The Limits Of This Analysis

Several caveats matter. This article reviews research on motivation. It is not compensation, employment law or psychological advice; pay design carries legal and contractual dimensions this article does not address. Everything is verified to August 2026. We did not obtain the full text of any of the four meta-analyses discussed, and rely on abstracts, on the authors' own summaries of their own work, and on a published comment comparing them; readers should note that authors summarising their own disputed work are not neutral. We cite Jenkins and colleagues (1998), the principal workplace meta-analysis on financial incentives, by citation only and state none of its findings, which is a material gap since it is the most directly relevant study to this article's subject. The same applies to Rynes, Gerhart and Minette (2004), whose title we report and whose findings we do not. We did not obtain the findings of Ariely and colleagues (2009) and state none, citing it only as the standard reference for a proposition a reader will encounter. We did not obtain Gneezy and Rustichini (2000a), Tang and Hall (1995), Frey and Götte (1999), Falk and Kosfeld (2006), Frey and Jegen (2001), or the charity collection study, and report them as described by others. Our sources give the daycare paper's page range as both 1–17 and 1–18. Several supporting claims come from a practitioner publication and are flagged where they appear. The replication study describes itself as not an exact replication and as a survey rather than a field experiment, and we report that caveat as its authors stated it. Our position on where the evidence lands, the frame-replacement mechanism, the asymmetry inference, the prosocial translation to workplace behaviours and the eight design rules are our own reasoning, not findings from the literature.

Behavioural Finance Thirty-Six Percent Where meta-analytic correction predicted a replication, and this article shows the other outcome. Behavioural Finance The Better Teams Reported More Errors Why self-report measures capture willingness to disclose alongside the thing measured. No. 5 The One That Held Estate agents pricing real houses were moved by numbers they knew were arbitrary, while insisting they were not. Warnings and incentives do not fix it....

Frequently Asked Questions

Is it true that money does not motivate people?
Nothing we located supports that. The field experiments in this article show money changing behaviour substantially, sometimes in the direction the payer did not intend. The defensible claims are narrower and concern the size of an incentive and the kind of behaviour it is attached to.
Do rewards damage intrinsic motivation or not?
It is genuinely contested. Four meta-analyses of overlapping literatures reached opposing conclusions, one calling the effect substantial and another calling it inconsequential. A third-party comment favoured the former, two reviews to one, while warning that meta-analysis itself can mislead on literatures of this kind.
What is the single most useful finding here?
Probably that small payments can produce worse performance than no payment. Students paid roughly three cents per correct answer are reported to have performed substantially worse than those paid nothing. The practical rule is in the paper's title: pay enough, or do not pay at all.
Why did fining late parents make them later?
The proposed explanation is that a fine converts a social obligation into a purchasable service. The finding is a real randomised field experiment, and a later replication attempt using surveys found the effect transitory and found that fines deterred, so treat it as a risk to consider rather than a mechanism to rely on.
Where should we be most careful?
With behaviour that is currently unpaid and socially motivated: helping colleagues, mentoring, flagging problems outside one's remit. That is where the paradoxical results concentrate in the literature, and it is exactly what organisations most often try to encourage with small rewards.
What is missing from this article?
The workplace meta-analysis. We obtained the citation for Jenkins and colleagues (1998) on financial incentives and performance but not its findings, so everything here comes from adjacent settings: education, laboratory tasks, daycare parents and charity collection. None of those is an employment relationship.
IB

About The Insight Bureau Research Desk

The Insight Bureau is GSH Financial's research publication, written for Canadian business owners and the students who will eventually advise them. This article declares that the workplace meta-analysis most relevant to its subject could not be obtained, and states its own position separately from the evidence.

References

  1. Deci, E. L., Koestner, R., & Ryan, R. M. (2001). Extrinsic Rewards and Intrinsic Motivation in Education: Reconsidered Once Again. Review of Educational Research, 71(1), on the finding that extrinsic rewards can undermine intrinsic motivation having been highly controversial since it first appeared in Deci (1971); on Cameron and Pierce (1994) having concluded the undermining effect was minimal and largely inconsequential for educational policy; and on the authors' position that the Cameron and Pierce meta-analysis was seriously flawed and its conclusions incorrect, with their own more recent meta-analysis showing tangible rewards do have a substantial undermining effect. Note: authors summarising their own disputed work; we did not obtain the full text. journals.sagepub.com
  2. Deci, E. L., Koestner, R., & Ryan, R. M. (2001), author-hosted copy, on Cameron and Pierce having advised resisting the implementation of incentive systems in the classroom and having advocated abandoning Deci and Ryan's (1980) cognitive evaluation theory; on the authors' meta-analysis having included 128 experiments, organised to test that theory much as Cameron and Pierce had done; on its showing that tangible rewards do significantly and substantially undermine intrinsic motivation; on there being reason for teachers to exercise great care when using reward-based incentive systems; and on Ryan and Deci (1996), When paradigms clash, Review of Educational Research 66, 33–38. Note: hosted by the authors' own research organisation. selfdeterminationtheory.org
  3. Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, hosted copy, on research reported in Deci (1971, 1972a, 1972b) having revealed that tangible rewards, specifically money, could undermine college students' intrinsic motivation. Note: we did not obtain the full text and report none of its effect sizes. home.ubalt.edu
  4. Comment on Deci, Koestner and Ryan (1999), Understanding the Effects of Extrinsic Rewards on Intrinsic Motivation: Uses and Abuses of Meta-Analysis, on three different meta-analytic reviews of the same literature having appeared and, despite their common focus, having offered dramatically opposed bottom-line conclusions; on the authors' assessment that the findings were more accurately captured by the reviews of Deci and colleagues and of Tang and Hall (1995) than by that of Cameron and Pierce (1994); and on their suggestion that there may be significant short- and long-term costs to the unthinking or automatic use of meta-analysis with theoretically derived, procedurally diverse and empirically complex literatures. Note: a published comment; we obtained the abstract. researchgate.net
  5. Cameron, J., Banko, K. M., & Pierce, W. D. Pervasive Negative Effects of Rewards on Intrinsic Motivation: The Myth Continues, on more than 100 experimental studies having been conducted on the topic over thirty years; on Cameron and Pierce (1994) having concluded that negative effects of reward were limited and could be easily prevented in applied settings; on Deci, Koestner and Ryan (1999) showing pervasive negative effects; and on the present article's purpose being to resolve differences in previous meta-analytic findings and to permit tests of competing theoretical explanations. Note: the opposing side's rejoinder; we obtained the abstract. pubmed.ncbi.nlm.nih.gov
  6. Gneezy, U., & Rustichini, A. (2000). A Fine is a Price. Journal of Legal Studies, 29(1), author-hosted copy, on the four periods of the study being before the fine (weeks 1–4), the first four weeks with the fine (weeks 5–8), the entire period with the fine (weeks 5–16) and the post-fine period; on the first stated fact that the effect of introducing the fine was a significant increase in late collections; and on the authors' statement that behaviour in the third period is still difficult to explain, that four weeks might be thought insufficient for behaviour to return, but that since two or three weeks had been enough to double the level of delays after introduction, the learning rate of the parents is too fast to justify so much persistence. Note: hosted copy of the paper; we obtained portions. ius.uzh.ch
  7. Ariely, D., Gneezy, U., Loewenstein, G., & Mazar, N. (2009). Large Stakes and Big Mistakes. The Review of Economic Studies, 76(2), 451–469, author-hosted copy. Note: we did not obtain this paper's findings and state none; it is cited as the standard reference a reader will encounter on very large incentives. people.duke.edu
  8. International Review of Law and Economics. (2020). Is a fine still a price? Replication as robustness in empirical legal studies, on Gneezy and Rustichini having used a randomised field experiment at Israeli daycares to test whether imposing negative consequences on a behaviour produces a reduction of that response, and having found instead that introducing a fine increased late behaviour; on the present authors seeking to replicate those results using experimental surveys administered on a crowdsourcing platform, described as not an exact replication but allowing control of design aspects difficult to replicate in the field; on their finding that the effects of fines on outcome behaviours and respondents' reasons are transitory, with respondents returning to baseline once fines are removed; on their results being consistent with intuitive judgments and standard rational-choice theory that fines deter; and on their caveat that a survey is not a field experiment and that more research is required to understand when and how any fine-is-a-price effect may arise. Note: a peer-reviewed replication study; we obtained the abstract. sciencedirect.com
  9. Academy of Business Sciences. Mo Money, Mo Problems: When and Why Financial Incentives Backfire, on any addition of money into an interaction, even a small amount, having the potential to change its nature, so that what was viewed as a favour may be transformed into a market interaction when even fractions of a dollar are on the table, attributed to Fiske (1991); on Gneezy and Rustichini (2000a) reporting that Israeli college students given the equivalent of three cents per correct IQ answer performed substantially worse than those offered no incentive; on the daycare fine having produced the opposite of the intended result rather than being merely ineffective; on prosocial behaviours being a subset to which people often react paradoxically because the motives are intrinsic; and on reports that schoolchildren collecting charitable donations collected less when given performance incentives, and that volunteers rewarded with pay reduced their volunteer hours, citing Frey and Götte (1999). Note: a practitioner publication; we did not obtain any of the underlying studies it describes. abacademies.org
  10. Gneezy, U., & Rustichini, A. (2000). Pay Enough or Don't Pay At All. The Quarterly Journal of Economics. Note: cited as listed in the reference list of reference 12; we did not obtain this paper and rely on the description at reference 9. people.duke.edu
  11. Springer Nature. Extrinsic Motivation: Using Reward and Punishment, reference list, identifying Jenkins, G. D. Jr, Mitra, A., Gupta, N., & Shaw, J. D. (1998), Are financial incentives related to performance? A meta-analytic review of empirical research, Journal of Applied Psychology, 83(5), 777; Rynes, S. L., Gerhart, B., & Minette, K. A. (2004), The importance of pay in employee motivation: discrepancies between what people say and what they do, Human Resource Management, 43(4), 381–394; Bonner and Sprinkle (2002); Irlenbusch and Sliwka (2003); and Falk and Kosfeld (2006), The hidden costs of control. Note: a reference list only. We obtained citations, not findings, for every study named here, and state no results from any of them. link.springer.com
  12. Federal Reserve Bank of Boston working paper no. 05-11, Large Stakes and Big Mistakes, reference list and text, giving Gneezy and Rustichini (2000b), A Fine is a Price, Journal of Legal Studies, XXIX, and stating that for many tasks, introducing incentives where there previously were none or raising existing ones is at issue. Note: a working-paper version; its page range for the daycare paper differs from that at reference 11 and we do not resolve it. bostonfed.org

This article reviews research on motivation and is not compensation, employment law or psychological advice. No meta-analysis discussed was obtained in full. The principal workplace meta-analysis on financial incentives, and two further directly relevant papers, are cited by citation only with no findings reported. Several claims rest on a practitioner publication that we identify. One key study has a replication attempt reaching the opposite conclusion, and both are reported. Sources give conflicting page ranges for one paper.