Intrinsic Motivation: a deep dive
Author: Frazer Mawson
Frazer Mawson is a writer from London (Substack: Erring Towards Answers). By drawing together insights from neuroscience, philosophy, and psychology, he is trying to understand what it means to live a Good life.
Soon after I entered into year 12, something crazy happened: I started studying.
Like – a lot.
Between the ages of 12 and 16, I’d had no motivation for school whatsoever. I didn’t do my homework. I didn’t revise for my exams. I diligently ignored everything my teachers said to me. While I was doing my GCSEs, my parents would force me to sit in my room with no distractions and study for two hours each day. Rather than put this time to good use, I would simply sit at my desk, stare out the window, and run down the clock.
But then suddenly, almost overnight, something changed: I started to care about doing well. I started to find my subjects interesting. I started to like – and listen to – my teachers. I developed absolute tunnel vision for my studies, and I eventually managed some of the best A Level grades in my entire school.
From the outside, it looked like I had suddenly decided to knuckle down and start taking my studies seriously – but from the inside, this isn’t what happened at all. There was no knuckling down going on; I wasn’t working hard in any real sense. I would often put in twelve hours of studying each day and at no point did I feel tired or strained or stressed. I felt curious, energised – excited.
This kind of experience – this sudden turning on (and off) – of motivation has happened to me a number of times throughout my life. Sometimes – for reasons that I’ve been trying to understand – I’m able to work happily and without rest for many weeks or months at a time. But at others, it’s as if the supply of motivation has run dry and getting anything done feels next to impossible.
I don’t think I’m alone in this.
In fact, to some extent or another, I think the above description probably applies to everyone. I also don’t think many of us have a good idea as to what’s going on here. This post is my attempt to find out.
This is a topic I’ve been researching in and around for quite a while, and as far as I can see, all roads lead to the idea of Intrinsic Motivation. Playing sport. Writing. Painting. Travelling. Exploring. Hiking. Reading. For the most part, we engage in these activities for their own sake – because they are inherently enjoyable. When we do something for its own sake – without regard for rewards, punishments, or really outcomes of any kind – we can be said to be intrinsically motivated. In contrast, when we’re motivated by external pressures and outcomes, we can can be said to be extrinsically motivated.
There’s good evidence to suggest that the more intrinsically motivated we are to do a task, the more we enjoy it, the better we learn, the better we perform, and the more likely we are to persevere in the face of obstacles and setbacks.
The hypothesis that I’ll be pushing throughout this piece is that whenever I’ve found myself in one of these high-motivation life periods, I’ve unwittingly stumbled upon a rich vein of intrinsic motivation. It therefore stands to reason that if I want to understand these experiences – and how to create more of them – I need to understand intrinsic motivation.
As it turns out, intrinsic motivation looks to be an extremely delicate thing. Under the right conditions, it can be encouraged and drawn out of us; in the wrong conditions, it can be suffocated, stifled – maybe even killed.
Thankfully, there’s been an absolute mountain of research done in this area – mostly carried out under the rubric of Self-Determination Theory – and in this piece, I’m going to do a deep-dive on it all.
Here goes. Main topics to be covered:
What is intrinsic motivation? A quick history + some context
How do psychologists measure intrinsic motivation?
What causes – and blocks – intrinsic motivation? A tour through the research
Discussion - concerns, criticisms, and miscellaneous ideas:
Dopamine: an alternative explanation of intrinsic motivation
Distraction vs. Autonomy: another possible alternative explanation
The largest meta-analysis looking at the effects of rewards and reward contingencies on intrinsic motivation
Mimetic desire: a missing piece of the puzzle maybe?
Western centrism: the most common criticism of SDT
Is the free-choice measure actually valid?
Should competition be outlawed?
Final thoughts + parting advice
1. What is intrinsic motivation? A quick history + some context
The general concept of intrinsic motivation is quite intuitive, so I’m hoping the sketch I’ve offered up so far will be enough to give you the general gist. Saying that, there are a number of possible points of confusion here, so it’s worth spending a little bit of time fleshing out this definition and distinguishing intrinsic motivation from other closely related concepts.
To do this, a whistle-stop tour through the history and wider research context of the concept is going to be useful (and hopefully interesting, too!):
When the research on intrinsic motivation first started, behaviourism was still the dominant school of thought within psychology. Operant Psychology – B. F. Skinner’s development on the original behaviourist position – was primarily concerned with how reinforcement and reinforcement contingencies influence the frequency of different behaviours. The upshot (with many qualifications) was that the more you reward a behaviour, the more frequently that behaviour will occur, e.g. give a mouse some cocaine for pulling a lever and over time the mouse will learn to start pulling that lever like it’s life depends on it. Particularly relevant to our discussion here is the fact that there is no space in this theory for intrinsic motivation. Before receiving reinforcement, behaviour is more or less random (e.g. maybe the mouse moves around randomly and knocks the lever, resulting in cocaine being administered). Once reinforcement has been dished out, motivation is then present to the extent that the behaviour has been reinforced.
That’s pretty much it.
Credit where it’s due: behaviourism and operant psychology were able to shed light on many previously enshadowed phenomena – but over time, a number of experimental findings began to emerge that put this paradigm under increasing pressure.
Here are a couple of them:
Nissen (1930) found that rats would cross an electrified grid in order to get to a novel maze area on the other side. Because neither the grid nor the novel space had been paired with a reinforcer, operant psychology would predict that the electrified grid would act as a negative reinforcer, deterring the exploratory behaviour. This obviously isn’t what happened.
Butler (1957) found that rhesus monkeys would learn discrimination problems solely for the opportunity to visually explore the environment. Again, this exploratory behaviour had not been previously reinforced, so this left behaviourists their heads.
Montgomery (1955) gave rats a chance to either return to their home base or explore a novel environment (that had never been reinforced in any way). The rats showed a strong preference for the latter.
Harlow (1953b) found that rhesus monkeys would solve discrimination tasks with the sole reward of being able to manipulate novel objects. Interestingly, these manipulation drives were extremely difficult to eradicate via the usual processes of extinction (e.g. if you stop rewarding the rat with cocaine when it pulls the lever, it will eventually stop pressing the lever. Not so with the drive to manipulate).
The trend here, as I’m sure you’ve noticed, is that there are certain kinds of behaviours that we – or, at least, animals generally – are reliably motivated to perform but that have never been reinforced. A certain class of these behaviours seems to be tied to exploration, and this class of behaviour is often far more resistant to extinction than the kinds of behaviours that have been learned via reinforcement, e.g. the lever pressing of a coke addicted mouse.
After a decade or so of discussion, nobody was able to find a way of convincingly integrating these findings into the dominant paradigms of the day. This culminated in a 1959 paper by Robert White, where he proposed that these behaviours were best viewed as being fuelled by innate psychological tendencies – tendencies associated with interest, curiosity, exploration, and play.
There’s a whole lot more to this story than is worth covering here, but the crux of this all is that White’s understanding of these behaviours ran in the face of the behaviourist theories of the day and ultimately paved the way for much of the research into intrinsic motivation that has happened since.
As it goes, almost all of the research into intrinsic motivation has been done under the rubric of Self-Determination Theory (SDT). Part of the reason for this is that SDT pretty much started out as THE study of intrinsic motivation (more on this shortly), before then expanding into a fully-fledged theory of human flourishing in its own right.
For good measure, here’s a fleshed out definition of intrinsic motivation from the founders of SDT themselves:
“[Intrinsic Motivation is] the primary and spontaneous propensity of some organisms, especially mammals, to develop through activity—to play, explore, and manipulate things and, in doing so, to expand their competencies and capacities. This natural inclination is an especially significant feature of human nature that affects people’s cognitive and emotional development, quality of performance, and psychological well-being. It is among the most important of the inner resources that evolution has provided (Deci & Ryan, 2000; Ryan & Hawley, 2016), and because it represents a prototypical manifestation of integrative organismic tendencies, SDT research began with it as a primary focus."
So, in other words: intrinsic motivation isn’t really about engaging in purely pleasurable activities. For example, I don’t think that SDT would say that eating delicious cookies would be an example of an intrinsically motivated activity. Rather, SDT’s conception of intrinsic motivation is more about engaging in activities that allow us to explore, expand, and develop our capacities (and that also happen to be deeply enjoyable!).
Up until fairly recently – when I started to make a habit of subsuming myself in different research literatures for many weeks at a time (as you do) – I had never come across SDT before. Given the amount and type of content I consume, I have no idea why this is the case. SDT is a massive area of study. There are thousands of researchers working in this space. Thousands (?) of studies have been conducted. Multiple millions of citations have been accrued.
As far as I can see, SDT has been developed slowly and patiently, with a sound scientific methodology. Its theories have been arrived at inductively, and they have been revised and refined to account for the best available empirical research. Many of its core findings do seem to replicate and show up consistently within large-scale meta-analyses. What’s more, if you put any stake in h-indices – a measure of a researcher’s impact and productivity – the founders of SDT theory are some of the most influential scholars alive. For reference, father of modern linguistics Noam Chomsky has a h-index of 196. Richard Dawkins has a h-index of 83. The founders of SDT, E. L. Deci and Richard Ryan, have h-indices of 180 and 229, respectively.
So, point being: SDT does seem to be a legitimate scientific enterprise, and yet outside of certain academic circles, its ideas don’t receive any air time at all. I don’t know why – and I’ve not been able to find anything online to explain this to my satisfaction. But regardless: The SDT view of motivation, which we’ll tentatively be adopting from here on out, views motivation as a spectrum ranging from most to least autonomous.
Here’s a diagram:
On the high autonomy end, you have intrinsic motivation, i.e. behaviours that are more or less entirely engaged in for their own sake.
Slightly lower down, you have behaviours that are extrinsically motivated but highly autonomous. For example, let’s say that I grew up in a Christian household and that I eventually came to endorse and enact a Christian ethic. By SDT’s lights, behaviours that flow from this ethic would be classed as extrinsically motivated, because they came from an external source – namely, my formative environment – but they would also be classed as highly autonomous, because I engage in them freely and willingly. They are self-endorsed.
A little further along the spectrum, you then get lower autonomy forms of extrinsic motivation. For example, let’s imagine that I grew up in a Christian household but that I did not come to assimilate the Christian ethic that I grew up around. Now, when I visit my family, I might find myself adhering to certain Christian norms, but I would only be doing so to avoid criticism or argument, rather than because I actively endorse these regulations. And then all the way off to the left of the spectrum, you have amotivation, which is where you simply have zero motivation – neither intrinsic or extrinsic – to do something.
One point worth making before we move on is this: SDT assumes that most behaviours are motivated by a mix of these motivation types. As a general rule, the more autonomous forms of motivation are better: they are experienced as being far less effortful, and they are associated with greater persistence and superior performance than controlled forms of motivation.
When I interpret my sudden turn to scholarliness through an SDT lens, here’s my best high-level guess as to what happened: I shifted from the far left side of the spectrum to somewhere much further to the right. I’m a naturally curious person who enjoys learning. It makes sense that studying interesting subjects would tap into my intrinsic motivation. But at the same time, I wasn’t simply studying for curiosity’s sake. I was also studying because I wanted to get good grades, I wanted to outperform my classmates, I wanted to get into a good university, etc.
So intrinsic and extrinsic forms of motivation were probably both in action here. The question I’m trying to answer with this piece is: why did this happen when it did?
This question breaks down into two subquestions
Why did I suddenly become intrinsically motivated to study?
Why did I suddenly shift from low to high-autonomy forms of extrinsic motivation?
As you will have gathered by now, this piece is primarily about intrinsic motivation, so you may think that I’m only addressing the first part of this question.
But as it goes, the conditions that SDT posits as being necessary for intrinsic motivation are also the same conditions that it posits as being necessary for high autonomy forms of extrinsic motivation. So by covering the intrinsic motivation research – which is, I should add, the seed from which all subsequent SDT research eventually sprung – we’ll be able to understand the conditions that are conducive to all forms of autonomous motivation.
I’ll circle back to this point later, but thought it was worth flagging now.
2. How do psychologists measure intrinsic motivation?
I promise: I will get into the meat of this piece shortly (next section), but before I do, I need to share a quick overview of the way intrinsic motivation is operationalised in lab experiments, because it’s interesting and also maybe not beyond criticism. In the 70s, E. L. Deci – one of the pioneers of intrinsic motivation research – developed the free-choice paradigm to investigate the impact of various rewards and reward contingencies on intrinsic motivation. I’ve read through a fairly wide range of papers in this space, and while the free-choice paradigm is often applied slightly differently in different contexts, there are a number of features that show up in pretty much all of them.
First, participants are broken up into at least two groups – the control group and the experimental group(s). In the initial period, both groups are encouraged to engage in some kind of intrinsically interesting task – for example, a puzzle (note: experimenters use a range of methods to make sure that these intrinsically interesting tasks are actually intrinsically interesting for the participants). During this initial period, the control group are generally left to engage with the task uninterrupted, while the experimental groups are interfered with in some way. This is where the independent variable gets deployed. So, for example, maybe the experimental group are told that they’re going to be rewarded for engaging with the task (independent variable = expectation of reward) – or maybe they’re told that they’re going to be evaluated based on task performance (independent variable = expectation of evaluation).
Once the initial period is complete, all participants then enter what is known as the free-choice period. This is where the experimenter leaves the room and participants are left on their own, apparently unobserved, with the task materials. Of note: there will also usually be a number of other activities or distractions intentionally left in the room – so, for example, in a study with preschoolers, a selection of toys were left in the room; in a study with adults, a selection of magazines.
Within this paradigm, then, intrinsic motivation is measured in two main ways:
Free choice measure – the amount of time participant’s spend with the ‘intrinsically motivating’ task in the free-choice period. The idea here is that the more intrinsically motivated the participant is to engaged with the task, the longer they will choose to spend on that task when nobody is watching or evaluating them in any way. This is generally taken to be the primary metric of interest.
Intrinsic Motivation Inventory – basically just a self-report questionnaire to try and get a read on how much the participant finds the task enjoyable and interesting in the free-choice period. This is usually taken to be a sort of supplementary measure.
As it turns out, these two measures are only modestly correlated with one another, which seems to have been a topic of some contention within this field. In the largest meta-analysis ever conducted on rewards and intrinsic motivation (see discussion section), these measures mostly moved in the same direction, but their effect sizes varied quite a lot. As far as I’ve been able to work out, the free-choice measure is generally viewed as the most important of the two because it avoids the well-documented shortcomings of self-report measures.
I have one or two concerns with the way SDT is measured, but i don’t want to get bogged down in the minutiae without first giving you the interesting stuff, so i’ve saved this for the discussion section. Read: is the free-choice measure actually valid?
This is probably all you need to know for now.
So – onto the research!
3. What causes – or blocks – intrinsic motivation? A tour through the research
The earliest experiments on intrinsic motivation were run in 1971 by E. L. Deci. He wanted to understand what would happen to a person’s intrinsic motivation for a task if you offered them a monetary reward for doing it. As mentioned already, the behaviourist assumption at the time was that when you reinforce a behaviour, its frequency increases – and also, importantly, that when you remove said reinforcement, the frequency of the behaviour returns to baseline.
Deci’s experiment was intended to see if these assumptions – particularly the second one – hold up.
In his first free-choice experiment, he asked both groups – the control and the treatment – to work on a Soma Cube puzzle. The treatment group were given a $1 reward for every puzzle they completed, whereas the control group were given nothing – they worked on the puzzle without any expectation of a reward. When all participants were then put in the free-choice condition and researchers measured the amount of time they spent with the puzzles, they found that the treatment group – i.e those that had been given the $1 rewards – actually engaged with the puzzles less.
So, in other words: when rewards were involved, intrinsic motivation for the task, as measured by the free-choice measure, went below the pre-reward baseline.
As I understand things, this was a pretty groundbreaking finding at the time – and behaviourists have been bitching about it ever since. Unfortunately for them – behaviourists, I mean – this was soon replicated with a whole host of different tasks, rewards, reward contingencies, participants, etc., so it does look to be a legitimate thing (although there are still some who hold out against it).
(FYI: the observation that extrinsic rewards tend to harm intrinsic motivation is known as the overjustification effect (also sometimes referred to as the undermining effect))
Before diving in any deeper, an interesting nuance should probably be drawn out here: SDT does not claim that rewards always harm motivation. In fact, the SDT view accepts that at the time rewards are being administered, they will often enhance motivation (with certain qualifications). This is why SDT generally holds that rewards can be an effective means of motivating menial, uninteresting tasks.
The key finding here was that rewards harmed subsequent intrinsic motivation for an intrinsically interesting task. This may seem like a pedantic point, but – if true – it has some pretty far reaching implications.
Take education: should we be rewarding kids for reading?
The prima facie conclusion of this research is that, no, we probably shouldn’t. Rewards may be an effective means of motivating reading at the time they are present, but when rewards are removed and kids are left to their own devices, they – the rewards – may actually have a deleterious impact on subsequent intrinsic motivation for reading.
Saying that: there are many qualifications and quirks to this picture, so let’s keep following the research.
(FYI: there have been a huge number of experiments run in this space, so I’m mostly just going to stick to research that’s in line with the general tenor. You’re getting the highlight reel, but rest assured: many replications and meta-analyses have been done here (some of which we’ll cover).)
In a bid to begin exploring the impact of different reward contingencies on intrinsic motivation, Deci (1972a) ran a followup study – once again with the same ‘interesting puzzles’ – but this time he paid the treatment group for simply showing up, rather than for each puzzle completed. The control condition was the same as in the original experiment. Here, it turned out that participants who received a reward for showing up exhibited just as much intrinsic motivation as those that received no reward at all. So, in other words: within this contingency, rewards did not harm intrinsic motivation at all.
Lepper, Green, and Nisbett (1973) ran an experiment looking at the impact of different reward contingencies on preschoolers’ intrinsic motivation for a drawing task using ‘attractive materials.’ The study included three groups:
Treatment group 1 were told that they would receive a ‘good player’ reward (a star and ribbon) if they completed the drawing task. All children in this group received a good player reward.
Treatment group 2 were not told about any rewards beforehand but were then all given a ‘good player’ reward after the drawing task.
The Control group were not told anything and did not receive a ‘good player’ reward.
In line with the previous Deci experiment, treatment group 1 showed less intrinsic motivation in the free choice period than the control group – so, once again, rewards undermined intrinsic motivation. But here’s the interesting thing: treatment group 2 showed just as much (if not more) intrinsic motivation in the free choice period as the control group.
So, in other words: if the reward was dangled beforehand, it harmed intrinsic motivation; if it was unexpected, it didn’t.
The plot thickens.
This finding was built upon by Ross (1975), who compared the impact of salient vs. non-salient rewards on children’s intrinsic motivation to play with a drum kit. One group of children was simply told that they would receive an unknown prize if they engaged with the drum (non-salient reward). The other group were given the same instructions, but they were told that the unknown prize was hidden inside a box that had been placed conspicuously in front of them (salient reward).
Turns out that the group with the salient reward showed significantly less intrinsic motivation in the free choice period than the group with the non-salient reward. Ross concluded that the salience of the reward seems to modulate its impact: the more salient, the stronger the deleterious effect on intrinsic motivation.
I’ll share one more experiment – because it adds texture to this picture, plus it touches on the topic of reading motivation – and then I’ll give a rundown of how SDT has accounted for all of these findings.
Marinek and Cambrell (2008) were interested in third-grade children’s reading motivation. They compared a no-reward condition with a token-reward condition – kids received something like a star or a sticker for reading – and found that the token rewards significantly reduced intrinsic motivation for reading.
Based on the other research above, this isn’t a major surprise – but here’s where it gets interesting: They also ran a third condition where they rewarded children with a book rather than a token. This time round, the reward had no negative effect on intrinsic motivation at all. The explanation put forward in the paper was that rewards that are more proximal to the desired behaviour – e.g. books for reading or, say, paint brushes for painting – do not have the same negative impact on intrinsic motivation as token rewards like money or stickers.
OK , so, so far we’ve got:
Rewards reduce intrinsic motivation
Rewards only reduce intrinsic motivation when they’re contingent on task completion but not when they’re contingent on simply showing up.
Salient rewards harm intrinsic motivation, but non-salient ones don’t (or, at least, not as much)
Token rewards harm intrinsic motivation, but rewards that relate closely to the task at hand don’t
How do we reconcile these findings? SDT’s answer, in a word: autonomy.
Rewards and reward contingencies that preserve or enhance an individual’s sense of autonomy have a positive impact on intrinsic motivation; those that diminish an individual’s sense of autonomy and that are experienced as an attempt to control their behaviour have a negative impact on intrinsic motivation.
Interpreting these findings through this lens, then:
In the first Deci experiment, where individuals were rewarded with $1 for each puzzle completed, it’s argued that the monetary reward was experienced by participants (whether consciously or not) as a means of controlling their behaviour. This resulted in a reduction in autonomy, which subsequently led to a decline in intrinsic motivation.
In the second Deci study, where individuals were given a monetary reward for showing up but not for completing the puzzles, it’s argued that the monetary reward was not experienced by participants as a means of controlling their behaviour – since they’d get the reward regardless of whether or not they completed any puzzles – which is why it didn’t impact intrinsic motivation.
In the Lepper, Green, and Nisbett study, where individuals in treatment 1 were told they would receive a ‘good player’ reward if they performed well, it’s argued that the expectation of evaluation and the dangling of a reward was interpreted as a means of controlling their behaviour. This likewise resulted in a reduction in autonomy and intrinsic motivation.
But in the same study, for treatment 2, where individuals were not forewarned about the ‘good player’ reward – but where they received the reward nonetheless – it’s argued that the activity was engaged in autonomously – without any external control – and that intrinsic motivation was therefore unaffected.
So, with the autonomy element of the theory in place – there are a few more key components that still need to be introduced – we can then begin taking a look at experiments that actively seek to manipulate this variable.
Here are two interesting examples, plus a meta-analysis for good measure:
Zuckerman, Porac, Lathin, Smith, and Deci (1978) used the standard free-choice paradigm, but they allowed one group of participants to choose which puzzles they engaged with – plus how they allotted their time between them – while the other group were yoked to the choices of a randomly selected counterpart from the first group. Turns out that participants from the first group, i.e. those who got to choose, showed significantly more intrinsic motivation in the free choice period. I’ve come across a handful of other studies that show the same thing: introducing choice often seems to improve performance.
This one isn’t strictly about the relationship between choice and intrinsic motivation, but I thought I’d throw it in anyway because it’s closely related and also interesting: Murayama et al (2015) had participants play a game where they were challenged to pause a stopwatch within 50 milliseconds of the 5-second mark. One group was allowed to choose which stopwatch they used (there were a number of different options); the other group were assigned a stopwatch at random. While this was going on, all participants were put in an fMRI to see what was going on in their brains. Turns out that performance was better in the choice condition. That participants in the choice condition were more resilient to negative feedback. That participants in the choice condition showed no drop in ventromedial PFC activity following failure, but there was a drop in vmPFC activity for participants in the no-choice condition.
Patall, Cooper, and Robinson (2008) ran a meta-analysis with 41 studies to examine the effects of choice on intrinsic motivation and related outcomes. They found robust effects showing that choice enhances intrinsic motivation in both adults and children.
So, there you have it:
As predicted, when we give people more choice – when you increase their sense of autonomy – it boosts their intrinsic motivation (and performance too).
I have a few qualms with the whole autonomy angle – to be discussed in the discussion section (where else!) – but when I try to match this up against my own experiences with intrinsic motivation, it fits pretty well.
Take, once again, the monomaniacal study spree discussed in the intro to this piece.
During my GCSEs, i.e. before I started working like a madman, my parents and teachers exerted constant pressure on me to study. As a result, my studies never felt self-motivated – autonomy was in the gutter – and I had zero intrinsic motivation. But when I moved into year 12 – which is when the monomania began – there was a big and definite uptick in the amount of autonomy I was being offered. To give a few examples:
Uniform – we were no longer assigned a uniform. There was a sixth form dress code, but within this dress code there was a fair amount of scope to choose what we wore.
Exams – at the start of sixth form, there was an extended respite from the pressure that came with our GCSEs. This armistice might have given me enough space to begin generating my own motivation for school.
Ambient climate – teachers started to treat us like adults that could be listened to and respected, rather than drones that needed to be coerced into compliance. We were allowed to leave the school grounds at lunch. We had free periods, in which we could do more or less whatever we wanted. Some of us even started getting cars.
So, yeah, in sum: my explosion in motivation does seem to coincide with a sharp uptick in autonomy. This still doesn’t explain everything though: for example, why did I specifically become monomaniacal about school? Why not something else – like, say, video games or football or puzzles? By introducing further elements in the SDT theory, I think we can start to explain some of this.
—
Assuming SDT is true, so far we’ve established that when rewards are experienced as controlling , i.e. when they infringe upon an individual’s sense of autonomy, they have a negative impact on intrinsic motivation. But what about verbal rewards? Do they count? This is important, since a massive percentage of the rewards we receive and dispense in life are of this kind.
The previous operant paradigm would have us believe that verbal rewards are likely to reinforce a behaviour and thereby increase its frequency. So, for example, if a kid does something good – like e.g. sharing their toys with their siblings – we ought to praise them. But maybe, as per the research above, this isn’t the case?
In much of the early research on verbal rewards and intrinsic motivation, participants were placed in the standard free-choice paradigm and given positive verbal feedback for working on an activity. For example, if they completed the activity, they were told ‘you did very well in completing the task; many people did not.’ If they failed to complete the activity, they were told ‘this was a very difficult task but you were progressing well with it.’ Turns out that participants who were given positive feedback generally displayed MORE free-choice persistence, i.e. intrinsic motivation, than those weren’t given any feedback at all.
We’ve just learned that token rewards have a negative impact on intrinsic motivation; now we’re finding out that verbal rewards apparently do the opposite. Again, given the research discussed so far, this isn’t what we’d expect at all. Does SDT have a means of reconciling these findings? You betcha – but it needed to introduce another component to its theory in order to do so: competence.
All things being equal, if a reward signals that its recipient is competent with respect to a given activity or task, it will have a positive impact on intrinsic motivation for that task/activity. So, for example, let’s say I’m learning to paint. I spend an hour brushing away at my canvas, and at the end of the session, my instructor says something to the effect of ‘you have a wonderful eye for colour’ (or whatever a painting instructor might say to a pupil!). This feedback serves as a signal that I am competent and therefore bolsters my intrinsic motivation for painting (which, I guess, is kind of what our folk psychology would predict?). But here’s the thing: the vast majority of rewards and reward contingencies affect competence and autonomy at the same time. Take this experiment:
Smith (1975) assigned three groups to a learning task about Art History. One group was told beforehand that they were going to receive a written evaluation; the other two weren’t. Of those who weren’t, one group received an unanticipated evaluation after the learning task – one didn’t. Unbeknownst to all participants, everybody that received an evaluation received a positive one. As per usual, they were all then observed in a free-choice period to see how much they engaged with the original task when left to their own devices.
The main finding here was that those in the first group – i.e. those who were told beforehand that they would be evaluated and who then received a positive evaluation – displayed significantly less intrinsic motivation than either of the other groups.
SDT would explain this result by saying that the anticipated evaluation negatively affected autonomy while positively affecting competence. This then netted out at a slight reduction in intrinsic motivation. In contrast, the unanticipated evaluation had no negative effect on autonomy, but a positive effect on competence – which netted out as an increase in intrinsic motivation.
This is generally how SDT understands most rewards and rewards contingencies: rewards that signal competence are conducive to intrinsic motivation, but unfortunately, many rewards also inadvertently impair autonomy – which is why they so often seem to harm intrinsic motivation.
Returning once again to the example of my monomaniacal study spree: this all tracks really nicely. Why did I become obsessive about studying rather than other things? Because I was good at it.
Once my teachers and parents stepped back, my sense of autonomy increased. This freed up some autonomous motivation for studying, and as soon as I started to put in the work, my results improved radically. Improved results were a strong competence signal, which freed up even more motivation. More motivation = even better results = even more motivation = even better results, etc.
But there was also another dynamic at play: the better my grades became, the more relaxed my teachers and parents became. This meant they gave me a lot more breathing space to do my own thing – read: more autonomy. Even more autonomy = even more intrinsic motivation = even more autonomy, etc.
As far as I can see, I was essentially locked inside multiple feedback loops.
—
If you’ve ever done any reading in and around SDT, you’ll know that the theory identifies three psychological needs – not two. Autonomy and competence are the first and the second. Relatedness is the third.
Relatedness is all about how connected we feel to those around us. Turns out that relatedness is another extremely important component in SDT’s model of intrinsic motivation. This particular element of the theory was stumbled upon accidentally, and the story’s pretty interesting, so I thought I’d tell it:
Anderson, Manoogian, and Reznick (1976) were examining the effects of rewards and feedback on young children’s intrinsic motivation. The usual fare. In a bid to create a condition with neither positive nor negative feedback, they instructed researchers within this condition to remain silent and unresponsive to any overtures from the children. As a no-reward, no-feedback condition, we’d expect this to have a fairly neutral impact on intrinsic motivation – but this isn’t what happened at all:
Children in this group showed less intrinsic motivation than any group in the study – including those who received negative feedback! It was theorised that these children felt rejected by the researchers, which thwarted their need for relatedness and thereby destroyed any intrinsic motivation they may have had for the task.
Plenty of later research supports this idea, e.g.
Ryan, Stiller, and Lynch (1994) surveyed a large group of junior high students. They found that students who felt security with teachers (and other figures in their lives) were more engaged with school work and exhibited higher levels of intrinsic motivation.
Bao and Lam (2008) looked at the importance of autonomy and relatedness for intrinsic motivation to study in Chinese students. They found that kids who make their own choices, rather than having them made for them show higher levels of intrinsic motivation, but that this was mediated by how close these children felt to their parents. If they were close, then choice was less important and they showed just as much intrinsic motivation as their higher autonomy peers.
This theory also lines up very neatly with Bowlby’s attachment theory: babies that are more securely attached to their primary caregivers demonstrate greater curiosity and tendency to explore.
So, there you go: Relatedness = super important. But one question you might still have is: how important?
After all, many activities that seem to be fuelled by intrinsic motivation – things like reading, hiking, etc. – are solitary. Solitary activities, by definition, do not involve relating to others, so how can relatedness be involved here?
SDT’s stance, so far as I can work it out, is that relatedness is often a background requirement for intrinsic motivation. Relatedness doesn’t necessarily have to be tied to each and every intrinsically motivated activity – although it sometimes helps. If your life generally meets your requirement for relatedness, then you will have readier access to intrinsic motivation across all activities.
So, bringing this back one last time to my own experience at school: I personally feel like autonomy and competence are more operative than relatedness in my case – but I would also say this:
During sixth form, almost everyone in my former friendship group left the school. This former friendship group, for reference, wasn’t great. If it was being scored on its ability to satisfy my need for relatedness, it would probably get a 3 out of 10. As a result of this disbanding, I was then forced to join a new friendship group, and I would say that those new relationships did a much better job of meeting my need for relatedness. Maybe 7 or 8 out of 10.
There’s also just the general point to make, which is that when I was studying hard, I was in harmony with my parents and my teachers. My aspirations for my life lined up closely with theirs’, and this improved the quality and closeness of those relationships. It seems, then, that competence may also have played an important role in my studies too – and that the feedback loop may have more layers to it than previously assumed.
—
For the record: all of the research and theory discussed in this section is actually taken solely from one of the six mini-theories that make up SDT. This particular mini-theory is called Cognitive Evaluation Theory (CET), and the other five are mostly just an extension of these core ideas.
So: according to CET, intrinsic motivation is determined by the interplay of three key factors: autonomy, competence, and relatedness.
There are some finer points and concerns that we’ll dig into in the next section, but this is the big picture theory – and actually, all things considered, I think it does quite an impressive job of mapping onto some of my own experiences with intrinsic motivation.
4. Discussion: concerns, criticisms, and miscellaneous ideas
As will hopefully be clear at this point, I’m fairly convinced that SDT is getting at something real – but there are a few parts of the theory that I’m not quite bought into – or that I think might need refining.
I’m no expert, so take these speculations with a hearty pinch of salt. It may well be the case that people have tackled and refuted these ideas and objections already. There are also a couple of other things that I didn’t touch on during previous sections but that I think are either important or interesting, so I’ve thrown them into this section too.
1. An alternative theory: dopamine’s impact on intrinsic motivation
As some of you will know, a while ago I did a bit of a deep dive on dopamine, the neurochemical linked to motivation, learning, and reward. I’m not a world authority on this subject by any means, so treat my speculation here with the level of skepticism it deserves. Nonetheless, I think the dopamine angle offers a slightly different – and, I think, quite compelling – way of interpreting some of the experimental findings from this literature.
First, a quick tl;dr on the dopamine piece:
When we receive a reward for completing a task, we experience a dopamine spike in certain ‘reward centres’ of the brain. Contrary to the lay view, this spike does not seem to be causing pleasure: it seems to be causing learning. In other words, dopamine seems to stamp in the relationship between stimulus, behaviour, and reward.
But dopamine is about more than just learning; it’s also about motivation. The next time we encounter the stimuli associated with a reward, dopamine spikes again, this time invigorating the behaviour that previously led to the reward.
So, to make this a bit more concrete: let’s say I walk past a sweet shop and I decide to go in and buy a bag of sherbet lemons. When I eat the sherbet lemons, there’s a dopamine spike in my brain, and this stamps in a relationship between the stimuli – the sight of the sweet shop – the behaviour – the act of going in there, buying the sweets, and eating them – and the reward – the enjoyment that came from eating the sherbet lemons.
Now imagine that a week later I’m walking past this sweetshop. My brain has come to associate these specific stimuli – the sight of the sweetshop – with reward, so it produces a dopamine spike. This dopamine spike then invigorates the behaviour that generated the reward last time – i.e. going into the shop and buying the sweets – and this ultimately results in the reward being obtained again.
But here’s the thing: let’s say that I encounter the stimuli, I perform the behaviour, but the reward is not forthcoming. Like, say: I see the sweetshop, I walk into it, but when I go to the sherbet lemon section, it turns out that there are none left.
What happens? Instead of spiking or staying constant, dopamine goes below baseline.
As I understand things, when this occurs, it is in some sense reducing the strength of the relationship between the stimuli, the behaviour, and the reward. This means that the next time I see the sweet shop, there will be a smaller dopamine spike than before, which will reduce the strength of the motivation I feel to go inside and buy the sweets.
I think we can use this understanding of dopamine to explain many of the findings from the intrinsic motivation literature without recourse to the core SDT concepts.
Consider once again the original study run by E. L. Deci (1971). I didn’t go into all of the details originally, because it wasn’t really needed, but it is now, so: Deci actually had three periods in his experiment. In the first period, all participants did the same thing. From the original paper:
‘After they entered the experimental room, they were seated at a table with the puzzle pieces in front of them, three drawings of configurations to the right of them, the latest issues of New Yorker, Time, and Playboy to their left, and the experimenter on the opposite side of the table. They were told that they would spend all three sessions using the pieces of plastic to form various configurations, such as the same they were shown.’
All participants were then asked to reproduce, using the puzzle pieces, the configurations shown in the drawings. In the second session, those in the experimental group were told that they would receive a $1 reward for every configuration completed within 13 minutes. Meanwhile, the control group did the exact same thing as in the first session. And then in the third session, both groups returned to the same conditions as in the first period. So for the control it was: no reward, no reward, no reward. For the experimental group is was: no reward, reward, no reward.
Now, in order to measure intrinsic motivation, Deci placed an eight minute free-choice period in the middle of each of these sessions. During this time, the experimenter left the room, telling participants ‘I shall be gone only a few minutes, you may do whatever you want while I am gone.’ They could work on the puzzles, read the magazines, stare around the room, etc.
Meanwhile, all participants were monitored to see how long they engaged with the puzzles during each of these free choice periods.
So, with all of that additional context, here are the results:
The interesting thing here – and the reason I think the dopamine explanation is plausible – is this: in the experimental group, free-choice engagement with the task goes way up in the second session, after rewards have first been introduced.
This is what the dopamine account would predict: an association between the puzzles and a monetary reward has been established. As a result, when the participants encounter the puzzle next time, dopamine levels rise to invigorate the behaviour that secured the reward last time, i.e. playing with the puzzles.
But in the following session, when the monetary rewards are removed, dopamine levels go below baseline, weakening the association between the behaviour (doing the puzzles) and the reward (receiving the money). As a result, the individual feels even less motivated to engage in the task than before, which results in free-choice engagement going way down. At least for this particular experiment, I find this explanation quite compelling. If anyone knows a reason as to why it’s wrong, I’m all ears!
2. Distraction vs. Autonomy: an alternative explanation
I think there’s good evidence that higher autonomy forms of motivation are superior – in terms of subjective experience, performance, persistence, etc. – than lower forms of autonomy. I’m completely on board with this
But when I look at some of the early experiments surrounding intrinsic motivation, I feel like autonomy as an explanation is a bit of a stretch. Take the Ross (1975) one, where salient rewards were compared to non-salient rewards.
This result is completely plausible to me: it makes sense that if you promise a kid an unknown reward for playing a drum, and then you hide the unknown reward in a box placed conspicuously in front of them, they will experience less enjoyment when they play the drum. And if they experience less enjoyment, it stands to reason that they will probably be less motivated to play with the drum the next time the opportunity presents itself, i.e. a reduction in intrinsic motivation.
But here’s the thing: SDT posits that the reason for this reduction in intrinsic motivation is that the highly salient reward is experienced as controlling – as impairing the individual’s sense of autonomy – and that this is ultimately what causes the downturn in intrinsic motivation.
But would these salient rewards really be experienced as a form of control? I’m not convinced. For me, a better explanation is simply that the reward functioned as a distraction, and that the distraction reduced the quality of the attention the kids brought to bear on the task.
Think of it like this: tasks that are intrinsically interesting are in some sense a reward unto themselves. But if you offer up highly salient extrinsic rewards for doing these tasks, then each participant’s attention is going to be divided between the task and the reward. This divided attention will impair the quality of the participant’s engagement with the task, which will in turn reduce the amount of enjoyment they’re able to extract from it.
Less enjoyment from the task = reduced time playing with the reward in the free-choice period. This seems to me like a much more plausible explanation than the autonomy one.
For another example, consider the Marinek and Cambrell (2008) experiment where they compared the effect of token-rewards, e.g. a gold star, with task-related rewards, e.g. a book, on reading motivation.
Would the reward of a book really be experienced as less controlling than the reward of a gold star? Again, I struggle to believe this. Instead, I think the token reward was probably more distracting than the task-related reward – which makes sense, since the task-related reward was really just a means of spending more time doing the task at hand anyway.
3. The largest meta-analysis on the effects of rewards and reward contingencies on intrinsic motivation
I was going to throw this into the previous section, but it all just got too cumbersome, so decided to save it for now.
Over the years, there have been mountains of experiments looking at the effects of different rewards and reward contingencies on intrinsic motivation. In 1999, Deci et al ran a huge meta-analysis on all of this research. Here are the headline results:
(Note: k = number of individual effect sizes included in the composite effect size ; d = composite effect size. Also, the numbers in brackets at the bottom of each box are the 95% confidence interval for each composite.)
Some of these categories are self-explanatory – e.g. tangible vs. verbal rewards, expected vs. unexpected rewards – but others are not. To help you interpret all of this, here’s a quick run-through of the categories I didn’t think were obvious:
Task-noncontingent rewards - rewards are given for simply being present; does not require engagement with the target activity.
Engagement-continent rewards - rewards given for spending time engaged with the target activity.
Completion-contingent rewards - rewards given for completing a target activity.
Performance-contingent - given for reaching a specific performance standard, e.g. doing better than 80% of people.
Upshot: verbal rewards seem to increase intrinsic motivation; almost all tangible rewards seem to undermine it, other than unexpected tangible rewards, which don’t have much of an effect either way.
So: SDT - 1; Behaviourism - Nil.
Here are a couple of bits in relation to all of this that I thought were interesting and worth drawing attention to:
Children vs. Students – notice that for verbal praise, the effect size is way bigger for college students than for children. One quite convincing explanation for this is that that much of the praise children receive from adults in their daily lives actually is intended to control their behaviour. They are therefore much more sensitive to the controlling aspect of feedback than college students, who interpret feedback as more straightforwardly competence affirming.
Praise – in addition to the findings from this meta-analysis, I thought it was worth adding that Henderlong and Lepper (2002) did a review of the praise research and found that when praise was viewed as informational, i.e. conveying genuine information about competence, it increased intrinsic motivation, whereas when praise was viewed as controlling, it harmed it.
Performance contingent rewards appear to have a negative impact on intrinsic motivation, but here’s the thing: a decent percentage of the experiments included in this category involved all participants receiving the reward. This means we can expect the competence affirming aspect of the reward to at least partially offset the controlling aspect. But in many real-world contexts, we would often only expect the upper X% of participants to meet the criteria for the reward. That means that the majority of people within this contingency would experience both the controlling AND the competence disconfirming aspects of the reward, which we would expect to absolutely hammer their intrinsic motivation for the activity. This whole point is interesting, because it suggests that performance contingent rewards might act as something like a gate-keeping function: anyone that performs well will receive enough competence affirming information to keep their intrinsic motivation for the activity alive. Meanwhile, everybody else’s intrinsic motivation will be completely crushed. More on this in the ‘Competition’ section below.
Delay time – in some of the experiments that were included in this meta-analysis, the free-choice measure was taken immediately after the reward period; in others, it was taken several days later. By comparing the two, we can get a read on whether the effects of rewards on intrinsic motivation are short-lived or whether they persist long after the rewards were administered. Turns out that the effect size for immediate and non-immediate rewards was more or less identical.
4. Mimetic desire: a missing piece of the puzzle maybe?
When I first started writing this piece, I was planning to criticise SDT on the grounds that it doesn’t really explain why I suddenly became obsessed with school and studying rather than other things.
But on reflection, I actually do think it offers up a fairly plausible explanation.
One factor, which I’ve mentioned already, is competence: I have a knack for school and learning, and this meant that when I started working hard, I received a strong competence signal. But why did I suddenly care obsessively about going to a good university when this hadn’t been on my radar at any point prior to then?
My assumption has always been that this was a simple case of mimetic desire: all of my classmates suddenly conceived a desire to go to a good university; my desire was in some sense learned from – or an imitation of – the desire I was seeing in them.
I think this is partially right – but one piece of the picture is missing: Up until this point, I’d been relatively resistant to this kind of mimesis. Prior to sixth form, many of my school friends wanted nothing more than to hang out in town, chase girls, and do the usual mindless teenager stuff – but I wasn’t really interested in any of it.
But then I enter year 12 and suddenly I become subject to the kinds of mimetic desire that I had been holding out against all this time. What’s up with that? As it turns out, one of SDT’s other mini-theories – Organismic Integrations Theory (OIT) – has something quite interesting to say here.
In essence, OIT says that I am much more likely to internalise the values and norms of those around me when two conditions are met:
When my adoption of those values and norms is volitional (read: autonomous).
When I feel connected with those around me (read: relatedness).
As mentioned already, when I was in sixth form, both of these conditions were met: I had a close group of friends, and my environment and relationships were much more supportive of autonomy than they had been at earlier stages of my schooling.
SDT’s explanation is that this uptick in relatedness and autonomy meant that I was much more likely to internalise the values of my surroundings – those presented by my parents, teachers, and peers – and this led to far more autonomous forms of motivation becoming active.
5. Western centrism: the most common criticism of SDT
When I was trying to understand why SDT was so rarely mentioned in mainstream circles, I did a bit of digging around the most common criticisms of the theory. Of those that I encountered, most weren’t particularly compelling, but one did make think: the theory – which claims to offer a universal account of human psychological needs – is excessively western-centric.
The autonomy component of the theory seems to be the main target of this criticism: in western culture, we place a strong emphasis on autonomy as a value, but in collectivist cultures, where the needs of the individual are generally subordinated to the needs of the group, this emphasis is not there.
Does this mean that SDT is only applicable in the west?
Proponents of SDT generally have a couple of responses to this criticism:
Autonomy vs. independence – the western centricism criticism conflates the ideas of autonomy and independence. Autonomy is about acting volitionally and endorsing one’s own behaviours; independence is about being reliant upon oneself and separate from others. These are two different things. SDT is about the former – the criticism relates to the latter. It is possible to act in a highly autonomous way – that is, to endorse one’s own behaviours willingly – and to also conform to the needs of the collective. This idea is brought out in clear relief when SDT talks about morals: if I reluctantly adhere to my community’s moral rules, dragging my feet all the way, I can be said to have a low autonomy form of motivation. If I see the value in these moral laws and willingly adhere to them, I can be said to possess a high autonomy form of motivation.
Universality of relatedness – SDT’s proponents also respond to this criticism by pointing out that relatedness – a value that would usually be more closely associated with collectivist cultures – plays a central role in their theory too. To zero in on autonomy and ignore relatedness feels a bit like cherry picking. What’s more, there’s no reason why individualist cultures should be more supportive of autonomy and collectivist culture more supportive of relatedness. There may well be cultures where the reverse is the case.
Don’t know about you, but I actually find both of these responses quite satisfactory.
6. Is the free-choice measure actually valid?
There are two niggling concerns that I’ve had while writing this piece. I wouldn’t feel right putting it out into the wild without at least flagging them, so here they are: First, does the free-choice measure actually measure intrinsic motivation? Does it, in other words, achieve construct validity? The measure basically relies on inferences to differentiate intrinsic and extrinsic motivation. In the free-choice period, we are inferring that the individual’s behaviour is fuelled by intrinsic motivation, but it seems perfectly conceivable that in some instances it may be fuelled by extrinsic factors like, e.g. expectation of future rewards.
This is a legitimate concern, and i’m far from the only one to raise it. SDT’s proponents generally respond in two main ways:
They emphasise the correlation between the free-choice measure and the self-report measure. As mentioned already, this correlation isn’t perfect, but it’s close enough to indicate that the free-choice measure probably is capturing some element of enjoyment and interest. From my perspective, i don’t think this really deals with the criticism, since the main objection here isn’t that the measure fails to capture intrinsic motivation – it’s that it captures both intrinsic and extrinsic motivation.
They also sometimes emphasise the importance of using multiple measures of intrinsic motivation to triangulate on this construct. The idea, as i understand it, is that taken alone none of these measures are perfect, but by raising the threshold of evidence – by saying that multiple measures need to point in the same direction for the result to be considered significant – we can more reliably zero in on intrinsic motivation. I find this response more convincing, though i’m not sure if anyone is actually putting it into effect.
My second concern relates to the free-choice paradigm’s external validity: are its findings generalisable to the real-world? These experiments are very controlled and sanitised, but the rewards and reward contingencies of real life are messy and complex and multi-layered. So – like – do the findings from these simplistic experiments carry over?
I’ve been able to satisfy myself with two main replies to this one:
This is just the nature of experimentation (or, at least, experimentation within the field of psychology). We need to simplify and sanitise in order to isolate our variables of interest. If we did not do so, it would be super difficult to detect a signal within all the noise. Also, as it goes: when SDT uses field experiments to test their findings from the lab, the results generally run in the same direction.
The findings and theory from this literature do a surprisingly good job of explaining my own real-life experiences with intrinsic motivaiton, e.g. my crazy study habits during A Levels. This obviously isn’t a particularly scientific way of evaluating the external validity of a paradigm, but for me at least, it does tilt me slightly in the direction of wanting to give the measure the benefit of the doubt.
7. Competition: should it be outlawed?
There’s been some tentative talk within this literature about the potential undermining effects of competition on intrinsic motivation.
Consider again, for example, the meta-analysis discussed above, where it was found that performance-contingent rewards are likely to undermine intrinsic motivation, even in those who actually fulfil the performance criteria. For those who don’t fulfil the criteria – the losers of the competition – this is expected to have an even greater undermining effect.
This line of reasoning has led some commentators to suggest that we ought to consider removing competition from certain domains, e.g. maybe schooling, exercise, work, etc. On one level, this makes sense – but I think we need to be careful here.
In some areas, it’s important that we ensure everybody succeeds to some extent or another. Take, for example, reading: we don’t want a handful of reading rockstars, while everybody else comes out of school illiterate. We want every student to leave school with a high level of literacy.
But if we introduce competition here, we may run the risk of harming the intrinsic motivation of the worse performing kids. This will mean that they fall even further behind – and that they leave school with a significantly lower literacy level than they would have done otherwise, i.e. without competition.
Meanwhile, as per the meta-analysis, we might find that the competition only has a neutral – if not slight undermining – effect on the intrinsic motivation of the high-performing kids.
This is a clear lose-lose.
But i think a further point needs to be made: in certain instances, competition inheres within the activity itself. That means the individual is never going to end up in a 'free choice' situation in real life, because the reward contingency is baked into the activity itself.
Take tennis, for example. Tennis is competition. The competition inheres within the activity. Sure, you can rally for a while, but if you’re not playing points – if you’re not competing – you’re not really playing tennis at all: you’re just practicing.
In some sense, this is just semantics – what does the word ‘play’ actually mean? Or the word ‘tennis’? – but in another sense, it’s not: you’re never going to enter a free-choice period where rewards are off the table entirely; performance-contingent rewards, however low stakes, are embedded within the very fabric of the game.
This is a good thing, and it’s probably part of what makes sport – and games in general – so compelling.
When I play competitive tennis, I become completely absorbed in the game in front of me. As far as I’m concerned, this kind of absorption is one of life’s great joys – being wholly immersed in the present moment, in the challenge at hand, and temporarily forgetting about every other worry in my life.
Without competition, this kind of narrowing of attention just would not happen. The enjoyment of the activity – it’s ability to elicit motivation and narrow attention – is inextricably tied up with the competition itself. As a result, I think that any blanket policy seeking to remove all forms of competition in order to militate against the potential alienating effects on poor performers would be a mistake.
Anyhow, there are lots of subtleties to this debate, and I’m not the guy to cut through all of the confusion. Hopefully this discussion at least gives a sense as to what’s at stake on both sides.
5. Final thoughts + parting advice
Throughout this piece, I’ve done my best to give a clear-eyed overview of the research on intrinsic motivation. I think this is a really important subject, not just because greater intrinsic motivation means greater output, but also because I think that greater intrinsic motivation, all things being equal, means a better life.
I don’t need any research to tell me this – it’s kind of self-evident – but as you’d expect, there actually are a load of papers showing a strong relationship between wellbeing and intrinsic motivation, e.g. 1, 2, 3, 4, etc.
This makes sense. We don’t want to be white knuckling everything we do. When we’re excited, inspired, energised, motivated – when the wind is in our sails – life is fun and wellbeing is abundant. Unfortunately, if the research literature is to be believed, there are many things that we’re currently doing – both to ourselves and to others – that undermine intrinsic motivation badly.
What I’ve presented here is really only the tip of the iceberg – the research literature is vast – but I’m hoping this sketch will be enough to give you a broad-strokes understanding of the main findings, plus the theories that have emerged to account for them.
Maybe I’ll do a followup piece or two to try and flesh out this picture some more. In the meantime, I’ll leave leave you with a parting piece of advice: If you want your kids to study – if you want them to actually enjoy learning – then you’re going to have to do something and you’re probably not going to like it:
You need to get off their backs!
Great piece. The relationship between autonomy and intrinsic motivation is often notable in sports. Agassi hated his life because he had a drill sergeant father and no amount of extrinsic rewards could compensate for the perceived lack of freedom.
I am curious about how we form goals. This is still a mystery to me. In my case I had no relatable peers to justify choosing the specific goals I chose in my life. Yet somehow I believed there was a clear relationship between effort and desired outcome and that gave me a significant amount of intrinsic motivation.
Fascinating essay.
Are you familiar with Randall Collins's theory of Emotional Energy and Interaction Ritual Chains
(https://en.wikipedia.org/wiki/Randall_Collins)?
His theories compliment what you've written.