The Effects of External Rewards on Intrinsic Motivation
Researched and Composed by Gabriel “Venom” Wilson, BSc. (Hons), CSCS
Abstract
For over 30 years now, there has been an intense debate on the effects of external
rewards on intrinsic motivation. The answer to this question appears to depend on
two properties of a reward: control and information. Though many professions have
utilized rewards to control behavior with students, athletes, and children, it appears
that this very controlling mechanism is what undermines intrinsic motivation.
Conversely, an informational message of competency can enhance intrinsic
motivation. Therefore, the way the reward is delivered will determine its effects on
intrinsic motivation. In this context, the purpose of this paper was to analyze the
effects of various rewards on intrinsic motivation. Rewards will be analyzed using the
well established Cognitive Evaluation Theory.
Introduction
We shall begin this article with a short story (McCullagh, 2005). Once there was a
man, who lived in a house and had a lawn. And kids would come to play on this
mans lawn to have fun. The man began to be annoyed by this, and decided to do
something about it. So, strangely enough...he paid them a dollar to come play on his
lawn. The kids happily took the dollar and played on his lawn. The next day, the man
told the kids that he did not have enough money, so he could only give them 50
cents to come play on his lawn. On the third day, he told them he could only give
them a nickel to come play on his lawn. The kids were displeased with this, and told
the man he could forget that, and that they would not play on his lawn for such a
cheap reward. What happened? These kids played on his lawn before for absolutely
nothing, but now, they quit playing, even though they were offered a nickel! Well, it
just so happens that this man understood an important concept in Sport Psychology.
That is, the effects of external rewards on intrinsic motivation.
Intrinsic & Extrinsic Motivation
Motivation can be defined as the intensity and direction of effort (McCullagh, 2005).
Intensity refers to the quantity of effort, while direction refers to what you are drawn
too. Evidence suggests that enhanced motivation promotes learning, performance,
enjoyment, and persistence in sport, among other benefits (McCullagh, 2005;
Wilson, 2005). Therefore, methods to enhance motivation have been thoroughly
investigated.
There are two forms of motivation: intrinsic and extrinsic motivation.
Intrinsic motivation can be defined as an individuals need to feel competency and
pride in something (McCullagh, 2005). Therefore, athletes who are intrinsically
motivated participate in sports for no apparent reward other than the satisfaction
and pleasure they get from the activity itself. There are 3 types of intrinsic
motivations (Weinberg & Gould, 2003):
• Knowledge
• Accomplishment
• Stimulation
Being intrinsically motivated for knowledge occurs when athletes participate in
activates because of the pleasure and satisfaction they get from learning, acquiring,
and studying something new in their sport. This would include learning how to squat,
or refining your pattern on deadlifts.
Being intrinsically motivated for accomplishments occurs when athletes participate in
activates because of the pleasure and satisfaction they get from mastering various
skills. For example, reaching a goal of squatting 400 pounds, or acquiring perfect
form on deadlifts.
Being intrinsically motivated for stimulation occurs when athletes participate in
activates because of pleasant sensations such as danger, pain, or excitement. For
example, the rush you get when lifting heavy in the weight room, or posing down on
stage.
Extrinsic Motivation can be defined as performance of an activity in order to attain
some separate outcome (Ryan & Deci, 2000). Therefore, athletes that are
extrinsically motivated participate in sport for external causes such as rewards,
positive feedback, recognition, etc.; rather than for the inherent satisfaction of
performing the activity itself (intrinsic motivation). There are 4 types of extrinsic
motivations (Weinberg & Gould, 2003):
• Integrated regulation
• Identified regulation
• Introjected regulation
• External regulation
Integrated regulation occurs when athletes perform activities to benefit different
aspects of life, rather than for the pleasure of participating itself. For example,
training and dieting hard in order to develop a healthy life style.
Identified regulation occurs when athletes participate in an activity because the
activity is considered of high value and important to the participant, even if they do
not enjoy the activity itself. The difference between this and integrated regulation, is
that this is limited to the particular activity itself. For example, lifting weights to
become huge. Or going to school, or doing your homework because it is a great way
to learn about things which may be useful to you.
One important concept to understand is that these two forms of extrinsic motivation
(more so with integrated regulation) and the three forms of intrinsic motivation all
involve the athlete participating in sports by their own initiative, because they want
to participate for some desired outcome (autonomous). Therefore, these 5 subtopics
of motivation have been found to positively influence affective, behavioral, and
cognitive functions (Vallerand, 1997; Vallerand and Rousseau, 2001).
Introjected regulation occurs when athletes participate in an activity because of
various pressures. For example, lifting heavy, or posing in the weight room, so you
can impress people in the gym. Or training so that you can get huge, and gain self
recognition and approval from others. Evidence suggests that Introjected regulation
increases tension and anxiety (Nikos, 2001). An outcome oriented mentality (click
Here for more information), in which ones worth is on the line, and their self esteem
is contingent on the outcome, is another example of introjection.
External regulation occurs when athletes participate in an activity only because they
feel they have to, or because they may get a reward. For example, playing for the
money, or telling someone if you do such and such, I’ll give you such and such. This
activity is performed entirely for the reward.
The last term to discuss is amotivation. This is when an athlete is neither intrinsically
or extrinsically motivated, and therefore, do not have any reason to participate in an
activity. For example, a kid in P.E. class, who sits down on the ground and writes
letters on the ground while his/her pears are playing baseball.
It is important to understand that people have various forms of motivation;
therefore, you can be both externally and internally motivated. However, results
indicate that individuals who are intrinsically motivated (and or have the two forms
of autonomous extrinsic motivation), compared to those who are controlled by others
to perform an activity (extrinsically motivated) have more interest, excitement, fun,
and confidence, which leads to enhanced performance, creativity, persistence, vigor,
general well-being, and self-esteem, among other benefits (Ryan & Deci, 2000). For
instance, Frederick and Ryan (1993) found that intrinsic motivation facilitated a
greater amount of hours and days per week of exercise and sport participation, and
greater levels of satisfaction and competency, compared to extrinsic motivation,
which facilitated greater anxiety, and decreased self-esteem. Similarly, Hodgins,
Yacko, Gottlieb, Goodwin and Rath (2002) found that intrinsically motivated rowers
had greater performance than extrinsically motivated rowers.
Therefore, ways to enhance intrinsic motivation are of the utmost importance for
athletes. In this context, the purpose of this paper was to discuss the effects of
rewards on intrinsic motivation.
Theories on the Effects of External Rewards on Intrinsic Motivation
It was originally thought by many that extrinsic rewards would enhance motivation
for a task that was already intrinsically motivating. This seemed as simple as 1
(external reward) +1 (already present intrinsic motivation) = 2 (more motivation).
But as usual, the concept is not this simple. For over 30 years, there has been an
intense debate over the effects of extrinsic rewards on intrinsic motivation;
particularly, its ability to undermine intrinsic motivation. First, it is important to
understand that all research is theory driven (Sawyer, 2005).
Theories are vital to research because they offer a framework for many areas of
research aimed at the nature of the topic, address possible mechanisms, and suggest
several specific interventions (McCullagh, 2005). Therefore, before we get into
studies and experiments on this topic, we need to discuss some theories.
Self Determination & Cognitive Evaluation Theory
Deci, Ryan, and others developed what today is the most well established theory on
the effects rewards have on intrinsic motivation—Self Determination Theory (SDT;
Deci, 1975; Deci & Ryan, 1985). This theory suggests that humans have three
central psychological needs, which are relatedness, effectance, and autonomy. Deci
and Ryan (1994) summarize these needs in the following quote: “people are
inherently motivated to feel connected to others within a social milieu (relatedness),
to function effectively in that milieu (effectance), and to feel a sense of personal
initiative in doing so (autonomy)” (p.7). Notice the term “inherently”. This theory
suggests that humans have an innate tendency to develop these needs.
Nevertheless, these needs do not develop automatically; they must be furnished by
the environment, which can either promote growth, or impede it.
While, this theory focuses on intrinsic motivation, it does not say what causes it. In
this context, Deci and Collegues developed a sub theory to SDT, known as Cognitive
Evaluation Theory (CET; Ryan & Deci, 2000).
CET focuses on factors which can increase or decrease intrinsic motivation. In
essence, this theory suggests that rewards have two basic properties that can
influence intrinsic motivation: information and control. And these properties can
increase or decrease intrinsic motivation depending on how they effect an individuals
self determination and competency.
The informational aspect of a reward relays information about a person’s
competency. MVP rewards, All Star Selections, Sandow Trophies, among other
rewards, all relay that the person receiving the reward is competent. CET suggests
that if the informational aspect of a reward relays that the person is competent, this
will enhance intrinsic motivation. Thus, a reward must be based on performance to
enhance intrinsic motivation; at least, from an informational view point. Conversely,
CET predicts that if a reward relays that the athlete is not competent, such as getting
a last place reward, this will decrease intrinsic motivation.
The second property of a reward is the controlling aspect. This has to do with a
person’s locus of causality. Locus of causality is the degree people perceive their
behavior to be freely determined (self determined) or caused by other people. If a
person feels their behavior is caused by outside pressures from others, they would
have an external locus of causality. If a person feels their behavior is self
determined, or initiated, they would have an internal locus of causality. CET predicts
that if a reward is perceived as controlling, people will attribute their behavior to an
outside source (an external locus of causality). Conversely, if people do not feel
controlled by the reward, they will attribute their behavior to self determination (an
internal locus of causality). CET predicts that if a reward is perceived as controlling,
it will decrease intrinsic motivation; but if a reward is not perceived as controlling,
and the person has an internal locus of causality, intrinsic motivation will be high.
To elaborate on control and intrinsic motivation, people who have an internal locus of
causality, feel that they participate in an activity because they want to (high intrinsic
motivation); whereas, people who have an external locus of causality, feel that they
participate in an activity because of an external cause (i.e. playing for the money).
Therefore, “paying people off” can decrease their intrinsic motivation, because they
may attribute their behavior to the money, rather than self determination. An
example of something that would promote an internal locus of causality would be a
coach who gives his athletes greater input on what drills are performed during
practice. An example of controlling someone through a reward would be telling
someone that you will give them such and such for doing such and such (external
regulation). For example, telling a kid, "if you clean your room, I'll give you five
dollars." In this scenario, the reward is controlling the persons behavior, rather than
self determination.
The key aspect of this theory is the perception of the receiver of the reward.
Someone who gives an athlete a reward may have the best intentions for the
athlete; but if it is perceived as controlling, it will decrease intrinsic motivation. For
example, if a player perceives that he/she is being rewarded with cars or money,
only so that he will stay on the team, this will decrease his/her intrinsic motivation,
because they will perceive the reward as controlling. Additionally, the reward giver
should stress the informational aspect of the reward, making sure the player knows
that the reward was given to them to show a sign of appreciation for their hardwork.
Rewards
Based on his research, the current author suggests that a reward can be defined as
an external agent administered when a desired act or task is performed, that has
controlling and informational properties. While rewards are typically delivered to
increase the probability of a response, they can increase or decrease the probability of an event occurring, depending on the saliency and direction of the controlling and
informational aspects of the reward. Saliency would refer to the intensity of either
the controlling or informational aspect of the reward. It is what stands out to you the
most. While direction would refer to whether the reward is perceived as increasing or
decreasing the athlete’s control; and whether the information is perceived as positive
(increasing the athlete’s competency) or negative (decreasing the athlete’s
competency). Rewards can come in the form of verbal rewards (i.e. telling someone
“good job!”), physical rewards (i.e. a pat on the back), or tangible rewards (i.e.
giving someone money, food, or a medallion), among others. There are 5 basic types
of rewards discussed in the literature as follows (Deci, Koestner, & Ryan, 1999):
• Task-Non contingent rewards
• Engagement contingent rewards
• Completion contingent rewards
• Performance contingent rewards
• Unexpected rewards
Task-Non contingent rewards are rewards given for just showing up for the study. In
an experiment, a participant may be paid to just show up for the experiment, but
they are not required to do anything. They could just sit around the entire time.
Engagement contingent rewards are rewards given for just participating in an
activity, and not necessarily completing it. For example, an experimenter may pay a
participant just to participate in an activity that involves making a puzzle, but they
do not have to complete or perform well on the puzzle.
Completion contingent rewards are rewards given for completing a task.
Performance contingent rewards are rewards given for performance, usually based
on a normative value. For example, doing better than 80% of the participants in a
study. A sub category of performance contingent rewards are competitively
contingent rewards. They involve rewarding individuals for defeating others.
Unexpected rewards occur when participants receive a reward after performing a
certain behavior, but were not expecting to receive a reward.
CET makes several predictions on these types of rewards.
This theory predicts that unexpected rewards would not decrease intrinsic
motivation, because the participant performed the task without knowledge of the
reward; therefore, the controlling aspect of the reward would not be as salient, and
participants would attribute their participation in the activity to an internal locus of
causality. Because of the informational aspect, unexpected rewards may also
enhance intrinsic motivation. But the administrator of the reward would have to give
it based on high performance, and stress the informational aspect for it to be
beneficial.
Non-contingent rewards do not require participating in the task, completing the task,
or performing well on the task, and they deliver no information about the person’s
competency. Therefore, CET predicts they will not effect intrinsic motivation.
Both engagement and completion contingent rewards are predicted to typically cause
the highest decrease in intrinsic motivation. This is because they contain a high
controlling aspect, but deliver no information about the competency of the individual.
For example, you could be paid for participating in an activity, but whether you
perform well or not, is irrelevant. Therefore, these types of rewards say nothing
about the person’s competency, and decrease their control.
Performance contingent rewards are a lot more complex. While they can decrease
control as engagement and completion contingent rewards do, they also relay a
sense of competency. Therefore, CET predicts that if the information aspect is more
salient for performance rewards, it may be able to counteract the controlling aspect
of the reward. Additionally, whether the message is portrayed as controlling or not
will also determine whether the reward decreases intrinsic motivation or not.
Therefore, it is predicted by CET that generally, these rewards will decrease intrinsic
motivation less than engagement and completion contingent rewards do.
Lastly, CET would suggest that verbal and physical rewards should enhance intrinsic
motivation. This is because the informational aspect is very salient, while the
controlling aspect is typically low. However, the controlling aspect can vary
significantly depending on the delivery. This will be discussed more in practical
applications.
Important note
This article is dealing with the effects external rewards have on intrinsic motivation
for activities that are already intrinsically motivating to the individual. Rewards have
an entirely different effect on activities that are boring to an individual. Therefore,
these results should not be generalized to other scenarios. It also primarily makes
predictions about when rewards are removed. Because while the rewards are being
administered, this is a test of both extrinsic and intrinsic motivation. Lastly, CET does
not make predictions on studies that combine verbal rewards with tangible rewards,
because they would contribute to intrinsic motivation in opposite directions.
Studies on the Effects of External Rewards on Intrinsic Motivation
In a revolutionary study on the effects of external rewards on intrinsic motivation,
Deci (1971) had participants play on an inherently interesting task, called the SOMA
puzzle. Participants were paid to play, were given verbal rewards (i.e. verbally
encouraged), or received no reinforcement for participating. Results found that
participants who were paid money to play spent a significantly less amount of time
(100 seconds) than participants who were not rewarded to play. Whereas, those who
received verbal rewards, played for a longer amount of time than both conditions
(intrinsic motivation increased).
Soon after this, Lepper and Greene (1975) performed a study called “Turning Play
Into Work.” Participants were nursery school children, who drew with felt pens, an
intrinsically motivating activity. Participants were divided into three conditions.
Condition one was the expected reward condition, in which the children agreed to
draw a picture in order to receive a good player certificate (completion contingent
rewards). In the unexpected reward condition, participants were unexpectedly given
a reward after they completed the task. In the control condition, children received no rewards for participating. One week later, participants were brought back in to play
with the felt pens without any rewards. Results found that children in the expected
reward condition decreased in intrinsic motivation; whereas, children in the other
two conditions maintained intrinsic motivation.
Limits must often be set for various activities, such as education. In this context,
Koestner, Ryan, Bernieri, and Holt (1984) investigated if limits could be set without
decreasing intrinsic motivation. Participants consisted of 44 first and second grade
students, engaging in a painting activity. Participants were given controlling verbal
limits, informational verbal limits, and no verbal limits. Controlling verbal limits
consisted of using various phrases such as, “I am going to tell you what you have to
do now” “We have rules you have to follow” “I want you to be a good boy or girl.”
Whereas, children in the information condition were just told what the experimenters
wanted them to do, instead of what they had to do. For example, “The smaller sheet
is for you to paint on, the larger sheet is a border to be kept clean.” Results found
that intrinsic motivation, enjoyment, creativity, and quality of the artist product were
decreased from the control condition, but not by the informational condition.
Therefore, informational constraints may be an effective way to regulate behavior.
This study supported cognitive evaluation theory. CET would predict that as control
decreased, intrinsic motivation would decrease. Other studies have also supported
these findings (Plant and Ryan, 1985; Deci, Koestner, & Ryan, 1999; Deci, Koestner,
& Ryan, 1999, b).
While the purpose of this article is to discuss the effects of rewards on intrinsic
motivation, CET has incredible implications for many areas of life. Rewards are just
one of a broad array of techniques in which we control people and decrease their
intrinsic motivation. Results show deadlines, evaluations, and imposed goals, all
decrease intrinsic motivation (Deci, Koestner, & Ryan, 1999; Deci, Koestner, & Ryan,
1999, b; Ryan & Deci, 2000). Whereas, provision of choice, and acknowledgement of
feelings, enhance intrinsic motivation (Deci, Koestner, & Ryan, 1999). These results
are all centered on one key concept—control. It appears that using controlling
language and actions consistently lowers intrinsic motivation. Plant and Ryan (1985)
suggest that control is “any vocalization that pressures a person to behave a certain
way.” It is suggested that contextual factors that tend to control someone—such as
pressure to feel, think or behave a certain way—lowers intrinsic motivation. For
example, a mother who says, “you know what you should do” would be controlling
compared to “what do you think you should do?” Or telling someone “Good job, you
should keep up the good work” or “Good job, you are doing as you should.” One
experiment found that saying (Deci, Koestner, & Ryan, 1999) “I have not been able
to use most of the data I got so far, but if you keep it up, I may be able to use
yours” puts pressure on the person, decreasing his/her control. While the other
participants were told, “compared to most of my subjects, you are doing really well”
which resulted in more control, and higher intrinsic motivation. Therefore, how
information is relayed is vital for how an individual perceives rewards and
instructions. Numerous future articles can be written on the implications these
theories have not only on bodybuilding, but everyday life—stay tuned.
Scholarships
The famous basketball player, Magic Johnson, was asked if he had received rewards
when being recruited by college basketball teams. He stated that (Weinberg & Gould,
2003) “I received my share of offers for cars and money. It immediately turned me
off. It was like they were trying to buy me, and I don’t like anyone trying to buy
me.” According to CET theory, what Magic was really annoyed with was the
controlling aspect of rewards. With hundreds of multi-million dollar contracts thrown
around in sports today, of immense interest would be the effect of rewards on
athlete’s intrinsic motivation. To investigate this question, there has been a
significant amount of research on the effects of athletic scholarships on intrinsic
motivation.
In one of the earliest studies on the effects of athletic scholarships on intrinsic
motivation, Ryan (1977, 1980; as cited in Frederick & Ryan, 1995) observed
collegiate Division I football players on scholarships and not on scholarships. Ryan
found that players who had scholarships did not enjoy football as much as nonscholarship
athletes; further, scholarship athletes had less intrinsic motivation every
year they had a scholarship, so that they had their lowest level of intrinsic motivation
during their senior year!
However, in a similar case, results found that male wrestlers and female athletes
from six different sports had higher intrinsic motivation than athletes who were not
on scholarships.
These results may seem contradictory at first glance. However, CET may offer some
helpful explanations.
As you recall, according to CET, there are two properties to a reward that influence
intrinsic motivation—control and information. Therefore, CET would predict that the
controlling and informational aspects of these rewards were different.
While scholarships can clearly be perceived as controlling, they also relay information
of competency. Now, what determines whether a scholarship would decrease or
increase intrinsic motivation is the saliency of the controlling and informative aspects
of the reward. If the controlling aspect was more salient, intrinsic motivation would
decrease; conversely, if the informative aspect of competency was more salient,
intrinsic motivation would increase.
Consider this. In the 1980’s, when this study was composed, athletic scholarships for
Division I football teams was common. However, it was rare for wrestlers and female
athletes to receive scholarships. Thus, the normality of a scholarship for Division I
football players, would have decreased the saliency of the informative properties of
the reward; however, the rarity of a scholarship for wrestlers and female athletes,
would highlight the informative message of competency.
This hypothesis was supported in a more recent study on athletic scholarships
(Amorose, Horne, & Miller, 1994). 440 male and female athletes in Division I sports
were analyzed. Results found that players on scholarship had lower intrinsic
motivation, enjoyment, and perceived choice than non-scholarship athletes. CET
would predict that the increased number of scholarships offered to female athletes
today, decreased the saliency of the informative aspect of scholarships for them,
resulting in a decrease in intrinsic motivation, similar to male football players.
Now, as the reader recalls, according to CET, it is the perception of the reward that
causes an athlete to increase or decrease in intrinsic motivation. In this context, if a
coach uses scholarships as leverage over their athletes—i.e. threatening that they
will lose their scholarships if they don’t perform well—then this would emphasize the
controlling aspect of the reward, decreasing intrinsic motivation. Therefore, coaches
should be extremely careful of how they treat the scholarships. Coaches should
emphasize the informational aspect, and downplay the controlling aspect, instead of
using scholarships as leverage over their athletes.
In this context, Amorose and Horn (2000) teamed up again to see if it was the
reward or coaching behaviors that decreased intrinsic motivation. What they found is
that the coaches’ behaviors were more influential on intrinsic motivation. For
example, if a coach was perceived as a tyrant over practice, players had lower
intrinsic motivation; whereas, coaches who were perceived as positive, and
democratic, increased intrinsic motivation. Therefore, this study suggests that the
coach has a bigger impact on intrinsic motivation than the scholarship itself.
Meta-Analyses
A meta-analysis is a statistical practice of combining the results of a number of
studies. Considering that literally hundreds of studies have been done on the effects
of external rewards on intrinsic motivation over the past 30 years, a summary of the
data through a meta-analysis would be of immense help to our research.
Thankfully, a number of scientists have done just this.
Rummel and Feinberg (1988) examined 45 studies between 1971 and 1985. They
found that extrinsic rewards significantly decreased intrinsic motivation. However,
there were some flaws in this meta-analysis. For example, they did not distinguish
between various rewards (i.e. expected vs. unexpected; contingent vs. noncontingent,
etc.).
Wiersma (1992) performed another meta-analysis on 20 published studies between
1971 and 1990. Results also found that rewards significantly decreased intrinsic
motivation. However, there were also flaws in this. First, the sample size of studies
was small; second, as with the former meta-analysis, they did not distinguish
between rewards; and third, many of the studies did not have control groups, and
instead only compared reward conditions to each other.
Following this, Tang and Hall (1995) performed another meta-analysis, which was
much more comprehensive than its predecessors. They examined 50 studies, and
distinguished between rewards. They found that task-contingent rewards decreased
intrinsic motivation the most, followed by performance contingent rewards; however,
unexpected rewards did not effect intrinsic motivation. They also found that verbal
rewards increased intrinsic motivation. Further, they found that rewards may
enhance intrinsic motivation for boring tasks. All of these findings supported CET.
Flaws in the experiment were that some of the measures were not purely of intrinsic
motivation. This is because they included results about the quality, quantity, and
ratings of how much participants enjoyed activities during the reward phase. Again,
this would be a measure of both intrinsic and extrinsic motivation. In order to assess
intrinsic motivation, you would need to test participants during another trial, when
rewards were not administrated.
The importance of this distinction is highlighted by the previously mentioned metaanalysis
by Wiersma (1992). Wiersma sought to clear up the supposed contradictory
studies on the effects of rewards on intrinsic motivation. It was hypothesized that
the conflicts seen could be attributed to differences in experimental designs. They
found that while the reward is being administered, intrinsic motivation does not
decrease (though intrinsic motivation is difficult to measure in this situation; this will
be discussed momentarily); however, when the reward is removed, intrinsic
motivation does decrease.
The problem is this: performance during the administration of the reward is not a
pure measure of intrinsic motivation. It is a measure of both intrinsic motivation and
the reward. When the reward is administrated, it would be predicted that if
performance increases, the extrinsic reward was stronger than intrinsic motivation; if
performance decreases, the extrinsic reward was weaker than intrinsic motivation;
and if performance remains the same, the extrinsic reward was equal to intrinsic
motivation. Therefore, a "free time" trial must be performed, in which the reward is
removed to accurately asses intrinsic motivation.
In summary, though these three meta-analyses had several methodological short
comings, they all supported CET.
So far, evidence strongly supports CET. But what would a scientific discussion be
without a little controversy? Don’t fret; this was supplied in a mighty way through
Cameron and Colleagues from the behaviorist camp.
Cameron and Pierce (1994) performed a meta-analysis on the effect of rewards on
intrinsic motivation. Two years later, Cameron and Eisenberger performed an almost
identical analysis, with a few slightly modifications. I will therefore refer to these
studies as Cameron, Eisenberger, and Pierce henceforth, unless stated otherwise.
There results were almost entirely contradictory to every shred of evidence shown in
this article. They concluded that rewards essentially do not effect intrinsic
motivation, and that there was no reason not to use rewards to regulate behaviors,
particularly in educational settings. And here is their boldest statement of all: they
suggested the “abandoning of cognitive evaluation theory”!
Since this meta-analysis was published, several scientists have harshly critiqued the
methods and conclusions made by Cameron, Eisenberger, and Pierce. In 1996, three
commentaries (Kohn, 1996; Lepper, Keavney, & Drake, 1996; Ryan & Deci, 1996)
argued that Cameron, Eisenberger, and Pierce’s meta-analysis was invalid and that
its conclusions false. More recently, the inventors of CET themselves stepped in and
critiqued this meta-analysis.
Deci, Koestner, and Ryan (1999) reported several flaws in this analysis. Some
notable comments were that this analysis combined the effects of boring tasks, with
intrinsically interesting task. As has been clearly displayed, CET makes no predictions
on boring tasks; only inherently interesting tasks. For, how would extrinsic rewards
undermine intrinsic motivation for tasks which have none, or very little intrinsic
motivation to begin with? Further, as show in the Tang and Hall (1995) metaanalysis,
rewards may actually be of benefit for boring activities. Therefore, this is
comparing apples and oranges, and would greatly skew the results. Another problem
was that they mixed some rewards together, and misclassified them. For example,
they included a reward from a study as a non-contingent reward, that actually was
administrated as an engagement contingent reward. They also eliminated 20% of the
studies as outliers, instead of isolating the cause of variability. Other critiques were
made, but clearly there were many flaws in this analysis.
A flaw in all these meta-analyses is that they did not include unpublished doctoral
theses. Doctoral theses are commonly the most comprehensive papers conducted;
thus, inclusion of these papers would increase the reliability of the analysis.
In light of the various flaws in these meta-analyses, Deci, Koestner, and Ryan (1999)
teamed up to compose the most comprehensive meta-analysis on rewards to date.
128 studies were examined. Participants ranged from college to pre-school students.
There were a wide array of rewards administrated, including marshmallows and
dollar bills.
Results found that verbal rewards, when not administered in a controlling manner,
increased intrinsic motivation. However, when they were administered in a
controlling manner, they decreased intrinsic motivation.
Both engagement and completion contingent rewards decreased intrinsic motivation.
Performance contingent rewards also decreased intrinsic motivation; but not as much
as the previous two rewards. However, participants in the performance contingent
condition, who got less than the maximum reward, showed a sharp decline in
intrinsic motivation—more so than any other group! This is because not only are they
being controlled by the reward, but they are being delivered information that they
are incompetent. Since in many cases, the majority do not achieve the standard of
excellence taken to obtain a performance contingent reward, these can be especially
dangerous to administer. Results also found a significant difference between the
effects of rewards on boring and intrinsically motivating tasks. This is in agreement
with the findings of Tang and Hall (1995).
Unexpected rewards and non-contingent rewards did not decrease intrinsic
motivation; presumably because the participants were not doing the task in order to
get those rewards, therefore, they did not feel controlled by them.
It has also been suggested that the effects of rewards on intrinsic motivation may be
transitory. But their results indicated they were long lasting.
One interesting finding was that rewards were more undermining to intrinsic
motivation for kids than college students. The authors suggested 3 explanations: 1.)
Kids are “bribed” more often than adults for performing activities, so it may have a
deeper effect. 2.) College students are more cognitively advanced, so they may be
able to distinguish the controlling aspect from the informational aspect of rewards
better than kids. 3.) College students may have different expectations during
experiments, because they have more experience in them than kids.
The authors suggested that verbal rewards are effective, in part because they are
unexpected rewards. In this context, if a coach always praises their athletes for
performing an activity, the athletes may come to expect praise, and CET would
predict that intrinsic motivation would decrease. In fact, they sourced 3 studies
which indicated that when participants were told they would be given performance
feedback, and then were given positive feedback (verbal rewards) after completion
of the task, intrinsic motivation significantly decreased! So evidence suggests that
you should give verbal rewards in an intermittent fashion. This is further supported
by studies on partial reinforcement (refer to Hull’s Quantitative Equation on Human
Performance).
In response to this meta-analysis, Cameron, Eisenberger, and Pierce (1999) teamed
up again to make 2 more meta-analyses. But it was more or less the same. Deci,
Koestner, and Ryan (1999, b) again responded with a convincing argument refuting
the claims made by them. The largest point that Cameron, Eisenberger, and Pierce
(1999) tried to get across was that performance-contingent rewards may enhance
intrinsic motivation. This was supported by a recent study by Cameron, Pierce,
Banko, Katherine, and Slyvia (2003) who found that giving rewards that are tied to
meeting increasingly demanding performance standards enhanced intrinsic
motivation. However, Deci, Koestner, and Ryan (1999, b) pointed out several flaws
in this meta-analysis. The primary differences were that they excluded 11 studies out
of 40 which indicated negative results, and had several other methodological flaws,
which had actually been pointed out in previous critiques, but were not corrected.
In summary, the results appear to strongly support CET, and suggest that we should
be very careful about how and if we use external rewards.
The Link Between Self Determination Theory, CET, & Goal Orientations
In another article in this issue of JHR, the current author addressed various goals. It
is a pre-requisite to study that article to be able to understand the topic that is about
to be discussed. Click Here, to study this article.
It has been suggested by several authors that there may be a link between self
determination theory and goal orientations (Nikos, 2001). Evidence suggests than an
ego (outcome) orientation can undermine intrinsic motivation; whereas, a mastery
goal orientation facilitates intrinsic motivation (Ryan, 1982; Brunel, 1999). This is
because athletes with an ego orientation are concerned more with the outcome of an
event, rather than the activity itself (Nicholls, 1989). For instance, Brunelt (1999)
found that an ego orientation increased introjected regulation (participating in an
activity because of various pressures) and external regulation (participating in an
activity for a reward) in 160 badminton athletes. Similarly, Ryan (1982) found that
an ego orientation decreased intrinsic motivation for participating in a puzzle game.
Conversely, a mastery orientation appears to enhance intrinsic motivation. This is
because a mastery orientation focuses on intrinsic properties of the activity, rather
than an external motivator. In this context, Butler (1987) found that a mastery goal
orientation enhanced intrinsic motivation, by promoting task persistence and
challenge seeking activities.
Various studies have supported this hypothesis. Duda, Chi, Newton, Walling, and
Catley (1995) found that a mastery goal orientation enhanced intrinsic motivation,
while an ego goal orientation maintained or decreased intrinsic motivation. Studies
on British and Romanian PE have found that students with the highest intrinsic
motivation were those with a high mastery orientation, regardless of their ego
orientation.
And a plethora of other studies have demonstrated this (refer to the aforementioned
goal setting article for more).
In order to summarize this data, Nikos (2001) performed a comprehensive analysis
on the link between Self Determination Theory and Goal orientations. Over 247
British University students were examined. Results indicated that a mastery
orientation was able to predict all three types of intrinsic motivation and identified
regulation (as you recall, this is an autonomous form of extrinsic motivation, and
related to enhanced self determination). However, a mastery orientation did not
predict introjected regulation (i.e. avoidance or participation in activities because of
guilt, pressure, etc.), external regulation (participation in activities purely for
rewards), or amotivation (a lack of motivation).
As expected, an ego orientation was able to predict introjected and external
regulation. This is most likely because athletes with an ego oriented mentality
participate in sport for extrinsic rewards such as acknowledgment, superiority, etc.
(Nichols, 1989). Additionally, an ego orientation was not a predictor of intrinsic
motivation, or other self determined measures (except, it did predict intrinsic motivation for sensations). This is attributed to the controlling nature of their
extrinsic goals. Lastly, ego orientations did not predict amotivation.
Nikos also examined the interaction between goal orientations, as individuals can be
both mastery and ego oriented. Results found that when participants master
orientation was high, external regulation was lower when ego orientation was lower.
Further, high mastery oriented individuals predicted high intrinsic motivation
regardless of ego orientation.
Thus, evidence suggests that a mastery goal orientation is optimal for facilitating
intrinsic motivation, while an ego orientation will tend to enhance extrinsic
motivation. These results are clearly explained using Self Determination Theory.
Now, an outcome oriented mentality often promotes competition (for information on
competition vs. cooperation Click Here). And CET would predict that competitive
mentalities can lead to a decrease in intrinsic motivation, as they focus on external
goals such as winning. In this context, Fortier, Vallerand, Briere & Provencher (1995)
tested CET with competitive athletes compared to recreational athletes. Results
indicated that recreational athletes had greater intrinsic motivation, while
competitive athletes had greater identified regulation and amotivation. Similarly,
Kavussanu and Roberts (1996) found that a competitive, outcome oriented
environment decreased the intrinsic motivation of tennis players.
The next question is, what is the effect of rewards on competitive situations. These
type of rewards are called competitively contingent rewards, and are a sub category
of performance contingent rewards. They involve rewarding individuals for defeating
others. In this context, Pritchard, Campbell and Campbell (1977) found that giving
participants a $5 reward for defeating participants in their group (approximately 6
participants per group) decreased intrinsic motivation relative to no reward. CET
predicts these types of rewards are very controlling, as they focus the participants
locus of causality on winning, which is an external motivator. Other studies have also
supported these findings (Ryan, Mims, & Koestner, 1983).
Thus, Self Determination Theory and CET are very helpful in explaining the results of
goal orientations and competition on intrinsic motivation.
Conclusion
The results clearly indicate that the rewards can greatly undermine intrinsic
motivation. For practical applications, click Here.
Researched and Composed by Gabriel “Venom” Wilson, BSc. (Hons), CSCS
Abstract
For over 30 years now, there has been an intense debate on the effects of external
rewards on intrinsic motivation. The answer to this question appears to depend on
two properties of a reward: control and information. Though many professions have
utilized rewards to control behavior with students, athletes, and children, it appears
that this very controlling mechanism is what undermines intrinsic motivation.
Conversely, an informational message of competency can enhance intrinsic
motivation. Therefore, the way the reward is delivered will determine its effects on
intrinsic motivation. In this context, the purpose of this paper was to analyze the
effects of various rewards on intrinsic motivation. Rewards will be analyzed using the
well established Cognitive Evaluation Theory.
Introduction
We shall begin this article with a short story (McCullagh, 2005). Once there was a
man, who lived in a house and had a lawn. And kids would come to play on this
mans lawn to have fun. The man began to be annoyed by this, and decided to do
something about it. So, strangely enough...he paid them a dollar to come play on his
lawn. The kids happily took the dollar and played on his lawn. The next day, the man
told the kids that he did not have enough money, so he could only give them 50
cents to come play on his lawn. On the third day, he told them he could only give
them a nickel to come play on his lawn. The kids were displeased with this, and told
the man he could forget that, and that they would not play on his lawn for such a
cheap reward. What happened? These kids played on his lawn before for absolutely
nothing, but now, they quit playing, even though they were offered a nickel! Well, it
just so happens that this man understood an important concept in Sport Psychology.
That is, the effects of external rewards on intrinsic motivation.
Intrinsic & Extrinsic Motivation
Motivation can be defined as the intensity and direction of effort (McCullagh, 2005).
Intensity refers to the quantity of effort, while direction refers to what you are drawn
too. Evidence suggests that enhanced motivation promotes learning, performance,
enjoyment, and persistence in sport, among other benefits (McCullagh, 2005;
Wilson, 2005). Therefore, methods to enhance motivation have been thoroughly
investigated.
There are two forms of motivation: intrinsic and extrinsic motivation.
Intrinsic motivation can be defined as an individuals need to feel competency and
pride in something (McCullagh, 2005). Therefore, athletes who are intrinsically
motivated participate in sports for no apparent reward other than the satisfaction
and pleasure they get from the activity itself. There are 3 types of intrinsic
motivations (Weinberg & Gould, 2003):
• Knowledge
• Accomplishment
• Stimulation
Being intrinsically motivated for knowledge occurs when athletes participate in
activates because of the pleasure and satisfaction they get from learning, acquiring,
and studying something new in their sport. This would include learning how to squat,
or refining your pattern on deadlifts.
Being intrinsically motivated for accomplishments occurs when athletes participate in
activates because of the pleasure and satisfaction they get from mastering various
skills. For example, reaching a goal of squatting 400 pounds, or acquiring perfect
form on deadlifts.
Being intrinsically motivated for stimulation occurs when athletes participate in
activates because of pleasant sensations such as danger, pain, or excitement. For
example, the rush you get when lifting heavy in the weight room, or posing down on
stage.
Extrinsic Motivation can be defined as performance of an activity in order to attain
some separate outcome (Ryan & Deci, 2000). Therefore, athletes that are
extrinsically motivated participate in sport for external causes such as rewards,
positive feedback, recognition, etc.; rather than for the inherent satisfaction of
performing the activity itself (intrinsic motivation). There are 4 types of extrinsic
motivations (Weinberg & Gould, 2003):
• Integrated regulation
• Identified regulation
• Introjected regulation
• External regulation
Integrated regulation occurs when athletes perform activities to benefit different
aspects of life, rather than for the pleasure of participating itself. For example,
training and dieting hard in order to develop a healthy life style.
Identified regulation occurs when athletes participate in an activity because the
activity is considered of high value and important to the participant, even if they do
not enjoy the activity itself. The difference between this and integrated regulation, is
that this is limited to the particular activity itself. For example, lifting weights to
become huge. Or going to school, or doing your homework because it is a great way
to learn about things which may be useful to you.
One important concept to understand is that these two forms of extrinsic motivation
(more so with integrated regulation) and the three forms of intrinsic motivation all
involve the athlete participating in sports by their own initiative, because they want
to participate for some desired outcome (autonomous). Therefore, these 5 subtopics
of motivation have been found to positively influence affective, behavioral, and
cognitive functions (Vallerand, 1997; Vallerand and Rousseau, 2001).
Introjected regulation occurs when athletes participate in an activity because of
various pressures. For example, lifting heavy, or posing in the weight room, so you
can impress people in the gym. Or training so that you can get huge, and gain self
recognition and approval from others. Evidence suggests that Introjected regulation
increases tension and anxiety (Nikos, 2001). An outcome oriented mentality (click
Here for more information), in which ones worth is on the line, and their self esteem
is contingent on the outcome, is another example of introjection.
External regulation occurs when athletes participate in an activity only because they
feel they have to, or because they may get a reward. For example, playing for the
money, or telling someone if you do such and such, I’ll give you such and such. This
activity is performed entirely for the reward.
The last term to discuss is amotivation. This is when an athlete is neither intrinsically
or extrinsically motivated, and therefore, do not have any reason to participate in an
activity. For example, a kid in P.E. class, who sits down on the ground and writes
letters on the ground while his/her pears are playing baseball.
It is important to understand that people have various forms of motivation;
therefore, you can be both externally and internally motivated. However, results
indicate that individuals who are intrinsically motivated (and or have the two forms
of autonomous extrinsic motivation), compared to those who are controlled by others
to perform an activity (extrinsically motivated) have more interest, excitement, fun,
and confidence, which leads to enhanced performance, creativity, persistence, vigor,
general well-being, and self-esteem, among other benefits (Ryan & Deci, 2000). For
instance, Frederick and Ryan (1993) found that intrinsic motivation facilitated a
greater amount of hours and days per week of exercise and sport participation, and
greater levels of satisfaction and competency, compared to extrinsic motivation,
which facilitated greater anxiety, and decreased self-esteem. Similarly, Hodgins,
Yacko, Gottlieb, Goodwin and Rath (2002) found that intrinsically motivated rowers
had greater performance than extrinsically motivated rowers.
Therefore, ways to enhance intrinsic motivation are of the utmost importance for
athletes. In this context, the purpose of this paper was to discuss the effects of
rewards on intrinsic motivation.
Theories on the Effects of External Rewards on Intrinsic Motivation
It was originally thought by many that extrinsic rewards would enhance motivation
for a task that was already intrinsically motivating. This seemed as simple as 1
(external reward) +1 (already present intrinsic motivation) = 2 (more motivation).
But as usual, the concept is not this simple. For over 30 years, there has been an
intense debate over the effects of extrinsic rewards on intrinsic motivation;
particularly, its ability to undermine intrinsic motivation. First, it is important to
understand that all research is theory driven (Sawyer, 2005).
Theories are vital to research because they offer a framework for many areas of
research aimed at the nature of the topic, address possible mechanisms, and suggest
several specific interventions (McCullagh, 2005). Therefore, before we get into
studies and experiments on this topic, we need to discuss some theories.
Self Determination & Cognitive Evaluation Theory
Deci, Ryan, and others developed what today is the most well established theory on
the effects rewards have on intrinsic motivation—Self Determination Theory (SDT;
Deci, 1975; Deci & Ryan, 1985). This theory suggests that humans have three
central psychological needs, which are relatedness, effectance, and autonomy. Deci
and Ryan (1994) summarize these needs in the following quote: “people are
inherently motivated to feel connected to others within a social milieu (relatedness),
to function effectively in that milieu (effectance), and to feel a sense of personal
initiative in doing so (autonomy)” (p.7). Notice the term “inherently”. This theory
suggests that humans have an innate tendency to develop these needs.
Nevertheless, these needs do not develop automatically; they must be furnished by
the environment, which can either promote growth, or impede it.
While, this theory focuses on intrinsic motivation, it does not say what causes it. In
this context, Deci and Collegues developed a sub theory to SDT, known as Cognitive
Evaluation Theory (CET; Ryan & Deci, 2000).
CET focuses on factors which can increase or decrease intrinsic motivation. In
essence, this theory suggests that rewards have two basic properties that can
influence intrinsic motivation: information and control. And these properties can
increase or decrease intrinsic motivation depending on how they effect an individuals
self determination and competency.
The informational aspect of a reward relays information about a person’s
competency. MVP rewards, All Star Selections, Sandow Trophies, among other
rewards, all relay that the person receiving the reward is competent. CET suggests
that if the informational aspect of a reward relays that the person is competent, this
will enhance intrinsic motivation. Thus, a reward must be based on performance to
enhance intrinsic motivation; at least, from an informational view point. Conversely,
CET predicts that if a reward relays that the athlete is not competent, such as getting
a last place reward, this will decrease intrinsic motivation.
The second property of a reward is the controlling aspect. This has to do with a
person’s locus of causality. Locus of causality is the degree people perceive their
behavior to be freely determined (self determined) or caused by other people. If a
person feels their behavior is caused by outside pressures from others, they would
have an external locus of causality. If a person feels their behavior is self
determined, or initiated, they would have an internal locus of causality. CET predicts
that if a reward is perceived as controlling, people will attribute their behavior to an
outside source (an external locus of causality). Conversely, if people do not feel
controlled by the reward, they will attribute their behavior to self determination (an
internal locus of causality). CET predicts that if a reward is perceived as controlling,
it will decrease intrinsic motivation; but if a reward is not perceived as controlling,
and the person has an internal locus of causality, intrinsic motivation will be high.
To elaborate on control and intrinsic motivation, people who have an internal locus of
causality, feel that they participate in an activity because they want to (high intrinsic
motivation); whereas, people who have an external locus of causality, feel that they
participate in an activity because of an external cause (i.e. playing for the money).
Therefore, “paying people off” can decrease their intrinsic motivation, because they
may attribute their behavior to the money, rather than self determination. An
example of something that would promote an internal locus of causality would be a
coach who gives his athletes greater input on what drills are performed during
practice. An example of controlling someone through a reward would be telling
someone that you will give them such and such for doing such and such (external
regulation). For example, telling a kid, "if you clean your room, I'll give you five
dollars." In this scenario, the reward is controlling the persons behavior, rather than
self determination.
The key aspect of this theory is the perception of the receiver of the reward.
Someone who gives an athlete a reward may have the best intentions for the
athlete; but if it is perceived as controlling, it will decrease intrinsic motivation. For
example, if a player perceives that he/she is being rewarded with cars or money,
only so that he will stay on the team, this will decrease his/her intrinsic motivation,
because they will perceive the reward as controlling. Additionally, the reward giver
should stress the informational aspect of the reward, making sure the player knows
that the reward was given to them to show a sign of appreciation for their hardwork.
Rewards
Based on his research, the current author suggests that a reward can be defined as
an external agent administered when a desired act or task is performed, that has
controlling and informational properties. While rewards are typically delivered to
increase the probability of a response, they can increase or decrease the probability of an event occurring, depending on the saliency and direction of the controlling and
informational aspects of the reward. Saliency would refer to the intensity of either
the controlling or informational aspect of the reward. It is what stands out to you the
most. While direction would refer to whether the reward is perceived as increasing or
decreasing the athlete’s control; and whether the information is perceived as positive
(increasing the athlete’s competency) or negative (decreasing the athlete’s
competency). Rewards can come in the form of verbal rewards (i.e. telling someone
“good job!”), physical rewards (i.e. a pat on the back), or tangible rewards (i.e.
giving someone money, food, or a medallion), among others. There are 5 basic types
of rewards discussed in the literature as follows (Deci, Koestner, & Ryan, 1999):
• Task-Non contingent rewards
• Engagement contingent rewards
• Completion contingent rewards
• Performance contingent rewards
• Unexpected rewards
Task-Non contingent rewards are rewards given for just showing up for the study. In
an experiment, a participant may be paid to just show up for the experiment, but
they are not required to do anything. They could just sit around the entire time.
Engagement contingent rewards are rewards given for just participating in an
activity, and not necessarily completing it. For example, an experimenter may pay a
participant just to participate in an activity that involves making a puzzle, but they
do not have to complete or perform well on the puzzle.
Completion contingent rewards are rewards given for completing a task.
Performance contingent rewards are rewards given for performance, usually based
on a normative value. For example, doing better than 80% of the participants in a
study. A sub category of performance contingent rewards are competitively
contingent rewards. They involve rewarding individuals for defeating others.
Unexpected rewards occur when participants receive a reward after performing a
certain behavior, but were not expecting to receive a reward.
CET makes several predictions on these types of rewards.
This theory predicts that unexpected rewards would not decrease intrinsic
motivation, because the participant performed the task without knowledge of the
reward; therefore, the controlling aspect of the reward would not be as salient, and
participants would attribute their participation in the activity to an internal locus of
causality. Because of the informational aspect, unexpected rewards may also
enhance intrinsic motivation. But the administrator of the reward would have to give
it based on high performance, and stress the informational aspect for it to be
beneficial.
Non-contingent rewards do not require participating in the task, completing the task,
or performing well on the task, and they deliver no information about the person’s
competency. Therefore, CET predicts they will not effect intrinsic motivation.
Both engagement and completion contingent rewards are predicted to typically cause
the highest decrease in intrinsic motivation. This is because they contain a high
controlling aspect, but deliver no information about the competency of the individual.
For example, you could be paid for participating in an activity, but whether you
perform well or not, is irrelevant. Therefore, these types of rewards say nothing
about the person’s competency, and decrease their control.
Performance contingent rewards are a lot more complex. While they can decrease
control as engagement and completion contingent rewards do, they also relay a
sense of competency. Therefore, CET predicts that if the information aspect is more
salient for performance rewards, it may be able to counteract the controlling aspect
of the reward. Additionally, whether the message is portrayed as controlling or not
will also determine whether the reward decreases intrinsic motivation or not.
Therefore, it is predicted by CET that generally, these rewards will decrease intrinsic
motivation less than engagement and completion contingent rewards do.
Lastly, CET would suggest that verbal and physical rewards should enhance intrinsic
motivation. This is because the informational aspect is very salient, while the
controlling aspect is typically low. However, the controlling aspect can vary
significantly depending on the delivery. This will be discussed more in practical
applications.
Important note
This article is dealing with the effects external rewards have on intrinsic motivation
for activities that are already intrinsically motivating to the individual. Rewards have
an entirely different effect on activities that are boring to an individual. Therefore,
these results should not be generalized to other scenarios. It also primarily makes
predictions about when rewards are removed. Because while the rewards are being
administered, this is a test of both extrinsic and intrinsic motivation. Lastly, CET does
not make predictions on studies that combine verbal rewards with tangible rewards,
because they would contribute to intrinsic motivation in opposite directions.
Studies on the Effects of External Rewards on Intrinsic Motivation
In a revolutionary study on the effects of external rewards on intrinsic motivation,
Deci (1971) had participants play on an inherently interesting task, called the SOMA
puzzle. Participants were paid to play, were given verbal rewards (i.e. verbally
encouraged), or received no reinforcement for participating. Results found that
participants who were paid money to play spent a significantly less amount of time
(100 seconds) than participants who were not rewarded to play. Whereas, those who
received verbal rewards, played for a longer amount of time than both conditions
(intrinsic motivation increased).
Soon after this, Lepper and Greene (1975) performed a study called “Turning Play
Into Work.” Participants were nursery school children, who drew with felt pens, an
intrinsically motivating activity. Participants were divided into three conditions.
Condition one was the expected reward condition, in which the children agreed to
draw a picture in order to receive a good player certificate (completion contingent
rewards). In the unexpected reward condition, participants were unexpectedly given
a reward after they completed the task. In the control condition, children received no rewards for participating. One week later, participants were brought back in to play
with the felt pens without any rewards. Results found that children in the expected
reward condition decreased in intrinsic motivation; whereas, children in the other
two conditions maintained intrinsic motivation.
Limits must often be set for various activities, such as education. In this context,
Koestner, Ryan, Bernieri, and Holt (1984) investigated if limits could be set without
decreasing intrinsic motivation. Participants consisted of 44 first and second grade
students, engaging in a painting activity. Participants were given controlling verbal
limits, informational verbal limits, and no verbal limits. Controlling verbal limits
consisted of using various phrases such as, “I am going to tell you what you have to
do now” “We have rules you have to follow” “I want you to be a good boy or girl.”
Whereas, children in the information condition were just told what the experimenters
wanted them to do, instead of what they had to do. For example, “The smaller sheet
is for you to paint on, the larger sheet is a border to be kept clean.” Results found
that intrinsic motivation, enjoyment, creativity, and quality of the artist product were
decreased from the control condition, but not by the informational condition.
Therefore, informational constraints may be an effective way to regulate behavior.
This study supported cognitive evaluation theory. CET would predict that as control
decreased, intrinsic motivation would decrease. Other studies have also supported
these findings (Plant and Ryan, 1985; Deci, Koestner, & Ryan, 1999; Deci, Koestner,
& Ryan, 1999, b).
While the purpose of this article is to discuss the effects of rewards on intrinsic
motivation, CET has incredible implications for many areas of life. Rewards are just
one of a broad array of techniques in which we control people and decrease their
intrinsic motivation. Results show deadlines, evaluations, and imposed goals, all
decrease intrinsic motivation (Deci, Koestner, & Ryan, 1999; Deci, Koestner, & Ryan,
1999, b; Ryan & Deci, 2000). Whereas, provision of choice, and acknowledgement of
feelings, enhance intrinsic motivation (Deci, Koestner, & Ryan, 1999). These results
are all centered on one key concept—control. It appears that using controlling
language and actions consistently lowers intrinsic motivation. Plant and Ryan (1985)
suggest that control is “any vocalization that pressures a person to behave a certain
way.” It is suggested that contextual factors that tend to control someone—such as
pressure to feel, think or behave a certain way—lowers intrinsic motivation. For
example, a mother who says, “you know what you should do” would be controlling
compared to “what do you think you should do?” Or telling someone “Good job, you
should keep up the good work” or “Good job, you are doing as you should.” One
experiment found that saying (Deci, Koestner, & Ryan, 1999) “I have not been able
to use most of the data I got so far, but if you keep it up, I may be able to use
yours” puts pressure on the person, decreasing his/her control. While the other
participants were told, “compared to most of my subjects, you are doing really well”
which resulted in more control, and higher intrinsic motivation. Therefore, how
information is relayed is vital for how an individual perceives rewards and
instructions. Numerous future articles can be written on the implications these
theories have not only on bodybuilding, but everyday life—stay tuned.
Scholarships
The famous basketball player, Magic Johnson, was asked if he had received rewards
when being recruited by college basketball teams. He stated that (Weinberg & Gould,
2003) “I received my share of offers for cars and money. It immediately turned me
off. It was like they were trying to buy me, and I don’t like anyone trying to buy
me.” According to CET theory, what Magic was really annoyed with was the
controlling aspect of rewards. With hundreds of multi-million dollar contracts thrown
around in sports today, of immense interest would be the effect of rewards on
athlete’s intrinsic motivation. To investigate this question, there has been a
significant amount of research on the effects of athletic scholarships on intrinsic
motivation.
In one of the earliest studies on the effects of athletic scholarships on intrinsic
motivation, Ryan (1977, 1980; as cited in Frederick & Ryan, 1995) observed
collegiate Division I football players on scholarships and not on scholarships. Ryan
found that players who had scholarships did not enjoy football as much as nonscholarship
athletes; further, scholarship athletes had less intrinsic motivation every
year they had a scholarship, so that they had their lowest level of intrinsic motivation
during their senior year!
However, in a similar case, results found that male wrestlers and female athletes
from six different sports had higher intrinsic motivation than athletes who were not
on scholarships.
These results may seem contradictory at first glance. However, CET may offer some
helpful explanations.
As you recall, according to CET, there are two properties to a reward that influence
intrinsic motivation—control and information. Therefore, CET would predict that the
controlling and informational aspects of these rewards were different.
While scholarships can clearly be perceived as controlling, they also relay information
of competency. Now, what determines whether a scholarship would decrease or
increase intrinsic motivation is the saliency of the controlling and informative aspects
of the reward. If the controlling aspect was more salient, intrinsic motivation would
decrease; conversely, if the informative aspect of competency was more salient,
intrinsic motivation would increase.
Consider this. In the 1980’s, when this study was composed, athletic scholarships for
Division I football teams was common. However, it was rare for wrestlers and female
athletes to receive scholarships. Thus, the normality of a scholarship for Division I
football players, would have decreased the saliency of the informative properties of
the reward; however, the rarity of a scholarship for wrestlers and female athletes,
would highlight the informative message of competency.
This hypothesis was supported in a more recent study on athletic scholarships
(Amorose, Horne, & Miller, 1994). 440 male and female athletes in Division I sports
were analyzed. Results found that players on scholarship had lower intrinsic
motivation, enjoyment, and perceived choice than non-scholarship athletes. CET
would predict that the increased number of scholarships offered to female athletes
today, decreased the saliency of the informative aspect of scholarships for them,
resulting in a decrease in intrinsic motivation, similar to male football players.
Now, as the reader recalls, according to CET, it is the perception of the reward that
causes an athlete to increase or decrease in intrinsic motivation. In this context, if a
coach uses scholarships as leverage over their athletes—i.e. threatening that they
will lose their scholarships if they don’t perform well—then this would emphasize the
controlling aspect of the reward, decreasing intrinsic motivation. Therefore, coaches
should be extremely careful of how they treat the scholarships. Coaches should
emphasize the informational aspect, and downplay the controlling aspect, instead of
using scholarships as leverage over their athletes.
In this context, Amorose and Horn (2000) teamed up again to see if it was the
reward or coaching behaviors that decreased intrinsic motivation. What they found is
that the coaches’ behaviors were more influential on intrinsic motivation. For
example, if a coach was perceived as a tyrant over practice, players had lower
intrinsic motivation; whereas, coaches who were perceived as positive, and
democratic, increased intrinsic motivation. Therefore, this study suggests that the
coach has a bigger impact on intrinsic motivation than the scholarship itself.
Meta-Analyses
A meta-analysis is a statistical practice of combining the results of a number of
studies. Considering that literally hundreds of studies have been done on the effects
of external rewards on intrinsic motivation over the past 30 years, a summary of the
data through a meta-analysis would be of immense help to our research.
Thankfully, a number of scientists have done just this.
Rummel and Feinberg (1988) examined 45 studies between 1971 and 1985. They
found that extrinsic rewards significantly decreased intrinsic motivation. However,
there were some flaws in this meta-analysis. For example, they did not distinguish
between various rewards (i.e. expected vs. unexpected; contingent vs. noncontingent,
etc.).
Wiersma (1992) performed another meta-analysis on 20 published studies between
1971 and 1990. Results also found that rewards significantly decreased intrinsic
motivation. However, there were also flaws in this. First, the sample size of studies
was small; second, as with the former meta-analysis, they did not distinguish
between rewards; and third, many of the studies did not have control groups, and
instead only compared reward conditions to each other.
Following this, Tang and Hall (1995) performed another meta-analysis, which was
much more comprehensive than its predecessors. They examined 50 studies, and
distinguished between rewards. They found that task-contingent rewards decreased
intrinsic motivation the most, followed by performance contingent rewards; however,
unexpected rewards did not effect intrinsic motivation. They also found that verbal
rewards increased intrinsic motivation. Further, they found that rewards may
enhance intrinsic motivation for boring tasks. All of these findings supported CET.
Flaws in the experiment were that some of the measures were not purely of intrinsic
motivation. This is because they included results about the quality, quantity, and
ratings of how much participants enjoyed activities during the reward phase. Again,
this would be a measure of both intrinsic and extrinsic motivation. In order to assess
intrinsic motivation, you would need to test participants during another trial, when
rewards were not administrated.
The importance of this distinction is highlighted by the previously mentioned metaanalysis
by Wiersma (1992). Wiersma sought to clear up the supposed contradictory
studies on the effects of rewards on intrinsic motivation. It was hypothesized that
the conflicts seen could be attributed to differences in experimental designs. They
found that while the reward is being administered, intrinsic motivation does not
decrease (though intrinsic motivation is difficult to measure in this situation; this will
be discussed momentarily); however, when the reward is removed, intrinsic
motivation does decrease.
The problem is this: performance during the administration of the reward is not a
pure measure of intrinsic motivation. It is a measure of both intrinsic motivation and
the reward. When the reward is administrated, it would be predicted that if
performance increases, the extrinsic reward was stronger than intrinsic motivation; if
performance decreases, the extrinsic reward was weaker than intrinsic motivation;
and if performance remains the same, the extrinsic reward was equal to intrinsic
motivation. Therefore, a "free time" trial must be performed, in which the reward is
removed to accurately asses intrinsic motivation.
In summary, though these three meta-analyses had several methodological short
comings, they all supported CET.
So far, evidence strongly supports CET. But what would a scientific discussion be
without a little controversy? Don’t fret; this was supplied in a mighty way through
Cameron and Colleagues from the behaviorist camp.
Cameron and Pierce (1994) performed a meta-analysis on the effect of rewards on
intrinsic motivation. Two years later, Cameron and Eisenberger performed an almost
identical analysis, with a few slightly modifications. I will therefore refer to these
studies as Cameron, Eisenberger, and Pierce henceforth, unless stated otherwise.
There results were almost entirely contradictory to every shred of evidence shown in
this article. They concluded that rewards essentially do not effect intrinsic
motivation, and that there was no reason not to use rewards to regulate behaviors,
particularly in educational settings. And here is their boldest statement of all: they
suggested the “abandoning of cognitive evaluation theory”!
Since this meta-analysis was published, several scientists have harshly critiqued the
methods and conclusions made by Cameron, Eisenberger, and Pierce. In 1996, three
commentaries (Kohn, 1996; Lepper, Keavney, & Drake, 1996; Ryan & Deci, 1996)
argued that Cameron, Eisenberger, and Pierce’s meta-analysis was invalid and that
its conclusions false. More recently, the inventors of CET themselves stepped in and
critiqued this meta-analysis.
Deci, Koestner, and Ryan (1999) reported several flaws in this analysis. Some
notable comments were that this analysis combined the effects of boring tasks, with
intrinsically interesting task. As has been clearly displayed, CET makes no predictions
on boring tasks; only inherently interesting tasks. For, how would extrinsic rewards
undermine intrinsic motivation for tasks which have none, or very little intrinsic
motivation to begin with? Further, as show in the Tang and Hall (1995) metaanalysis,
rewards may actually be of benefit for boring activities. Therefore, this is
comparing apples and oranges, and would greatly skew the results. Another problem
was that they mixed some rewards together, and misclassified them. For example,
they included a reward from a study as a non-contingent reward, that actually was
administrated as an engagement contingent reward. They also eliminated 20% of the
studies as outliers, instead of isolating the cause of variability. Other critiques were
made, but clearly there were many flaws in this analysis.
A flaw in all these meta-analyses is that they did not include unpublished doctoral
theses. Doctoral theses are commonly the most comprehensive papers conducted;
thus, inclusion of these papers would increase the reliability of the analysis.
In light of the various flaws in these meta-analyses, Deci, Koestner, and Ryan (1999)
teamed up to compose the most comprehensive meta-analysis on rewards to date.
128 studies were examined. Participants ranged from college to pre-school students.
There were a wide array of rewards administrated, including marshmallows and
dollar bills.
Results found that verbal rewards, when not administered in a controlling manner,
increased intrinsic motivation. However, when they were administered in a
controlling manner, they decreased intrinsic motivation.
Both engagement and completion contingent rewards decreased intrinsic motivation.
Performance contingent rewards also decreased intrinsic motivation; but not as much
as the previous two rewards. However, participants in the performance contingent
condition, who got less than the maximum reward, showed a sharp decline in
intrinsic motivation—more so than any other group! This is because not only are they
being controlled by the reward, but they are being delivered information that they
are incompetent. Since in many cases, the majority do not achieve the standard of
excellence taken to obtain a performance contingent reward, these can be especially
dangerous to administer. Results also found a significant difference between the
effects of rewards on boring and intrinsically motivating tasks. This is in agreement
with the findings of Tang and Hall (1995).
Unexpected rewards and non-contingent rewards did not decrease intrinsic
motivation; presumably because the participants were not doing the task in order to
get those rewards, therefore, they did not feel controlled by them.
It has also been suggested that the effects of rewards on intrinsic motivation may be
transitory. But their results indicated they were long lasting.
One interesting finding was that rewards were more undermining to intrinsic
motivation for kids than college students. The authors suggested 3 explanations: 1.)
Kids are “bribed” more often than adults for performing activities, so it may have a
deeper effect. 2.) College students are more cognitively advanced, so they may be
able to distinguish the controlling aspect from the informational aspect of rewards
better than kids. 3.) College students may have different expectations during
experiments, because they have more experience in them than kids.
The authors suggested that verbal rewards are effective, in part because they are
unexpected rewards. In this context, if a coach always praises their athletes for
performing an activity, the athletes may come to expect praise, and CET would
predict that intrinsic motivation would decrease. In fact, they sourced 3 studies
which indicated that when participants were told they would be given performance
feedback, and then were given positive feedback (verbal rewards) after completion
of the task, intrinsic motivation significantly decreased! So evidence suggests that
you should give verbal rewards in an intermittent fashion. This is further supported
by studies on partial reinforcement (refer to Hull’s Quantitative Equation on Human
Performance).
In response to this meta-analysis, Cameron, Eisenberger, and Pierce (1999) teamed
up again to make 2 more meta-analyses. But it was more or less the same. Deci,
Koestner, and Ryan (1999, b) again responded with a convincing argument refuting
the claims made by them. The largest point that Cameron, Eisenberger, and Pierce
(1999) tried to get across was that performance-contingent rewards may enhance
intrinsic motivation. This was supported by a recent study by Cameron, Pierce,
Banko, Katherine, and Slyvia (2003) who found that giving rewards that are tied to
meeting increasingly demanding performance standards enhanced intrinsic
motivation. However, Deci, Koestner, and Ryan (1999, b) pointed out several flaws
in this meta-analysis. The primary differences were that they excluded 11 studies out
of 40 which indicated negative results, and had several other methodological flaws,
which had actually been pointed out in previous critiques, but were not corrected.
In summary, the results appear to strongly support CET, and suggest that we should
be very careful about how and if we use external rewards.
The Link Between Self Determination Theory, CET, & Goal Orientations
In another article in this issue of JHR, the current author addressed various goals. It
is a pre-requisite to study that article to be able to understand the topic that is about
to be discussed. Click Here, to study this article.
It has been suggested by several authors that there may be a link between self
determination theory and goal orientations (Nikos, 2001). Evidence suggests than an
ego (outcome) orientation can undermine intrinsic motivation; whereas, a mastery
goal orientation facilitates intrinsic motivation (Ryan, 1982; Brunel, 1999). This is
because athletes with an ego orientation are concerned more with the outcome of an
event, rather than the activity itself (Nicholls, 1989). For instance, Brunelt (1999)
found that an ego orientation increased introjected regulation (participating in an
activity because of various pressures) and external regulation (participating in an
activity for a reward) in 160 badminton athletes. Similarly, Ryan (1982) found that
an ego orientation decreased intrinsic motivation for participating in a puzzle game.
Conversely, a mastery orientation appears to enhance intrinsic motivation. This is
because a mastery orientation focuses on intrinsic properties of the activity, rather
than an external motivator. In this context, Butler (1987) found that a mastery goal
orientation enhanced intrinsic motivation, by promoting task persistence and
challenge seeking activities.
Various studies have supported this hypothesis. Duda, Chi, Newton, Walling, and
Catley (1995) found that a mastery goal orientation enhanced intrinsic motivation,
while an ego goal orientation maintained or decreased intrinsic motivation. Studies
on British and Romanian PE have found that students with the highest intrinsic
motivation were those with a high mastery orientation, regardless of their ego
orientation.
And a plethora of other studies have demonstrated this (refer to the aforementioned
goal setting article for more).
In order to summarize this data, Nikos (2001) performed a comprehensive analysis
on the link between Self Determination Theory and Goal orientations. Over 247
British University students were examined. Results indicated that a mastery
orientation was able to predict all three types of intrinsic motivation and identified
regulation (as you recall, this is an autonomous form of extrinsic motivation, and
related to enhanced self determination). However, a mastery orientation did not
predict introjected regulation (i.e. avoidance or participation in activities because of
guilt, pressure, etc.), external regulation (participation in activities purely for
rewards), or amotivation (a lack of motivation).
As expected, an ego orientation was able to predict introjected and external
regulation. This is most likely because athletes with an ego oriented mentality
participate in sport for extrinsic rewards such as acknowledgment, superiority, etc.
(Nichols, 1989). Additionally, an ego orientation was not a predictor of intrinsic
motivation, or other self determined measures (except, it did predict intrinsic motivation for sensations). This is attributed to the controlling nature of their
extrinsic goals. Lastly, ego orientations did not predict amotivation.
Nikos also examined the interaction between goal orientations, as individuals can be
both mastery and ego oriented. Results found that when participants master
orientation was high, external regulation was lower when ego orientation was lower.
Further, high mastery oriented individuals predicted high intrinsic motivation
regardless of ego orientation.
Thus, evidence suggests that a mastery goal orientation is optimal for facilitating
intrinsic motivation, while an ego orientation will tend to enhance extrinsic
motivation. These results are clearly explained using Self Determination Theory.
Now, an outcome oriented mentality often promotes competition (for information on
competition vs. cooperation Click Here). And CET would predict that competitive
mentalities can lead to a decrease in intrinsic motivation, as they focus on external
goals such as winning. In this context, Fortier, Vallerand, Briere & Provencher (1995)
tested CET with competitive athletes compared to recreational athletes. Results
indicated that recreational athletes had greater intrinsic motivation, while
competitive athletes had greater identified regulation and amotivation. Similarly,
Kavussanu and Roberts (1996) found that a competitive, outcome oriented
environment decreased the intrinsic motivation of tennis players.
The next question is, what is the effect of rewards on competitive situations. These
type of rewards are called competitively contingent rewards, and are a sub category
of performance contingent rewards. They involve rewarding individuals for defeating
others. In this context, Pritchard, Campbell and Campbell (1977) found that giving
participants a $5 reward for defeating participants in their group (approximately 6
participants per group) decreased intrinsic motivation relative to no reward. CET
predicts these types of rewards are very controlling, as they focus the participants
locus of causality on winning, which is an external motivator. Other studies have also
supported these findings (Ryan, Mims, & Koestner, 1983).
Thus, Self Determination Theory and CET are very helpful in explaining the results of
goal orientations and competition on intrinsic motivation.
Conclusion
The results clearly indicate that the rewards can greatly undermine intrinsic
motivation. For practical applications, click Here.
Comments