Content uploaded by Amber Todd
Author content
All content in this area was uploaded by Amber Todd on Jul 03, 2015
Content may be subject to copyright.
1
May 20, 2015Journal of Emerging Investigators
Journal of
Emerging Investigators
Received: June 22, 2014; Accepted: February 5, 2015;
Published: May 20, 2015
Copyright: (C) 2015 Greene and Todd. All JEI articles
are distributed under the attribution non-commercial, no
derivative license (http://creativecommons.org/licenses/
by-nc-nd/3.0/). This means that anyone is free to
share, copy and distribute an unaltered article for non-
commercial purposes provided the original author and
source is credited.
Introduction
The effects of positive and negative reinforcement
on different forms of performance have been studied
and debated for years. Positive reinforcement is
generally dened as reinforcing a specic behavior
by providing a positive reward for that behavior (1).
Negative reinforcement is generally dened as removing
a negative experience that causes discomfort or
harm by performing a specic behavior (1). Negative
reinforcement could be argued as a serious instigating
factor, but also as a discouraging factor, while views
on positive reinforcement could var y from seeing it as
a natural aspect of human decency to unnecessary
coddling.
Reward versus punishment as a means of motivation
is both a psychological and biological question and
has had varying results, depending on the subject and
situation. In one study, reactions and performances
were different, even between very close age groups.
Researchers studied ‘social comparison,’ a process
where a person nds where they stand compared to
those around them. Positive social comparison led
to greater success than negative social comparison,
and the study reported higher self-reinforcement after
positive comparisons, no matter the age, sex, or reward
(2). These positive inuences could suppor t the idea that
optimism, where through compliments or the desire for
candy with no consequences, benets performance. In
one experiment testing substance-dependent individuals
(SDI), negative reinforcement motivated subjects during
a decision-making task, and it was found that SDI
were more sensitive to the frequency, rather than the
magnitude, of loss while control subjects responded to
both (3). Another study found that subjects had better
academic success after seeing their previous work
in prosperous projects, which could ser ve as positive
reinforcement (4). However, in another study, neural
activity in three sections of the brain in subjects in the
case of reward or loss was very similar. Additionally,
this neural activity did not change with the magnitude
of the reinforcement (5). This is a surprising discovery,
given the repeated ndings that positive and negative
reinforcements seem to have at least marginally different
effects on performance. Unfortunately, with conicting
data, a solid conclusion has yet to be reached. These
previous studies have suggested that while positive
reinforcement in groups does have a positive effect
on the subjects, brain activity, as monitored by fMRI,
displays little difference (5). Negative reinforcement also
may have a positive effect on subjects compared to the
control treatment, but it is not fully understood why it
may be less successful than reward tactics. As a result,
researchers remain in search for a more denitive,
comprehensive understanding of how different forms of
reinforcement affect performance, and why each form
does so.
This study attempted to nd if and how positive and
negative reinforcement affected the mental performance
of sixth graders on a math test. This could aid teachers
in planning their class style and motivation techniques,
as well as how they handle student achievement or
failure. How students learn is a major component of
the learning process, and nding ways to increase
The Eect of Positive and Negative Reinforcement on Sixth
Graders’ Mental Math Performance
Summary
We examined the relationship between dierent forms of
reinforcement and subjects’ mathematical performance.
Thirty sixth-grade students were asked to mentally
calculate answers to fraction problems, while holding a
heart rate monitor, after receiving a positive, negative,
or neutral form of reinforcement. While the results
suggest that any motivation, positive or negative, may
result in higher grades, the positive reinforcement group
displayed the largest gain. However, the results were not
statistically signicant, though the trend was consistent
with prior research. The heart rates of the positively-
reinforced group were signicantly lower than those
of the negatively-reinforced and control groups. These
results suggest that motivation through reward, rather
than punishment, has a greater eect on the performance
of students academically and may increase their grades,
as well as lower heart rates during assessments.
Tess M. Greene1and Amber Todd2
1Dayton Regional STEM School, Dayton, OH
2Wright State University, Dayton, OH
2
May 20, 2015Journal of Emerging Investigators
Journal of
Emerging Investigators
student success is an investment in the next generation.
Given that previous studies found positively-reinforced
subjects had higher performance (2, 4) and that negative
reinforcement did have a smaller but noticeable positive
effect (3), we hypothesized that students given positive
reinforcement would have the highest average score
and negatively-reinforced students would do better on
average than the control group. Although the data of this
study are consistent with the hypothesis, the increases
were not statistically signicant. These results suggest
that motivation through reward, rather than punishment,
has a greater effect on the academic performance of
students and may increase their grades.
Results
Since the impact of positive and negative
reinforcement on students has not yet been established,
this study attempted to determine how reinforcement
type affected the mental math performances of sixth
grade students at a Midwest suburban STEM (science,
technology, engineering, and mathematics) school. The
students were each given a twenty-question fraction test
(Table 1A). During the test, the students held a heart
rate monitor with their hands and answered the math
questions orally. Prior to the test, the thirty students were
divided into three groups: control (no reinforcement),
positive (positive reinforcement), and negative (negative
reinforcement). Prior to test administration, students in
the positive-reinforcement group were told that if they
earned an “A” on the test, they would receive candy;
students in the negative-reinforcement group were told
that if they did not meet their teacher’s expectations,
they would come in at recess and after school to study
fractions; students in the control group were told that
their results were anonymous and would not affect
their academic career. Each group represented a mix
of ethnicities, genders, backgrounds, and academic
performances (Table 1B). To ensure similar ability
levels between groups, each group had a similar (within
0.36%) average on their current math class grade
(Figure 1). We hypothesized that students given positive
reinforcement would have the highest average score,
followed by negatively-reinforced students and then the
control group.
The average score of the control group on the oral math
exam was 56.5% correct (Figure 1). For the negative-
reinforcement group, the average score was 61.5%
correct (Figure 1); the average score of the positive-
reinforcement group was 71.5% correct (Figure 1). The
positive-reinforcement group had the highest average
score, followed by the negative-reinforcement group.
Both reinforcement groups had a higher score than the
control group, which received no type of reinforcement.
These ndings are consistent with our hypothesis that
students given positive reinforcement would have the
highest average score, followed by negatively reinforced
Table 1: Oral Test Questions and Student Demographics. A) Oral test questions given to students and the
correct answers. B) Student Demographics. Students listed by number, displaying reinforcement type (green =
positive, pink = negative, yellow = control), gender, ethnicity, in-class math class grade (percent), and oral exam
performance compared to current math class grade.
3
May 20, 2015Journal of Emerging Investigators
Journal of
Emerging Investigators
students, and then the control group.
While a trend consistent with the hypothesis was
found in the data, there was no signicant difference
between the performance of the groups (ANOVA, F =
0.89, p = 0.422), perhaps due to the small sample size
(n = 30) and high variability in the groups, as shown
by the large range of the oral exam scores (Figure 1).
The control group’s scores ranged from 100% correct
to 5% correct; the negative-reinforcement group’s
scores ranged from 80% to 25% correct; the positive-
reinforcement group’s scores ranged from 100% to
40% correct. The high variability in each of the groups,
combined with the small number of test subjects, could
have resulted in the statistical insignicance, despite the
trend being consistent with the hypothesis.
Looking at individual student per formance, only four
students of the thirty performed at the same level or
better on the oral math exam as in their current math
class (Table 1B). Two students who performed at the
same level both had a class grade of 100% and received
100% on their oral exam (Table 1B, students 1, 2). Two
students had oral exam scores that were higher than
their class grade (current class grades of 88.71% and
90.32%; their oral math exam scores were 90% and
95%, respectively) (Table 1B, students 21, 20). All four of
these students were in either the positive-reinforcement
group or the control group (one same, one better in each
group). All other students, including all the students in
the negative reinforcement group, had oral exam scores
that were lower than their current math class grade
(Table 1B).
Students’ heart rates during the oral exam were
also recorded. The average heart rate of the control
group while taking the oral exam was 98.28 BPM (beats
per minute); the average heart rate of the negative-
reinforcement group was 96.72 BPM; the average heart
rate of the positive-reinforcement group was 87.17 BPM
(Figure 2). The positive reinforcement group had the
lowest average heart rate, followed by the negative
reinforcement group. Both groups had a lower average
heart rate than did the control group (Figure 2). Average
heart rates for the groups were found to be statistically
signicantly different (ANOVA, F = 4.05, p = 0.029).
The positive reinforcement group had a signicantly
lower average heart rate than did the control (t-test, p
= 0.005) and the negative-reinforcement groups (t-test,
p = 0.042). Due to students completing the oral exam at
different speeds, heart rates for specic questions could
not be compared across the groups.
Discussion
While not statistically signicant (ANOVA, F = 0.89,
p = 0.422), the trend in the data suggests that with
both positive and negative reinforcement, students are
more likely to perform better on an oral math exam than
controls, potentially with positively-reinforced individuals
having the greatest performance. This indicates that
these students may have responded with more drive to
succeed when there was a reward given for reaching
a certain mark. These results matched the expected
results: that students given positive reinforcement
would have the highest average score and negatively-
reinforced students would still do better on average than
the control group.
Surprisingly, many of the students who had the
highest scores in class were not the top scorers in
the experiment. Two subjects in the control group,
whose scores placed them at the top of their class,
had experimental scores 40% and 45% less than their
class grade. One negatively-reinforced subject, who
possessed a perfect score in class, actually had a lower
score in the experiment than did a student whose class
score was nearly failing. In the positively-reinforced and
control groups, only two subjects scored higher than
or equal to their class score (Table 1B) and none in
Figure 1: Type of Reinforcement Versus Average Score
of Subject Groups. Student scores for in-class math grades
(blue) and oral exam grades (green) were averaged for each
group. Error bars represent range of scores for group.
Figure 2: Type of Reinforcement Versus Average Heart
Rate of Subject Groups. The mean hear t rates of students
were averaged by group. Error bars represent standard
deviation. * denotes a signicant difference from the positively-
reinforced group.
4
May 20, 2015Journal of Emerging Investigators
Journal of
Emerging Investigators
the negatively-reinforced group did (Table 1B). These
radically different scores could be due to the different
motivations, or the different form the test was taken
in: orally versus written. These ndings may indicate
students nd oral math exams more difcult than written
work.
Oral testing could also have af fected students’ hear t
rates. While the heart rates of the positively-reinforced
subjects were signicantly lower than those of the other
groups (Figure 2), the difference between them (or
the similarity of the negatively-reinforced and control
subjects) may be due to the fact that the students had
not been tested this way before. Nervousness may have
arisen in all groups and shifted the rates, not because
of the actual questions, but by the way they were given.
However, considering the large difference between the
positively-reinforced subjects and the two other groups,
it is possible that the possibility of reward, as well as
having nothing to lose, kept those subjects more calm
than their counterparts. Students being encouraged
through fear of punishment may exhibit nervousness or
be distracted by their worries, while students who are
being rewarded for success may feel that they have less
to lose and may be able to focus more easily. However,
students not given any reason, punishment or reward,
to perform to the best of their abilities may do worse
because they do not feel any motivation to succeed. This
data could potentially be used by teachers to show why
positive reinforcement has a greater ef fect on students’
success and used by child psychologists to describe
how children are motivated.
In future studies, a larger sample size would aid
in counteracting statistical insignicance. Students
could be tested multiple times, to assess dif ferent
reactions pertaining to the situation. Heart rates
could also be carefully monitored to correspond with
individual questions (allowing heart rates to return to
resting between questions), in order to determine if
certain questions make students’ heart rates increase.
Depending on the subjects, different rewards and
punishments may also be issued. Candy may not be
the healthiest reward, but our results indicate that it may
have been an incentive for children, and while extra
studying may not be an incredibly joyful experience, it
be may be able to help students succeed academically.
In conclusion, this study’s ndings on mental math
performance were not statistically signicant, and thus,
cannot be the basis of massive educational reform.
However, the trend found matches previous experiments
and may indicate how different forms of motivation
affect students’ mental math performance. Findings did
indicate that students in the positively-reinforced group
had signicantly decreased heart rates while taking
the oral exam and that most students (26/30 or 86.7%)
performed worse on the oral exam than their current
in-class math grade. These results could be used by
teachers or administrators in lesson and assessment
planning to possibly increase students’ motivation and
reduce heart rate in class or on standardized tests.
Methods
Sixth grade students at a Midwest suburban STEM
school were chosen as test subjects. A twenty-question
test involving adding, subtracting, multiplying, and
dividing fractions was written for the study (Table 1A).
The test contained varying levels of difculty and was
checked by a sixth-grade math teacher for accuracy and
relevance. Fractions were chosen as the topic because
the students had recently completed a unit on fractions,
so the content would be familiar to them. Consent forms
were dispersed amongst the entire sixth grade class at
the Midwest suburban STEM school. Thirty students
who returned the consent forms were chosen for the
study. These students represented a mix of ethnicities,
genders, backgrounds, and academic performances
(Table 1B). The thirty subjects were divided into three
groups of ten students based on their math class grades
shortly after their nal fractions test, which was provided
by the sixth grade math teacher (Table 1B). The groups
were divided to have the closest average grade in their
math class, and the group’s averages were all within
.36% of each other (Figure 1). Each group was classied
as control, positive, or negative reinforcement.
Every student was brought to a small ofce to take
the test during their study hall period. Fraction problems
were shown on index cards to the students. A Vernier
heart-rate monitor was held by each student as they
mentally calculated their answers. Percent correct was
recorded in a spreadsheet that calculated individual and
group scores. Students in the control group were told that
the scores were completely anonymous and would have
no effect on them or their academic career. Students in
the negative-reinforcement group were told that their
teacher would see their scores, and should they not
meet their teacher’s standards, they would be asked to
come in during recess and after school to study fractions.
Students in the positive reinforcement group were told
that if they scored an A, they would receive candy after
all of the scores were collected. These reinforcements
were chosen due to their possible immediate effect and
how much they would motivate students.
References
1. Skimmer, B. F. (1938). The Behavior of Organisms:
An Experimental Analysis. Appleton-Century: Oxford,
England.
2. Appleeld, J. M. (1975). The Effects of Social
Comparison, Contingency of Reward, and Age of
5
May 20, 2015Journal of Emerging Investigators
Journal of
Emerging Investigators
Subject on Self-reinforcement, Self-condence, and
Task Persistence.
3. Thompson, L., Claus, E., Mikulich-Gilbertson, S.,
Banich, M., Crowley, T., Krmpotich, T., & ... Tanabe,
J. (2012). Negative reinforcement learning is affected
in substance dependence. Drug And Alcohol
Dependence, 123(1-3), 84-90.
4. Boyer, W. (2006). Accentuate the Positive: The
Relationship between Positive Explanatory Style and
Academic Achievement of Prospective Elementary
Teachers. Journal Of Research In Childhood
Education, 21(1), 53.
5. Weis, T., Puschmann, S., Brechmann, A., & Thiel,
C. (2013). Positive and negative reinforcement
activate human auditory cortex. Frontiers In Human
Neuroscience, 7842. doi:10.3389/fnhum.2013.0084
