Physical Fitness and Academic Achievement
11

Journal of Exercise Physiologyonline

(JEPonline)




Volume 8 Number 1 February 2005




Pediatric Exercise Physiology
Managing Editor
Tommy Boone, Ph.D.

Editor-in-Chief
PHYSICAL FITNESS AND ACADEMIC ACHIEVEMENT
Robert Robergs, Ph.D.

Review Board
JAMES B GRISSOM.
Todd Astorino, Ph.D.

Julien Baker, Ph.D.

Tommy Boone, Ph.D.
Lance Dalleck, Ph.D.

Dan
Drury,
DPE.
ABSTRACT
Hermann Engals, Ph.D.

Eric Goulet, M.Sc.
Grissom JB. Physical Fitness And Academic Achievement. JEPonline
Robert Gotshall, Ph.D.
Len
Kravitz,
Ph.D.
2005;8(1):11-25 . The purpose of this study was to evaluate the
James Laskin, Ph.D.
relationship between physical fitness and academic achievement. To do
Jon Linderman, Ph.D.
so, scores on the FITNESSGRAM®, a physical fitness test, were
Derek Marks, Ph.D.
compared to reading and mathematics scores on the Stanford
Cristine Mermier, Ph.D.
Achievement Test 9th edition, a standardized norm-referenced
Daryl Parker, Ph.D.
Robert Robergs, Ph.D.
achievement test. Subjects were all 5th, 7th, and 9th grade California school
Brent
Ruby,
Ph.D.
children enrolled in public school in 2002 for whom there was complete
Jason Siegler, Ph.D.
data on both the physical fitness and academic achievement tests. The
Greg Tardie, Ph.D.
sample size was 884,715 students. Results indicate a consistent positive
Ben Zhou, Ph.D.
relationship between overall fitness and academic achievement. That is,



as overall fitness scores improved, mean achievement scores also
Official Research Journal
improved. This relationship between fitness and achievement appeared to
of The American Society of
be stronger for females than males and stronger for higher socio-
Exercise Physiologists
economic status (SES) than lower SES students. Results should be
(ASEP)
interpreted with caution. It cannot be inferred from these data that physical

ISSN 1097-9751
fitness causes academic achievement to improve. It is more likely that
physical and mental processes influence each other in ways that are still
being understood.

Key Words: School Children, Activity, Grades, Learning

---

Physical Fitness and Academic Achievement
12

INTRODUCTION

The health benefits of regular physical exercise are widely acknowledged. For example, a front-page
story in the Sacramento Bee (1) reported that poor diet and physical inactivity might soon overtake
tobacco as the leading cause of death (2). The Health section of CNN.com reported that for cancer
survivors the key ways to prevent the return of the disease included a healthy diet and exercise (3).

Even though the benefits of physical exercise are acknowledged, physical education in public schools
is viewed as an extracurricular activity and physical education teachers have experienced first hand
when money is tight and/or when there is pressure to improve test scores, physical education is one
of the first activities to be cut back or eliminated. However, if there is evidence that physical education
has a direct positive effect on important educational domains such as reading and mathematics, it
could be argued that physical education is not extracurricular. Rather, it is a vital component in
students’ academic success.

It is not the intention of this paper to argue that the importance of physical education is its benefit to
academic achievement. The overall health benefits of organized physical activity are probably much
more important than possible academic benefits. However, when policy makers need to make difficult
decisions about where to spend public funds and administrators need to make decisions about where
to focus resources in a climate of academic accountability, a proven relationship between physical
fitness and academic achievement could be used as an argument to support, retain, and perhaps
even improve physical education programs.

At this time little research has examined the relationship between physical fitness and academic
achievement. One study that reported a consistent and significant relationship between fitness and
achievement (4) had several threats to validity. First, there were validity concerns with the academic
achievement indicator. Academic achievement was based on a non-standardized, subjective five-
point rating scale. Therefore, the meaning of achievement could vary from site to site. Second, the
reported correlations, although statistically significant (i.e., at 0.001, 0.01, & 0.05 levels of
significance) were not impressive. For example, the average correlation between sit-ups and
achievement for girls age 7 to 15 was 0.13. Third, there were unexplained inconsistencies. Sit-ups
showed a statistically significant relationship with achievement for 8, 9, 12, 13, 14, & 15-year old girls
but not for 7, 10, & 11-year old girls. The study’s conclusion of a consistent and significant
relationship between fitness and academic achievement was based on the fact that there were more
statistically significant correlations than there were not.

One review of research that concluded a positive relationship between physical and mental skills
expressed concern that reviewed studies did not demonstrate causality (5). Most reviewed studies
used correlation designs. Reviewed experimental studies had design weaknesses. One study with an
experimental design in that the independent variable was manipulated did not use random
assignment or matching to control for preexisting group differences (6). Another experimental study
employed random assignment but failed to find a statistically significant difference in academic
achievement between experimental and control subjects (7).

Another review of research, that concluded a positive relationship between fitness and achievement,
expressed concern with the methodological weaknesses in studies that used correlation designs (8).
For example, the Vanves study, one of the reviewed studies was an unpublished manuscript where
the sample size was small and the abstract failed to explain how experimental and control subjects
were matched. As for experimental studies, the reviewer argued the failure to find a statistically

---

Physical Fitness and Academic Achievement
13

significant difference in academic achievement between experimental and control subjects was not
problematic. Even though experimental subjects were receiving less instructional time, because they
were being pulled out of class to engage in physical education, they performed as well as, if not
better, than controls.

The failure of experimental designs to find statistically significant differences between experimental
and control subjects is due in part to the difficulty in raising academic achievement. It is very difficult
to raise student achievement, beyond what might be expected, even when that is the specific focus
(9). A study intended to affect achievement indirectly would encounter even more difficulty.

Little research has examined the relationship between physical fitness and academic achievement,
and the research evidence that exists is not strong. One reason research evidence may be scarce is
that it is difficult to obtain valid and reliable measures for both physical fitness and academic
achievement on the same subjects. Another reason is that it is difficult to obtain large representative
samples of students who have both achievement and fitness scores.

The purpose of this study was to evaluate the relationship between physical fitness and academic
achievement. The study attempted to address some of the weaknesses of earlier studies. For
example, this study had valid and reliable measures of both physical fitness and academic
achievement and a large sample size. However, the criticism that correlation studies do not
demonstrate causality is a weakness of the current study.

METHODS

California Law’s Education Code (EC) mandates
…during the month of February, March, April, or May, the governing board of each
school district maintaining any of grades five, seven and nine shall administer to each pupil in
those grades the physical performance test designated by the State Board of Education (10).

In February 1996, the California State Board of Education designated the FITNESSGRAM® as the
required physical fitness test to be administered to California students. The Cooper Institute for
Aerobics Research in Dallas, Texas developed the FITNESSGRAM®.

In spring 2001, the California Department of Education (CDE) began reporting the Physical Fitness
Test (PFT) results each year. All students in grades five, seven, and nine are expected take the PFT.
Students who are physically unable to take the entire test are to be given as much of the test as
possible.

The FITNESSGRAM® provides a number of options for each performance task so all students,
including those with special needs, have the maximum opportunity to complete the test. The PFT
measures five aspects: 1) aerobic capacity, 2) body composition, and 3) trunk strength, 4) upper body
strength, and 5) flexibility. The FITNESSGRAM® requires students to complete the following six tests:
•
one of the options from aerobic capacity
•
one of the options from body composition
•
the curl-up test
•
the trunk lift test
•
one of the options from upper body strength
•
one of the options from flexibility

---

Physical Fitness and Academic Achievement
14

The FITNESSGRAM® uses standards established by the Cooper Institute for Aerobics Research to
evaluate fitness performance. Performance is classified as: (1) in the healthy fitness zone (HFZ)
which means students met the fitness target or (2) needs improvement which means students failed
to meet the target. It is possible on some tests to exceed the HFZ target. Scores exceeding the target
are included with students that have scored within the HFZ.

Since the HFZ standard is considered a minimal level of achievement, students must meet all of the
fitness standards to be considered fit. Overall test scores range from zero, none of the standards
were met, to six, all standards were met or exceeded.

For academic achievement testing, California Law’s Education Code (EC) states
…there is hereby established the Standardized Testing and Reporting Program, to be known as
the STAR Program.
(b) Commencing in the 1997-98 fiscal year and each fiscal year thereafter, and from the funds
available for that purpose, each school district, charter school, and county office of education shall
administer to each of its pupils in grades 2 to 11, inclusive, the achievement test designated by the
State Board of Education (11).

Each spring the STAR program administers four distinct standardized achievement tests: the
California Standards Tests (CST), a standardized norm-referenced test, a Spanish assessment of
basic skills, and the California Alternative Performance Assessment (CAPA). STAR tests were
administered first in the spring of 1998. From 1998 to 2002, the standardized norm-referenced test
was the Stanford Achievement Test 9th edition (SAT/9), Form T. In 2003, the standardized norm-
referenced test was changed to the California Achievement Test, Sixth Edition Survey (CAT/6). The
Spanish assessment of basic skills has been the Spanish Assessment of Basic Education, Second
Edition (SABE/2). Tests are administered to all public school students enrolled in grade 2 through 11.

Collecting demographic information, such as birth date and gender, is part of the test administration
process for both PFT and STAR. This study uses data from tests administered in the spring of 2001
and 2002. Demographic information from the PFT and STAR programs was used to create matched
files. The county/district/school (CDS) code, grade level, birth date, and gender were used to match
student records. Each California school has a unique CDS code. This matched file contained both the
PFT scores and standardized achievement test scores for individual students. As such, these data
were used to evaluate the relationship between overall scores on the PFT and the standardized
achievement tests. There could be errors in the matching process but there was no reason to believe
matching errors biased the results.

Matched files were created for 2001 and 2002. Sample sizes for 2001 and 2002 were 634,112 and
884,715 students, respectively. The sample size increased in 2002 because of increased PFT
participation. Because sample sizes were larger, this study reports 2002 results.

The analysis had two parts. First, for each overall PFT score, which ranged from zero to six, the
average SAT/9 reading and mathematics scores were calculated in normal curve equivalent (NCE)
units. The NCE score is a normalized standard score with a mean of 50 and a standard deviation of
21. NCE scores range from 1 to 99. The average test score by overall PFT score provides an
indicator of the relationship between fitness and achievement. Second, analysis of variance (ANOVA)
was used to test the relationship between the overall PFT score and achievement scores.

---

Physical Fitness and Academic Achievement
15

RESULTS
Cohort Sample Compared to Population
To evaluate whether the matched cohort was representative of the larger population who took the
PFT and the SAT/9, analyses compared PFT and SAT/9 scores for the two groups. Figure 1 shows
the percent of 2002 students achieving the different number of fitness standards for the statewide
population compared to the matched cohort sample. The number of fitness standards achieved and
the overall PFT score are interchangeable terms.
PFT Statewide
PFT Matched Cohort
35.0%
29.7%
28.7%
30.0%
27.7%
27.2%
25.0%
20.8%20.5%
20.0%
13.5%
15.0%
12.6%
10.0%
7.2% 6.2%
Percent of Students
5.0%
2.9% 2.0%
0.6% 0.2%
0.0%
0
1
2
3
4
5
6
Number of Fitness Standards Achieved

Figure 1. Summary of fitness standards achieved for 2002 grades 5, 7, & 9; n = 1,090,248
(statewide) 884,715 (cohort)

To insure comparability of overall PFT scores, only students with complete data were included. If
missing data were included, the overall PFT score, except for a score of six, could have multiple
meanings. A score of less than six could mean (1) the test was incomplete, (2) the student was
absent, or (3) the student failed to achieve the minimum standard on one to six tests. When there was
no missing or incomplete data, the meaning of scores was clear: students attempted all of the tests
and were able to demonstrate minimal competency on the number of tests indicated by the score.

Figure 1 indicates a discrepancy between the 2002 statewide population and the matched cohort
sample. The matched cohort demonstrated a slightly higher performance on the PFT. For example,
only 27% of the population achieved all six PFT standards compared to 30% of the matched cohort.

Figure 1 is negatively skewed. That is, more students achieved higher than lower scores. Even so,
the fitness standard was considered a minimal level of achievement. Therefore, only students who
achieved all six PFT standards were considered minimally fit.

Next, average reading and mathematics NCE scores for the sample and population were compared
to evaluate whether the sample differed markedly from the population in terms of academic

---

Physical Fitness and Academic Achievement
16

achievement. Table 1 compares academic achievement scores for the population to the matched
cohorts.

Table 1. SAT/9 mean test scores for 2002
SAT/9 Reading NCE
SAT/9 Mathematics NCE
Grade Level Statewide Matched Cohort Statewide Matched Cohort
5
47
48
53
55
7
47
48
51
53
9
41
43
52
54

Although the mean achievement scores were close, the matched cohort demonstrated slightly higher
achievement on the SAT/9 reading and mathematics tests than the statewide population. Further
analyses should be interpreted with this score difference in mind.

The Relationship Between PFT and Mean SAT/9 Scores
To evaluate the relationship between physical fitness and academic achievement, average
achievement scores were calculated for each PFT score. Figure 2 shows the average reading and
mathematics NCE scores on the 2002 SAT/9 by the number of fitness standards achieved (i.e., the
overall PFT score) on the 2002 PFT.
SAT/9 Reading
SAT/9 Mathematics
75
70
65
60
60
55
55
51
48
50
46
44
52
42
45
47
40
44
42
40
35
SAT/9 Mean NCE Score
37
38
30
25
0
1
2
3
4
5
6
Number of Fitness Standards Achieved

Figure 2. 2002 grade 5, 7, & 9 SAT/9 mean reading and mathematics NCE scores by overall
PFT scores; n = 890,280 (reading) 888,241 (mathematics).

The average SAT/9 reading NCE score for students who did not achieve any of the fitness standards
was 37. This same NCE score for mathematics was 42. As the overall PFT score improved the mean
reading and mathematics NCE scores also improved. Results indicate a positive relationship between
academic achievement and physical fitness.

---

Physical Fitness and Academic Achievement
17

Although evidence suggests a relationship between physical fitness and academic achievement,
there are more formal tests. Analysis of Variance (ANOVA) was used to test this relationship. Table 2
shows these results.

Table 2. ANOVA results when the dependent variable was SAT/9 reading scores and the
independent variable was the overall PFT score: 2002 Grades 5, 7, & 9
Source
df
SS
MS
F
Pr > F
R Square
Regression
6
9990420.1
1665070.0
4146.9
< .0001
0.0362
Residual
661885 265758622.5
401.5



Total
661891 275749042.6





ANOVA results suggest that some, or all, of the mean reading scores are statistically different from
each other. Tukey’s studentized range test for multiple comparisons revealed that each mean reading
score for each PFT score was significantly different from all other reading score means at P < 0.01.
For example, students who had a PFT score of six had a mean reading score of 52. The mean score
of 52 was statistically greater than the mean score of 47, the mean reading score for students who
had a PFT score of five. Students who had a PFT score of five had a mean reading score that was
statistically greater than students who had a PFT score of four or less. Students who had a PFT
score of four had a mean reading score that was statistically greater than students who had a PFT
score of three or less and so on.

Table 3 shows these same results for SAT/9 mathematics scores.

Table 3. ANOVA results when the dependent variable was SAT/9 mathematics scores and the
independent variable was the overall PFT score: 2002 Grades 5, 7, & 9.
Source
df
SS
MS
F
Pr > F R Square
Regression
6
14822504.8 2470417.5
5617.3 < .0001
0.0487
Residual
658658 289670477.1
439.8
Total
658664 304492981.9





As with the reading scores, the mean scores are statistically different from each other and Tukey’s
studentized range test for multiple comparisons revealed that each mean for each of the PFT scores
was significantly different from all other means at P < 0.01.

Even though the means were statistically
significant from each other the R2 for each of
Table 4. Correlation of 2002 PFT scores and
the ANOVA analyses was small; indicating
SAT/9 scores: Grades 5, 7, & 9.
considerable unaccounted test score
SAT/9
variance. The R2 for reading and
Variables
SAT/9 Reading
Mathematics
mathematics was 0.04 and 0.05 respectively.
PFT score
0.186
0.217

SAT/9 reading

0.768
The correlations between overall PFT scores
and SAT/9 scores were not high. Results are
shown in Table 4.

---

Physical Fitness and Academic Achievement
18

Even though there was a relationship between PFT scores and average achievement scores, SAT/9
scores have a lot of variance within the overall PFT scores, which explains why the R2 data from the
ANOVA analyses were so low.

Figure 3 graphically illustrates the variance of SAT/9 reading NCE scores for students who had an
overall PFT of 6.
8000
7000
6000
5000
4000
3000
Number of Students 2000
1000
0
1 10 15 19 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 59 62 65 68 70 73 78 83 87 93
Reading NCE Score


Figure 3. Distribution of SAT/9 reading NCE scores for students with a PFT of 6; n = 247,538.

Students with an overall PFT score of 6 had an average reading NCE score of 52. Figure 3 shows
many of these students had scores higher and lower scores than 52. The standard deviation was 20.

The PFT student record indicated whether the students failed, passed, or exceeded the standard. If
there is a relationship between physical fitness and academic achievement, then students who
exceed the standard should have higher average scores than students who merely pass.

Figure 4 shows the SAT/9 mean NCE scores by whether students exceeded, met, or failed on five
different PFT tests.

---

Physical Fitness and Academic Achievement
19

Exceed
Pass
Fail
54
52
50.5
50.4
49.6
50
48.7
49.8
48
48.2
47.5
47.0
47.1
46
45.3
44
44.5
43.5
42
42.9
43.1
42.8
40
41.3
SAT/9 Mean Reading NCE Score 38
Aerobic
Body
Curl-Ups
Trunk Lifts
Upper Body
Flexibility
Capacity
Composition
Strength
Physical Fitness Test

Figure 4. 2002 grades 5, 7, & 9 SAT/9 mean reading NCE score by five different PFTs; n =
890,280.

For PFT tests, except Trunk Lifts and Flexibility, students who exceeded on the PFT had higher
average SAT/9 reading scores than students who passed and students who passed the PFT had
higher average SAT/9 reading scores than students who failed. Physically fitter students, even when
fitness was defined in different ways, had higher achievement test scores.

The exceed data for Trunk Lifts were not included in Figure 4 because the criteria defining exceed
were unclear. For Trunk Lifts, while lying face down, students attempted to lift their upper body 12
inches off the floor and hold that position long enough to measure the distance. Recorders assigned a
value of 0 to 99. High values in this test (e.g., 50 inches and higher) made no sense. In reality,
students either met the standard by lifting their body 12 inches or they didn’t. High score values were
classified as exceed but these high values were nonsensical and were not included in this analysis.

There was no exceed category for the Flexibility test. There were only two categories, (1) met the
standard or (2) failed to meet the standard.

The Relationship Between PFT and Mean SAT/9 Scores by Subgroups
Subgroup analyses were conducted to evaluate how the relationship between fitness and
achievement was modified by other student characteristics. First, the relationship between physical
fitness and academic achievement was evaluated by gender. Figure 5 shows at the relationship
between 2002 PFT scores and SAT/9 mean reading scores by gender.

---

Physical Fitness and Academic Achievement
20


Female
Male
Female
Male
70
50%
65
45%
60
55
40%
54
50
49
49
35%
45
45
46
42
40
30%
40
41
42
44
37
37
38
38
35
25%
Score 30
20%
25
29.3%
30.0%
20
28.1%
29.5%
15%
15
21.0%20.0%
10%
Percent of Students
10
SAT/9 Mean Reading NCE
12.3%13.0%
5
5%
6.8%
0.2% 0.3% 1.7% 2.3% 5.6%
0
0%
0
1
2
3
4
5
6
Number of Fitness Standards Achieved

Figure 5. 2002 grade 5, 7, & 9 SAT/9 mean reading NCE score by overall PFT score and gender,
n = 436,139 (female), 448,576 (male)

Figure 5 shows that female and male students demonstrated comparable levels of fitness. For
example, 30 percent of females and 29 percent of males had an overall PFT score of 6. There were
modest test score differences by gender. The overall mean reading NCE score of 48 for females was
slightly higher than overall average of 45 for males. Even though there were test score differences,
the relationship between fitness and achievement was consistent across gender. For females and
males, as the overall PFT score increased so did the mean reading achievement scores.

Even though there were no differences in fitness and little difference in test scores, there appeared to
be an interaction effect. The rate of change in achievement scores was higher for females than
males. The difference in mean reading NCE scores between the least and most fit students was 17
NCE points for females and 12 NCE points for males.

To evaluate whether results were consistent across socio-economic status (SES), the relationship
between physical fitness and academic achievement was evaluated by NSLP. NSLP is the national
school lunch program. NSLP indicates whether or not students received free or reduced lunch and
serves as a proxy for SES. Participation in NSLP is an indicator of lower SES. No NSLP participation
is an indicator of higher SES. Figure 6 shows these results.

---

Document Outline

• ÿ
• ÿ
• ÿ
• ÿ
• ÿ

---