Journal of Physical Activity & Health, 2006, 3(Suppl. 2), S47-S57
© 2006 Human Kinetics, Inc.
Comparison of Two Versions of the
PACER Aerobic Fitness Test
James J. McClain, Gregory J. Welk, Michelle Ihmels,
and Jodee Schaben
Background: The PACER test is a valid and reliable assessment of aerobic
capacity in children. However, many schools lack adequate space to administer
the test. This study compared the utility of the standard 20m PACER test with
an alternative 15m PACER protocol in 5th and 8th grade students. Methods:
A total of 171 students completed both PACER protocols in a counterbalanced
design. Agreement between the two protocols was assessed with correlations,
repeated-measures ANOVA, and classification agreement into the FITNESSGRAM® healthy fitness zones. Results: The difference in estimated VO2max
between the two tests was slightly larger for boys (5th grade, 1.32 ml/kg/min;
8th grade, 1.72 ml/kg/min) than girls (5th grade, 0.14 ml/kg/min; 8th grade,
1.11 ml/kg/min), but these differences are probably not of practical significance.
Classification agreement was 88% for boys and 91% for girls. Conclusions:
Collectively, the results suggest that the 15m and 20m PACER provide similar
information about aerobic fitness in youth. The 20m test is recommended when
possible, but the 15m provides a useful alternative for schools with smaller
gymnasiums.
Key Words: aerobic capacity, field test, children
Fitness testing is a common component of most physical education programs because
it provides schools and teachers with a way to document and track important dimensions of health related fitness.1 It also provides a way to offer feedback to students
and parents that may help children develop more active lifestyles. While it is important to have adequate levels of fitness in each component of health related fitness,
aerobic capacity (an indicator of cardiovascular fitness) is generally considered to be
the most important fitness measure due to its association with coronary heart disease
risk in adults and chronic disease risk factors in children and adolescents.2
Aerobic capacity can be measured accurately under laboratory conditions using
a maximal treadmill or bike-test protocol. Due to the high cost of the equipment and
the time required for individual assessment, this type of procedure is not practical for
McClain is with the Dept. of Exercise and Wellness, Arizona State University, Polytechnic Campus,
7350 E. Unity Ave., Mesa, AZ 85212-0180. Welk and Ihmels are with the Dept. of Health and Human
Performance, Iowa State University, Ames, IA 50011. Schaben is with the Dept. of Exercise Science,
Central College, Pella, IA 50219.
S47
S48
McClain et al.
school-based fitness assessments. Field tests allow aerobic capacity to be estimated
outside of the laboratory and in group settings such as physical education class. To be
effective, field tests must be valid, reliable, easily administered, allow simultaneous
assessment, require minimal equipment, and be noninvasive in nature.
The Progressive Aerobic Cardiovascular Endurance Run (PACER) is the recommended test for aerobic capacity in the FITNESSGRAM® assessment battery.3
The PACER test consists of running a 20-meter shuttle course with 1-minute paced
stages at an initial running speed of 8.5 km h–1 with an increasing speed of 0.5 km
h–1 per stage.4 The PACER assessment protocol has many advantages over other
commonly used field tests of aerobic fitness such as the mile run. The protocol can
be performed inside, which eliminates environmental limitations and influences on
testing. The PACER test avoids problems with pacing since the tones on the audio
CD indicate the required running pace. The design characteristics of the PACER
test also provide a more motivational environment for testing.5 A final advantage
of the PACER protocol is that the progressive nature allows much of the test to
be performed at a sustainable, submaximal effort. Only the final stages require a
maximal effort from the participant. These factors help to reduce stress and motivational concerns associated with fitness testing.
The validity of the PACER test has been examined by comparing scores on
the test (maximum running speed, number of levels completed, number of laps,
or estimated VO2) with measured VO2max from a treadmill-based protocol. The
validity has been well supported, as studies have reported correlation coefficients
ranging from r = 0.69 and r = 0.87.4, 6–9 The reported test-retest reliability of laps
performed for the PACER protocol in school-age children is also high, with intraclass reliability coefficients of R = 0.82 to R = 0.93.4,5,8,10
The PACER test has been shown to be an effective assessment tool for many
physical education programs. However, one problem limits its overall utility: many
schools have gyms or physical education classrooms that are less than 20 meters in
length. The test could be administered outdoors, but this makes testing contingent
on weather. The development of a shorter length PACER protocol would result
in a more versatile field test because it would allow the test to be administered in
smaller facilities.
The variables included in the PACER’s prediction equation for aerobic capacity
are age and maximum running speed.4 By changing the time intervals on the audio
CD, it would be possible to create a shorter test (i.e., < 20m) that replicates the key
features of the original test. By maintaining the 1-minute intervals and the same
running speeds at each level, it should be possible to employ the same prediction
equation for estimation of aerobic capacity. However, there are several issues that
could influence results using a shorter protocol. The number of laps per level has to
increase with a shorter distance and this would increase the number of turns. More
turns would imply more acceleration/deceleration during the test and this could lead
to premature fatigue. In addition, the increased number of laps could potentially
result in premature dropout due to boredom. Before a shorter test can be promoted
for use in schools, it is important to determine whether the shorter protocol yields
similar estimates of aerobic fitness as the longer version.
This study compared the utility of the standard 20-meter PACER test with an
alternative 15-meter PACER protocol in 5th and 8th grade students. Specifically,
this study compared the aerobic capacity estimates (VO2max) obtained from the
two protocols and the classification agreement between protocols for attaining the
Two Versions of the PACER Test
S49
Healthy Fitness Zone (the criterion referenced standard in the FITNESSGRAM test
battery).11
Methods
Participants
Students (N
N = 204) from selected physical education classes in a local school district were recruited to participate in the project. The sample included 82 fifth-grade
students (49 M, 33 F) and 122 eighth-grade students (75 M, 47 F). All aspects of
this research were reviewed and approved by the Iowa State University Institutional
Review Board before the project was initiated. Informed consent was obtained from
parents and assent was obtained from the students prior to testing.
Development of the 15m PACER CD
The goal in the study was to use similar 1-minute levels and the same running
speeds in the 15m protocol to allow the use of the same prediction equations. To
accomplish this, the same computer program used in developing the 20m PACER
CD was used to create the 15m PACER CD. There are some minor differences in
actual running speed between the protocols due to the way in which the number of
laps per level are computed on the audio CD (see Table 1). The number of seconds
per level in the original 20m protocol ranges from 60 to 66 seconds. The compaTable 1 Comparison of Laps/Level and Total Laps for the 15m and
20m PACER
Running
speed (kph)
8.5
9
9.5
10
10.5
11
11.5
12
12.5
13
13.5
14
14.5
15
15.5
Level
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Laps/Level
20m
15m
Total Laps
20m
15m
7
8
8
9
9
10
10
11
11
11
12
12
13
13
13
7
15
23
32
41
51
61
72
83
94
106
118
131
144
157
9
10
11
12
12
13
13
14
14
15
15
16
17
17
18
Note: Running speed based on published values from Leger et al. (1988).
9
19
30
42
54
67
80
94
108
123
138
154
171
188
206
S50
McClain et al.
rable values for the 15m protocol were 58 to 65 seconds. The mean difference
between running speeds for the first 8 levels is 0.05 ± 0.06 km h–1 and the overall
difference for the complete 15-level comparison is 0.12 ± 0.11 km h–1. The speed
difference is slightly higher in the upper levels. However, this is not likely to be a
major concern, as the majority of schools lacking a 20m physical education facility
are elementary schools, and few elementary school students are able to reach the
upper levels of the PACER test.
Testing Procedures
Testing was performed during normal physical education classes. Students in the
study were already familiar with the PACER protocol, but the purpose and procedures for the testing were explained prior to each test. A researcher demonstrated
the testing procedure and identified the change of lap and change of level tones on
the audio CD. Participants performed the 15m and 20m protocols approximately
one week apart in a counterbalanced design by class section. They wore a numbered
jersey to assist with recording of results by the research team. Participants completed
the testing in groups of 10 or 15 students and a maximum 5:1 subject to recorder
ratio was maintained. Participants were encouraged to give maximal effort during
testing to ensure an accurate estimate of aerobic capacity for both tests.
Data Processing
A total of 180 students (76 fifth graders; 104 eighth graders) completed both versions of the PACER test. Observations during the trials suggested that some students
did not give equivalent effort on both versions of the test. To avoid error due to
incomplete effort, students who had differences of four or more levels between the
two assessments were removed from the analyses. There were nine students with
differential effort on the two tests. They consisted of three 5th grade boys and six
8th grade boys. There was no apparent pattern between the protocols, with four
incomplete effort trials on the 20m protocol and five incomplete effort trials on
the 15m protocol. Because the order of administration was counterbalanced in the
study, there would likely be no systematic bias introduced by the overall group level
estimates of aerobic capacity. However, correlations and classification agreement
results would be biased if these data were included. Therefore the data from these
students were removed from all analyses. The final sample included the remaining
171 students (5th grade: 37 M, 31 F; 8th grade: 60 M, 43 F) who exhibited similar
effort on both tests.
Another issue that could confound the test results is the determination of the
participant’s number of completed laps. The FITNESSGRAM test protocol allows
students to self-select out of the test but also specifies that students should be
stopped if they fail to reach the line for two consecutive tones on the audio CD. For
a subsample of participants, two independent observers were used to record student
lap numbers to ensure that this information was recorded accurately. Correlations
between the primary and secondary recorder’s lap counts were computed for the
15m and the 20m PACER. Primary and secondary recorder data were available
for 68 participants doing the 15m version and for 75 participants doing the 20m
version. Correlations were high between recorders, with r = 1.0 and r = 0.99 for
the 15m and 20m PACER, respectively. Primary recorder lap count values were
used for all further analyses.
Two Versions of the PACER Test
S51
Analysis
Means and standard deviations for the number of laps and levels were computed
for each test and these are reported separately for both age and gender groups. The
estimated VO2max values for each participant were computed using the Leger et
al.4 equation: VO2max = 31.025 + (3.238 * max speed) – (3.248 * age) + (0.1536
* max speed * age). Descriptive data were also reported for estimated VO2max and
these are also reported separately for each age and gender group.
Differences in VO2max estimates from the two protocols were analyzed using
a two-way (age × gender) repeated-measures analysis of variance (ANOVA). The
use of a repeated-measures analyses provides a more stringent test of possible differences since it controls for individual variability in the tests. An alpha level of 0.05
was used to evaluate the significance of the interaction term and the main effects.
Correlations were computed between the scores from both tests to provide
an indication of overall agreement between the two protocols. Correlations were
computed for laps, levels, and estimated VO2max. Values are reported separately
for each age and gender group. A predetermined criteria level of 0.8 was used to
designate good agreement between the two tests.
Bland-Altman plots were used to examine the agreement between the two
assessments across the range of VO2max estimates. The Bland-Altman plots display
the mean of the 15m and 20m VO2max estimate for an individual on the X-axis
and the difference between the estimates (20m–15m) on the Y-axis. Correlations
between data points on the Bland-Altman plots were used to determine whether
there was any form of systematic bias across the range of fitness levels.
Classification agreement for achieving the FITNESSGRAM criterion referenced
standards were computed using 2 × 2 contingency tables. Participants were classified into either Needs Improvement (NI) or Healthy Fitness Zone (HFZ) categories.
Percent correct classifications were computed as a descriptive statistic. Kappa values
were also reported to provide an overall indicator of agreement after adjusting for
the agreement that would occur simply due to chance.
Results
Descriptive statistics for the participants are presented in Table 2. The overall F
test for the repeated-measures ANOVA test was significant, F(1, 167 = 15.7), p
< .001. The main effect for gender was significant ((p < 0.05), with boys scoring
higher on both the 15m PACER and 20m PACER test. The age main effect was
not significant ((p > .05), but the values for the 8th graders were slightly higher than
for the 5th graders.
The primary comparisons of interest were with the possible differences between
the two PACER tests. There was a significant gender-by-test interaction, F(1, 167)
= 4.20, p < .042, and this was due to differences in the test comparisons for boys
and girls. The difference in estimated VO2max between the two tests was slightly
larger for boys (5th grade, 1.32 ml/kg/min; 8th grade, 1.72 ml/kg/min) than for
girls (5th grade, 0.14 ml/kg/min; 8th grade, 1.11 ml/kg/min). The main effect for
test was also significant but the effect sizes for the actual differences were small
(<0.30) and probably not of practical significance.
The correlations between laps, levels, and VO2 estimates from the two tests
are presented in Table 3. Correlations between VO2max estimates were slightly
higher for boys (rr = 0.79) than for girls (r = 0.67). Correlations were higher for
S52
McClain et al.
Table 2 Descriptive Statistics on Laps, Levels, and VO2max
Estimates by Gender and Grade Level for the 15m and 20m PACER
(n)
Male
Mean
SD
(n)
Female
Mean
SD
5th Grade
P15 Laps
P20 Laps
P15 Levels
P20 Levels
P15 VO2 (ml/kg/min)
P20 VO2 (ml/kg/min)
37
37
37
37
37
37
33.84
31.24
3.81
4.35
44.33
45.65
15.69
12.93
1.30
1.53
3.40
3.96
31
31
31
31
31
31
31.81
23.61
3.52
3.45
43.7
43.56
16.59
10.75
1.45
1.28
3.78
3.25
8th Grade
P15 Laps
P20 Laps
P15 Levels
P20 Levels
P15 VO2 (ml/kg/min)
P20 VO2 (ml/kg/min)
60
60
60
60
60
60
55.02
49.07
5.47
6.10
43.64
45.36
21.54
23.43
1.67
2.33
4.32
6.01
43
43
43
43
43
43
38.16
32.6
4.07
4.49
40.08
41.19
14.96
13.93
1.33
1.53
3.31
3.95
Note: P15 = 15m PACER; P20 = 20m PACER.
Table 3 Pearson Correlations Between Laps, Levels, and VO2max
Estimates by Gender and Grade Level for the 20m and 15m PACER
Boys
Girls
5th Grade
Laps P20 vs. P15
Level P20 vs. P15
VO2 P20 vs. P15
0.58
0.56
0.61
0.54
0.48
0.53
8th Grade
Laps P20 vs. P15
Level P20 vs. P15
VO2 P20 vs. P15
0.91
0.87
0.85
0.83
0.74
0.71
Note: Correlations significant at the p < 0.01 level.
Two Versions of the PACER Test
S53
Figure 1A and 1B — Bland-Altman plots for 5th grade girls (A) and boys (B).
the 8th grade students (M, r = 0.85; F, r = 0.71) than for 5th grade students (M, r
= 0.61; F, r = 0.53).
Bland-Altman plots were used to examine agreement between tests across the
range of fitness levels. Separate graphs were produced for each of the four genderby-grade comparisons. The plots are shown in Figures 1 and 2. The solid line depicts
the mean difference in the two tests while the dotted lines show the boundaries
for the 95% confidence interval. Nearly all of the data points fell within the 95%
confidence intervals. Correlations between the individual means and differences in
VO2max estimates showed a significant correlation of r = 0.54 for 8th grade boys,
S54
McClain et al.
Figure 2A and 2B — Bland-Altman plots for 8th grade girls (A) and boys (B).
suggesting that for this subsample, less fit individuals performed better in the 15m
test and more fit individuals performed better in the 20m test. Correlations for the
8th grade girls, 5th grade boys, and 5th grade girls were not significant, indicating
no fitness related bias between protocols in the remaining subsamples.
Classification agreement between the 15m and 20m test was examined by looking at passing rates, based on achieving the FITNESSGRAM criterion referenced
standards for VO2max. Classification agreement and kappa statistics are presented
in the 2 × 2 contingency table shown in Table 4. Classification agreement for age
and gender groups ranged from 85% to 93%. The misclassification was evenly
NI
21.7 (13)
10.0 (6)
31.7 (19)
Kappa value
0.64*
NI
10.81 (4)
5.4 (2)
16.2 (6)
Kappa value
0.63*
HFZP15
HFZ
5.0 (3)
63.3 (38)
68.3 (41)
HFZP15
HFZ
2.7 (1)
81.1 (30)
83.8 (31)
Total
26.7 (16)
73.3 (44)
100.0 (60)
% Agreement
85.0
Total
13.5 (5)
86.5 (32)
100.0 (37)
% Agreement
91.9
HFZP20
NI % (n)
HFZ % (n)
Total % (n)
8th Grade Girls
HFZP20
NI % (n)
HFZ % (n)
Total % (n)
5th Grade Girls
NI
7.0 (3)
4.7 (2)
11.6 (5)
Kappa value
0.68*
NI
0.0 (0)
6.5 (2)
6.5 (2)
Kappa value
NA
HFZP15
HFZ
2.3 (1)
86.1 (37)
88.4 (38)
HFZP15
HFZ
6.5 (2)
87.1 (27)
93.6 (29)
Total
9.3 (4)
90.7 (39)
100.0 (43)
% Agreement
93.1
Total
6.5 (2)
93.6 (29)
100.0 (31)
% Agreement
87.1
Note: HFZP20 = Healthy Fitness Zone for 20m PACER; HFZP15 = Healthy Fitness Zone for 15m PACER; NI = needs improvement; HFZ =
Healthy Fitness Zone classification; NA = Value could not be correctly computed due to data distribution. *Significant at p < 0.001
HFZP20
NI % (n)
HFZ % (n)
Total % (n)
8th Grade Boys
HFZP20
NI % (n)
HFZ % (n)
Total % (n)
5th Grade Boys
Table 4 Classification Agreement for Reaching the Healthy Fitness Zone by Gender and Grade Level for the
20m and 15m PACER
Two Versions of the PACER Test
S55
S56
McClain et al.
distributed for girls, but boys were more likely to pass the 20m than the 15m. Eight
boys passed the 20m test and failed the 15m test, while only four boys passed the
15m test and failed the 20m test.
Discussion
This study examined the utility of a 15m version of the well-established 20m PACER
test. The 15m PACER protocol was designed to maintain the critical properties of
the original 20m PACER test establish by Leger et al.4 The results suggest that the
15m protocol yields similar information about aerobic fitness as the well-established
20m protocol. The differences in VO2max estimates between the 15m protocol and
20m protocol were generally small and not of much practical significance. The differences varied somewhat by grade and gender (range: 0.7 to 1.8 ml/kg/min), but
the effect sizes of the differences for all comparisons were less than 0.30.
The differences in estimates of aerobic capacity from the two versions of the
PACER are comparable to differences that are evident between other field tests of
fitness in children. Mahar et al. examined differences between VO2max estimates
for the 20m PACER and the Mile Run in 266 4th and 5th grade students.10 Differences of 0.9 and 1.6 ml/kg/min were reported for boys and girls, respectively.
Recent work in our laboratory reported differences between VO2 max estimates for
the 20m PACER and the Mile Run in 473 7th and 8th grade students.12 Differences
of 1.6 and 1.8 ml/kg/min were reported for boys and girls, respectively. The similar
differences between the 15m and 20m protocols indicate that the two versions are
at least as similar as comparisons between the PACER and the Mile Run.
The increased number of laps per level and the proportionate increase in the
number of turns in the 15m protocol were considered as potential limitations to
the utility of the modified assessment. However, the small differences in VO2max
estimates between protocols suggest that the modification to the protocol had little
effect on the functional performance of the field test. The utility of the 15m protocol is further supported by the balanced distribution of difference scores on the
Bland-Altman plots. The only plot that exhibited a tendency toward differential
test performance by fitness level was for the 8th grade boys.
The 8th grade boys with higher fitness levels demonstrated better performance
on the 20m protocol compared to the 15m protocol. It is possible that early withdrawal due to boredom contributed to this trend, but it could also have been due
to other factors such as fatigue from turns. The 8th grade boys with lower fitness
levels performed better on the 15m protocol than the 20m protocol. This may be
the result of a psychological perception that the shorter length protocol required
less work or was easier than the longer length protocol.
The impact of any potential boredom withdrawal at higher levels of the 15m
PACER does not significantly impact the overall utility of the test. The decision
to use the 15m PACER protocol would likely be driven by the type of physical
education facilities available to teachers. The majority of smaller facilities that lack
adequate room to utilize the 20m PACER are located in elementary school buildings.
Children in these schools likely would not reach higher lap counts where turns or
boredom could have an impact on test performance. Because the differences appear
to be larger among the older sample, it is prudent to recommend the use of the 20m
protocol if space permits.
Two Versions of the PACER Test
S57
One potential limitation of this study was the lack of a laboratory-based criterion
test. However, the validity of the PACER assessment in school-age individuals has
been well established in the literature. Leger et al. demonstrated a moderate correlation (rr = 0.71) between predicted VO2max from the 20m PACER with measured
VO2max by retro extrapolation of the O2 recovery curve in boys and girls ages 8
to 18.4 Several other studies, with samples comparable to the present study, have
reported significant correlations between cycle or treadmill measured VO2max and
either maximum running speed or laps.6,7,8 The overall correlations in these studies
ranged from r = 0.69 to r = 0.87.
In conclusion, both the 15m and 20m PACER protocols provide similar
information about aerobic capacity in students. The differences between aerobic
capacity estimates from the 15m and 20m PACER protocols found in the present
study are similar to differences typically found in studies comparing other fieldbased aerobic capacity estimates such as the 20m PACER and the Mile Run. The
results of the study suggest that the 15m PACER protocol could be used in place
of the 20m PACER protocol when available facilities cannot accommodate the
longer testing protocol.
References
1. National Association for Sport and Physical Education. Moving Into the Future National
Standards for Physical Education. 2nd ed. Reston, VA: NASPE Publications; 2004.
2. Cureton KJ, Plowman SA. Aerobic capacity assessments. In: Welk GJ, Morrow J, Falls
H, eds. FITNESSGRAM Reference Guide. Dallas, TX: The Cooper Institute; 2001.
3. Meredith MD, Welk GJ. (Eds). FITNESSGRAM Test Administration Manual. Champaign, IL: Human Kinetics; 2004
4. Leger LA, Mercier D, Gadoury C, Lambert J. The multistage 20 metre shuttle run test
for aerobic fitness. J Sports Sci. 1988; 6:93-101.
5. Vincent SD, Barker R, Clarke M, Harrison J. A comparison of peak heart rates elicited
by the 1-mile run/walk and the Progressive Aerobic Cardiovascular Endurace Run. Res
Q Exerc Sport. 1999; 70:75-78.
6. Van Mechelen W, Hlobil H, Kemper HCG. Validation of two running tests as estimates
of maximal aerobic power in children. Eur J Appl Physiol. 1986; 55:503-506.
7. Boreham CAG, Paliczka VJ, Nichols AK. A comparison of the PWC170 and 20-MST
tests of aerobic fitness in adolescent schoolchildren. J Sports Med Phys Fitness. 1990;
30:19-23.
8. Liu NY-S, Plowman SA, Looney MA. The reliability and validity of the 20-meter shuttle
test in American students 12 to 15 years old. Res Q Exerc Sport. 1992; 63:360-365.
9. Mahoney C. 20-MST and PWC170 validity in non-Caucasian children in the UK. Br J
Sports Med. 1992; 26:45-47
10. Mahar MT, Rowe DA, Parker CR, Mahar FJ, Dawson DM, Holt JE. Criterion-referenced
and norm-referenced agreement between the mile run/walk and PACER. Meas Phys
Ed Exerc Science. 1997; 1:245-258.
11. Chun DM, Corbin CB, Pangrazi RP. Validation of criterion-referenced standards for
the mile run and progressive aerobic cardiovascular endurance tests. Res Q Exerc Sport.
2000; 71:125-34.
12. Welk GJ, McClain JJ, Schaben JA, Mahar MT. Method agreement between two field
measures of aerobic fitness in youth. Med Sci Sports Exerc. 2004; 36:S134.
Note: Data were collected as part of a Masters Degree research project completed by the first author
while at Iowa State University.