The Reliability and Validity of Traditional and Modified Pull-ups
With Children in Grades 3 - 5
Mary E. Engleman & James R. Morrow, Jr.
University of Houston

Overview

Three nationally available youth fitness tests (AAHPERD's Physical Best; Fitnessgram; President's Challenge) contain an item of upper body musculoskeletal strength. The upper body strength/endurance measure has generally been added because of its perceived importance for functioning in daily activities and not from a 'health related' perspective. There have been, however, assertions that performance on this item could be related to bone density which is a health related factor. Once a decision has been made to include a measure of upper body strength, considerable discussion occurs regarding the particular test item to use (pull up; chin-up; flexed arm hang; modified pull-ups; chins, dips, etc.). Each of these tests has been used in various test batteries over the past twenty years. Generally, the assumption is made that the performance on these test items is related and that using one or another has little impact on test results. However, recent evidence suggests that the traditional pull-up test does not discriminate well, scores are positively skewed, many students are unable to complete even one pull-up, and that test performance is highly related to body weight (Pate, 1987). The purpose of the present research was to determine the reliability and concurrent validity of the traditional and modified pull-ups with elementary age children in grades three through five and to investigate the influence of body composition on test item performance. Results indicated that the sum of two trials of each pull-up measure is very reliable for both boys (R = .938 and .869) and girls (R = .954 and .895) for the traditional and modified pull-up respectively. Concurrent validities were .60 for the girls and .63 for the boys (.63 for total group), indicating that less than 40% of the variance is common between the two upper body test items. Both pull-up measures were negatively related to skinfold thickness indicating that neither test has a distinct advantage over the other in terms of negating the effects of body weight. For girls, the correlations were -.35 and -.43 for the pull-up and modified pull-up respectively. The values were similar for the boys (-.36 and -.45). These results indicated that while both measures are highly reliable, concurrent validities illustrate that the two tests measure different aspects of upper body strength. These results point to the specificity of the strength/endurance measurement obtained with a particular test. Neither test totally negates the effect of body weight. However, a large proportion of boys and girls cannot complete a single pull-up. Therefore, from a motivational perspective, the modified pull-up might be more appropriate.

Study details

Purpose: Determine the reliability and concurrent validity of the traditional and modified pull-up tests with elementary age children in grades three through five and investigate the influence of body composition on test item performance.

Subjects: Students in grades 3,4, and 5. 242 boys, and 227 girls.

Methods: Alpha reliability; Correlation between traditional and modified pull-up scores for concurrent validity.

Results:

Conclusions

• Both the traditional and modified pull-up tests are very reliable with children in grades 3,4, and 5.
• The two test items share little variance.
• Neither test totally negates the effects of body size as determined by skinfold thickness.
• Fitness test developers who utilize the modified pull-up test should be aware that the musculoskeletal function assessed this way shares little variation with the traditional pull-up test (different factors are being measured) and that the effects of body fatness are not negated by use of the modified pull-up test.