A Preliminary Study of One Year of Pedometer Self-Monitoring
Catrine Tudor-Locke, Ph.D.
Department of Exercise and Wellness
Arizona State University
David R. Bassett, Jr., Ph.D.
Department of Health and Exercise Science
The University of Tennessee
Ann M. Swartz, Ph.D. and Scott J. Strath, Ph.D.
Department of Human Movement Sciences
University of Wisconsin-Milwaukee
Brian B. Parr, Ph.D.
Exercise and Sports Science
University of South Carolina Aiken
Jared P. Reis, Ph.D.
Graduate School of Public Health
San Diego State University
Katrina D. DuBose, Ph.D.
Center for Physical Activity & Weight Management
Schiefelbusch Institute for Lifespan Studies
University of Kansas
Barbara E. Ainsworth, Ph.D.
Department of Exercise & Nutritional Sciences
San Diego State University
ABSTRACT
Background: Long-term pedometer monitoring has not
been attempted. Purpose: The purpose of this project was to col-
lect 365 days of continuous self-monitored pedometer data to
explore the natural variability of physical activity. Methods:
Twenty-three participants (7 men, 16 women; M age = 38 ± 9.9
years; M body mass index = 27.7 ± 6.2 kg/m
2
) were recruited by
word of mouth at two southern U.S. universities. Participants
were asked to wear pedometers at their waist during waking
hours and record steps per day and daily behaviors (e.g.,
sport/exercise, work or not) on a simple calendar. In total, par-
ticipants wore pedometers and recorded 8,197 person-days of
data (of a possible 8,395 person-days, or 98%) for a mean of
10,090 ± 3,389 steps/day. Missing values were estimated using
the Missing Values Analysis EM function in SPSS, Version
11.0.1. Results: A mean of 10,082 ± 3,319 steps/day was com-
puted. Using the corrected data, differences in steps/day were
significant for season (summer > winter, F = 7.57, p = .001),
day of the week (weekday > weekend, F = 3.97, p = .011), type
of day (workday vs. nonworkday, F = 9.467, p = .008), and par-
ticipation in sport/exercise (day with sport/exercise > day with-
out sport/exercise, F = 102.5, p < .0001). Conclusions: These
data suggest that surveillance should be conducted in the
spring/fall or that an appropriate correction factor should be
considered if the intent is to capture values resembling the
year-round average.
(Ann Behav Med 2004, 28(3):158–162)
INTRODUCTION
Walking is the most frequently reported type of leisure-time
physical activity (1–3). Although walking is also fundamentally
important to our daily lives, we have little understanding about
how much people walk outside leisure-time contexts. Although
most self-report approaches to physical activity assessment in-
quire about walking frequency and duration (specifically, brisk
walking or walking for exercise), ubiquitous ambulatory activ-
ity is more difficult to capture with questionnaires (4,5). Regard-
less, walking is likely the most important activity to assess accu-
rately, especially in the majority of individuals who do not
regularly participate in sport or exercise (6).
158
This study was conducted while Dr. Tudor-Locke, Dr. DuBose, Dr.
Ainsworth, and Mr. Reis were at the Prevention Research Center, Ar-
nold School of Public Health, University of South Carolina and while
Drs. Swartz, Strath, and Parr were at the Department of Health and Ex-
ercise Science at The University of Tennessee.
We are indebted to Cary Springer (University of Tennessee Statistical
Consulting Services) for conducting the statistical analyses and to Pat-
rick Schneider for assisting with data entry and processing. Finally, we
thank our participants for their daily contributions to this study.
Reprint Address: C. Tudor-Locke, Ph.D., Department of Exercise and
Wellness, Arizona State University, 7001 East Williams Field Road,
Mesa, AZ 85212. E-mail: Tudor-Locke@asu.edu
© 2004 by The Society of Behavioral Medicine.