The collection of children’s and adolescents’ pedometer-determined physical activity (typically expressed as steps per day) is growing more and more common place in both research and practice. Until recently, no large representative set of pedometer data existed to provide standards for interpreting values of ambulatory physical activity. Therefore, it could not be determined whether any specific sample of 10-yr-old boys recruited to a physical activity intervention (by way of a single example) were less active than other 10-yr-old boys. A source of expected or normative values is required. An early attempt to describe such values (22) was based on a single relevant study published between 1980 and 2000; the conclusion reached was that 8- to 12-yr-olds were expected to average 12,000 (girls) to 16,000 (boys) steps per day (14).
Since that time, two other reviews (1,21) have compiled the quickly growing body of step-defined physical activity literature, publishing collectively assembled normative data representing the age span of 5- to 19-yr-olds. For children, girls average between 10,000 and 13,000 steps per day and boys average between 12,000 and 16,000 steps per day. In general, the highest values for steps per day culminate before the age of 12 yr, and then appear to deteriorate with each year of adolescence, closing at approximately 8000 to 9000 steps per day before the transition to young adulthood (18). Although useful, these assembled values were culled from literature that might be best described as fragmented (i.e., assorted and missing age groups, various regions and countries, etc.) and disjointed (i.e., diverse instrumentation and technologies used, distinct methods of data collection and treatment, etc.). What is missing from this evolving landscape is a broad and inclusive sex- and age-specific source of normative data collected from a large national sample, in the same manner, using the same instrument. With such a resource, both group-level and individual-level child/adolescent data could be confidently compared with that of representative peers, facilitating interpretation. Such a large data source has recently become available in Canada, and its findings revealed similar age-related distribution of steps per day for boys and girls to that assembled earlier (6).
The Canadian Physical Activity Levels Among Youth (CANPLAY) study is an ongoing child and adolescent surveillance study that has collected pedometer data consistently using the Yamax SW-200 pedometer on more than 34,000 5- to 19-yr-old participants from across Canada between 2005 and 2011. The purpose of this analysis of the CANPLAY data is to provide a comprehensive and sex- and age-specific normative data for children’s and adolescents’ pedometer-determined physical activity. The potential applicability of such data is broad and includes surveillance, screening, comparison purposes, planning strategies, prioritizing efforts and distributing resources, evaluating intervention effects, and tracking change.
The CANPLAY data collection protocol has been published previously (7), so only a summary is provided here. Children between 5 and 19 yr were selected by random digit dialing (contacting the household, selecting a parent or legal guardian respondent of children and youth throughout the year, including holiday periods) and a computer-assisted telephone interview was conducted with the parent. After presenting the study requirements during the interview, parents who agreed to their child’s participation in the pedometer portion of the study (≅6000 families annually) were sent a data collection kit and prompted as necessary to return the data by mail in a timely manner (see additional details ). Specifically, participants were asked to wear the pedometer and log steps for up to seven consecutive days, weekdays and weekend days included. As previously reported, weekday steps were higher than weekend day steps (6). Most logging sheets included 7 d of recorded data with only 3% of boys and 4% of girls reporting 5 d or less (7). In total, 21,271 children and 12,956 adolescents returned completed sheets between 2005 and 2011. Extensive data checking, detailed previously (6,7), was conducted to assure data quality. Written informed consent was received from all participants. All protocols were approved by the Human Participants Review Committee of York University for all survey years and by the Health Canada ethics review board for years 2010–2011.
Data treatment analysis
Data treatment follows procedures that were described in greater detail previously in the original process (6) and descriptive epidemiology papers (7). Briefly, daily step counts less than 1000 and more than 30,000 steps per day were truncated to these limits. For this analysis, all values were included regardless of the number of days logged (as indicated earlier, in >95% of the cases, the pedometer was worn for at least 5 d ). The mean and 95% confidence interval (CI) were computed for steps per day (averaging steps taken over logged days) by each single-year age and for both sexes separately and combined.
Percentile values (5 percentile increments) were computed for each single-year age by sex. The percentiles in 5% increments were calculated using the LMS method (LMS Chartmaker Pro Version 2.3; Harlow Printing Limited, South Shields, UK), which uses a Box-Cox power transformation to normalize the data at each year of age. Splines were fitted by maximum penalized likelihood to create three age-specific, smoothed curves termed L (λ), M (median), and S (coefficient of variation). Equivalent degrees of freedom for L, M, and S measured the complexity of each fitted curve, and Q statistics tested for normality in the location, scale, and skewness of the z-scores. Percentile curves were created using the values of L, M, and S at each age along with the normal equivalent deviate. Smoothing constants were chosen to be as small as possible to bring the Q tests for each moment close to normality, therefore creating a parsimonious model. The model was fitted for the entire sample, and the L, M, and S values chosen for the entire sample were then applied to each gender separately. Finally, quintiles (cut by 20th, 40th, 60th, and 80th percentiles) were computed for four combined age groups (5–7, 8–10, 11–14, and 15–19 yr), stratified by sex.
All data were weighted by the sample weights in calculating means, LMS-derived percentiles, and quintiles to reflect the complex sample design and the age–sex distribution of the population. CIs were computed using the Statistical Package for the Social Sciences Complex Sample procedures (version 18; SPSS Inc., IBM, Chicago, IL) to account for the sample design.
In total, 34,227 children and adolescents (17,314 boys and 16,913 girls) participated in CANPLAY. As detailed in Table 1, participation was skewed toward children of younger ages, living in higher education and income level households, who were of normal weight and had parents who reported that they were just as active or more active than their peers. Being classified as overweight or obese and having parents who rated themselves as being substantially more active was more prevalent among boys. A minimum sample size was allocated to each province/territory in the sample design, and participation rates by region reflected this design feature.
Table 2 displays mean, CI steps per day for both sexes by age, and also by each single-year age stratified by sex. Mean steps per day increased from 11,602 steps per day among 5-yr-olds, to a sample peak mean value of 12,348 steps per day among 10-yr-olds, and then declined to 9778–10,073 among 15- to 19-yr-olds. Similarly, the mean values of steps per day increased from 12,049 steps per day among 5-yr-old boys to a sample peak mean value of 13,030 steps per day among 10-yr-olds, and then declined to 10,189–10,818 among 15- to 19-yr-old boys. For girls, the mean steps per day ranged from 11,095 to 11,638 among 5- to 10-yr-olds and was lower (9231–9476 steps per day) among 15- to 19-yr-olds. With the exception of 19-yr-olds, the mean steps per day were higher among boys than girls at every age. Table 3 presents the five quintile-defined categories (in ascending order: lowest, lower than average, average, higher than average, and highest) for the four combined age groups, stratified by sex. Consistent with the data presented in Table 2, the highest quintile of steps per day for both boys and girls increased between ages 5–7 and 8–9 yr and then appeared to decrease after ages 10–11 yr (Table 3).
LMC curves and detailed single-year age incremental percentiles values for steps per day are presented in Figure 1 and Table 4, stratified by sex. The LMS-generated 20th percentiles (smoothed within and across years) were higher than the “lowest” quintiles in Table 3 for boys 7, 11, 12, and 15 yr and for girls 11, 12, and 15 yr. The LMS-generated 80th percentiles were lower than the “highest” quintiles in Table 3 for boys 6–11 and 15 yr and girls of all ages. These differences reflect the LMS assumptions that differences between consecutive years of age occur uniformly and that the data are distributed normally, whereas a slightly skew quintile values are actually apparent in Table 3.
These CANPLAY data represent the largest and most comprehensive set of sex- and age-specific normative values for children’s and adolescents’ pedometer-determined physical activity to date and hence serve as an important reference source. They provide a significant improvement over previous attempts to describe expected values, aggregated from the extant literature that is largely based on disparate studies. That being said, the well-known differences (18) between boys’ and girls’ pedometer-defined physical activity was confirmed, with a single exception apparent for 19-yr-olds. Furthermore, previous reports (1,18,21) illuminating a childhood increase in steps per day, before peaking and subsequently decreasing across adolescence up until the transition to adulthood, was also substantiated with these CANPLAY data.
The mean values of accelerometer-determined steps per day have been previously published for U.S. boys and girls ages 6–19 yr based on 2610 participants with at least 1 d of monitoring in the 2005–2006 cycle of the National Health and Nutrition Examination Survey (20). The ActiGraph 7164 accelerometer was used; however, because it is known to be more sensitive to lower force accelerations than pedometers (10,16), the data were adjusted to provide values congruent with these instruments. These lower technology and lower cost instruments are more likely to be used in clinical and practical applications so adjustments were made in that original study to make the values meaningful on this level. Although useful in terms of providing mean expected values, a full range of percentiles has not yet been published from these data.
The mean accelerometer-determined steps per day collected on 1613 Canadian boys and girls 6–19 yr have also been reported for Canadian children (5), but again, a full range of percentiles have not been published. That smaller study used the ActiCal accelerometer, which has been validated in adults and children as a step counter (9,13); however, the applicability of its output relative to that collected by the Yamax SW-200 pedometer used in CANPLAY is unknown.
Just like there is no single universal value of curl ups, vertical jump, or grip strength performance by which all children or adolescents should be evaluated (3,4,12), these detailed normative reference pedometer data clearly demonstrate that there is no single value of steps per day that can be conveniently used to say what is “average” across both sexes and all ages. Instead, in keeping with the conventional provision of normative reference values for these fitness tests in children and adolescents (3,4), we offer similarly organized pedometer-determined physical activity data. These normative data will be valuable to a range of users (from researchers, to practitioners, to lay people) for a myriad and variety of purposes including surveillance, screening, comparative evaluations, planning strategies, prioritizing efforts and distributing resources, evaluating intervention effects, and tracking change. Uniquely, these data offer, for the first time, child and adolescent-specific values that can be considered indicative of relatively low levels of ambulatory physical activity. For example, from a research perspective, it will be possible to track and compare proportions of children/adolescents who are in the lowest quintile (i.e., falling below the 20th percentile defined by this distribution), just like we track and compare proportions defined as overweight/obese relevant to the 85th percentile of a BMI distribution (11). Determinants research can be used to identify characteristics of individuals and contextual factors external to the individual, which explain the relative position among quintiles. Intervention research can shed light on what behavioral and program elements work best for different ages of boys and girls categorized into different quintiles.
As just a few examples of practical application, researchers planning an intervention to increase steps per day among relatively inactive 10-yr-old boys could compare the baseline steps per day of those recruited to the normative values in Table 2 and determine eligibility of participants according to whether they fell in the 1st or 2nd quintile relative to their peers, or “lowest” or “lower than average.” At a population level, characteristics of children and youth who report lower than “average” steps per day could be identified. Strategic plans could then be made to improve steps per day among this segment, and the effectiveness of these plans could be evaluated, tracked, and refined to further improve population levels of ambulatory activity of children and youth. Finally, these data are also available to clinicians and any child/adolescent who has ever worn a pedometer and was curious to know how their data compare relative to normative standards.
Previous attempts have been made to organize children’s step data (1,21). Tudor-Locke et al. (23) proposed a sex-specific graduated step index for children 6–12 yr based on their earlier work (17,19), attempting to determine BMI-referenced criterion values for steps per day in 1954 in U.S., Australian, and Swedish children. Values for girls were 1) <7000; 2) 7000–9499; 3) 9500–11,999; 4) 12,000–14,499; and, 5) ≥14,500 steps per day. Values for boys were 1) <10,000; 2) 10,000–12,499; 3) 12,500–14,999; 4) 15,000–17,499; and 5) ≥17,500 steps per day (19). The descriptors assigned to these ranges were “copper” through “platinum.” These previous data values are inconsistent primarily in the higher end values relative to the quintile-defined categories defined here, but this likely reflects the difference between a criterion-referenced approach (what children “should” be doing relative to a selected criterion, in this case a healthy BMI) and a normative reference approach (what children actually do, and relative only to each other). Both approaches have their unique advantages and disadvantages (2,8), and thus providing that these additional normative data are not only justified but needed.
Defining categories of activities by steps per day requires some sort of qualitative label to facilitate easy communication. Congruent with an adult graduated step index (20), each of the increasing categories (within each sex) could be interpreted as “sedentary,” “low active,” “somewhat active,” “active,” and “highly active.” In keeping with calls to use the term “sedentary” only in regard to nonmovement and specifically in reference to sitting behaviors (15), it makes sense to relabel the lowest category “inactive.” An alternative choice is to follow convention practiced by the Canadian Physical Activity, Fitness, and Lifestyle Approach to categorize anthropometric and performance-testing results: “needs improvement,” “fair,” “good,” “very good,” and “excellent.”(4) There were concerns over potential psychological effects related to such labels, therefore grounded in prevalent approaches to sports awards, there have also been attempts (20) to label children’s increasing step-defined physical activity categories as “copper,” “bronze,” “silver,” “gold,” and “platinum.” Furthermore, value-laden terms such as “good” imply having met a standard such as current physical activity guidelines (emphasizing time and intensity), whereas a consensus regarding step-based guidelines have yet to be achieved. Having carefully considered all options, we settled on a set of qualitative labels that we believe correctly and clearly communicate relative position in a distribution compared with peers: lowest, lower than average, average, higher than average, and highest. We believe that this strategy attempts to describe relative status across levels of a modifiable behavior (that which an individual has the ability to control, at least to some degree). However, we acknowledge that it may be difficult to avoid any potential for stigmatization whatever labels are applied; a lower level naturally implies a neglected behavior, for whatever reasons. Prudent application is advised when used at the individual level. However, we also firmly believe that such data-driven interpretation is necessary on all levels and should not be disregarded in an attempt to be excessively polite instead of proactive.
These data are based on a large representative sample of the household-based Canadian children and adolescents. Although physical ability, learning disabilities, and other such information was not asked in the recruitment interview, it is likely that the data were gathered only from those children/adolescents able to perform ambulatory physical activities. Furthermore, CANPLAY participants are more likely to be younger (5–10 yr), are less likely to live in lower income households, and have parents who are more educated (university graduates) and more active than their peers. Therefore, the application of these data will be most directly appropriate to similar populations. The data were collected using a research-grade (10,16) pedometer, and again, data obtained using similar instrumentation is most comparable. The Yamax SW-200 that has been adopted by CANPLAY is among the most commonly used pedometers in child and adolescents samples (21). There was no need to adjust these data in any manner to make them more applicable. The pedometer used in CANPLAY is a relatively inexpensive instrument (<Can$30), making it an accessible choice to nonresearch users including a full range of childcare workers, parents, and children/adolescents themselves. Although the grand majority of participants provided >5 d of data, we considered any number of days logged. We previously demonstrated no data patterns indicative of reactivity and that even a single day of monitoring is defensible in terms of population monitoring (7).
These normative pedometer-determined physical activity data were systematically collected for 6 yr and represent 21,129 children and 12,859 adolescents from across Canada. No other data source is comparable in terms of size or rigor at this time. The percentiles offered herein are of great value to researchers, clinicians/practitioners other childcare workers, and families interested in comparing children’s and adolescents’ pedometer-determined data to that of their peers.
This study has been made possible through a financial contribution from the Public Health Agency of Canada and the Interprovincial Sport and Recreation Council. The views expressed herein do not necessarily represent the views of the Public Health Agency of Canada or the Interprovincial Sport and Recreation Association. Furthermore, the results of the present study do not constitute endorsement by the American College of Sports Medicine.
The authors declare no conflicts of interest.
1. Beets MW, Bornstein D, Beighle A, Cardinal BJ, Morgan CF. Pedometer-measured physical activity
patterns of youth: a 13-country review. Am J Prev Med
. 2010; 38: 208–16.
2. Berk RA. Determination of optimal cutting scores in criterion referenced measurement
. J Exp Educ
. 1976; 45: (4–9).
3. Canadian Fitness and Lifestyle Research Institute. Fitness Results from the Campbell Survey. 1988 Campbell Survey on Well-being in Canada. Canadian Fitness and Lifestyle Research Institute. Ottawa, ON [cited 2012 April 17]. Available from: http://cflri.ca/media/node/489/tables/88norms.pdf
4. Canadian Society for Exercise
Physiology. The Canadian Physical Activity, Fitness and Lifestyle Approach (CPAFLA) 3rd edition
. Ottawa, ON: Canadian Society for Exercise
Physiology; 2003. p. 7–31.
5. Colley RC, Garriguet D, Janssen I, Craig CL, Clarke J, Tremblay MS. Physical activity
levels of Canadian children and youth: Results from the 2007–2009 Canadian Health Measures Survey. Health Rep
. 2011; 22: 1. Statistics Canada Catalogue 82-003-XPE.
6. Craig CL, Cameron C, Griffiths JM, Tudor-Locke C. Descriptive epidemiology of youth pedometer-determined physical activity
: CANPLAY. Med Sci Sports Exerc
. 2010; 42 (9): 1639–43.
7. Craig CL, Tudor-Locke C, Cragg S, Cameron C. Process and treatment of pedometer data collection for youth: the CANPLAY study. Med Sci Sports Exerc
. 2010; 42 (3): 430–5.
8. Cureton KJ, Warren GL. Criterion-referenced standards for youth health-related fitness tests: a tutorial. Res Q Exerc Sport
. 1990; 61: 7–19.
9. Esliger DW, Probert A, Gorber SC, Bryan S, Laviolette M, Tremblay MS. Validity of the Actical accelerometer step-count function. Med Sci Sports Exerc
. 2007; 39 (7): 1200–4. PubMed PMID: 17596790.
10. Le Masurier GC, Tudor-Locke C. Motion sensor accuracy under controlled and free-living conditions. Med Sci Sports Exerc
. 2004; 36 (5): 867–71.
11. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity and trends in body mass index among US children and adolescents, 1999–2010. JAMA
. 2012; 307: 483–90. Epub 2012 Jan 17. PubMed PMID: 22253364.
12. Payne N, Gledhill N, Katzmarzyk PT, Jamnik VK, Keir PJ. Canadian musculoskeletal fitness norms. Can J Appl Physiol
. 2000; 25: 430–42. PubMed PMID: 11098155.
13. Rosenkranz RR, Rosenkranz SK, Weber C. Validity of the Actical accelerometerstep-count function in children. Pediatr Exerc Sci
. 2011; 23: 355–65. PubMed PMID: 21881156.
14. Rowlands AV, Eston RG, Ingledew DK. Relationship between activity levels, aerobic fitness, and body fat in 8- to 10-yr-old children. J Appl Physiol
. 1999; 86: 1428–35.
15. Sedentary Behaviour Research Network. Letter to the Editor. Standardized use of the terms “sedentary” and “sedentary behaviour.” Appl Physiol Nutr Metab
. 2012; 37: 540–2.
16. Tudor-Locke C., Ainsworth BE, Thompson RW, Mathews CE. Comparison of pedometer and accelerometer measures of free-living physical activity
. Med Sci Sports Exerc
. 2002; 34 (12): 2045–51.
17. Tudor-Locke C, Bassett DR Jr. How many steps/day are enough? Preliminary pedometer indices for public health. Sports Med
. 2004; 34: 1–8.
18. Tudor-Locke C, Craig CL, Beets MW, et al.. How many steps/day are enough? For children and adolescents. Int J Behav Nutr Phys Act
. 2011; 8: 78.
19. Tudor-Locke C, Hatano Y, Pangrazi RP, Kang M. Revisiting “how many steps are enough?” Med Sci Sports Exerc
. 2008; 40 (7 Suppl): S537–43.
20. Tudor-Locke C, Johnson WD, Katzmarzyk PT. Accelerometer-determined steps per day in U.S. children and youth. Med Sci Sports Exerc
. 2010; 42 (12): 2240–50.
21. Tudor-Locke C, McClain JJ, Hart TL, Sisson SB, Washington TL. Expected values for pedometer-determined physical activity
in youth. Res Q Exerc Sport
. 2009; 80: 164–74.
22. Tudor-Locke C, Myers AM. Methodological considerations for researchers and practitioners using pedometers to measure physical (ambulatory) activity. Res Q Exerc Sport
. 2001; 72: 1–12.
23. Tudor-Locke C, Pangrazi RP, Corbin CB, et al.. BMI-referenced standards for recommended pedometer-determined steps/day in children. Prev Med
. 2004; 38: 857–64.
Keywords:©2013The American College of Sports Medicine
WALKING; PHYSICAL ACTIVITY; EXERCISE; ASSESSMENT; MEASUREMENT