Walking is increasingly recognized as a straightforward and effective way of implementing regular, health-enhancing physical activity into the daily routine of the general population (12,15). It is inexpensive, easy, convenient, low impact, and can be performed in social settings (i.e., group walks). Walking promotion interventions, often used in conjunction with pedometers (step counters), can incorporate goals expressed in terms of time spent walking, steps taken during walking, or total daily step count (11).
A variety of recommendations has been published to prompt and promote physical activity. Common recommendations for adults are to accumulate a daily total of 10,000 steps or to add 30 min of walking to a normal sedentary daily routine (17). Recent guidelines have emphasized that walking should be of at least moderate intensity to improve health (6). Pedometers have been widely used in physical activity research, but previous research has demonstrated that daily step counts do not necessarily correspond to time spent in moderate-intensity physical activity (13,14). While walking interventions that use goal setting have been effective in increasing daily steps (2), these rarely monitor the intensity of walking because of technological or practical challenges. Some guidance may suggest a specific step rate target (e.g., try to complete 3000 steps in 30 min), but many pedometers are not able to provide feedback on walking intensity. Some recently developed models record steps and/or time above a specific intensity in addition to the total step count. For example, various models of the Omron pedometer (Omron Corp., Schaumburg, IL) use an "aerobic steps" function corresponding to a step rate of 60 steps per minute. However, there is no existing research evidence to support that this stepping rate leads to improved health and increased aerobic fitness or is associated with moderate-intensity walking.
Step rate corresponding to moderate-intensity walking has been investigated in two previous studies (9,18). Both studies led to the conclusion that a step rate of approximately 100 steps per minute corresponded to moderate-intensity walking across the whole sample. However, separate guidelines were recommended for men and women (in both studies, the recommended step rate for men was lower than that for women). The gender difference of 10 (9) and 11 steps per minute (18) was unexplained, but this could have been due to a gender difference in height and, therefore, in stride length. For example, from secondary analyses of the descriptive data of these studies, it was determined that the male-to-female ratio of height was approximately inversely proportional to the male-to-female ratio of recommended stride rate.
Both of these previous studies were conducted in well-controlled laboratory conditions on a treadmill. Walking (and running, in the study of Tudor-Locke et al. ) speeds were discrete, that is, these did not cover a range of intermediate speeds between the prescribed speeds of 2.4 mph (3.9 kph), 3.0 mph (4.8 kph), 3.5 mph (5.6 kph), and 4.1 mph (6.6 kph) (9) and 3.0 mph (4.8 kph), 4.0 mph (6.4 kph), and 6.0 mph (9.7 kph) (18). The contribution of height (and/or stride length) was not addressed in either study. In addition, no overground walking data have been presented in studies of stride rate and energy expenditure to determine stride rate guidelines for moderate-intensity overground walking.
Therefore, the purpose of the current study was to replicate the general study protocols of Tudor-Locke et al. (18) and Marshall et al. (9), incorporating the following design features: (a) include a range of randomly assigned speeds; (b) focus on walking (i.e., exclude running); (c) include overground walking trials; and (d) determine the influence of height and/or stride length by including several stride length measures and height. The overall goal was to develop stride rate guidelines corresponding to moderate-intensity walking overground, adjusted for height and/or stride length.
The study was conducted at three different sites. The same protocol was followed at each site, using similar equipment. Data were collected from each participant in a single testing session, consisting of the following four sets of procedures, in the order listed: (a) anthropometric data collection, (b) three 6-min walking trials on a treadmill, (c) three indoor overground walking trials of at least 4 min, and (d) two stride length tests.
Participants (N = 75; mean age = 32.9 ± 12.4 yr; n = 38/37 females/males) volunteered at three different university sites (n = 25 at each site) in the United States. Recruitment was via flyers and a university LISTSERV, targeting university employees and their families. All participants provided written informed consent for all procedures, which were approved by the institutional review boards of the three participating universities. In compensation for their time, all participants were given US $10 and a pedometer.
On arrival at the laboratory, the study was verbally explained to participants, and they were given a study information sheet to read. After being given the opportunity to ask questions, they were asked to sign the informed consent form if they wished to participate in the study. Each participant was tested by two researchers. In all cases where more than one measure was taken, the average was used for data analysis. Before data collection, the metabolic systems and pedometers were calibrated using standard procedures. All pedometers were 100% accurate during a 100-step walk test.
Height in centimeters (HtCm) was measured to the nearest 0.1 cm from two trials, with shoes removed. Iliac height (IH) was measured using a standard tape measure, from the iliac crest to the floor, with shoes removed (two trials, measured to the nearest 0.1 cm), as an indicator of leg length. Weight was measured from two trials, using a physician's weighing scale, to the nearest 0.1 kg. Sitting height was measured while the participant was sitting on a solid table, from two trials, measured to the nearest 0.1 cm. Sitting height was subsequently used to calculate another leg length estimate (LL), by subtracting sitting height from standing height.
Treadmill walking trials.
Participants completed three 6-min controlled treadmill trials at slow, moderate, and fast walking speeds. For each participant, a set of three evenly spaced walking speeds was randomly assigned. Slow speeds were between 2.0 mph (3.2 kph) and 2.6 mph (4.2 kph) in 0.1-mph (0.16-kph) increments. Similarly, moderate speeds were between 2.7 mph (4.3 kph) and 3.3 mph (5.3 kph), and fast speeds were between 3.4 mph (5.5 kph) and 4.0 mph (6.4 kph). Examples of two randomly assigned sets of walking speeds would therefore be 2.0 mph/3.2 kph (slow), 2.7 mph/4.3 kph (medium), and 3.4 mph/5.5 kph (fast) or 2.6 mph/4.2 kph (slow), 3.3 mph/5.3 kph (medium), and 4.0 mph/6.4 kph (fast). In addition, the order of trials (e.g., slow, moderate, then fast; fast, slow, then moderate) was randomly counterbalanced among participants.
Before the trials, a brief warm-up was given, followed by practice stepping onto and off the treadmill at the three assigned speeds. Participants were then fitted with (a) an Accusplit AH120-M9 pedometer (Accusplit, Livermore, CA) worn on the waistband above the midline of the right knee, (b) an HR monitor (Polar Electro, Oulu, Finland), and (c) the mouthpiece or facemask and metabolic system (makes and models differed across testing sites). For each trial, the treadmill was set to the appropriate speed, and after a 5-s countdown, the participant stepped onto the treadmill and began walking. The event marker on the metabolic system was pressed immediately before and after each trial, for later reference in the V˙O2 data, and the pedometer was reset to zero before each trial. During the first minute, treadmill speed was measured twice with an electronic tachometer for subsequent data analysis.
Stride rate was measured using three different methods. Total step counts during each 6-min trial were measured using the pedometer. On two occasions after the first minute, time taken for 10 strides was recorded, using a digital stopwatch (to replicate and test the protocol previously used by Heil et al. 8). Also, the number of steps taken during the fifth minute for each speed was counted using a hand tally counter and was used as the criterion measure of the 60-s step rate in the current study. Toward the end of each trial, the participant was given a 5-s countdown to notify when he/she should step off the treadmill. Within each trial, HR was recorded during the last 15 s of each minute, for determination of steady state, as per the recommendations of the American College of Sports Medicine (1). Two minutes of static rest was taken between trials.
Overground walking trials.
Similar to the treadmill trials, participants completed three overground walking trials at three different speeds (in the same order as the treadmill trials). To simulate the conditions of the treadmill trials, the treadmill stride rate obtained from the 60-s hand tally count was prescribed for the overground trials. This was accomplished by setting a clip-on metronome to the treadmill stride rate and by asking participants to match their stride rate to the metronome. Because of logistic challenges, it was not possible to hand-count the steps during the overground trials. However, throughout the overground trials, a researcher walked behind each participant, and it was observed that almost all participants matched the metronome stride rate perfectly. Of those who did not, only one or two missteps occurred during a complete overground trial.
Unlike the treadmill trials, overground trials were not limited to 6 min for logistical reasons (because participants were required to walk a complete number of laps to provide a known distance for calculating average walking speed). Although the track distance varied among the three sites, all tracks were indoors in the same building as the laboratory where the treadmill trials were completed and were oval (i.e., with long straightways and broad, sweeping curves). To obtain steady-state data, participants walked at least 4 min (i.e., at the 4-min point, participants were asked to complete the current lap). Similar to the treadmill trials, the event marker was pressed immediately before and after each trial, for reference when analyzing the V˙O2 data, and HR was recorded during the last 15 s of each of the first 4 min and during the last 15 s of the trial to determine steady state.
Stride length tests.
Two trials each of two overground stride length tests were conducted after all walking trials, so that participants were fully warmed up when performing the stride length tests. For both tests, participants were asked to walk at their normal walking pace and begin at a start line with feet together. The 10-m stride test (10-mT) involved walking toward a target point approximately 15 m from the start line. As participants walked toward the target point, steps were counted by a researcher up to a marked 10-m point (therefore, the resulting score was the complete steps taken before crossing the 10-m point). For the 10-stride test (10-sT), participants were asked to walk toward a distant target point until asked to stop. A measuring tape was laid on the ground alongside the walking area, and participants started with their heels in line with the zero point of the measuring tape. Researchers counted the steps and visually noted the heel strike of the 10th stride. The recorded score was therefore the distance in meters (to the nearest centimeter) that was taken to walk 10 single steps.
Data Processing and Analysis.
After each test, calorimetry data were downloaded, and V˙O2 was determined for the final 2 min of each treadmill and overground walking trial. Descriptive statistics were calculated for the study variables, and multiple regression analysis was used to develop a regression equation to predict overground V˙O2 from stride rate and the various stride length indicators (height, two different measures of leg length, and two different stride length tests). A standard multiple regression analysis was run initially to replicate the analyses used by Tudor-Locke et al. (18) and Marshall et al. (9). Subsequently, mixed model regression (otherwise known as a random coefficients model) was used to develop the equation used for determining the stride rate cut points, as in Marshall et al. (9), to account for nonindependence of observations (i.e., multiple data points obtained from each participant). In the mixed model regression, individual intercepts and slopes are estimated separately for each participant, and the repeated factor is modeled as a random rather than a fixed factor (3,16,19). All analyses were conducted using SPSS version 16.0.1 (SPSS, Inc., Chicago, IL). The testing site was also tested as a potential predictor to determine whether any systematic bias was introduced across sites.
Descriptive data are presented in Table 1. Participants covered a broad range of age, height, weight, and body mass index (BMI). On the basis of commonly used criteria (20), no participants were classified as underweight, 22 (29%) were classified as overweight, and 11 (15%) were classified as obese. All variables were relatively normally distributed (skewness and kurtosis < |2.0|), justifying the use of parametric data analysis methods.
Multiple regression analyses
Results of the multiple regression analysis are presented in Table 2. Each of the stride length-related variables (HtCm, IH, 10-mT, and 10-sT) explained significant (P < 0.05) additional variance in V˙O2 when added to stride rate, except for LL, which was nonsignificant (P = 0.052). Because the amount of additional variance was not meaningfully different among the different stride length-related variables and because height is the most easily measured, most readily available, and the most intuitively interpretable variable in everyday situations, stride rate guidelines were subsequently developed for people of different heights.
Development of height-related stride rate recommendations.
To adjust for the random effects of nonindependent observations, a mixed model regression was used to develop the regression equation for determining stride rate recommendations and to test for the influence of research site. In agreement with the multiple regression analysis, stride rate and HtCm were again significant predictors of V˙O2 (P < 0.05). However, research site did not add significantly to prediction accuracy (P = 0.45), indicating that there was no systematic effect of site on V˙O2 beyond the combined effect of stride rate and HtCm. For comparison with the previous studies of Tudor-Locke et al. (18) and Marshall et al. (9), the equation using only stride rate is provided below:
The equation including stride rate and HtCm is as follows:
Notably, the unstandardized slope for HtCm (0.093274) indicated a meaningful relationship between height and V˙O2 at any given stride rate, such that for every additional 10 cm (4 inches) in height, at a given stride rate V˙O2 would be 0.9 mL·kg−1·min−1 (or 0.3 METs) higher. Equation 1 was solved to obtain 3 METs (using the standard definition of 1 MET = 3.5 mL·kg−1·min−1) stride rate cut points for comparison with previous studies (9,18), resulting in a cut point of 103 steps per minute. Using equation 2, 3-MET stride rate cut points were developed for various heights between 60 inches (152.4 cm) and 78 inches (198.1 cm). Fractional results (e.g., 106.4 steps per minute) were rounded up to the next integer, on the premise that the cut points should correspond to a minimum of 3 METs. These ranged from 90 to 113 steps per minute and are presented in Table 3. On the basis of common health recommendations of 30 min of moderate-intensity physical activity, these correspond to 30-min step count targets ranging from 2700 to 3390 steps. Stride rate cut points for 1-cm height increments are available from the first author. For reference, 4-MET and 5-MET cut points are also provided.
To further evaluate the validity of the height-related 3-MET stride rate cut points, we investigated the accuracy of using stride rate as an indicator of whether the walker is engaging in moderate-intensity walking (≥3 METs, from measured V˙O2). Based on the regression-derived stride rate guidelines, participants were coded for each of the three overground walking trials as walking either below or above the 3-MET stride rate cut point recommended for their height. They were also coded as being either below or above 3 METs based on measured V˙O2. Using these codings, and using measured MET intensity as the criterion, sensitivity and positive predictive value were 0.88 and 0.82, respectively, whereas specificity and negative predictive value were 0.51 and 0.62, respectively.
The main purpose of this study was to determine the influence of stride length-related parameters on stride rates corresponding to walking at moderate intensity (defined as 3 METs) during overground walking. The study builds on two previous treadmill studies, one in ostensibly healthy adults (18) and the other in predominantly overweight or obese adults (9). In both previous studies, stride rate recommendations were different for men and women. In the study by Tudor-Locke et al. (18), the values were 96 and 107 steps per minute for men and women, respectively, with a gender-weighted average of 102 steps per minute. In the study of Marshall et al. (9), the corresponding values were 101, 111, and 106 steps per minute, respectively. In both previous studies, the researchers suggested a convenient population cut point of 100 steps per minute.
In general agreement with the two previous studies, the generic stride rate associated with walking at 3 METs in the current study was 103 steps per minute. The replication of the previous treadmill-based findings in overground walking supports the use of this approximate rule of thumb in many practical situations. For many reasons, the use of a single recommendation is attractive (not least because it is easy to remember, and easy to use in calculations of total step targets for fixed durations, e.g., 1000 steps in 10 min) and can form the basis for a simple, general public health recommendation. In addition, general recommendations of 120 steps per minute for 4 METs and 140 steps per minute for 5 METs can be inferred from the current study. As recognized by both previous sets of researchers, however, the general recommendation of 100 steps per minute was associated with considerable interindividual deviation from the 3 MET intensity that is the recommended minimum for health-enhancing physical activity, and our study has shed light on potential reasons for this. Especially when used for prescriptive purposes, whether the goal is to achieve the 3-MET minimum or a specified higher intensity, two potential problems may arise from a single recommendation for individuals who differ in height. First, if the goal is to achieve 3 METs, shorter individuals will not achieve the moderate-intensity threshold, thereby potentially not achieving health benefits. Second, in progressive programs where the goal may be to achieve higher intensities such as 4 or 5 METs, the one-size-fits-all approximation may lead to taller individuals (who typically walk at lower cadences compared with shorter individuals) finding the prescription difficult to maintain, which could influence adherence. This underlines the importance of the primary additional finding of the current study, namely, that stride rates corresponding to specific intensities of walking should be adjusted for height. For example, the 3-MET stride rates recommended for adults 60 inches tall and 78 inches tall are different by 23 steps per minute (113 vs 90 steps per minute, respectively). In addition, applying the 100-steps-per-minute generic guideline to people of all heights would result in a 3-MET walking intensity in an adult 70 inches tall, but only 2.3 METs in an adult 60 inches tall, and 3.6 METs in an adult 78 inches tall. Because the measurement of height is straightforward, this additional information is valuable and easily incorporated into individualized walking recommendations. In addition, it leads to scientifically more accurate monitoring of walking intensity that is extremely useful outside the general public health recommendation (e.g., for scientific research purposes or for physical activity professionals or members of the public who wish to use a more accurate rubric with minimal added inconvenience or calculations).
From the coefficient (slope) for stride rate in the current regression equation, we can determine that an increase in stride rate of 19 steps per minute would result in an increased walking intensity of 1 MET (this was true for the generic equation and the height-dependent equation). This contrasts with the two previous studies, which suggested that an increase of 9 to 10 steps per minute (18) or 8 to 10 steps per minute (9) would result in an increase of 1 MET. The differences may be partially explained by the inclusion of a running speed in the study of Tudor-Locke et al. (18). From the secondary analysis of these data on the two walking trials only, the resultant slope corresponds to a 21-steps-per-minute increase in walking cadence to achieve an increase of 1 MET. Interestingly, the slopes from the regular regression models in the study of Marshall et al. (9) correspond to a 24-steps-per-minute (women) or a 16-steps-per-minute (men) increase in walking cadence to achieve an increase of 1 MET. The considerable difference in slope between the regular regression equation and mixed model equation was unexplained and contrasted with the current study, in which the slopes were relatively similar for the regular regression (b = 0.040) and mixed model regression (b = 0.052) analyses (slopes were converted to MET outcome for direct comparison with Marshall et al. ). No equation was provided in the study of Marshall et al. (9) for all participants, but one must assume that the common slope would correspond to a step rate increase somewhere between 16 and 24 steps per minute for each 1-MET increase in walking intensity.
In terms of the strength of the relationship between stride rate and measured MET walking intensity, the equation from the current study explained 38% of the variance in measured MET compared with 35% (men) and 23% (women) in the study of Marshall et al. (9) and 80% (men) and 83% (women) variance in the study of Tudor-Locke et al. (18). The SEE in the current study (0.67 MET) was lower than the 1.30 METs (men) and 1.52 METs (women) in the study of Marshall et al. (9) and the 1.27 METs (men) and 1.14 METs (women) in the study of Tudor-Locke et al. (18). The regression equation in the current study therefore produced slightly higher relative predictive ability (in terms of percent variance) compared with the study of Marshall et al. (9) but lower compared with the study of Tudor-Locke et al. (18). In terms of absolute prediction accuracy, however (SEE), the current study improved considerably on the results of both previous studies.
Criterion-referenced prediction accuracy (below/above 3 METs) of stride rate cut points was only reported previously in the study of Marshall et al. (9). Using their stride rate cut points, 52% of truly moderate-intensity walking bouts were correctly classified (sensitivity) and 48% were incorrectly classified (100% − sensitivity). In the current study, 88% of bouts >3 METs were correctly classified, and 12% were incorrectly classified. In the study of Marshall et al. (9), true-positive rate was 67%/43% for men/women, and false-positive rate was 27%/22% men/women (no data were presented across the whole sample). In the current study, true-positive rate was 88%, and false-positive rate was 49%. This indicates that sensitivity (the ability of the height-related stride rate recommendations to accurately identify walking at >3 METs) was high and that the positive predictive value (the likelihood that a stride rate greater than the recommendations would be >3 METs) also was high, improving considerably on the findings of Marshall et al. (9). However, the false-positive rate was worse (i.e., a greater number of bouts above the recommended stride rate were actually <3 METs) compared with that of Marshall et al. (9).
The current study has several strengths. The sample was heterogeneous (notably in terms of age and body fatness), making the results generalizable to the population. Data were collected over a continuum of randomly assigned treadmill speeds from slow (2.0 mph/3.2 kph) to fast (4.0 mph/6.4 kph). Oxygen consumption was measured during overground walking, making the results generalizable to recreational walking.
The collection of data at three different sites was an additional strength. Data collected at a single site may be subject to systematic, site-specific biases (e.g., procedural bias, test administrator bias, or equipment bias), and so the combination of data from different sites (following standardized procedures) strengthens the study. Different metabolic carts were used at the different sites (Cosmed K4b2 (Cosmed SRL, Rome, Italy), Oxycon Mobile (Viasys Healthcare, Inc., Yorba Linda, CA), and Cosmed Quark b2 (Cosmed SRL)). One could therefore question the equivalence of the V˙O2 data collected at the three sites. These concerns can be addressed in two ways. The Cosmed has been shown to compare similarly to four other laboratory metabolic systems in previous research (4,5,7,10), which indicates that any interinstrument bias can probably be assumed to be minimal. More importantly, the addition of site as a predictor in the mixed model analysis did not add significantly or meaningfully to the prediction of overground walking energy expenditure, indicating the lack of a site-specific or instrument-specific bias.
A limitation of the current study was that steps were not hand-counted during the overground trials. From the treadmill trials, we determined there were several outlying pedometer scores compared with the hand-counted criterion, most notably during slower speeds and in participants with higher BMI. We therefore decided against using the pedometer data from the overground trials, particularly because our observations indicated that participants followed the metronome-directed stride rate correctly.
The results of the current study, in addition to the basic secondary calculations of the previous study results described in the introduction, indicate that gender differences in stride rate recommendations found in the two previous studies may be explained by gender differences in height. We should note that 3 METs is the minimum recommended intensity for improving health and that walking at a higher intensity would bring additional health benefits (hence, we have also provided step rate cut points for 4 and 5 METs). Although recent physical activity recommendations (6) incorporate adjusted guidelines to take account of vigorous physical activity (>6 METs), this may not be feasible for walking, especially in some populations. From our data, stride rates corresponding to 6 METs (161 steps per minute overall, and a range of height-related cut points of 147-170 steps per minute) would be impractical for the general population, and in our overground trials, few participants walked at an intensity >6 METs. In addition, one participant expressed concern during the treadmill practice regarding the fastest walking speed and was therefore reassigned a slower set of speeds. In the study of Marshall et al. (9), 16% of participants were unable to maintain a walking speed of 3.5 mph (5.6 kph), and 34% were unable to maintain 4.1 mph (6.6 kph) for 6 min. Obese participants, in particular, were unable to maintain the faster speeds (22% and 57% drop-out at 3.5 mph/5.6 kph and 4.1 mph/6.6 kph, respectively). Although all participants in the study of Tudor-Locke et al. (18) were able to maintain a walking speed of 4.0 mph/6.4 kph (mean ± SD = 5.2 ± 0.8 mph, 8.4 ± 1.3 kph, and 5.3 ± 0.6 METs in men and women, respectively), the sample consisted of healthy young adults.
The use of 3 METs (or moderate intensity) as the threshold for health-enhancing physical activity is based on a public health message that is fairly consistent internationally. There are two potential dangers of emphasizing this message too strongly. The first is that these recommendations are intended for improving or maintaining health in the general population, and they may not apply to populations such as frail older adults and individuals with chronic medical conditions. Similarly, there is a growing body of evidence that long periods of sedentary behavior, such as sitting, carry a health risk that is independent of physical activity. Engaging in brief intermittent episodes of physical activity (regardless of intensity) may therefore carry a preventive health benefit that is independent of the "30 min of moderate-intensity physical activity" message. The current cadence guidelines do not apply to these types of populations or to the message of avoiding long sedentary periods, and we would caution against the perception that walking is only beneficial if it is conducted in longer bouts and above moderate intensity.
Future directions for related research include investigating how walkers can be prompted in "real time" to habitually walk at a prescribed cadence. Although participants in the current study were able to match the metronome cadence, it is unclear whether this would be possible over longer bouts (e.g., 10-30 min of continuous walking), especially for previously sedentary adults who are unaccustomed to walking for longer periods, or whether the metronome would eventually become a nuisance. In addition, newer pedometer models provide feedback on time spent in moderate-intensity physical activity or higher, and the agreement between this output and the stride rates recommended here warrant investigation. Perhaps more importantly in the long term, the relationship between walking cadence and health outcomes should be investigated in randomized control trials.
In conclusion, the current study clarifies the use of stride rate guidelines to ensure that overground walking is conducted at a moderate intensity (defined as 3 METs). The current study confirms the overall generic recommendation of approximately 100 steps per minute (or 3000 steps in 30 min) arising from two previous treadmill-based studies (9,18), which adds confidence to this recommendation as a general heuristic for moderate-intensity overground walking. Our slope of 19 steps per minute, which was replicated in our secondary analysis of the walking data of Tudor-Locke et al. (18) (21 steps per minute) and was similar to the gender-averaged regular regression slope of the study of Marshall et al. (9), suggests the use of heuristic guidelines of 120 and 140 steps per minute corresponding to 4 and 5 METs, respectively. Importantly, we have also clearly demonstrated the need to consider stride length when more precise estimation of walking intensity is needed. This is conveniently accomplished by measuring height because height explained the stride rate/MET relationship similarly to the more complicated leg length and stride length tests investigated in the current study.
The project was partly funded by a monetary and equipment donation from Accusplit, Livermore, CA, which placed no restrictions on use of the data. The authors declare that they have no competing interests.
The authors thank Ashley Guerieri and Jennifer Aycock for helping with data collection at the East Carolina University site as part of an undergraduate special project class. The authors also thank Catrine Tudor-Locke and Simon Marshall, who provided additional information and/or data for secondary analysis, to more accurately contextualize our results with the two previous studies.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. American College of Sports Medicine. ACSM's Guidelines for Exercise Testing and Prescription
. 8th ed. Philadelphia (PA): Lippincott Williams & Wilkins; 2010. p. 81.
2. Baker G, Gray SR, Wright A, et al. The effect of a pedometer-based community walking intervention "Walking for Wellbeing in the West" on physical activity levels and health outcomes: a 12-week randomized controlled trial. Int J Behav Nutr Phys Act
3. Blackwell E, de Leon CF, Miller GE. Applying mixed regression models to the analysis of repeated-measures data in psychosomatic medicine. Psychosom Med
4. Duffield R, Dawson B, Pinnington HC, Wong P. Accuracy and reliability of a Cosmed K4b2 portable gas analysis system. J Sci Med Sport
5. Eisenmann JC, Brisko N, Shadrick D, Welsh S. Comparative analysis of the Cosmed Quark b2 and K4b2 gas analysis systems during submaximal exercise. J Sports Med Phys Fitness
6. Haskell WL, Lee IM, Pate RR, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc
7. Hausswirth C, Bigard AX, Le Chevalier JM. The Cosmed K4 telemetry system as an accurate device for oxygen uptake measurements during exercise. Int J Sports Med
8. Heil DP, Higginson BK, Keller CP, Juergens CA. Body size as a determinant of activity monitor output during overground walking. J Exerc Physiol Online
9. Marshall SJ, Levy SS, Tudor-Locke CE, et al. Translating physical activity recommendations into a pedometer-based step goal: 3000 steps in 30 minutes. Am J Prev Med
10. McLaughlin JE, King GA, Howley ET, Bassett DR Jr, Ainsworth BE. Validation of the Cosmed K4 b2 portable metabolic system. Int J Sports Med
11. Merom D, Rissel C, Phongsavan P, et al. Promoting walking with pedometers in the community: the step-by-step trial. Am J Prev Med
12. Morris JN, Hardman AE. Walking to health. Sports Med
13. Rowe DA, Kemble CD, Robinson TS, Mahar MT. "Healthy stepping": accuracy of the 10,000 steps/day recommendation for identifying sufficiently active older adults. In: Proceedings of the Centers for Disease Control and Prevention International Congress on Physical Activity and Public Health
. 2006. p. 139.
14. Rowe DA, Mahar MT, Kemble CD, Robinson TS, Raedeke TD. Associations between total daily step count and minutes of "healthy"physical activity. Med Sci Sports Exerc
. 2006;38(5 suppl):S378.
15. Scottish Executive. Let's Make Scotland More Active: A Strategy for Physical Activity
. Edinburgh (UK): Scottish Executive; 2003. p. 35.
16. Treuth MS, Schmitz K, Catellier DJ, et al. Defining accelerometer thresholds for activity intensities in adolescent girls. Med Sci Sports Exerc
17. 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.
18. Tudor-Locke C, Sisson SB, Collova T, Lee SM, Swan PD. Pedometer-determined step count guidelines
for classifying walking intensity in a young ostensibly healthy population. Can J Appl Physiol
19. Welk GJ. Principles of design and analyses for the calibration of accelerometry-based activity monitors. Med Sci Sports Exerc
. 2005;37(11 suppl):S501-11.
20. World Health Organization. Obesity: Preventing and Managing the Global Epidemic
. Geneva (Switzerland): World Health Organization; 2000. p. 9.