Valid methods for assessing physical activity patterns are essential for accurate evaluation of intervention programs and population surveillance. Validity of physical activity assessment instruments has typically been examined by correlating self-reported behavior with either physiologic criteria, e.g., cardiovascular fitness, adiposity, heart rate, or blood pressure, or other established measures of physical activity, e.g., accelerometers (3,9,11,12,19,21,22). Numerous physical activity (PA) measures have been validated in white adults (3,9,11,12,21,22); however, few studies have reported the validity of such measures in African-Americans (20). African-Americans may have unique perceptions of health behaviors that can influence how they respond to physical activity assessment instruments, suggesting a need for validation and adaptation of instruments before their use in African-American populations (8,16,17). Validity of activity measures appears to vary by ethnicity (24).
In prior validity studies, correlation coefficients between self-reported measures of activity and indices of cardiorespiratory fitness have been highly variable, ranging from 0.11 to 0.76 (3,9,11,12,19,21,22). Correlations have generally been stronger when the range of activities have been limited to those of moderate and/or vigorous intensity or recreational sports (3,7,9,11,12,19–21,24). In the one study with exclusively African-American participants, this pattern was evident for women but not men, indicating that this relationship is neither well defined nor understood in African-Americans (20).
In this article, we examine the validity of a modified version of the CHAMPS physical activity questionnaire (9,22) in a sample of African-American adults participating in the Healthy Body/Healthy Spirit intervention trial. Validity is examined by comparison of the modified CHAMPS measure to cardiorespiratory fitness (estimated maximal V̇O2) and physiologic parameters related to fitness, used in prior validity studies, e.g., blood pressure and body mass index (BMI) (7,9,18,22,24) as well as by comparison to other established measures of self-reported activity. Previous CHAMPS validity studies have been conducted in predominantly white or Asian-American populations (9,22).
METHODS
Study design and participants.
Data were obtained during the baseline assessment from the Healthy Body/Healthy Spirit Trial, a federally funded study to increase fruit and vegetable intake and physical activity among African-American adults recruited from African-American churches in the Atlanta metropolitan area (17). Informed written consent was obtained from all participants, and the project was approved by the institution’s human investigations committee. Sixteen churches were randomly assigned to three intervention conditions. Group 1 received standard nutrition and PA intervention materials, group 2 received culturally tailored self-help nutrition and PA intervention materials, and group 3 received the same intervention as group 2, plus four telephone counseling calls based on Motivational Interviewing (13). Data from a 17th church that participated in a pilot of the intervention were also included, as the relevant measures and data collection procedures were identical to those used in the full trial. Baseline data were collected at health fairs at each church before initiation of any intervention activities, so results here are reported without regard to experimental group. A total of 1078 participants were recruited into the trial, of which a subsample of 169 completed one of two submaximal cardiovascular fitness tests (this subsample is described in more detail below). Additional details regarding the main trial can be found elsewhere (17).
Measures.
The primary physical activity measure in the study was an adaptation of the CHAMPS physical activity recall (9,22), a self-administered instrument developed for underactive populations and tested primarily among older adults. The original measure included 41 items. The frequency of activity is assessed in times per week, and duration is classified using six categories ranging from “less than 1 h·wk−1” to “9 or more h·wk−1.” Harada et al. (9), using the original CHAMPS measure, found 2-wk test-retest correlations for total activity and moderate-intensity activity of 0.62 and 0.76, respectively. Stewart et al. (22) found 6-month reliability coefficients for total and moderate-intensity activities of 0.66 and 0.76, respectively.
Based on focus groups, pilot testing of the instrument, and review of the literature, some modifications were made to the instrument for our population. First, items from the social activities and recreation/hobbies sections were excluded, e.g., use a computer and woodworking, both to reduce response burden and because these activity domains were not targets of our intervention. Second, we modified the recall time frame from the past month to the past 2 wk, as formative work indicated that individuals may be better able to accurately recall activity during this more proximal recall period. Finally, a single item was added that assessed the frequency of dancing during church with the following wording: “How often in the past 2 wk did you dance during church services or sway to the music of the choir?”
The modified CHAMPS instrument can be scored to yield indices of several activity dimensions including frequency of activity and caloric expenditure with and without adjustment for a participant’s weight. To avoid confounding of weight, our primary measure of expenditure was based on kilocalories per kilogram per week, without consideration for participants’ weight. Separate estimates were obtained for moderate to vigorous physical activities (MVPA), defined as those with a MET value ≥ 3.0; vigorous activities, defined as those with a MET value ≥ 5.0; as well as all activities. MET values for activities other than praise dancing were based on the 1993 Ainsworth et al. compendium (1), with the downward adjustments recommended by Stewart et al. (22). The MET value for praise dancing is 5.0 in the revised Ainsworth et al. compendium (2). However, consistent with the recommendation of Stewart et al., we used a MET value of 3.0 in our computations to account for the fact that individuals are not active the entire time they report the activity, and middle-age and elder populations may exert less vigorous effort.
As done in several prior validity studies with similar instruments (20,21), we constructed from the modified CHAMPS items, an index of “Intentional Sports and Recreational Activities,” comprised of 20 exercise or sports-related activities, not part of daily functioning. The activities included in this index were: fast walking, jog/run, biking, aerobic machines, stair/step machine, swimming (gently or moderate/fast), water exercises, stretching, yoga/tai-chi, aerobics, dance, strength training (light, moderate/heavy), general conditioning, basketball, soccer, racquetball, golf (cart and walking), and tennis (singles and doubles). We hypothesized that validity coefficients would be relatively higher for this index, as these activities are intentionally undertaken for the purpose of exercise, and therefore they may be more accurately recalled.
Excluded from the analyses were 34 individuals who failed to respond to at least 16 of the 31 items (frequency or duration) on the modified CHAMPS or whose weekly caloric expenditure for moderate to vigorous-intensity activities exceeded 14,000 kcal. These criteria were invoked to exclude those respondents who either did not completely attend to the questionnaire or who may have misunderstood the instructions and overestimated their activity levels. For those answering 16 or more items, missing values were coded as 0 for both frequency and duration. The 34 excluded individuals did not differ significantly from those included in the analyses with regard to estimated V̇O2max, blood pressure, cholesterol, BMI, any of the Yale activity indices, or any of the walking items, with the exception of flight of stairs climbed per day. Those excluded reported climbing significantly (P = 0.04) fewer flights of stairs per day, 2.4, than those included in the analyses, 3.2.
Other measures of physical activity.
In addition to the modified CHAMPS questionnaire, participants completed the Activity Dimensions section of the Yale Physical Activity Survey (YPAS) (7,15,24). This entails five individual frequency items for vigorous activity, leisurely walking, moving around on your feet, standing, and sitting. In addition, for vigorous activity and leisurely walking, duration is assessed. Frequency (and duration when assessed) items are multiplied by a weighting factor provided by the instrument developers to yield five dimension scores: vigorous activity, leisurely walking, moving, standing, and sitting as well as a total activity index, based on the sum of the five individual indices (7). The weighting factors are based on the relative intensity of each activity category. For example, the weight for vigorous activity is 5, whereas the weight for sitting is 1. Previous studies using the YPAS have found 2-wk test-retest correlations of the total activity index of r = 0.55 (18) and r = 0.65 (7).
Because a major emphasis of the Healthy Body/Healthy Spirit intervention was to increase walking and moderate-intensity activity, in addition to the walking items on the CHAMPS and YPAS measures, we included selected items from the Paffenbarger Physical Activity Questionnaire (15). We inquired about city blocks usually walked per day, pace of walking, flights of stairs climbed per day, usual minutes of walking per day, and degree of exertion when physically active (rated 0–10). Each of these is assessed with single items. Test-retest correlations for flights climbed and blocks walked in prior studies have ranged from 0.23 to 0.68 (15).
We also included global activity items assessing: “... how active you have been during the past 2 wk, that is had hobbies, work, social activities, or other activities that keep you busy?” “... how physically active you have been during the past 2 wk, that is, did activities such as brisk walking, swimming, dancing, general conditioning or recreational sports?” “Which of these statements best describes your view? (answered, I get enough exercise to keep me healthy or I ought to do more exercise), and “at least once a week do you engage in regular physical activity like brisk walking, jogging, bicycling, or swimming long enough to work up a sweat, get your rate thumping or get out of breath?” (If yes, respondent was asked to indicate times per week).
Other variables assessed.
Income was assessed with an eight-category ordinal item, with answers ranging from <$10,000 to >$149,999. Responses were collapsed into four categories: <$30,000; $30,000–49,999; $50,000–69,000; and ≥$70,000 for the current analyses. Education was categorized as “ high school/vocational school or less,” “some college,” and “completed college or higher.” Self-reported age was collapsed into two categories: 21–39 and 40 and older.
Physiologic measures.
Blood pressure, height, and weight were obtained from participants at the baseline health fair. Height and weight were converted to BMI, computed as weight (kg)/height2 (m). Systolic and diastolic blood pressure were assessed using a manual aneroid sphygmomanometer according to the JNC-VI protocol (5). Height was measured with a stadiometer, and weight was obtained by trained staff with shoes and heavy outer clothing removed, using the Tanita office scale (model TBF 300). Total cholesterol was measured in nonfasting capillary samples using the Johnson and Johnson/Kodak DT60. Response rates for these measures are difficult to estimate as recruitment was done on a first come first serve “quota” basis, with a goal of recruiting 45 participants per church.
Cardiovascular fitness.
Two treadmills (Bodyguard Inc.) were transported to each church, where we were able to test approximately 10–16 persons per health fair. Before attending the health fair, as part of the consent form, participants were asked to complete a brief screening form (a modified version of the Physical Activity Readiness Questionnaire (PARQ)) (4) to determine their eligibility for fitness testing. Based on contraindications from the PARQ, we excluded from the treadmill test 361 individuals. Individuals excluded due their PARQ response did not differ significantly from those completing the treadmill test on any of the modified CHAMPS indices, Paffenbarger items, or any the YPAS indices with the exception of the sitting index. They were however significantly (P < 0.01) older, 47.3, than those completing the treadmill test, 40.8. In addition, we excluded participants > 69 yr of age and those with blood pressure > 140/90 mm Hg. Finally, we also excluded individuals reporting regular aerobic activity. This exclusion was based on participants’ responses to two items regarding their participation in aerobic activity. Frequency was assessed by the question: “How many days do you currently participate in aerobic activity in a typical week?” Response options for this item ranged from 0 to 7 d. Duration was assessed by asking respondents to indicate how many minutes per day they were active in the above aerobic activity using the following five categories: 0–15, 15–30, 30–45, 45–60, and >60 min. This information was used to classify participants as having either met or not met the Centers for Disease Control (CDC) and American College of Sports Medicine (ACSM) guidelines (4 d, 30 min·d−1) for regular physical activity (14,23). The exclusion of highly active participants was invoked to maximize the likelihood of detecting change in fitness levels among intervention participants, as the intervention targeted low to moderately active individuals. Those excluded from the treadmill due to meeting CDC/ACSM guidelines also reported significantly higher rates of activity on each of the modified CHAMPS indices, two of the YPAS indices (leisurely walking and vigorous activity), and two of the Paffenbarger items (blocks walked and minutes walked per day).
From those eligible for the fitness test, convenience quota sampling was used at each church, with a goal of testing 10–16 participants per health fair. Individuals completed one of two treadmill protocols, based on their self-reported level of activity and/or age. Participants with moderate levels of activity or those age 60 or less completed the Balke protocol (6), which began at 3 mph at a 2.5% grade. The grade was increased by 2.5% every 2 min with treadmill speed remaining at 3 mph for the duration of the test. Participants with low levels of activity or age 60 or higher completed a less strenuous modified Balke protocol. The treadmill was set at 2 mph and 0 grade, and grade increased 2.5% every 2 min.
Compared with those assigned to the standard Balke treadmill protocol, individuals assigned to the modified Balke protocol had significantly (ANOVA P value < 0.05) lower estimated V̇O2max, 28 versus 36, as well as significantly lower values on all major CHAMPS indices. This suggests our triage process was effective in assigning individuals to their appropriate fitness test.
Heart rate and RPE were recorded at the end of every stage. Tests were terminated when the participant reached 85% of their age predicted maximal heart rate (220 − age), when the treadmill reached its maximal elevation of 15%, or when participants chose to terminate the test. Thirteen individuals who began the test but did not complete at least two stages were excluded from the analyses. Estimated maximum V̇O2 for both protocols was computed using the following equation: V̇O2max: SM2 + b(HRmax − HR2), with b computed as (SM2 − SM1)/(HR2) − (HR1) (4), where SM corresponds to the stage-specific workload; HR, the stage-specific achieved steady state heart rate, and HRmax, the predicted maximum heart rate.
A total of 169 subjects completed at least two stages of the treadmill protocol. Twenty-eight of these 169 were excluded because they did not complete at least half of the modified CHAMPS questionnaire, and an additional three were excluded because their questionnaire indicated >14,000 kcal of activity per week. This resulted in a final sample of 138 participants for the validity study. Of these, 70 completed the standard Balke protocol, and the remaining 68 completed the modified Balke protocol.
Analyses.
Self-reported activity variables were correlated with estimated V̇O2max, resting blood pressure, BMI, and total cholesterol. Partial correlations (Pearson) were used to control for age and gender when examining the entire sample, and age, when conducting gender-specific analyses. Two variables not normally distributed were square root transformed, city blocks walked and minutes walked. Correlations for modified CHAMPS indices are presented separately by gender, income, education, and age.
RESULTS
Sample description.
As shown in Table 1, the baseline sample was predominantly female (78%) with a mean age of 41 yr (Tables 1 and 2). Approximately two thirds of the sample was married or living with a partner, and approximately 50% reported income > $50,000 and a college education or higher.
;)
TABLE 1: Sample Description (N =138).
TABLE 2: Summary of physical activity measures (N =138).
Correlation between activity measures and physiologic variables.
As shown in Table 3, with the exception of moderate to vigorous activities, each of the modified CHAMPS indices were significantly correlated with estimated V̇O2max (Table 3). Highest correlations between the modified CHAMPS indices and estimated V̇O2max were observed for the indices of vigorous and sports-related activities (r = 0.19 and r = 0.32, respectively). Two Yale indices were significantly associated with estimated V̇O2max, leisurely walking, and sitting (negative correlation). Of the remaining activity measures, only the two global items: “how physically active are you” and “do you get enough exercise” were significantly correlated (r = 0.20 and r = 0.19, respectively) with estimated maximum V̇O2. With the exception of the moving index from the YPAS, which was correlated (r = −0.19) with serum cholesterol, none of the activity measures, including the modified CHAMPS, were significantly correlated with blood pressure, BMI, or total cholesterol.
TABLE 3: Correlationa between physical activity measures and physiologic validity measures (N =138).
Correlation between modified CHAMPS indices and physiologic variables by gender.
For the modified CHAMPS indices, correlations with estimated V̇O2max were generally higher for males than females (Tables 4 and 5) For females, only the sports index (r = 0.19) was significantly correlated with estimated V̇O2max. For males, all four indices were significantly correlated with estimated V̇O2max, with the highest values observed for the index of vigorous (r = 0.61) and sports-related activities (r = 0.50). For the remaining physiologic variables, correlations with the modified CHAMPS indices, although nonsignificant, were generally higher for males than females.
TABLE 4: Correlationa between physical activity measures and physiologic validity measures for males only (N =29).
TABLE 5: Correlation a between physical activity measures and physiologic validity measures for females only (N =109).
Correlation between other activity measures and physiologic variables by gender.
Among males, only the YPAS leisurely walking index and two of the global items assessing usual activity were significantly correlated with V̇O2max. Among females, none of the other activity measures were significantly correlated with estimated V̇O2max. There were scattered significant correlations between the other activity measures and the remaining physiologic measures, although several were opposite to the criterion direction, e.g., “How active” was positively correlated (r = 0.39) with cholesterol among males and the vigorous activity, leisurely walking, and summary indices from the YPAS were positively correlated (r = 0.21) with diastolic blood pressure among females.
Correlation between CHAMPS indices and physiologic variables by socioeconomic variables.
The correlations between the modified CHAMPS indices and estimated V̇O2max were considerably higher for participants with income below $30,000 (Tables 6–8). With regard to educational attainment, the correlations between the modified CHAMPS indices and estimated V̇O2max were generally highest for participants who did not complete college. With regard to age, correlations with estimated V̇O2max were higher among those 21–39 yr of age.
TABLE 6: Correlation a between modified CHAMPS indices and physiologic measures by income (N =122).
TABLE 7: Correlation a between modified CHAMPS indices and physiologic measures by education (n =138).
TABLE 8: Correlation a between modified CHAMPS indices and physiologic measures by age (N =138).
DISCUSSION
The correlation of responses from the modified CHAMPS self-reported physical activity recall with V̇O2max in this sample of African-American adults ranged from 0.16 to 0.32, with higher validity coefficients observed for indices of vigorous activities and sports. Correlations with estimated V̇O2max were generally stronger for the modified CHAMPS indices than the other activity measures. Two prior studies, one by Stewart et al. (22) and the other by Harada et al. (9), examined the validity of the original CHAMPS in elder, non-African-American populations. Both used the 6-min walk as the fitness criterion, and both examined only the indices comprised of total activities and moderate/vigorous activities (>3 METs). We constructed the vigorous activities (>5 METs) and sports activities indices for this study. Our correlations for indices of total activities and moderate/vigorous activities with estimated V̇O2max are comparable to those obtained by Stewart et al., 0.22 and 0.27, respectively, but lower than those reported by Harada et al., 0.46 and 0.54, respectively. In the one prior study with African-Americans (all Seventh-Day Adventists), Singh et al. (20) compared values from a self-report questionnaire, somewhat similar to the original CHAMPS with cardiovascular (CV) fitness estimated by the Bruce treadmill protocol. They found correlations of vigorous and sports activities and CV fitness of 0.20 and 0.28, respectively, for women and −0.05 and −0.06, respectively, for men. This is opposite to the pattern observed in the present study (as well as at least one prior study (21)), where men showed higher correlations than women.
Correlations between the modified CHAMPS indices and estimated V̇O2max were stronger for individuals with lower income and educational attainment. The SES findings were somewhat surprising and we could find no prior studies that have observed this pattern. One possible explanation is that upper SES individuals may be more aware of public health recommendations and therefore more subject to social desirability bias, which could decrease correlations with an objective criterion. Another explanation is that there was more variability in the self-report and/or physiologic variables in these groups, which would strengthen correlations. Examination of variable ranges and variance across groups for the modified CHAMPS and estimated V̇O2max values suggest this was a partial explanation (data not shown).
Consistent with several prior studies, the correlation of physical activity with cardiovascular fitness was stronger for more vigorous activities and recreational sports (3,7,9,11,12,19–21,24). One reason for this finding may be that individuals are better able to accurately recall more intensive activities or their involvement in recreational sports compared with lighter-intensity activity or activities of daily living. Additionally, cardiorespiratory fitness is more easily impacted by physical activities of higher intensity.
In the present study, the generally weak correlation of the modified CHAMPS indices with BMI and blood pressure is consistent with Stewart et al. (22), although when frequency of involvement in moderate and all activities were used in that study rather than caloric expenditure, correlations with BMI were higher (r = −0.17 and r = −0.21). Conversely, a validity study of the YPAS (7) found generally higher correlations of self-reported activity with blood pressure (inversely correlated) than with estimated V̇O2max. The weak correlations between YPAS activity dimensions with CV fitness and other physiologic variables are inconsistent with at least three prior studies (7,18,24). In the study by Young et al. (24), validity coefficients were generally lower for African-Americans in their sample than for non-African-Americans, which could explain the lower correlations here, as our sample was entirely African-American. The sitting index of the YPAS was negatively correlated with estimated V̇O2max. Prior studies have found the sitting index to be uncorrelated with cardiovascular fitness, although this is the first study to find an inverse association (7,18,24).
The study has several limitations that may have decreased our validity coefficients. Other strategies that can be used to establish instrument validity beyond submaximal treadmill testing were not employed here. For example, data from pedometers and accelerometers have been used as a criterion reference. Correlations of self-report activity were higher with pedometers/accelerometers readings relative to CV fitness in one study (20) but not others (7,9). Activity diaries and 24-h activity recalls can also be used as validity criterion, although they are subject to some of the same biases as recall instruments and may therefore not be the preferred reference criteria. There is considerable error in estimating V̇O2max via submaximal treadmill tests without direct measurement of oxygen exchange. This method assumes a constant for mechanical efficiency among individuals and therefore tends to overestimate V̇O2max for trained individuals and underestimate V̇O2max for untrained individuals. Individual variation in HR maximum and age-predicted maximal heart rate further contribute error to V̇O2max uptake prediction (10). The other physiologic variables, e.g., BMI, blood pressure, and lipids, although often used in validity studies, are only indirect markers of fitness, not direct measures of activity. Therefore, the correlation between activity measures and these parameters is not expected to be strong. They were included here more to examine their association with the modified CHAMPS relative to the other activity measures than as direct validation of self-reported activity.
Another factor that may have attenuated validity coefficients was that the study questionnaire used here was extremely long, 34 pages including nine pages containing the modified CHAMPS items (situated approximately midway in the instrument). Participants may have become tired or bored completing such a long instrument, and this may have decreased their ability and/or willingness to accurately recall their activity. The primary purpose of conducting the submaximal treadmill test was to detect intervention effects among individuals with initially low to moderate levels of baseline physical activity not to validate the modified CHAMPS measure. Therefore, our treadmill sample was intentionally proscribed by limiting those meeting CDC/ACSM activity guidelines, in addition to excluding these more activity individuals because of PARQ ineligibility. The latter group was somewhat less active than those completing the treadmill test. Thus, our sample was truncated at both the lower and upper end of the distribution and comprised what was likely a generally moderately active sample. As a result of this reduced variability, the validity coefficients reported here could be considered a lower end estimate of actual validity. Including a more diversely active sample would likely have resulted in higher correlations. Strengths of the study include use of previously validated measures, standardized, objective validity criteria, and a diverse socioeconomic sample of African-Americans.
Validity of physical activity using the modified CHAMPS physical activity questionnaire, based on comparison to treadmill estimates of V̇O2max, appears higher for more intensive physical activities. This suggests that surveillance or intervention studies targeting low- to moderate-intensity activities should consider other methods for establishing instrument validity. Report of low-intensity activities in the modified CHAMPS or similar instruments may nonetheless be valid, though uncorrelated with CV fitness and other physiologic parameters. Additional research testing and refining self-report activity measures in socioeconomic and ethnic subpopulations is needed.
Development of this manuscript was supported in parts by National Heart, Lung, Blood Institute grants HL64959 and HL62659 to the first author.
REFERENCES
1. Ainsworth, B. E., W. L. Haskell, A. S. Leon, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med. Sci. Sports Exerc. 25: 71–80, 1993.
2. Ainsworth, B. E., W. L. Haskell, M. C. Whitt, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exerc. 32 (Suppl. 9): S498–S504, 2000.
3. Ainsworth, B. E., B. Sternfeld, M. T. Richardson, and K. Jackson. Evaluation of the kaiser physical activity survey in women. Med. Sci. Sports Exerc. 32: 1327–1338, 2000.
4. American College of Sports Medicine. ACSM’s Resource Manual for Guidelines for
Exercise Testing, and Prescription, 2nd Ed. Philadelphia: Lea & Febiger; 1993, pp. 220–225.
5. Anonymous. The sixth report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure. Arch. Intern. Med. 157: 2413–2446, 1997.
6. Balke, B., and R. Ware. An experimental study of physical fitness of Air Force personnel. U.S. Armed Forces Med. J. 10: 675–688, 1959.
7. Dipietro, L., C. J. Caspersen, A. M. Ostfeld, and E. R. Nadel. A survey for assessing physical activity among older adults. Med. Sci. Sports Exerc. 25: 628–642, 1993.
8. Flack, J. M., H. Amaro, W. Jenkins, et al. Epidemiology of minority health. Health Psychol. 14: 592–600, 1995.
9. Harada, N., V. Chiu, A. King, and A. Stewart. An evaluation of three self-report physical activity instruments for older adults. Med. Sci. Sports Exerc. 33: 962–970, 2001.
10. Heyward, V.
Advanced Fitness Assessment and Exercise Prescription. Champaign, IL: Human Kinetics, 1997.
11. Jacobs, D. R., B. E. Ainsworth, T. J. Hartman, and A. S. Leon. A simultaneous evaluation of 10 commonly used physical activity questionnaires. Med. Sci. Sports Exerc. 25: 81–91, 1993.
12. Kohl, H. W., S. N. Blair, R. S. Paffenbarger, C. A. Macera, and J. J. Kronenfeld. A mail survey of physical activity habits as related to measured physical fitness. Am. J. Epidemiol. 127: 1228–1239, 1988.
13. Miller, W., and S. Rollnick. Motivational Interviewing: Preparing People to Change Addictive Behavior. New York: Guilford Press, 1991, pp. 1–90.
14. Pate, R. R., M. Pratt, S. N. Blair, et al. Physical activity and public health: a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 273: 402–407, 1995.
15. Pereira, M. A., S. J. Fitzergerald, E. W. Gregg, et al. A collection of Physical Activity Questionnaires for health-related research. Med. Sci. Sports Exerc. 29 (Suppl. 6): S1–205, 1997.
16. Resnicow, K., R. Braithwaite, J. Ahluwalia, and C. Diiorio. Cultural sensitivity in public health. In: Health Issues in the Black Community, 2nd Ed, R. Braithwaite and S. Taylor (Eds.). San Francisco: Jossey-Bass, 2001, pp. 516–542.
17. Resnicow, K., A. Jackson, R. Braithwaite, et al. Healthy body/healthy spirit: design and evaluation of a church-based nutrition and physical activity intervention using motivational interviewing. Health Educ. Res. 17: 562–573, 2002.
18. Schuler, P. B., M. T. Richardson, P. Ochoa, and M. Q. Wang. Accuracy and repeatability of the Yale physical activity survey in assessing physical activity of older adults. Percept. Mot. Skills 93: 163–177, 2001.
19. Siconolfi, S. F., T. M. Lasater, R. C. Snow, and R. A. Carleton. Self-reported physical activity compared with maximal oxygen uptake. Am. J. Epidemiol. 122: 101–105, 1985.
20. Singh, P. N., G. E. Fraser, S. F. Knutsen, K. D. Lindsted, and H. W. Bennett.
Validity of a physical activity questionnaire among African-American Seventh-day Adventists. Med. Sci. Sports Exerc. 33: 468–475, 2001.
21. Singh, P. N., S. Tonstad, D. E. Abbey, and G. E. Fraser.
Validity of selected physical activity questions in white Seventh-day Adventists and non-Adventists. Med. Sci. Sports Exerc. 28: 1026–1037, 1996.
22. Stewart, A., K. Mills, A. King, W. Haskell, D. Gillis, and P. Ritter. Champs physical activity questionnaire for older adults: outcome for interventions. Med. Sci. Sports Exerc. 33: 1126–1141, 2001.
23. U.S. Department of Health and Human Services. Physical activity and health: a report of the Surgeon General. Atlanta, GA: Centers for Disease Control and Prevention, 1996, pp. 11–37.
24. Young, D., S. Lee, and L. Appel. A comparison of the Yale Physical Activity Survey with other physical activity measures. Med. Sci. Sports Exerc. 33: 955–961, 2001.
