Time spent sedentary (too much sitting as distinct from lack of exercise) such as daily sitting time, watching television, driving in cars, or sitting during leisure time is independently associated with an increased risk of premature all-cause and cardiovascular disease mortality (10,18,28,37). Although it has been shown via accelerometer-derived assessment that older adults have the highest levels of sedentary time (23), these objective measures do not provide the contextual information necessary for the identification of intervention targets and public health messages on how to reduce sedentary time (27).
Because the use of objective measures is not always feasible in epidemiological and health behavior intervention studies, high-quality self-report measures are needed, yet there is limited evidence on the measurement properties of existing questionnaires, particularly with older adults. In a review of 60 articles reporting nonoccupational sedentary behavior in adults, nine studies reported on reliability, and three studies reported on validity (7); only one reported on the measurement properties of a measure administered with older adults (24). Two previous studies have reported good test-retest reliability with older adults using a measure of global sedentary time designed for use in the general adult population (9,19). However, in these studies, validity was assessed against weekly pedometer step counts (9) or accelerometer-assessed moderate- to vigorous-intensity physical activity and total accelerometer counts (19) and not directly against an objective measure of sedentary time.
Decreasing sedentary time is a potentially important preventive health target. To evaluate interventions, measures that are responsive to change are required. By comparing the responsiveness to change in sedentary time of several different measures, we can evaluate which instruments can detect significant changes with the smallest number of participants, i.e., provide the most power for a given sample size (34). In contrast with the evidence for responsiveness to change of several physical activity measurement tools (12,29,36), no studies to date have examined the responsiveness of either self-report or accelerometer measures of sedentary time. Thus, it is not clear which type of sedentary time measure would be more responsive: the self-report measures may be better able to capture the specific behaviors targeted by intervention; on the other hand, the accelerometer measures are likely to have lower background variability because of their good reliability (5).
We examined the measurement properties of a self-reported total sedentary time measure and specific sedentary time items in nonworking older adult participants of a sedentary behavior intervention trial. The specific measurement properties examined were test-retest reliability, responsiveness to change, and validity compared with accelerometer-derived sedentary time. We also assessed the responsiveness to change of accelerometer-derived sedentary time.
Nonworking older adults were recruited for a study examining the feasibility of an intervention (Stand Up for Your Health) designed to reduce sedentary time (13). Sixty participants were recruited through community-based organizations for older adults and retirement villages within urban areas of Brisbane, Australia. Eligibility criteria included aged 60 yr or older, self-reported television (TV) viewing time of ≥2 h·d, not in paid employment, able to ambulate without assistance from another person, and residence within 50 km of the research center. Participants were enrolled in the study between May and December 2009 and were not paid for their involvement. The study was approved by The University of Queensland's Behavioural Social Sciences Ethics Review Committee.
Study Design and Protocols
This before to after study involved multiple baseline assessments during a preintervention period, followed immediately by an intervention and a postintervention assessment period. Self-reported sedentary time (described below) was assessed using an interviewer-administered questionnaire (face to face) during three home visits conducted approximately 1 wk apart. During the first visit, participants provided written informed consent, had their height and weight measured, and completed a questionnaire to assess demographic characteristics and sedentary time (T1 assessment). Participants were instructed to wear a uniaxial accelerometer (ActiGraph model GT1M; http://www.theactigraph.com/) fitted firmly around their waist over the right hipbone during all waking hours (except when bathing or participating in water-based activities) and to complete a log to record wear times. During the second visit, the questionnaire was readministered (T2 assessment), data from the accelerometer were downloaded, and the battery was recharged. A short intervention session was conducted, and participants were asked to adhere to the accelerometer wearing protocol for a further week. During visit 3, the accelerometer and log were collected, and the questionnaire was readministered (T3 assessment). Data from T1 and T2 and the first accelerometer assessment (preintervention period) were used to assess test-retest reliability and validity. Data from T3 and the second accelerometer assessment (after intervention) were used to assess responsiveness to change from T2 and the first accelerometer assessment.
The inclusion criteria for this measurement study were at least four valid days of accelerometer data during each assessment period (valid day = at least 10 h of wear established from both accelerometer and log data; data from days on which the researcher conducted the home visits were not included) and complete responses to the interviewer-administered questionnaire at the three assessment time points (T1, T2, and T3). All participants provided valid accelerometer data for at least half of the assessment period covered by self-report.
Demographic measures and anthropometry.
Participants were asked to report their date of birth (from which age was calculated), highest level of education completed, the number of people in their household, and how much time they usually spent sleeping each night during the past week. Participants' height (via stadiometer, nearest 0.5 cm) and weight (via electronic scales, nearest 0.1 kg) were measured to derive body mass index (kg·m−2).
Self-report measure of sedentary time.
Self-reported sedentary time, in the specific domains of leisure time and transport, was assessed using a seven-item 1-wk recall questionnaire adapted from a previous measure of leisure time sitting developed for the general adult population (32). The original summary measure was shown to have good test-retest reliability (intraclass correlation coefficient (ICC) = 0.79, 95% confidence interval (CI) = 0.71-0.85, 145 participants with mean ± SD age of 50.8 ± 13.5 yr) and a modest validity assessed against a behavior log (Spearman ρ = 0.30, 130 participants with mean ± SD age of 38.8 ± 15 yr); however, some individual items showed poor reliability and validity (32). Our adapted questionnaire asked participants to report on activities they did during the last week while they were sitting or lying down (not including time spent in bed) and to report the total time spent in each activity. The seven individual sedentary items were (a) TV or video/DVD watching, (b) computer use, (c) reading, (d) socializing with friends or family, (e) time traveling in a motor vehicle or on public transport, (f) doing hobbies, and (g) any other activities they did while they were sitting or lying down (see Appendix, Supplemental Digital Content 1, http://links.lww.com/MSS/A93, Questionnaire to measure sedentary time in older adults).
We adapted the original measure of Salmon et al. (32) to meet the needs of our intervention trial and to increase their relevance to our target population, by (i) ensuring that time in the individual behaviors was mutually exclusive because participants were instructed to only count time when it was their main activity; for example, they would count time spent knitting while watching TV as either TV time or time doing hobbies but not as both; (ii) addressing problems of low test-retest reliability in the original measure by removing two activities (listening to music/CD/radio and relaxing, thinking, and resting); (iii) revising two items to reflect the activities of older adults (talking on the telephone was included as part of socializing, and going for a recreational drive was expanded to include all transport (2)); (iv) administering the questionnaire in a face-to-face interview (which allowed for prompting of participants, e.g., to identify activities undertaken on specific days), in contrast to studies with the original measure that were completed without any contact with an interviewer; (v) reducing participant burden by only asking about total time spent in activities during the last week, instead of asking separately about weekdays and weekend days as per the original measure; and (vi) ensuring that the individual behaviors were suitable targets for behavior change. As per the original measure, responses for the items were reported as continuous time, in hours and minutes per week. Weekly time in each activity was converted to hours per day. Total sedentary time was calculated as the sum of daily time in each activity and reported as hours per day.
Accelerometer-derived sedentary time.
Data were collected in 1-min epochs. The commonly used cut point of <100 counts per minute (cpm) (8,23) was used to derive sedentary time. Days on which the researcher conducted home visits and when the accelerometer was worn for <10 h were excluded. Wear time was determined from a combination of data from the accelerometer and wearing logs completed by participants. Average sedentary time was calculated as total sedentary time/number of valid days and was expressed as hours per day. When examining responsiveness to change, sedentary time was standardized to 16 h of waking time (participants reported a median sleeping time of 8 h per night) to account for differences in accelerometer wear time in each assessment period. Data were summarized using SAS 9.1 (SAS Institute, Inc., Cary, NC) via a modified version of the program available from the National Cancer Institute (http://riskfactor.cancer.gov/tools/nhanes_pam/ ).
Analyses were conducted using Stata Statistical Software Release 11.0 (StataCorp LP, College Station, TX) and SPSS version 17.0 (SPSS, Inc., Chicago, IL). All sedentary variables had nonnormal distributions, with the exception of total self-report sedentary time. Statistical significance was set at P < 0.05.
Reliability was assessed using a 1-wk (before intervention) test-retest protocol. Correlation between T1 and T2 measures of time in each sedentary behavior and total sedentary time was assessed with the Spearman rank correlation coefficient (ρ), with 95% CI calculated using the Fisher transformation. To allow comparison with the original measure (32), we also assessed test-retest using single-measures ICC with 95% CI, with an absolute agreement definition, which were calculated using two-way mixed-effects models (33).
The relative validity of self-report total sedentary time was assessed with the Spearman rank correlation coefficient (ρ) with 95% CI using data from T2 and accelerometer-derived sedentary time during the preintervention period. Bland-Altman (3,4) plots were used to assess absolute agreement between the two measures. Regressing average self-report sedentary time/accelerometer-derived sedentary time on the differences between the two measures revealed that the mean difference increased significantly as average values increased. The variability, however, remained constant across average values. Therefore, the Bland-Altman plot presents the trend line for mean difference obtained from the regression and limits of agreement (±2 SD).
Responsiveness to change.
Responsiveness to change was assessed using the responsiveness statistic (RS) (14), which quantifies the minimum clinically important difference, or, if this is unknown, the difference observed in an intervention, in relation to variability over a stable period. We calculated RS as mean change (Δ) within participants during the intervention period (T2 to T3) divided by the square root of two times mean squared error (MSE), our measure of background variability during a stable period (before intervention, T1 to T2). Repeated-measures ANOVA were used to determine MSE. Thus, RS was calculated as [Δ/√(2MSE)] using the change and MSE for each individual behavior and total sedentary time (both self-report and accelerometer-derived), with the direction of the change removed to ease interpretation of the magnitude of the RS. We also examined the proportion of participants who made substantial changes (15 min·d−1 for individual behaviors and 30 min·d−1 for total sedentary time).
Forty-eight (80%) of the participants in the intervention study met the inclusion criteria and were included in the current investigation. The majority of participants in this measurement study were women (n = 35, 72.9%), had completed postsecondary-level (e.g., university) or professional-level (e.g., teaching certificate) education (n = 35, 72.9%), lived with others (n = 27, 56.3%), and were community dwelling (n = 39, 81.3%). The participants' mean ± SD age and body mass index were 72.8 ± 8.1 yr and 27.2 ± 4.8 kg·m−2.
The median (minimum-maximum) number of days from T1 to T2 and from T2 to T3 was 7 (6-11) and 7 (6-14), respectively. The duration for each individual sedentary item (median (25th-75th percentile)) and total self-reported sedentary time (mean ± SD)) at each assessment (i.e., T1, T2, and T3) are shown in Table 1. TV viewing time comprised the largest component of total sedentary time (37.6%, 37.1%, and 36.8% at T1, T2, and T3, respectively). A large proportion of participants reported no time in computer use, hobbies, or "other" sedentary behavior.
The median (minimum-maximum) number of valid days the accelerometer was worn was 6 (4-10) in both assessment periods. The mean ± SD accelerometer wear time was 14.6 ± 0.97 h·d−1 in the first assessment period and 14.4 ± 1.10 h·d−1 in the second assessment period. Total sedentary time median (25th-75th percentile) was 10.2 (9.5-10.9) and 9.8 (8.5-10.5) h·d−1 during the pre- and postintervention assessment periods, respectively. Accelerometer-derived sedentary time significantly decreased from before to after intervention (P < 0.001), as assessed using the Wilcoxon signed rank test. Total self-reported sedentary time did not exceed accelerometer wear time for any participant.
Table 1 presents the test-retest reliability of individual sedentary behavior items and total self-reported sedentary time. The Spearman correlations were high for computer use, TV viewing time, and reading (ρ > 0.75), modest to acceptable for the other individual items (ρ = 0.23-0.61), and acceptable for total sedentary time (ρ = 0.56). ICC were similar to the Spearman correlations for total sedentary time and most individual items but were considerably lower for other sedentary time and hobbies.
The correlation between total self-report and accelerometer-derived sedentary time was statistically significant but modest (ρ (95% CI) = 0.30 (0.02-0.54)). Figure 1 shows the Bland-Altman plot for total self-reported and accelerometer-derived sedentary time. Linear regression showed a significant negative association between the difference in the two measures (self-reported minus accelerometer-derived sedentary time) and the average of these two measures (B = −0.67, SE = 0.21, P = 0.003). Thus, the mean difference is estimated at −9.20 h + 0.67 × average of the two measures. At mean levels of average self-reported/accelerometer-derived sedentary time (8.36 h), the mean difference indicated self-reported sedentary time was 3.60 h·d−1 lower than accelerometer-derived sedentary time with wide limits of agreement (mean difference ± 3.82 h).
Responsiveness to change.
Table 2 shows the variability in the preintervention period [√(2MSE)], the change from before to after intervention, and the responsiveness to change for the self-report and accelerometer-derived measures of sedentary time. The reduction detected was greater for self-report (0.85 h·d−1) than for accelerometer-derived sedentary time (0.50 h·d−1). The variability for self-report was greater than for accelerometer-derived sedentary time; however, they were both similarly responsive to change (0.47 and 0.39, respectively) because the amount of change was assessed to be larger by self-report than by accelerometer. Of the specific sedentary behavior items, the greatest reductions were seen in TV viewing time and hobbies, and these also had better responsiveness (0.34 and 0.33, respectively) than the other items. Table 2 also shows the proportion of participants making substantial changes from before to after intervention. A larger proportion of participants made a substantial reduction in sedentary time as assessed by self-report than derived by accelerometer, whereas more participants reported making a substantial reduction in TV viewing time than the other individual sedentary behaviors.
This study examined the measurement properties of a self-report sedentary time questionnaire adapted specifically for use in epidemiological and health behavior intervention studies with nonworking older adults. Our questionnaire is unique among measures designed for older adults in that it assesses continuous time spent in specific sedentary behaviors as well as the total time spent in these behaviors. Other novel aspects of this measure are the ability to capture waking time spent lying down, not just sitting time, and also the mutually exclusive nature of the items, which may improve the face validity of the summed measure of total sedentary time. This study is also unique in that it is the first in older adults to report validity for self-report sedentary time compared with accelerometer-derived sedentary time. Importantly, we provided the novel evidence that accelerometer-derived and self-reported sedentary time measures are both responsive to change. Coupled with finding adequate, although not ideal, validity and reliability for our self-report measure of sedentary time, this suggests that both our self-report and accelerometer-derived measures of sedentary time are suitable for use in interventions with older adults.
The test-retest reliability of the individual sedentary time items was similar to that reported for measures previously used with a general adult population sample (32) and for similar sedentary measures in varied populations (7). The reliability of total self-report sedentary time (ICC = 0.52) was fair to good but lower than that for the earlier study with the general adult population sample (ICC = 0.82 ). This is broadly consistent with previous studies that have shown measures used with older adults have lower test-retest reliability than those used with general adult populations, including for screen time (24) and for physical activity (25). This may be reflective of differences in the amount of true variation in sedentary behaviors during 1 wk, possibly because of the age and working status of the participants. Other possible explanations are differences in the amount of measurement error that may be related to factors specific to older adults such as deficits in concentration/cognition/memory (30). The higher test-retest reliability in the study by Salmon et al. (32) could reflect the greater between person variability in the general adult population, which increases the ICC. To explore this, we repeated analyses excluding participants who reported that their sedentary time had changed during the preintervention period. At T1 and T2, participants were asked whether their sitting was comparable to a typical week (five-point scale from much less than normal to a lot more than normal). Test-retest reliability was examined for participants (n = 32) who responded that their sitting at T2 was at a similar level to that at T1. The reliability of all measures improved: self-report sedentary time had good reliability (ICC (95% CI) = 0.74 (0.44-0.88) and ρ (95% CI) = 0.71 (0.48-0.85)); the Spearman correlations for individual items ranged from ρ = 0.41 (socializing) to ρ = 0.93 (computer use). The increase in test-retest reliability in these subanalyses suggests there may be substantial true variability in sedentary behaviors in older adults. Thus, the measure of total sedentary time has acceptable repeatability when considering both random and systematic errors.
The validity of this self-report measure of total sedentary time was less than ideal. However, our findings were comparable to what has been seen for global and composite measures in adult populations (assessed against accelerometer-derived sedentary time) in terms of correlation (Spearman ρ ranging from −0.01 to 0.61 [6,8,11,31]), mean difference (ranging from −2.2 to 0.18 h·d−1 [15,22]), and limits of agreement (ranging from ±5.53 to ±6.90 h·d−1 [15,22]). Participants tended to report less sedentary time than was recorded by the accelerometer, with the discrepancy decreasing at higher levels of sedentary time. Previous studies have shown that adults underreport time spent in individual sedentary behaviors such as watching TV (24,26), which may exacerbate the differences between self-report and comparison measures when using a summary score of total sedentary time, as is the case for our study. Further, it is possible that other salient aspects of sedentary time were not captured by the items in this questionnaire such as time spent eating.
Alternatively, the discrepancy could relate to overdetection by the accelerometer, e.g., some time standing still might have been classified as sedentary. The accelerometer is not considered a gold-standard criterion for sedentary time because it does not detect posture. Recent evidence suggests that devices that measure body position (such as the activPAL™ (PAL Technologies Ltd., Glasgow, UK), which distinguishes between time spent sitting/lying, standing, or walking) may be a better comparison measure for validity studies and for assessing changes in sedentary time. One study reported a smaller discrepancy between self-report sedentary time and sedentary time as measured by the activPAL™ than between self-report sedentary time and sedentary time derived from the ActiGraph GT1M accelerometer (16). Another study reported that the activPAL™ detected greater reductions in sedentary time than the ActiGraph GT3X accelerometer (20). Furthermore, the ideal accelerometer cut point for sedentary time for older adults is not known. The <100-cpm cut point has been shown to be detrimentally associated with biomarkers of cardiometabolic health and inflammation in population-based studies (17) and was thus used as the primary comparison value. Conclusions regarding validity were only minimally affected by the choice of cut point. In a sensitivity analysis where sedentary time was classed as <50 cpm, results were similar to those reported in the article in terms of correlations, limits of agreement, and responsiveness. Mean differences were smaller than when using the commonly cited <100-cpm cut point. It may be that cut points lower than <100 cpm might be more suitable in this older adult population for distinguishing sedentary time from activity.
A pertinent finding of our study was that both the self-report and accelerometer measures of sedentary time were able to detect reductions after the intervention and were similarly responsive to change. This would suggest that the individual items we chose to include in the self-report measure were among the behaviors that people changed during the intervention. Notably, the amount of change detected was greater via self-report than via accelerometer (51 vs 30 min·d−1). The greater change via self-report could be due to biased reporting; alternatively, the accelerometer may have failed to detect some sitting reductions (e.g., if sitting or lying down was replaced with standing still). The individual sedentary items were less responsive to change than total self-report sedentary time, with time watching TV and doing hobbies being the most responsive compared with the other items. Notably, most items had lower background variability than total sedentary time (both self-report and accelerometer-derived), so the lower responsiveness could relate to participants not changing these behaviors, rather than poor measurement quality. The lack of a control group in our before to after study limits our ability to assess whether changes in individual items and total sedentary time are as a result of the intervention or if the participants are more accurately reporting their behavior. In our study, a larger RS (in absolute terms) reflects a greater magnitude of change observed during the intervention period, not necessarily a better ability to detect a minimum clinically meaningful change. Caution with these findings is recommended because the minimum clinically meaningful change is as yet unknown for sedentary time and could be smaller or larger than what was achieved in this intervention.
The strengths of this study are our ability to report on responsiveness to change in addition to reliability and validity and the use of a self-report measure designed specifically for older adults that assessed time in individual sedentary behaviors and total sedentary time. The nonnormal distribution of individual sedentary time items and accelerometer-derived sedentary time limits the usefulness of the ICC and RS reported for these items (which rely on assumptions of normality); however, this was not a limitation for total self-report sedentary time. We did not have an appropriate criterion to assess validity of the individual sedentary items and recommend that future studies use a combination of log and objective measures for this purpose (35). Although participants wore the accelerometer for at least half of the period covered by self-report, there was not complete overlap between assessment periods, which may have contributed to the less than ideal correlation and agreement between the self-report and accelerometer-derived sedentary time measures. We are also limited in our ability to generalize the findings given the small nonrandom sample that was more highly educated than the general population of older Australian adults (1) and possibly had other differences related to motivation to participate in the intervention (38).
Our findings suggest that the summary measure of total sedentary time has good reliability but may be assessing behaviors that are not stable from week to week. The responsiveness is sufficiently high that the numbers of participants required to detect behavior change are achievable within the typical sample sizes of many behavioral interventions. Hence, in terms of responsiveness, the measure is suitable for detecting change after interventions; however, the interpretation of changes that would be detected is questionable, given that we did not identify strong validity. This may be improved by providing specific prompts in the other sedentary time item, such as for time spent eating. We administered the questionnaire via a face-to-face interview with nonworking older adults. The viability of alternate delivery methods such as via the telephone or self-completion and the utility of this questionnaire to evaluate change after interventions in younger adult populations across the leisure and transport domains of sedentary time should also be investigated.
This study was funded by a National Health and Medical Research Council (NHMRC) Program Grant (569940) and Supplementary Research Funding from the School of Population Health at The University of Queensland.
All authors are supported by an NHMRC Program Grant (569940) and a Queensland Health Core Research Infrastructure Grant. In addition, Paul Gardiner is supported by a National Heart Foundation Postgraduate Scholarship (PP 06B 2889), Bronwyn Clark is supported by an Australian Postgraduate Award, Elizabeth Eakin is supported by an NHMRC Senior Research Fellowship (511001), Genevieve Healy is supported by an NHMRC (569861)/National Heart Foundation (PH 08B 3905) Postdoctoral Fellowship, and Neville Owen is supported by an NHMRC Senior Principal Research Fellowship (1003960).
The authors report no conflict of interest.
The authors would like to sincerely thank the participants involved in this research.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
3. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
4. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res
5. Buman MP, Hekler EB, Haskell WL, et al. Objective light-intensity physical activity associations with rated health in older adults. Am J Epidemiol
6. Chinapaw MJ, Slootmaker SM, Schuit AJ, van Zuidam M, van Mechelen W. Reliability and validity of the Activity Questionnaire for Adults and Adolescents (AQuAA). BMC Med Res Methodol
7. Clark BK, Sugiyama T, Healy GN, Salmon J, Dunstan DW, Owen N. Validity and reliability of measures of television viewing time and other non-occupational sedentary behaviour of adults: a review. Obes Rev
8. Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc
9. Deng HB, Macfarlane DJ, Thomas GN, et al. Reliability and validity of the IPAQ-Chinese: the Guangzhou Biobank Cohort study. Med Sci Sports Exerc
10. Dunstan DW, Barr EL, Healy GN, et al. Television viewing time and mortality: the Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Circulation
11. Fjeldsoe BS, Marshall AL, Miller YD. Measurement properties of the Australian Women's Activity Survey. Med Sci Sports Exerc
12. Forsén L, Loland NW, Vuillemin A, et al. Self-administered physical activity questionnaires for the elderly: a systematic review of measurement properties. Sports Med
13. Gardiner PA, Eakin EG, Healy GN, Owen N. Feasibility of reducing older adults' sedentary time. Am J Prev Med
14. Guyatt G, Walter S, Norman G. Measuring change over time: assessing the usefulness of evaluative instruments. J Chronic Dis
15. Hagströmer M, Ainsworth BE, Oja P, Sjöström M. Comparison of a subjective and an objective measure of physical activity in a population sample. J Phys Act Health
16. Hart TL, Ainsworth BE, Tudor-Locke C. Objective and subjective measures of sedentary behavior and physical activity. Med Sci Sports Exerc
17. Healy GN, Matthews CE, Dunstan DW, Winker EA, Owen N. Sedentary time and cardio-metabolic biomarkers in US adults: NHANES 2003-06. Eur Heart J
18. Katzmarzyk PT, Church TS, Craig CL, Bouchard C. Sitting time and mortality from all causes, cardiovascular disease, and cancer. Med Sci Sports Exerc
19. Kolbe-Alexander TL, Lambert EV, Harkins JB, Ekelund U. Comparison of two methods of measuring physical activity in South African older adults. J Aging Phys Act
20. Kozey-Keadle S, Libertine A, Lyden K, Staudenmayer J, Freedson P. Validation of wearable monitors for assessing sedentary behavior. Med Sci Sports Exerc
21. Deleted in proof.
22. Marshall AL, Miller YD, Burton NW, Brown WJ. Measuring total and domain-specific sitting: a study of reliability and validity. Med Sci Sports Exerc
23. Matthews CE, Chen KY, Freedson PS, et al. Amount of time spent in sedentary behaviors in the United States, 2003-2004. Am J Epidemiol
24. Matton L, Wijndaele K, Duvigneaud N, et al. Reliability and validity of the Flemish Physical Activity Computerized Questionnaire in adults. Res Q Exerc Sport
25. Norman A, Bellocco R, Bergström A, Wolk A. Validity and reproducibility of self-reported total physical activity-differences by relative weight. Int J Obes Relat Metab Disord
26. Otten JJ, Littenberg B, Harvey-Berino JR. Relationship between self-report and an objective measure of television-viewing time in adults. Obesity (Silver Spring)
27. Owen N, Bauman A, Brown W. Too much sitting: a novel and important predictor of chronic disease risk? Br J Sports Med
28. Patel AV, Bernstein L, Deka A, et al. Leisure time spent sitting in relation to total mortality in a prospective cohort of US adults. Am J Epidemiol
29. Reeves MM, Marshall AL, Owen N, Winkler EA, Eakin EG. Measuring physical activity change in broad-reach intervention trials. J Phys Act Health
30. Rikli RE. Reliability, validity, and methodological issues in assessing physical activity in older adults. Res Q Exerc Sport
. 2000;71(2 suppl):S89-96.
31. Rosenberg DE, Norman GJ, Wagner N, Patrick K, Calfas KJ, Sallis JF. Reliability and validity of the sedentary behavior questionnaire (SBQ) for adults. J Phys Act Health
32. Salmon J, Owen N, Crawford D, Bauman A, Sallis JF. Physical activity and sedentary behavior: a population-based study of barriers, enjoyment, and preference. Health Psychol
33. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull
34. Tuley MR, Mulrow CD, McMahan CA. Estimating and testing an index of responsiveness and the relationship of the index to power. J Clin Epidemiol
35. van der Ploeg HP, Merom D, Chau JY, Bittman M, Trost SG, Bauman AE. Advances in population surveillance for physical activity and sedentary behavior: reliability and validity of time use surveys. Am J Epidemiol
36. van Poppel MN, Chinapaw MJ, Mokkink LB, van Mechelen W, Terwee CB. Physical activity questionnaires for adults: a systematic review of measurement properties. Sports Med
37. Warren TY, Barry V, Hooker SP, Sui X, Church TS, Blair SN. Sedentary behaviors increase risk of cardiovascular disease mortality in men. Med Sci Sports Exerc
38. Waters LA, Galichet B, Owen N, Eakin EG. Who participates in physical activity intervention trials? J Phys Act Health