Health Education Research Advance Access originally published online on September 20, 2006
Health Education Research 2007 22(3):438-449; doi:10.1093/her/cyl107
Social desirability is associated with some physical activity, psychosocial variables and sedentary behavior but not self-reported physical activity among adolescent males
1 Department of Exercise and Health Sciences, Centre for Sport, Exercise and Health, University of Bristol, Tyndall Avenue, Bristol, BS8 1TP, UK
2 Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, 1100 Bates Street, Houston, TX 77030-2600, USA
* Correspondence to: R. Jago. E-mail: russ.jago{at}bris.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
This study examined whether controlling for social desirability improved the association between self-reported and objectively measured physical activity among adolescent males and the extent that psychosocial variables predict physical activity after controlling for social desirability. Participants (n = 447) were 10- to 14-year old Houston Boy Scouts. Participants completed self-reports of physical activity, sedentary behavior, preferences, self-efficacy and social desirability and wore an MTI accelerometer for 3 days. Correlations were conducted among variables. Regression models were performed to examine the relationships between objectively measured (accelerometer) and self-reported physical activity, objectively measured physical activity and psychosocial variables and self-reports of physical activity and psychosocial variables. All models controlled for social desirability. There were weak associations between self-reported and objectively measured physical activity measures that were slightly improved after controlling for social desirability. Psychosocial variables were strongly associated with self-reports of physical activity, but weakly associated with accelerometer physical activity. Social desirability was positively associated with physical activity preferences (r = 0.169) and self-efficacy (r = 0.118) and negatively associated (r = –0.158) with self-reported sedentary behavior. Differences in the strength of relationships between self-reported and objectively measured physical activity and psychosocial variables were not a function of social desirability.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
US children and adolescents do not currently meet physical activity recommendations [1, 2]. Thus, there is a need to identify the factors that influence their physical activity so that effective intervention programs can be designed [3]. Correlations of psychosocial variables (e.g. self-efficacy) with physical activity have been higher with self-reported than objectively measured physical activity [4, 5]. The higher correlations between self-reported physical activity and self-reported psychosocial variables could be due primarily to shared variance which was a function of social desirability [5–7]. Social desirability is a tendency for individuals to provide responses that they believe to be consistent with social norms [8] and expectations. Physical activity is well documented as a health enhancing behavior and thus, in the context of the current research social desirability could be defined as the extent to which participants overreport engaging in health enhancing physical activity and under-report time spent engaged in non-active pursuits.
Social desirability has been reported to affect self-reported physical activity, although the evidence is not conclusive. Klesges et al. [5] reported that self-reported physical activity was associated with social desirability among 8- to 10-year old African–American girls. Warnecke et al. [6] reported that social desirability predicted self-reported physical activity among a racially diverse sample of adults. Adams et al. [8] reported that social desirability was associated with physical activity among adult women when physical activity was assessed using a 7-day physical activity recall, but not when a 24-day recall was used. This suggests that associations may differ by method of assessment. Similarly, Motl et al. [7] reported a weak association between social desirability assessed by the Marlowe–Crowne social desirability scale and self-reported physical activity, but no association with the same measures of physical activity when the Lie scale of social desirability was used in the same sample. Further, social desirability has been associated with under-reporting of diet in women, but not men [9]. Therefore, the association between social desirability and self-reported physical activity could be gender specific or instrument specific.
If self-reports of physical activity among youth are influenced by social desirability, it is possible that controlling for social desirability may remove some of the error in the self-report measures. Removing error might increase the statistical associations between self-reported and accelerometer-measured physical activity. Equally self-reports of sedentary behavior, the opposite end of the activity continuum, could also be affected by respondents' willingness to report engaging in more physical activity thereby under-reporting sedentary behavior. Controlling for social desirability may therefore improve the validity of the less-expensive assessment methods.
Since activity-related psychosocial variables could also be biased by social desirability, it is possible that our understanding of the factors that influence adolescent physical activity behaviors is hindered by social desirability. Thus, a clear understanding of the role of social desirability and its relation to self-reported physical activity and activity-related psychosocial variables would increase measurement precision. The increased precision in measurement may aid our understanding of the variables that influence physical activity and how they could be changed. An examination of whether psychosocial variables provide stronger predictions of self-reported physical activity than accelerometer-measured physical activity after controlling for social desirability would also provide insight into whether the increased predictiveness of these models is a function of shared social desirability variance.
As social desirability has been associated with body mass index (BMI) among 8- to 10-year old African–American girls [5], it could be that heavier individuals are more likely to provide socially desirable answers, thereby reducing the validity of these measurements. Currently, the only information about the association between social desirability and self-reports of physical activity among youth is from African–American girls [5]. Thus, there is a need for comparable data from adolescent males.
This paper addresses the research needs outlined above by examining among adolescent males: (i) the association between self-reported and objective assessments of physical activity before and after controlling for social desirability; (ii) the associations between social desirability and physical activity psychosocial variables (self-efficacy and preferences) and (iii) the extent to which psychosocial variables and demographics predict self-reported and objectively assessed physical activity after controlling for social desirability and differences in the predictiveness of these models.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
Participants
Participants (n = 447) were 10- to 14-year old Boy Scouts recruited from the greater Houston area. Data were from the baseline assessment of a Boy Scout physical activity intervention trial and were collected during the spring and fall of 2003 and it was assumed that participants were representative of similar aged male adolescents in the measured socio-economic and ethnic groups in Houston. Ethical approval was obtained from the local institutional review board and written informed parental consent and child assent were obtained for all participants. Participants' ethnicity and the highest education achieved within the household (an indicator of socio-economic status) were obtained by parental self-report. Participant's stature was measured to the nearest 0.1 cm using a stadiometer (Shorr Height Measuring Board, Olney, MD, USA) in accordance with a standard stature protocol [10]. Body weight was measured to the nearest 0.1 kg using a calibrated scale (Seca 770 Model Scale, Vogel and Halke, Hamburg, Germany) in accordance with standard assessment procedure [10]. BMI (kilogram per meter square) and body mass percentile were computed using Centers for Disease Control and Prevention age and gender specific percentiles [11].
Physical activity
Physical activity was monitored for 3 days using the previously validated [12] MTI accelerometer (Manufacturing Technologies Inc., Fort Walton Beach, FL, USA). The MTI accelerometer is a uniaxial matchbox-sized device that detects vertical acceleration, which is then filtered to produce readings or counts for a set period of time [13]. The monitors in this study were set to record at 1-min intervals and consequently counts were obtained for each minute in which the device was worn to provide an indication of the magnitude of activity in which the participants engaged. Each monitor was attached to an elastic belt at the waist above the right hip during a troop meeting and programmed to begin recording at midnight. Participants were instructed to wear the monitor at all times (including during sleep) except when bathing or swimming. Monitors were removed on the fourth morning after data collection and returned by courier.
Self-reported physical activity and sedentary behavior were assessed using the boys activity questionnaire [14] that asked whether participants engaged in 22 common activities at all, for <15 min, or 
15 min on the previous day. Responses to each question were coded as 0, 7.5 (midpoint between none and 15 min) or 15 min (lower end of possible scores) of physical activity. The coded responses of all activities were then summed to provide an estimate of the minutes of physical activity in which the participant engaged on the ‘previous day’. The same questions were also asked for ‘usual activity’ with three response categories: none, a little or a lot, which were coded also as 0, 7.5 or 15 min of habitual activity to facilitate comparison with the yesterday responses, and data were summed. Self-reports of physical activity and sedentary behavior were completed when the activity monitors were distributed, i.e. the yesterday measurement corresponds to the day before the first day of activity monitoring. While this prevented an overlap in measurements, it is hypothesized that all measures provide an indication of habitual physical activity, and should therefore be associated.
A correlation of 0.8 between two administrations of this activity questionnaire taken 5 days apart has been reported, suggesting good test–retest reliability [15]. Higher social desirability scores were associated with higher self-reported physical activity among 8- to 10-year old African–American girls [5]. Controlling for social desirability could improve the validity of this instrument. Further, adoption of these physical activity and social desirability measures facilitated comparison with the only other study conducted among youth (African–American girls) which used the same questionnaires [5].
Sedentary behavior for the ‘previous day’ was assessed by asking scouts to report whether they spent no time, <30 min, 30–60 min, 1–3 hours or >3 hours day–1 in six different sedentary pursuits. The six different activities were watching television/videos, playing computer games, playing board games, homework/reading, talking on the phone/hanging out with friends and listening to music/playing a musical instrument. These activities were included as adolescents could underestimate time spent engaged in sedentary behavior as a function of the social desirability trait to be more active and/or under-report time spent in less socially desirable activities such as sedentary behaviors. Since we are not aware of previous studies that have examined this possibility among adolescent males, the sedentary behavior questions were added to this study. Time spent in these activities was coded as 0, 15, 45, 90 and 180 min day–1, respectively. The coded responses for each participant were then summed. Participants were also asked to report the amount of time that they usually engaged in each behavior and the totals were summed to provide an estimate of sedentary behavior.
Psychosocial variables
Physical activity preferences were assessed using the previously validated 22-item questionnaire which asked whether participants (i) did not like participating in the activity, (ii) were not sure, (iii) liked participating in the activity a little or (iv) liked participating a lot and responses summed [16]. Sedentary preferences were assessed using a nine-item [16] questionnaire, which was coded using the same methods and responses summed. Physical activity self-efficacy was assessed using a previously validated adaptation of the self-efficacy scale developed by Saunders et al. [17]. Participants were asked whether they (i) disagreed a lot, (ii) disagreed, (iii) were not sure, (iv) agreed or (v) agreed a lot with 19 physical activity self-efficacy statements. Responses were summed across items to provide a self-efficacy rating. Social desirability was assessed using a modified version of the ‘lie scale’ from the revised Manifest Anxiety Scale [18], a nine-item questionnaire that asks whether socially desirable characteristics are always true of me (5), sometimes true of me (4), not sure (3), not often true of me (2) or never true of me (1). Some items included in this scale are ‘I never lie’, ‘I never get angry’, ‘I tell the truth every single time’ and ‘I am always good’. This scale is one of the most frequently used measures of social desirability, particularly among children and a previous paper has justified its use for children [5].
Data reduction
The number of accelerometer minutes per day was determined using an SPSS program with 20 min of continuous zeros indicating that the monitor was not worn [19]. As a result of a technical error in which the sampling period was incorrectly set to record every second as opposed to every minute, MTI data were not obtained from 96 participants. Remaining participants were included if the scout wore the instrument for at least 800 min between 6.00 a.m. and midnight (maximum 1080 min) for two out of the 3 days [14]. Adolescent cutpoints developed by Puyau et al. [12] were used to categorize the physical activity recorded in each minute as sedentary (<800 counts), light (800–3199 counts) or moderate-to-vigorous intensity (3200 counts) and mean minutes per day in each category was calculated. There is currently no consensus [20, 21] on the best cutpoints to use to interpret accelerometer data obtained from adolescents. However, the chosen cutpoints were generated using whole-body calorimetry [12]. This is the most sophisticated approach that has been employed and the cutpoints obtained are similar to those obtained from field-based energy consumption among adolescent girls [22]. To handle missing data, minutes were adjusted by multiplying the original values by the inverse of the percent total time reported [1]. Minutes of self-reported habitual physical activity and sedentary behavior were calculated for each participant.
Statistical analyses
Descriptive statistics including means and standard deviations (SDs) were calculated for all variables. Bivariate Pearson correlations were conducted among self-reported physical activity and sedentary behaviors, MTI physical activity and sedentary minutes, psychosocial variables, BMI and social desirability. Linear regression models that accounted for the clustering of scouts within troops were performed using the ‘XTREG’ procedure in STATA (version 9, College Station, TX, USA) and the within troop and between troop R2s obtained. To examine the relationships between self-reported and accelerometer data, models were run with the self-report variables as the dependent variables and the accelerometer variables as the independent variables. Social desirability was then added to the models. To examine the extent to which psychosocial variables were associated with physical activity and sedentary behaviors six models in which the physical activity and sedentary behaviors were the dependent variables and psychosocial variables were the independent variables were run while controlling for social desirability. Demographic characteristics that have previously been associated with physical activity (BMI, age, parental education and ethnicity) [23] were included in the models. Since >72% of the participants were Anglo-American, ethnicity was coded as Anglo-American or Other. As analyses were exploratory, psychosocial variables that did not approach significance (P < 0.10) were removed in a backward elimination process. Significance was set at 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
Participant characteristics are shown in Table I. Participants were predominately from households in which at least one parent was college educated. The mean age of the participants was 12.8 years with a mean BMI of 21.1 and a mean BMI percentile of 65.4.
|
Means and SDs for all physical activity, sedentary behavior, psychosocial variables are shown in Table I. Participants reported spending considerably more time (86–124 min day–1) engaged in moderate-to-vigorous physical activity (MVPA) than the level documented by the accelerometer (24.4 min day–1). Conversely, more time was spent engaged in sedentary behavior according to the accelerometer (917 min day–1) than the self-reported (244–294 min day–1) data.
Correlations among physical activity, sedentary behavior, psychosocial variables, BMI and social desirability are shown in Table II. Social desirability was positively associated with physical activity preferences (r = 0.169, P < 0.001) and self-efficacy (r = 0.118, P = 0.023) and negatively associated with self-reported sedentary behavior yesterday (r = –0.158, P = 0.001). Self-reported physical activity yesterday was weakly associated (r = 0.103, P = 0.086) with accelerometer-obtained minutes of MVPA per day. There were low to moderate associations between self-reported physical activity and psychosocial variables (r ranged from 0.318 to 0.647, P < 0.001) with weaker, but significant, associations between accelerometer minutes of MVPA, self-efficacy (r = 0.185, P = 0.001) and physical activity preferences (r = 0.123, P = 0.041). BMI was negatively associated with accelerometer minutes of MVPA (r = –0.185, P = 0.002), but not self-reported MVPA.
|
Regression models in which self-reported physical activity or sedentary behavior was the dependent variable and either accelerometer minutes of MVPA or sedentary behavior was the independent variable are shown before and after controlling for social desirability (Table III). The associations between the self-reported physical activity and accelerometer MVPA were low (standard ß all
0.13), with weaker association between self-reported sedentary behavior and accelerometer sedentary behavior (standard ß all 0.10). Social desirability was a significant predictor (standard ß = –0.15, P = 0.008) of self-reported sedentary behavior yesterday, only. Although the standard ßs for the associations between the objective and self-report report measures were largely unchanged after the addition of social desirability to the model, the within troop R2s did improve for self-reported usual MVPA as well as both measures of sedentary behavior. This suggests that the addition of social desirability had a small impact on the explainable variable variance in self-reported physical activity and sedentary behavior among these male adolescents.
|
The regression models in which accelerometer-obtained minutes of MVPA or sedentary behavior were the dependent variables are presented in Table IV. Physical activity self-efficacy was positively associated (standard ß = 0.162, P = 0.006) while BMI was negatively associated (standard ß = –0.137, P = 0.025) with minutes of MVPA. The model accounted for 6% of the variance in activity within troops and 17.8% of the variance between troops. Age was positively associated (standard ß = 0.233, P < 0.001) with minutes of sedentary behavior and self-efficacy negatively associated (standard ß = –0.133, P = 0.020) in a model that a within troop R2 of 0.10 and a between troop R2 of 0.345.
|
The regression models in which self-reported physical activity (yesterday or usual) was the dependent variable are shown in Table V. Self-efficacy was positively associated (standard ß = 0.319, P < 0.001) with self-reported physical activity yesterday. The within troop R2 was 0.119 and between troop R2 was 0.079. Physical activity self-efficacy (standard ß = 0.237, P < 0.001) and physical activity preferences (standard ß = 0.309, P < 0.001) were positively associated with usual self-reported physical activity in a model that accounted for 47.8 of the variance within troops and 41.9% of the variance between troops.
|
The regression models in which self-reported sedentary behavior (yesterday or usual) was the dependent variable are also shown in Table V. Sedentary behavior preferences (standard ß = 0.156, P = 0.001), and age (standard ß = 0.126, P = 0.007) were positively associated with self-reported sedentary behavior yesterday while social desirability was negatively associated (standard ß = –0.156, P = 0.001). The model accounted for 6.4% of the variance within troops and 12.2% of the variance between troops. Sedentary behavior preferences (standard ß = 0.381, P < 0.001) and age (standard ß = 0.176, P < 0.001) were positively associated with usual self-reported sedentary behavior while social desirability was negatively associated (standard ß = –0.099, P = 0.023). The model accounted for 18% of the variance within troops and 25.4% of the variance between troops.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
Associations between physical activity and psychosocial variables were higher for self-reported physical activity than objectively measured physical activity, but social desirability did not appear to be a strong contributor to this association. It is not clear why there is a closer association between self-reported physical activity and psychosocial variables, but given the weak associations between the self-reported and objectively assessed physical activity measures, it is likely that these two methods did not measure the same construct. Some other, third variable may account for the correlations between self-reported physical activity and self-reported psychosocial variables. For example, children and adolescents' perceptions of sporting ability [24, 25] have been associated with physical activity. It is possible that participants' self-concept of athletic ability influenced their self-report of physical activity and self-efficacy [24]. This is supported by the correlation (r = 0.434) between self-efficacy and physical activity preferences, which suggests that participants who perceive a greater ability to engage in activity also had higher physical activity preferences. Other psychosocial variables such as self-worth [26] and body satisfaction [27] may be worthy of further investigation particularly among girls, who were not the focus of this study.
Alternatively, it could be that psychosocial variables predict habitual activity which is less accurately captured by accelerometer monitoring over a limited number of days. Support for this possibility is provided by the regression models in which usual self-reported MVPA or usual sedentary behavior was the dependent variable which accounted for the highest proportion of the within troop variance (47 and 18%). It could also be possible that the tendency to provide socially desirable responses exerted a significant impact on self-reported physical activity but our measure of social desirability might not have been most appropriate for this group. Thus, in order to further explore this issue there is a need for replication using multiple measures of social desirability.
Social desirability was positively associated with two psychosocial variables: physical activity preferences and self-efficacy, but was only associated with one behavioral variable: self-reported sedentary behavior yesterday, and this was a negative association. This is in contrast to Klesges et al. [5] who reported a positive association between social desirability and self-reported physical activity, but no association between social desirability and physical activity self-efficacy and preferences among 8- to 10-year old African–American girls. Further, although Klesges et al. [5] reported a negative association between BMI and social desirability, we did not find such a relationship. This may suggest that the influence of social desirability changes as children age or those relationships differ by ethnicity (the current sample was predominantly Anglo-American while Klesges' sample was comprised of African–Americans). Alternatively, since this study only included boys while the study by Klesges et al. [5] included only girls, the differences between the studies could be a function of gender.
Self-reported physical activity yesterday was only very weakly associated (r = 0.103, P = 0.086) with accelerometer-determined minutes of MVPA with non-significant associations between all other objective and self-reported paired variables. Although this low correlation is likely to be influenced by the lack of overlap in measurements the results also suggest that the two different measurement approaches are not providing a comparable assessment of habitual physical activity. These findings are similar to a previous study comparing data from an earlier version of this questionnaire, which also reported very low correlations with accelerometer-determined MVPA [14]. Therefore, it is possible to conclude that although the questionnaire may have provided some insights into the activities in which children and adolescents engaged, it did not provide an accurate representation of habitual physical activity, to the extent that this is measured by the MTI.
Although the results of this study do not provide support for social desirability accounting for differences between self-report and objective assessments of physical activity, the data presented here indicate that physical activity self-efficacy and preferences predict physical activity among male adolescents. It is important to note, however, that the regression models explained <10% of the within troop variance in objectively measured physical activity and sedentary behavior. Thus, other variables are likely predicting physical activity in this group. Understanding the role of these variables may provide further insights into the disparity between self-report and objective assessments of physical activity and may help design more effective strategies to increase adolescent physical activity.
Although the data presented in this paper have provided insights into the role of social desirability in self-reported physical activity and related psychosocial variables, further examination in different age, gender and ethnic groups is warranted. As adult research suggests that relationships may be specific to a particular scale, further research that examines associations with multiple measures of social desirability among children and adolescents is required [7, 9]. Finally, since the findings suggest that other variables such as self-perceptions of sporting ability could account for the stronger associations between self-reported physical activity and other psychosocial variables, more research in this area is also required.
Strengths and limitations
This study combined objective and self-report measures of physical activity, psychosocial variables and social desirability in a population that has not been previously studied. The study is limited by the use of a physical activity self-report instrument that has low validity against an accelerometer, and by the male, predominantly middle class, Anglo-American sample that limits the ability to generalize to other gender and ethnic groups. It is not clear if different associations might have been detected with other self-report measures of physical activity that provide actual self-reported minutes of activity or the momentary time sampling of physical activity [28], or with a different questionnaire; and thus examining associations with alternative self-reports of physical activity is warranted. The accelerometer physical activity data may also have limited our ability to detect relationships as data were included if a participant possessed at least 2 days of data, when 4 days have been recommended [13].
|
Conclusions |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
These data show that among 10- to 14-year old males, physical activity was associated with psychosocial variables, but stronger associations were detected when self-reports of physical activity were employed and this difference in relationship was not a function of social desirability. The study also shows that controlling for social desirability did not improve the relationship between self-reports and objective assessments of physical activity in this sample.
|
Conflict of interest statement |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
None declared.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
This study was funded by a grant from the American Cancer Society (ACS TURSG-01). This work is also a publication of the United States Department of Agriculture/Agricultural Research Service (USDA/ARS) Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children's Hospital, Houston, TX, USA. This project has been funded in part by federal funds from the USDA/ARS under co-operative agreement 58-6250-6001. The contents of this publication do not necessarily reflect the views or polices of the USDA, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsConflict of interest statementAcknowledgementsReferences |
|---|
1. Jago R, Anderson C, Baranowski T, et al. Adolescent patterns of physical activity: differences by gender, day and time of day. Am J Prev Med (2005) 28:447–52.[CrossRef][Web of Science][Medline]
2. Pate RP, Freedson PS, Sallis JF, et al. Compliance with physical activity guidelines: prevalence in a population of children and youth. Ann Epidemiol (2002) 12:303–8.[CrossRef][Web of Science][Medline]
3. Baranowski T, Jago R. Understanding mechanisms of change in children's physical activity programs. Exerc Sport Sci Rev (2005) 33:163–8.[CrossRef][Web of Science][Medline]
4. Trost SG, Pate RR, Dowda M, et al. Psychosocial correlates of physical activity in white and African-American girls. J Adolesc Health (2002) 31:226–33.[Web of Science][Medline]
5. Klesges LM, Baranowski T, Beech B, et al. Social desirability bias in self-reported dietary, physical activity and weight concerns measures in 8- to 10-year-old African-American girls: results from the Girls Health Enrichment Multisite Studies (GEMS). Prev Med (2004) 38(Suppl.):S78–87.[Web of Science][Medline]
6. Warnecke RB, Johnson TP, Chavez N, et al. Improving question wording in surveys of culturally diverse populations. Ann Epidemiol (1997) 7:334–42.[CrossRef][Web of Science][Medline]
7. Motl RW, McAuley E, Distefano C. Is social desirability associated with self-reported physical activity? Prev Med (2005) 40:735–9.[CrossRef][Web of Science][Medline]
8. Adams SA, Matthews CE, Ebbeling CB, et al. The effect of social desirability and social approval on self-reports of physical activity. Am J Epidemiol (2005) 161:389–98.
9. Novotny JA, Rumpler WV, Riddick H, et al. Personality characteristics as predictors of underreporting of energy intake on 24-hour dietary recall interviews. J Am Diet Assoc (2003) 103:1146–51.[CrossRef][Web of Science][Medline]
10. Gordon CC, Chumlea WC, Roche AF. Stature, recumbent length, and weight. In: Anthropometric Standardization Reference Manual—Lohman TG, Roche AF, Martorell R, eds. (1988) Champaign, IL: Human Kinetics. 3–8.
11. National Center for Health Statistics. CDC Growth Charts: United States (2004) [cited 2005]. Available at: http://www.cdc.gov/growthcharts/.
12. Puyau MR, Adolph AL, Vohra FA, et al. Validation and calibration of physical activity monitors in children. Obes Res (2002) 10:150–7.[Web of Science][Medline]
13. Trost SG, Pate RR, Freedson PS, et al. Using objective physical activity measures with youth: how many days are needed. Med Sci Sports Exerc (2000) 32:426–31.
14. Treuth M, Sherwood N, Baranowski T, et al. Physical activity self-report and accelerometry measures from the Girls Health Enrichment Multi-site Studies. Prev Med (2004) 38(Suppl.):S43–9.[CrossRef][Web of Science][Medline]
15. Treuth MS, Sherwood NE, Butte NF, et al. Validity and reliability of activity measures in African American girls for GEMS. Med Sci Sports Exerc (2003) 35:532–9.
16. Sherwood N, Taylor WC, Treuth M, et al. Measurement characteristics of activity related psychosocial measures in 8-10 year old African American girls in the Girls Health Enrichment Multi-site Study (GEMS). Prev Med (2004) 38(Suppl.):S60–8.[CrossRef][Web of Science][Medline]
17. Saunders RP, Pate RR, Felton G, et al. Development of questionnaires to measure psychosocial influences on children's physical activity. Prev Med (1997) 26:241–7.[CrossRef][Web of Science][Medline]
18. Reynolds CR, Paget KD. National normative and reliability data for the Revised Children's Manifest Anxiety Scale. School Psychol Rev (1983) 12:324–6.
19. Jago R, et al. Observed environmental features and the physical activity of adolescent males. Am J Prev Med (2005) 29:98–104.[CrossRef][Web of Science][Medline]
20. Reilly JJ, Jackson DM, Montgomery C, et al. Total energy expenditure and physical activity in young Scottish children: mixed longitudinal study. Lancet (2004) 363:211–2.[CrossRef][Web of Science][Medline]
21. Ekelund U, Brage S, Wareham NJ. Physical activity in young children. Lancet (2004) 363:1163.[Web of Science][Medline]
22. Treuth M, Scmitz K, Catellier D, et al. Defining accelerometer thresholds for physical activity intensities in adolescent girls. Obes Res (2003) 11(Suppl.):A33.
23. Sallis JF, Zakarian JM, Hovell MF, et al. Ethnic, socioeconomic, and sex differences in physical activity among adolescents. J Clin Epidemiol (1996) 49:125–34.[CrossRef][Web of Science][Medline]
24. Crocker PR, Eklund RC, Kowalski KC. Children's physical activity and physical self-perceptions. J Sports Sci (2000) 18:383–94.[CrossRef][Web of Science][Medline]
25. Anderson CB. Athletic identity and its relation to exercise behavior: scale development and initial validation. J Sport Exerc Psychol (2004) 26:39–56.[Web of Science]
26. Neumark-Sztainer D, Story M, Hannan PJ, Tharp T, et al. Factors associated with changes in physical activity: a cohort study of inactive adolescent girls. Arch Pediatr Adolesc Med (2003) 157:803–10.
27. Ferron C, Narring F, Cauderay M, et al. Sport activity in adolescence: associations with health perceptions and experimental behaviours. Health Educ Res (1999) 14:225–33.
28. Murdey ID, Cameron N, Biddle SJ, et al. Short-term changes in sedentary behaviour during adolescence: project STIL (Sedentary Teenagers and Inactive Lifestyles). Ann Hum Biol (2005) 32:283–96.[CrossRef][Web of Science][Medline]
Received on August 22, 2005; accepted on August 22, 2006
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||