Measuring physical activity in public open space--an electronic device versus direct observation

Using automated active infrared counters to estimate footfall on urban park footpaths: behavioural stability and validity testing

BACKGROUND

Using infrared counters is a promising unobtrusive method of assessing footfall in urban parks. However, infrared counters are susceptible to reliability and validity issues, and there is limited guidance for their use. The aims of this study were to (1) determine how many weeks of automated active infrared count data would provide behaviourally stable estimates of urban park footfall for each meteorological season, and (2) determine the validity of automated active infrared count estimates of footfall in comparison to direct manual observation counts.

METHODS

Three automated active infrared counters collected daily footfall counts for 365 days on three footpaths in an urban park within Northampton, England, between May 2021 - May 2022. Intraclass correlation coefficients were used to compare the behavioural stability of abbreviated data collection schedules with total median footfall within each meteorological season (Spring, Summer, Autumn, Winter). Public holidays, events, and extreme outliers were removed. Ten one-hour manual observations were conducted at the site of an infrared counter to determine the validity of the infrared counter.

RESULTS

At least four-weeks (28 days) of infrared counts are required to provide 'good' to 'excellent' (Intraclass correlation > 0.75, > 0.9, respectively) estimates of median daily footfall per meteorological season in an urban park. Infrared counters had, on average, -4.65 counts per hour (95% LoA -12.4, 3.14; Mean absolute percentage error 13.7%) lower counts compared to manual observation counts during one-hour observation periods (23.2 ± 15.6, 27.9 ± 18.9 counts per hour, respectively). Infrared counts explained 98% of the variance in manual observation counts. The number of groups during an observation period explained 78% of the variance in the difference between infrared and manual counts.

CONCLUSIONS

Abbreviated data collection schedules can still obtain estimates of urban park footfall. Automated active infrared counts are strongly associated with manual counts; however, they tend to underestimate footfall, often due to people in groups. Methodological and practical recommendations are provided.

Opportunities for Integrating Public Health and Urban Planning Approaches to Promote Active Community Environments

The growing emphasis on promoting environmental change as a means to increase physical activity has motivated conversation and collaboration between researchers and practitioners in the fields of public health and urban planning. Although these fields share similar objectives, their methodological approaches for examining the association between the environment and behavior often differ in significant ways. To facilitate communication, this article discusses ways these fields can collaborate in developing and applying conceptual frameworks, adopting behavioral and environmental measures, and strengthening study designs. By collaborating to build synergism in research and dissemination, public health and urban planning professionals can enhance efforts to increase the number of communities that promote active living.

Adolescents' sports and exercise environments in a U.S. time use survey

BACKGROUND

Studies examining environmental influences on adolescent physical activity largely measure the presence and availability of social resources and built environment facilities. Unfortunately, this research approach provides limited information about adolescents' social company during exercise or the extent to which adolescents actually use physical settings for physical activity.

PURPOSE

The current study used data from the nationally representative American Time Use Survey (2003-2006) to describe demographic and temporal patterns in the social and physical contexts of physical activity among adolescents.

METHODS

The sample consisted of high school students (aged 15-18 years) reporting at least one bout of sports or exercise (N=867). During the interview, participants reported where (e.g., outdoors, home, work) and with whom (e.g., alone, family, friends) each bout occurred. Sample-weighted multinomial logistic regression analyses compared the proportion of bouts occurring in each environment by age, gender, family income, season, weekend/weekday, and time of day, controlling for race/ethnicity. Data were analyzed in 2009.

RESULTS

Girls were more likely to exercise with family (22% vs 16%), and less likely to exercise with friends/acquaintances/others (47% vs 52%) and outdoors (18% vs 24%) than boys. Compared with those aged 15 years, a larger proportion of exercise bouts among those aged 18 years occurred alone (23% vs 18%); and a smaller proportion occurred at home (14% vs 20%), at someone else's house (5% vs 12%), and at school (14% vs 27%) (p's<0.001).

CONCLUSIONS

Information about the social and physical contexts of adolescents' sports and exercise can help guide the selection of future environmental targets for investigation and intervention.

Social and physical environments of sports and exercise reported among adults in the American Time Use Survey

OBJECTIVE

Demographic and temporal patterns in the social and physical environments of sports and exercise in the American Time Use Survey (years 2003-2006) are described.

METHOD

The sample consisted of adult respondents (ages 21+) reporting at least one bout of sports or exercise (N=8844). During the interview, participants reported where (e.g., outdoors, home, work) and with whom (e.g., alone, family, coworkers) each bout occurred. Sample-weighted multinomial logistic regression analyses estimated the adjusted proportion of bouts occurring in each environment by gender, age, education, race/ethnicity, season, weekend/weekday, and time of day.

RESULTS

Among members of the oldest age group (ages 60+), more exercise bouts occurred alone and outdoors compared to younger age groups. Among college graduates, more exercise bouts occurred at a gym/health club compared to groups with lower levels of education. Exercise bouts occurring alone were most likely to happen in the winter, on weekdays, and in the morning. Exercise bouts occurring outdoors were most likely to happen in the summer, on weekend days, and in the morning.

CONCLUSION

Future research and intervention efforts exploring where, when and with whom exercise bouts occur may prove beneficial to addressing public health concerns about physical inactivity.

Physical activity in preschoolers: understanding prevalence and measurement issues

Accurate physical activity quantification in preschoolers is essential to establish physical activity prevalence, dose-response relationships between activity and health outcomes, and intervention effectiveness. To date, best practice approaches for physical activity measurement in preschool-aged children have been relatively understudied. This article provides a review of physical activity measurement tools for preschoolers, an overview of measurement of preschoolers' physical activity, and directions for further research. Electronic and manual literature searches were used to identify 49 studies that measured young children's physical activity, and 32 studies that assessed the validity and/or reliability of physical activity measures with preschool-aged children. While no prevalence data exist, measurement studies indicate that preschool children exhibit low levels of vigorous activity and high levels of inactivity, boys are more active than girls, and activity patterns tend to be sporadic and omnidirectional. As such, measures capable of capturing differing activity intensities in very short timeframes and over multiple planes are likely to have the most utility with this population. Accelerometers are well suited for this purpose, and a number of models have been used to objectively quantify preschoolers' physical activity. Only one model of pedometer has been investigated for validity with preschool-aged children, showing equivocal results. Direct observation of physical activity can provide detailed contextual information on preschoolers' physical activity, but is subjective and impractical for understanding daily physical activity. Proxy-report questionnaires are unlikely to be useful for determining actual physical activity levels of young children, and instead may be useful for identifying potential correlates of activity. Establishing validity is challenging due to the absence of a precise physical activity measure, or 'criterion', for young children. Both energy expenditure (EE) and direct observation have been considered criterion measures in the literature; however, EE is influenced by multiple variables, so its use as a physical activity 'criterion' is not ideal. Also, direct observation is inherently subjective, and coding protocols may result in failure to capture intermittent activity, thereby limiting its utility as a physical activity criterion. Accordingly, these issues must be taken into account where EE or direct observation are used to validate physical activity instruments. A combination of objective monitoring and direct observation may provide the best standard for the assessment of physical activity measurement tools. Ideally, the convergent validity of various physical activity tools should be investigated to determine the level of agreement between currently available measures. The correlational approaches commonly employed in the assessment of physical activity measures do not reveal this relationship, and can conceal potential bias of either measure.