Physical activity assessment – Observation

Direct observation has been used in children as a criterion method to assess physical activity. The main strength of this method is the detailed contextual information it provides but its subjective nature does not allow intensity or energy expenditure to be assessed. Considerable time and effort is required to conduct observation studies but advances in technology have increased the potential of this method to assess physical activity. The increasing interest in the determinants of physical activity and the social environmental influences of activity in young people has also caused an increase in the use of this method (Trost, 2007).
Typically an observer will watch the children using a specific observational system and record a rating of physical activity level into a laptop computer or coding form. A number of systems are available and have been comprehensively reviewed (McKenzie, 2002) and recently specifically in pre-schoolers (Oliver, 2007). McKenzie (2002) reviewed nine protocols, for observing physical activity, eight of which had been validated concurrently with accelerometry, heart rate monitoring, or energy expenditure assessed by indirect calorimetry. Inter-observer reliability was also found to be high in these studies with coefficients of agreement ranging from 84% to 99%.

A summary of the established observation systems is given:

The Fargo Activity Timesampling Survey (FATS) – was the first attempt to develop a research instrument for describing children’s activity and related parent behaviour. The FATS used an interval time sampling procedure; observing behaviour for 10 seconds and then record behaviour for 10 seconds (Klesges et al, 1984).

Activity Patterns and Energy Expenditure (APEE) – observes free play by 15 second momentary time sampling and rates activity level into one of 5 categories (Epstein et al, 1984).

Children’s Physical Activity Form (CPAF) – was designed to measure physical activity during physical activity classes using partial time sampling at 1 minute intervals. There are four activity level categories (O’Hara, 1989).

Behaviors of Eating and Physical Activity for Children’s Health: Evaluation System (BEACHES) – measures physical activity in a range of settings. It uses momentary time sampling at 1 minute intervals and 5 activity level categories (McKenzie et al, 1991).

Children’s Activity Rating Scale (CARS) – is a five level scale (resting, low, medium, medium-to-high, vigorous), designed to categorise the intensity of physical activities and discriminate between levels of energy expenditure in young children. The system uses partial time sampling at 1 minute intervals (Puhl et al, 1990).

Studies of Children’s Activity and Nutrition: Children’s Activity Timesampling Survey (SCAN CATS) – measures physical activity in a range of locations with an interval time sampling procedure to observe behaviour for 10 seconds and then record behaviour for 10 seconds. There are four activity categories: stationary, minimal activity, slow movement, rapid movement (Klesges et al, 1990).

System for Observing Fitness Instruction Time (SOFIT) – measures activity during physical education classes using momentary time sampling; 10 second observe/record intervals. Five activity categories are used (McKenzie et al, 1991).

Level and Tempo of Children’s Activity (LETO) – measures physical activity in natural settings including the home using 3 second time sampling. Activity is categorised as 14 postures with 3 intensity levels (Bailey et al, 1995).

More recently, two observational systems have been developed which incorporate the previously validated CARS and SOFIT systems:

System for Observing Play and Leisure Activity in Youth (SOPLAY) – designed to capture behavioural and contextual information in groups of (rather than individual) children and adolescents. The target area is scanned to record the number of boys and girls, their activity level and additional information including, the time, temperature, equipment provision, level of organisation and supervision. This system was validated with self-reported physical activity (r=0.35-0.73); inter-observer reliability was found to be high, with intraclass reliability coefficients for activity levels ranging from 0.76-0.99 and percentage agreement for physical activity context ranged from 88-97% (McKenzie et al, 2000).

System for Observing Play and Recreation in Communities (SOPARC) has recently been developed from SOPLAY. The aim is to capture physical activity and contextual information in people attending community-based settings such as parks and recreation areas. An area is scanned systematically from left to right. The activity levels and types are recorded for boys and girls, concurrently with contextual information (Brown et al, 2006).

Observational System for Recording Physical Activity in Children – preschool version (OSRAC-P) – has recently been developed to collect contextual and behavioural information at an individual level in a range of pre-school settings. The OSRAC-P measures: physical activity level using the CARS system; type of activity; location of activity; social context; prompts for activity. The data are entered directly into a personalised digital assistant (PDA) or Pocket PC (McKenzie et al, 2006). Inter-observer agreement for OSRAC-P in three preschoolers was assessed and percentage agreements for each of the major variables ranged from 89% to 100% (Brown et al, 2006).