Cohort analysis of children's sports participation
Many experts recognise that children's participation in sports is beneficial to health and physical fitness and also facilitates the development of social/interpersonal skills and psychological well-being. This is why information on the range, frequency and duration of sporting activities, the characteristics of children who participate in them and changes in these characteristics over time, are of particular interest to researchers and policymakers. Also of interest is any information on changes in participation during the childhood lifecycle. For example, do children continue to play the same sports or increase/decrease or rearrange their sporting activities as they grow older? Detecting patterns in children's sports participation over their lifecycle and reasons behind these patterns could inform policymaking.
While changes in children's sports participation over their lifecycle can be best examined using longitudinal data, which follows the same child over time, such data are not always available and are often costly to collect. One alternative is to pool successive cross-sectional data and construct a 'longitudinal' or 'pseudo-panel' data to study this phenomenon. A research project currently underway in Analytical Services Branch, in collaboration with the National Centre for Culture and Recreation Statistics, aims to do exactly this. The study pools together data from the three (2000, 2003, 2006) cross-sectional surveys on Children's Participation in Cultural and Leisure Activities to construct a 'longitudinal' or 'pseudo-panel' data to study children's participation in sports over time. From this pooled dataset, cohorts of children were identified, and each cohort's sports participation over time was traced. Cohorts have been defined on the basis of the year of birth of the child, and both three-year birth cohorts and single-year birth cohorts are being studied in this project. A modelling technique called age-period-cohort analysis is also being tested, which is a regression-based decomposition technique that disentangles the confluence of age, birth cohort, and period effects on children's sports participation.
The sorts of questions the study will attempt to provide answers to include:
- How do the rates of participation in sport by 6-8 year olds identified in the 2000 survey compare with those of 9-11 years olds in the 2003 survey and 12-14 year olds in the 2006 survey?
- Is there a cohort effect in sports participation?
- Are there any changes in participation in sport during the childhood lifecycle i.e., do young children who play sport, continue to play sport or participate in similar or different activities as they grow?
- Do these participants (in sports) have similar socio-demographic characteristics?
- Are there any changes in the mix of activities e.g. do 6-8 year old children who have music lessons, play sport and watch television for below average hour continue this trend as 9-11 and 12-14 year olds?
For more information about the analysis, please contact Anil Kumar on (02) 6252 5344.