Malaysian schoolchildren study illustrates relative age effect
The relative age effect (RAE) describes the difference in children who are the same chronological age but have different developmental ages. The technical definition is "an observable effect resulting from a relationship between chronological age and an eligibility cut-off date used for cohort selection." In sport development this difference can be a problem when we don't account for it in team selection, and opportunities for practice, coaching, and competition.
RAE leads to a selection bias for older athletes because maturity is mistaken for ability. When younger athletes gets less attention and fewer opportunities for participation these youngsters either dropout of the activity or suffer from an overall lack of development when compared to their older counterparts. This reduces the talent pool, which is felt years later when the number of candidates for national teams is small and real talent becomes hard to find.
Older athletes are bigger and seem to have higher levels of strength, speed, and coordination. All of which are related to maturity.
The effect is quite normal and can be found in almost any cohort of children based on an annual selection period. Those born early in the period will have an advantage over those born later. Parents and teachers can relate stories about varying rates of growth between youngsters in the same family or classroom. We can't battle the RAE but we can and should implement steps to mitigate its effects
The RAE can be found in any annual selection period. Those born early in the period will have an advantage over those born later. For example, if a selection period runs from 1 September to 31 August, as many sport and school calendars do, then those born in Q1 (September, October, and November) are older than those born in Q4 (June, July, and August).
RAE lingo uses quarters (Q1, Q2, etc.) when discussing age cohorts. Q1 children are older than Q4 children because they were born early in the selection period i.e. during the first quarter. These differences disappear as children age but the damage is done in many cases early on because the RAE is not considered when children are selected for sport teams, praised for their ability, or offered advanced coaching. The "younger" athletes are sidelined because they are perceived to be less talented than their "older" peers when ability has nothing to do with theirs skills at all.
The effect of the RAE on sport performance can be profound thus it is often thought of as a sport issue. But the RAE results from a child's normal growth process and appears in any kind of data set that measures growth factors like height and weight.
Malaysian schoolchildren study
USSA Malaysia's Performance Data Project has collected an extensive database of over 40,000 measurements for Malaysian children ranging in age from 8 to 17. These records were collected by students in the 2014 cohort of the International Diploma in Physical Education and Sports Coaching program. Each student was tasked with collecting biometric measurements and performance data from 50 children between the ages of 8 and 17.
The full study analyzed cohorts for: 8-, 9-, 10, 15-, and 16-year-olds. For the 8s, 9s, and 10s we expected to find evidence of the RAE. For the 15s and 16s we hypothesized that the effect would be far less noticeable since growth patterns would have leveled out by these ages. Both biometric and performance data was analyzed. Biometric data included height and weight. Performance data included times for a 50 m run and a flexed-arm hang. The performance tests give us measures of speed and upper body strength.
The full report is not complete but we do have enough data to illustrate the RAE at young ages.
Method
In this study ages were calculated directly from a child's Malaysian IC number which has both birthdate and sex encoded into it. Subjects were then ranked by age within each single year grouping and divided into quarters based on age on the date of testing. Q1 included those born in January, February, and March and represented the youngest group; Q4 consisted of those born in October, November, and December and were the oldest in each cohort.
It should be noted that this analysis is based a child's actual age and is not relative to any artificial selection period. The idea remains the same however and it may be easier to understand the problem inherent with the RAE if the mental gymnastics of calculating ages based on rolling calendars can be eliminated by looking at actual ages and analyze children according to birth rank within the single-year cohort.
Means were calculated for height and 50 m run times. These means were then charted and the results can be seen in the figures below.
Results
The relative age effect is easily noticeable for 8-year-olds for both height and 50 m running times. Figures 1 and 2 show the average height of 8-year-old girls and boys. The tallest children are clustered in the Q3 and Q4 quarters; these are also the oldest children.
8-year-old girls showed an average difference in height of 2 cm between Q1 and Q4 while 8-year-old boys showed over 3 cm difference.
Figure 1: Mean height for 8-year-old girls
Figure 2: Mean height for 8-year-old boys
Figure 3 shows faster performance times in Q4 for 8-year-old girls and Figure 4 shows performance times for 8-year-old boys clustered in Q3 and Q4. For this group the effect is quite noticeable. In both boys and girls the fastest times were recorded for the oldest athletes.
Figures 3: Mean times for 8-year-old girls 50 m run
Figure 4: Mean performance times for 8-year-old boys in 50 m run
If you have the data and the time the relative age effect is very easy to show on a chart. And it is not unusual, it happens everywhere that groupings are based on chronological age.
The effect of the RAE can be extreme at young ages but lessens as children get older until it eventually disappears. But development that occurs in the younger years is the most critical and knowing that the relative age effect exists and is an a priori part of life means that coaches, administrators, and parents must try to mitigate its effects on young child athletes.
--------
Bill Price (price@ussa-my.com) is the Chief Information Officer at USSA Malaysia in Kuala Lumpur.