This Sporting Life

Football position and atopy—both subject to the birth order effect?

BMJ 2003; 327 doi: (Published 18 December 2003) Cite this as: BMJ 2003;327:1473
  1. Michael R Perkin, clinical lecturer (m.perkin{at}
  1. Department of Child Health, St George's Hospital Medical School, London SW17 0RE

    In 1989 Strachan published a paper observing that the prevalence of hay fever varied with family size.1 He proposed that allergic diseases may be prevented by infection in early childhood, transmitted by unhygienic contact with older siblings. This protection from atopy is also related to order of birth; younger siblings are less likely to be atopic.2 To explain this observation Strachan suggested that scientists should be seeking other influences of development, lifestyle, or environment that vary strongly with birth order.2

    With this in mind, I (the author), as the father of three young boys, was concerned that, when it came to playing football, the youngest was always going to be nominated the goalkeeper and the eldest the striker. This led to my primary hypothesis that perhaps not only atopy varied with birth order but also football position, with goalkeepers being more likely to be the youngest in the family and forwards the eldest (fig 1). The secondary, politically incorrect hypothesis, was that perhaps it was birth order among male siblings that was the deciding factor.

    fig 1
    fig 1

    Pele (inside left forward, now called striker) fits the birth order hypothesis (eldest of three children)

    Credit: EMPICS

    Participants, methods, and results

    A questionnaire went to all premiership clubs, explaining the underlying hypothesis and asking each player about the number of younger and older siblings and their sex. The stresses of playing in the premiership were clearly too great because only three of the 20 teams replied. Two refused to participate, and one (Middlesbrough) said it would ask its players, but I heard nothing further. Intriguingly, one of the refusers (Sunderland) wrote that “Due to the volume of requests we get on a daily basis similar to yours we are not able to help on this occasion.” Changing tactics, I wrote to all 24 clubs in Nationwide Division Three. Fourteen clubs replied, supplying sibling details of 232 players (23 goalkeepers, 72 defenders, 68 midfielders, and 69 forwards).

    The mean family size varied significantly by football position (mean number of siblings of goalkeeper 1.13 (95% confidence interval 0.78 to 1.48), defender 1.79 (1.47 to 2.11), midfield 2.40 (1.95 to 2.84), forward 2.00 (1.53 to 2.47)). One explanation for goalkeepers being from smaller families may be that in such families individual children spend longer in goal, whereas children in bigger families can rotate through other positions (fig 2).

    fig 2
    fig 2

    David Seaman (goalkeeper) matches the family size observation (one brother)

    Credit: EMPICS

    Overall 68 players were eldest children, 54 middle siblings, and 110 youngest children. For the statistical analysis I treated each position in turn as a dichotomous outcome (for example, goalkeeper or any other position) in a logistic model with family position (eldest, middle, or youngest sibling) included as the dependent variable. I used being the youngest sibling as the baseline group (odds ratio 1.0) as this was the most prevalent family position among the football players. I adjusted family position for family size. Twelve players were only children and were excluded from the “all siblings” analysis. For the analysis that was restricted to birth order of male siblings only, I excluded 82 football players who were only children.

    No trend became obvious with goalkeepers, defenders, and midfield players (table). There was an indication that forwards were less likely to be eldest siblings, which was stronger among male siblings only, but the relationship, to use football parlance, missed statistical significance by some way.

    Logistic analysis of football position and birth position (all siblings and male siblings only), adjusted for family size

    View this table:


    The results tend towards the opposite of the primary hypothesis in that forwards were less likely, not more likely, to be eldest siblings. In contrast, perhaps the secondary hypothesis, that birth order among siblings decides football position, has some foundation in that the relation seemed stronger when male siblings alone were considered. Overall, with broad confidence intervals, it seems safe to conclude that male siblings should not be dissuaded from adopting any football position that appeals to them.


    • Contributors MRP is the sole author.

    • Competing interests None declared.

    • Ethical approval Not required.


    View Abstract

    Sign in

    Log in through your institution

    Free trial

    Register for a free trial to to receive unlimited access to all content on for 14 days.
    Sign up for a free trial