A case-control study was carried out to investigate many of the proposed causes of glue ear in childhood. One hundred and fifty cases with two matched controls were found to be remarkably similar in nearly all medical and social aspects of their past and present lives, thus providing no support for many of the currently held views on the aetiology of glue ear. Of the 5 factors which were found to increase the risk of a child undergoing surgery for glue ear, only one of these is thought to be related to the development of the condition, rather than to the chances of its detection. This factor was parental smoking (RR 1.64). The 4 other risk factors appear to influence the chance of glue ear being detected, diagnosed and referred for surgical treatment - the child's mother being employed outside the home, but only if the father is employed in non-manual work (RR 3.0); attending pre-school day-care (RR 2.00); having an older sibling who had been diagnosed as suffering from glue ear (RR 1.84); and having been born locally (in Oxfordshire) (RR 1.89). Possible explanations for these social and behavioural factors are discussed.