The relation of measured intelligence to birth order and maternal age.

One of the difficulties in interpreting the relation between fertility and measured intelligence is the possible influence of birth order. Penrose (1963), among others, has recognized that a negative correlation between sibship size and intelligence would result if birth order and intelligence were also negatively correlated. For various reasons most observations on birth order and intelligence have been inconsistent or inconclusive. No conclusion can be drawn from comparisons (first with second, second with third born, etc.) between school-children, for since children in low birth ranks are likely to come from small families, it is unknown whether observed differences are due to the ordinal position or to the size of family identified by it. Attempts have been made to overcome this difficulty by comparing children from families of similar size, for example, first born from two child families with second born from two child families. But at the time-usually age II-when school-children are examined, many families are incomplete, and in the example quoted the second born is more likely to come from a completed and smaller family (since 11 years have passed without the birth of a surviving sib) than the first born (who has a sib born in the 11-yearperiod). So again it is uncertain whether an observed difference is attributable to ordinal position or family size. This difficulty was recognized in the Report on the Trend of Scottish Intelligence by the Scottish Council for Research in Education (1949). It would seem possible to assess the significance of birth order by examination of intelligence of children within the same sibships. Some investigations of this type have been based on small numbers or selected material, or have failed to correct adequately for age differences between sibs at the time of testing. Unfortunately age correction raises problems noted by Roberts (1947), and it is undoubtedly better to avoid them by a standard test of sibs as they attain the same age. This was the method used in the 1947 Scottish survey, and from examination of 1036 sib pairs it was concluded that there was no evidence of a relationship between measured intelligence and birth order. In most investigations of intelligence and birth order little attention has been given to the possible significance of maternal age, and results from the few reports (summarized by Locke & Goldstein, 1937) are inconsistent. Roberts (1947) concluded that age was unimportant, but a recent investigation of 11,280 children born in Aberdeen showed that test scores increase with increasing mother's age and decrease with increasing birth rank (Illsley, 1967). This result raises the familiar difficulties of interpretation in population studies where differences within families cannot be separated from differences between families. Our own data also comprise intelligence scores for a population of children identified at birth. Numbers are large enough to provide a substantial number of fraternities with two or more