The purpose of this study is to compare obstetric and birthweight data of twin children from the two largest databases in Japan to estimate the difference between them in sample collection. The first group consisted of 1131 twin-pair school applicants, and the second group consisted of members' children from several maternal associations devoted to twins and included 951 pairs. All data were gathered by questionnaire. The mean birth years of the twins in these two databases were 1979 and 1995 respectively. The percentage of mothers treated with ovulation-stimulating drugs or in-vitro fertilization was markedly higher in the maternal associations group. Gestational age was around 1 week less in the maternal associations group, whereas birthweight according to gestational weeks and intrapair relative birth weight difference as a percentage according to zygosity showed little difference between both groups. We conclude that the obstetric and birthweight feature data from both groups should be considered to construct twin growth charts based on the methods of sample selection.