Birth order studies: some sources of bias.

The article investigates some sources of bias in birth order studies, which may distort the observed distribution of patients by birth-rank. Various biases stemmed when (1) there are biases in the sample due to biases in the population from which the sample is drawn, (2) there is the distortion introduced when sibships are not complete, (3) there are biases which the investigator may introduce in his manipulation of the data and (4) there is the correlation between birth order and the proportion of missing or unreliable information. Practical suggestions for the planning and reporting of birth order studies are presented, and a method is proposed for overcoming the problem of incomplete sibships.