Sir.—Given the recent emphasis of AJDC on the need for accurate statistical analysis and understanding in the presentation of scientific data,1,2I was most surprised to read the recent article by Shepherd et al,3 which purports to demonstrate a causal relationship between nuchal cords and neonatal anemia.
The authors' basic methodology is to compare the hematologic values of term infants with nuchal cords to a supposed control group. Unfortunately, the study is deficient for a variety of methodological reasons:
The criteria for selection of the "normal" control group are obscure. Apparently, both sick and well infants were included (in the authors' own words, some underwent "medical workup").
Factors such as birth weight, Apgar score, and incidence of small-for-gestational-age or large-for-gestational-age infants—all factors that affect neonatal hematologic values—are not mentioned. Thus, one cannot assume that the groups were matched appropriately to study the independent variable, ie,