EXPERIMENTAL DESIGN OF THE CHINA STUDY 357- In each of the sixty-five counties in the study, two villages were
selected for the collection of the information. Having two villages
in each county rather than one gives a more reliable county av-
erage. When the values of two villages are more similar to each
other than to all the other counties, then this means higher-quality
data.^3 - When possible, variables were measured by more than one kind
of method. For example, iron status was measured in six different
ways, riboflavin (vitamin B 2 ) in three ways, and so forth. Also, in
many cases, we could assess the quality and reliability of data by
comparing variables known to have plausible biological relation-
ships. - The populations under study proved to be very stable. An average
of 93-94% of the men in the survey were born in the county where
they lived at the time of the survey; for women it was 89%. Also,
according to data published by the World Bank,4 the diets at the
time of our survey were very similar to those consumed in earlier
years. This was ideal because those earlier years represented the
time when the diseases were initially forming.
UNIQUENESS OF DATA
One idea that makes our study unique is our use of the ecologic study
design. Critics of the ecologic study design correctly assume that it is a
weak design for determining cause-and-effect associations when one is
interested in the effects of Single causes acting on Single outcomes. But
this is not the way that nutrition works. Rather, nutrition causes or pre-
vents disease by multiple nutrients and other chemicals acting together,
as in foods. An ecologic study is almost ideal if we wish to learn how
an array of dietary factors act together to cause disease. It is the com-
prehenSive effects of nutrients and other factors on disease occurrence
where the most important lessons are to be learned. To investigate these
comprehensive causes of disease, it was therefore necessary to record
as many dietary and other lifestyle factors as pOSSible, then formulate
hypotheses and interpret data that represent comprehenSiveness.
Perhaps the most unique characteristic that set this study apart con-
cerned the nutritional characteristics of the diets consumed in rural
China. Virtually every other human study on diet and health, of what-
ever deSign, has involved subjects who were consuming a rich Western