Explaining the Explanations of 100 Million Missing Women - Mechanism and Causality in Biology and Economics - page 225

This description seems to be further justified by Oster's remarks about the idea of
science:
I'd be lying if I told you it wouldn't be great if I was right all the time If you work like

this, especially if it's something that people care about, and you get to collect some more
data that is maybe going to be even more informative than what you had before, it's your
responsibility to do that. This is the way science works. 5
On the face of it, Oster's statement might be regarded as representing two
methodological points. One is the doctrine that follows the tradition of Popperian
falsificationism; the other is that the size of data does matter. The aim of science,
including economics, is to identify stable connections—or regularity laws —among
variables (or factors) of interest so that, by using these connections or laws, scientists
are able to explain the phenomena that are thought to be governed by the connections
or laws. If, in the process of hypothesis testing, the scientists somehow find out that
there are no stable connections or lawlike relations among the targeted variables, the
correct action is for scientists to repudiate their hypotheses about those connections
or laws. In our case, the hepatitis B hypothesis is that there is a positive connection
between the prevalence of hepatitis B among mothers of newborns and the increased
ratio of newborn males to females and thus the number of missing women. In her
2005 paper, Oster, by using various quantitative strategies and a larger dataset than
Blumberg's, concluded that there is indeed a robust connection between hepatitis B
and the missing women, so she suggested that hepatitis B could explain the phe-
nomenon. Later, Lin and Luoh's study of three million newborns rejected the
hepatitis B hypothesis. In order to respond to Lin and Luoh, Oster et al. ( 2010 )
used new empirical data and found no effect of hepatitis B carrier status on the sex
ratio. This finding led her to reject the hepatitis B hypothesis and claim that hepatitis
B does not explain male-biased sex ratios in China. Oster's practice of rejecting her
previous hypothesis thus illustrates that her action following her doctrine of science.
However, if we interpret the missing-women debate in this methodological
sense, then the following short paragraph, which is quoted from the 2008 working
paper version of Oster et al. ( 2010 ), may seem puzzling (Oster et al. 2008 , p. 6):

...

[After showing] that hepatitis B carrier rates cannot explain male-biased sex ratios or the
'missing women' in China ... [a]n important remaining issue is whether it is possible to
reconcile the biological results in the original paper (Oster 2005 ) with these [2008] results
and, in particular, how the individual-level data from outside of China and the evidence
from vaccination campaigns in Alaska can coexist with new results from China ... We re-
visit the original individual-level data from Greece and the Philippines and continue to find
support for the connection between paternal hepatitis B carrier status and offspring sex
ratio. Moreover, in the data from China discussed here, we also find some interaction
between hepatitis B, gender and fertility: women with the hepatitis B e antigen (carriers
who are also replicating an additional viral antigen) seem to have fewer male children.
Further, women who are carriers of the virus have fewer children overall, even with
extensive controls. Together, this evidence suggests that there may still be some interaction
between hepatitis B and fertility outcomes (in general) but that clearly the pathways are
much more complicated than the simple carrier-male offspring connection.

5 Quoted in Justin Lahart, “Economist Scraps Hepatitis Theory on China's 'Missing Women,'”
The Wall Street Journal , May 22, 2008 .