Biology Reference
In-Depth Information
If we read between the lines, the quotation seems to suggest that even though the
new empirical result does not support the hepatitis B hypothesis, the hepatitis
hypothesis should not be rejected immediately; economists would rather at one
stage hang on to the refuted hypothesis and try to find a way to reconcile the old
findings with the new ones. In this sense, we might define two methodological views
of science. On the one hand, science is traditionally regarded as being constituted by
a great number of theories. The aim of each theory is to provide an explanation for a
phenomenon in question. Each theory in turn contains a number of hypotheses, each
of which posits a general law that is supposed to govern or regulate a corresponding
part of the targeted phenomenon if that part can be logically derived from the general
law. Once all the relevant hypotheses—that is, the relevant regularity laws—of the
theory can be used to derive the corresponding parts of the phenomenon, the theory
is said to provide an explanation of the phenomenon. When a new phenomenon of a
similar kind can no longer be derived from—or explained by—the same theory
(the same set of regularity laws), then the theory and its component laws should be
substituted or replaced by other new theories and their component new laws.
We may call the idea that involves the description of scientists' practices the
substitutive conception of scientific practices. This is Popperian in spirit.
On the other hand, our case study shows that the substitutive framework does not
necessarily appear in economists' practices. The empirical result of Lin and Luoh is
thought to reject the hepatitis B hypothesis by their findings of little effect of HBV on
the sex ratio at birth and a significant correlation between higher birth order and the
sex ratio at birth. Yet we also witness the so-called complementary conception of
scientific practice as followed in Oster's work: when scientists want to check
whether there is a significant correlation between any two targeted events, they
usually apply empirical tools, such as regression analysis, to run a significance test.
When they find that the estimate that represents the correlation does indeed fall
within the prescribed confidence interval, they accept the correlation hypothesis and
confirm that there is a correlation between the two events. In contrast, if the estimate
falls outside the confidence interval, it is normally presumed that such a hypothesis
should be rejected and no correlation is acknowledged. Yet in practice it is intended
to append additional conditions to the failed hypothesis to explain why the hypothe-
sis conflicts with the testing result. In other words, some explanation is provided to
reconcile what is stated in the hypothesis and the contradictory testing result. In fact,
Oster herself used the term complementary that fits precisely into our observation.
In her reply to Das Gupta's comments, Oster stated (Oster 2006 , pp. 325-326),
The key to thinking about the relative potential of culture and biology to explain the over-
representation of men in a population is understanding that marginal effects may be seen to
operate and still tell us relatively little about the average. In the end, it seems better to think
of these two explanations as complementary . The issue of gender imbalance in Asia—the
causes and consequences—is an important one; we should endeavor to have a complete
understanding, not just a partial one. (our emphasis)
In order to have a complete understanding, it needs first to understand the fact
that the observed correlation between any two targeted events is in fact the net
result of a complicated interaction among a great many relevant factors involved in
Search WWH ::




Custom Search