|
year ago, I wrote an op-ed
for the Wall Street Journal exposing a case of egregious
political bias at Harvard University Press. In an unusual move,
the board of Harvard Press declined to publish The
Case For Marriage — a lively, rigorous, and path-breaking
study of the advantages of marriage coauthored by respected University
of Chicago sociologist Linda Waite and syndicated columnist Maggie
Gallagher. Although The Case for Marriage had garnered two
positive reviews from Harvard's own scholarly referees, the Press's
Board of Syndics rejected the book at the last minute on the grounds
that Waite and Gallagher had failed to prove a causal relationship
between marriage and the many benefits that they claimed for the
institution.
Harvard's stated
reasons for rejecting The Case for Marriage were utterly
unconvincing. For one thing, Harvard had already published feminist
tracts with scandalously thin empirical grounding by Catherine MacKinnon
and Carol Gilligan. (The shaky empirical foundations of Gilligan's
In
A Different Voice were exposed, to considerable public attention,
by Christina Hoff Sommers's book The
War Against Boys.) So why not publish Waite and Gallagher's
extraordinarily well-researched study? Was this a case of bias against
a book that challenged feminist orthodoxy by showing the unique
advantages of marriage? You bet it was. (For more on bias at Harvard
Press, see "Harvard's
Book Problem.") But now, a spectacular new piece of research
has provided stunning vindication for the Waite-Gallagher thesis
on the benefits of marriage.
What the Harvard
Press board was asking Waite and Gallagher to do was next to impossible.
The only sure way to make causal judgments about the effects of
marriage would be to run a controlled experiment, randomly assigning
young people to marriage and singlehood and then following their
progress throughout life. But human beings are not guinea pigs.
That is why even the very best sociological research generally fails
to provide concrete causal proof. Since we cannot easily run controlled
experiments on real human beings, we generally have to make causal
judgments through inference. Harvard was obviously holding The
Case for Marriage to an impossible standard — and a double
standard — in order to suppress the book.
Yet, lo and
behold, as reported in the Washington Post, a year later,
two creative researchers, Donna Ginther and Madeline Zavodny of
the Federal Reserve Bank of Atlanta, have actually found a way to
do the seemingly impossible. For the first time in the history of
research on marriage, Ginther and Zovodny appear to have successfully
shown that the "marriage premium" — the tendency
of married individuals to make more money than single people —
is an actual effect of marriage, and not just a function of a preference
shown by both employers and potential spouses for people with qualities
likely to bring about success.
Ginther and
Zavodny pulled of this neat little trick by studying "shotgun
weddings" — marriages that took place after the woman
was already pregnant. Ginther and Zovodny reasoned that couples
marrying under such pressured circumstances were likely to include
many individuals who might not otherwise marry. If the men in these
"shotgun" marriages ended up with greater income than
single men with the same sort of background, then the "marriage
premium" would be real, and not simply the result of a "selection
effect." (The marriage premium applies to working wives as
well, but Ginther and Zavodny only studied men.)
It turns out
that even men stampeded into marriage by a pregnancy earn about
16 percent more than single guys. Almost 90 percent of the marriage
premium remains, even for a group in which selectivity has been
substantially short-circuited by the advent of a pregnancy. Of course,
even here, a degree of selection bias is bound to exist. Not every
couple marries when there is a pregnancy, and it's reasonable to
suppose that those who are already most suited to each other, and
to marriage itself, are more likely to marry on the discovery of
a pregnancy. But the white men in the "shotgun" group
earned less, were younger, and had less education than other white
men getting married. Clearly, these men were less desirable as husbands,
and the marriages were substantially precipitated by the pregnancies.
Yet the marriage premium remained. So Ginther and Zavodny appear
to have found the "holy grail" of sociological research
on the effects of marriage — a way to eliminate selection bias
and provide causal proof of marriage's beneficial effects.
When I called
Linda Waite for comment on the Ginther and Zavodny study, she was
obviously excited. Waite called the study, "pretty amazing,"
and characterized the results "powerful evidence for a causal
effect over selectivity." According to Waite, it was "almost
shocking" that a full 90 percent of the marriage premium remained
in effect for the "shotgun" couples.
Why the marriage
premium? The mutual advice, emotional support, and concrete help
that married partners provide to one another seems to free up and
strengthen both husbands and wives to succeed at what they do. When
married women work, they make more money than single women. When
married women mother, on the other hand, their personal financial
premium disappears. Yet mothers benefit from something far more
valuable — the support and protection of a husband who himself
seems to strive (and succeed) that much more for the sake of —
and with the help of — his wife and child.
Now that this
important causal evidence in support of Waite and Gallagher's The
Case for Marriage has emerged, their erstwhile feminist critics
will no doubt fall over themselves in the rush to retract their
skeptical attacks. And with causal proof at last secure, surely
Harvard University Press will offer to publish the sequel to the
rejected book, just as Harvard Press has continued to publish book
after book by Catherine MacKinnon and Carol Gilligan. At least,
that's what would happen if Harvard Press is motivated — as
they say they are — not by ideology, but simply by the highest
standards of scholarship.
|
