Sometimes, she was able to confirm the general wisdom: tobacco, for instance, is indeed dangerous for pregnant women
Emily Oster has been crusading for years against what she calls the onslaught of bad advice directed at new parents, chastising everyone from grandmothers to charlatans to health authorities in her first book about pregnancy. And there's more to come in her second book about early childhood.
"There are still many terrible studies being done," said the 38-year-old professor of economics at Brown University.
It all began out of frustration when, pregnant with her first daughter, she ran into a brick wall of universal recommendations by doctors, but was unable to find actual statistics about the risks of things like drinking coffee or undergoing amniocentesis.
Then a professor at the University of Chicago, who had trained in statistics while obtaining her PhD in economics from Harvard, Oster went back to the studies that had been done on the risks posed by alcohol, tobacco, deli meats, sushi and even gardening.
Sometimes, she was able to confirm the general wisdom: tobacco, for instance, is indeed dangerous for pregnant women.
But Oster more often rejected the common advice, or found it wasn't conclusive.
She found no evidence to prove that drinking coffee in moderation is dangerous. Eating sushi does expose one to the risk of salmonella, yes. But it is no riskier for a pregnant woman than for anyone else.
And gardening is to be avoided due to a parasite called toxoplasmosis, found in animal waste, which risks harming the developing fetus.
A passage in her book about drinking alcohol stoked plenty of controversy, but the economist was unmoved by the emotional responses.
Drinking a lot is dangerous for the fetus, she said, but no one had proven the same for light drinking, despite US and French recommendations for "zero alcohol while pregnant."
Oster's book, "Expecting Better: Why the Conventional Pregnancy Wisdom Is Wrong-and What You Really Need to Know," has sold 80,000 copies since 2014 and has been translated into Hebrew, Chinese, Japanese and Korean (the Japanese version cut a chapter on epidurals, which are very rare in that country).
Oster says pregnancy and child-raising are fields in which bad studies abound.
The number of participants in these studies is often too small, making generalizations impossible.
And rare are the randomized, controlled trials – the gold standard in scientific research because they allow scientists to isolate the effect of a single element or medication.
Even more, some doctors are unable to translate the latest findings into medical advice. For instance, why do many doctors continue to recommend bed rest, when the evidence shows it is useless, and possibly even harmful?
"It can be difficult for all of us, not just doctors, to use data over anecdotes or over our own experiences," said Oster.
Fear of lawsuits can also make some US doctors hyper-cautious.
Oster never says something is "true" or "false." But she often says, "there is no evidence," which can be a frustrating conclusion for parents on the hunt for certainty.
Her next book, due out in April 2019 and focused on children aged one to three, no offers no more reassurances.
Oster even takes on breastfeeding, a practice recommended by the World Health Organization, which touts its benefits for intelligence and against obesity.
In the United States, women are often encouraged to breastfeed as long as possible.
But Oster says she found just one rigorous study, large scale and randomized, comparing two groups of mothers in Belarus in the 1990s.
Half were encouraged to breastfeed, the other half were not.
Most studies compare the IQ of children who were breastfed to children who were not – but women who breastfeed are often wealthier, more educated and possess a higher IQ themselves, which could contribute to the difference.
"The trouble is that the evidence they are based on is often seriously biased by the fact that women who breastfeed are typically different from those who do not," said Oster.
"Breastfeeding rates differ dramatically across income, education and race."
To really uncover the effect of breastfeeding, a randomized trial is necessary.
What is clear from the Belarus trial, she says, is that mother's milk reduces diarrhea and eczema in the short term.
"But many of the long term benefits that people claim are probably well not supported in the evidence," such as on IQ and obesity.
Topics in her book include infant sleep issues, vaccines, disciplinary methods, educational philosophies and the impact of children on a couple's relationship.
She did not reveal her findings, but gave a broad outline of what readers can expect.
"More so than in the first book, it becomes really clear that the preferences of the family and what works for the family should influence a lot of the choices that you make," said Oster.
"Something that works for some people is not going to work for other people," she added.
"There's really not one recipe."
There is one exception, she says, for parents who are hesitant to praise their children too much – particularly little ones.
"It's fine, you can tell your baby that they're great as much as you want!"
At least until proven otherwise.