Only a few generations ago, it made sense to ignore expert parenting advice. Most of it was nonsense.
In the early 20th century, parents were instructed by books and manuals to always keep their childrens’ heads pointed north, so as to somehow be in line with electrical currents traveling the globe.
Pregnant mothers were told to “avoid thinking of ugly people,” because their thoughts might somehow transform the appearance of the child.
On the advice of experts, parents fed their children kerosene and turpentine to cure croup and the common cold.
That wasn’t that long ago. There are still nearly half a million people alive today who were alive then (although they probably weren’t the kids drinking fuel oils and volatile solvents). But we’ve surely come a long way since then in terms of scientific standards, and the ability for anyone to quickly research and fact check information themselves.
Nevertheless, a long history of misguided recommendations and misleading media has created the perception that “parenting expert” is an utter misnomer. Again and again, supposedly professional advice on how to raise children has done nothing but lead parents astray, chasing dubious trends only to be turned 180 degrees a short while later.
“America’s parenting experts, in short, have fared no better or worse than the rest of us in the quest for calm consistency in child-rearing technique and theory,” writes Ann Hulbert in “Raising America: Experts, Parents, and a Century of Advice About Children.” “The story of the popular advisers’ search for clarity about children and for authority with mothers is marked by controversies, contradictions, and unintended consequences. Among the most ironic of those consequences has been to leave parents, teachers, policymakers, ministers and the media – to say nothing of the experts themselves – convinced that expert counsel is precisely what it was not supposed to be: constantly shifting and conflicting, throwing both grown people and children here and there like balls.”
Today, everyone knows someone, a friend, family member or coworker, who has said something along these lines: “I don’t listen to any of that expert advice. What’s the point? It changes all the time anyway.”
That attitude may be causing many parents to throw out their babies with the bathwater, so to speak. Recommendations for raising children do change often, but there are many legitimate, reasonable reasons that recommendations evolve. It’s not just haphazard and reckless.
The problem is, researchers and authorities often take for granted, or don’t appreciate the importance of, parents understanding the full context of why recommendations are made, and why they change. Breaking this down can help make what looks like nonsense start to make a lot more sense.
Why the science changes
First things first: science itself has changed dramatically in the very recent past. If it seems like researchers can’t make up their minds, a major reason is because modern standards have only recently evolved.
In the 1960s and ‘70s, when modern psychological studies were beginning to take off, the characteristics of a “good” study were much different than they are today. Consider Mischel & Ebbesen’s 1970 study, which, along with a series of similar experiments, became known as the “Marshmallow Test.” The researchers put two tiny cookies or pieces of pretzel (or sometimes marshmallows) in front of a child, asked the child to indicate their favorite, and then said “you can eat all of the [non-preferred snack] if you like, but I have to leave the room now and if you can wait until I come back, you can have the [preferred snack].” Then they timed how long the child was willing to wait to get the preferred snack.
The problems with this study are many, so we will list only some of them here:
- The study participants were 32 children – a tiny sample size by any definition.
- These children were what is called a “convenience sample” – the researchers worked at Stanford University, and the children all attended Stanford’s Bing Nursery School, and were thus the children of Stanford professors and graduate students – hardly racially or economically representative of the broader population.
- The age range of the children studied was quite large: 3 years, 6 months to 5 years, 8 months. A massive amount of cognitive development occurs in those two years that could have impacted the results; no analysis was conducted on this factor.
- Subsequent researchers found that a shift in the way the rewards are presented can impact the results: if one sticker is presented as the immediate reward, and then four others are added to the same pool to identify the delayed reward, 3-year-old children actually outperformed 4-year-old children on their desire to wait for five stickers instead of take one now. This means that apparent changes in cognitive processes may just be an artifact of the study design, an impact that was not considered by Mischel & Ebbesen.
- While we do still use laboratory-based experiments to try to understand behavior, it is possible that what we learn may not have “ecological validity” – in other words, people might behave differently in real life. Professor Angela Duckworth found that the ability to wait for sticker rewards in the lab were only weakly related to children’s performance in real-life delays.
We say all that not to pick on that particular study, but just to highlight that even this widely referenced research has some obvious flaws from today’s perspective. Today, there’s a greater focus in the scientific community to eliminate, or at least account for, these issues.
Then again, it’s not as if experiments done today can be executed flawlessly in every case.
In an ideal world, we might start by developing a hypothesis that we could test in the lab. If it works, we could apply it in naturalistic settings, and test if it works again. Then we could further test to see whether the effect holds over a period longer than the few weeks that typical studies normally run.
In reality, a variety of factors confound the goal to see whether an intervention really works.
For example, let’s say an early-career professor is interested in asking The Big Questions, but might only be able to get funding to study a very small, limited question. They might want to study a large, nationally-representative sample but the majority of parents and schools who signed up to participate in experiments at their institution are middle class and white. They still might want to claim they answered a Big Question, and those findings might be hailed as successful in secondary sources, but it isn’t until years later that further analysis finds those findings unconvincing.
“Too often, intervention studies are characterized by very small experimental samples, the lack of control groups, and the absence of appropriate statistical tests for whether attitude changes occur,” one meta-analysis found, in examining years of research on programs meant to counter racism in children. “Furthermore, when well-documented evidence of attitude change is presented, intervention effects are weak at best.”
This problem is compounded when we consider that journals want to publish new, exciting findings, which leaves us with a significant publication bias. That means that researchers finding “no effect” typically can’t get their study published, so following analyses may falsely conclude that most interventions do actually produce the desired outcomes when more often than not, they don’t.
Unfortunately (or fortunately, depending on your perspective), scientific findings do get reversed, in many different disciplines. Dr. Vinay Prasad and his colleagues conducted an analysis of around 1,300 articles in the New England Journal of Medicine and found that 77% of the articles discussing some kind of new treatment or therapy found that the new way was better. But 40% of the 360 articles that looked at something doctors were already doing found that the current practice was contradicted. (The study is behind a paywall but Dr. Prasad discusses his results in this short video.)
Dr. Prasad has argued elsewhere that we can reduce the number of reversals by conducting randomized controlled trials on new technologies and procedures before performing them routinely at scale. Randomized controlled trials are the gold standard in psychology research as well, with the added complications that we would need long-term studies that help us to understand whether an intervention holds over time and or washes out eventually, as well as whether an effect size is small but cheap to implement and thus interesting in spite of the small effect size (as may be the case regarding the impact of fantasy play on children’s executive function).
A true long-term study may take decades. But we honestly haven’t even been studying these issues with consistent rigor, for that long. Looked at this way, it’s completely expected that the science will change, for a long time to come.
Why the recommendations change … even if the science doesn’t
So, there aren’t nearly as many definitive answers as parents might think, and even some really solid work can be legitimately overturned over time. That’s merely the scientific method in action. But it’s still tempting to think that once we get good science, we’ll finally have everything we need.
Unfortunately, communication is a further fundamental problem that often makes parents more confused than they should be. Just like all science isn’t equal – there is proper science and junk – the same goes for how information is spread by different sources. There are some sources that offer honest, accurate and thorough information, and many others that are misleading, careless and generally unreliable.
These lesser sources are highly motivated to make it seem like their messages are new, different, and more important than anything else out there. Just like academic journals, even respected media outlets are essentially biased toward anything new – it’s called the news, after all – so anything that isn’t somehow surprising doesn’t become a story. As it relates to research and ideas about raising children, this creates the impression that it’s changing all the time, when in reality the large majority of reality hasn’t changed at all.
The problems with the parenting media alone take up a lengthy article. The problems with marketing targeted at parents take up a book. Suffice it to say here that parents should not take mainstream news and advertising at face value, and instead seek out high-quality sources of parenting information.
It’s tempting to instead simply say that you should talk to your doctor, but doctors often simply regurgitate recommendations from health authorities, and that advice can come with a number of caveats as well. That’s because, even if you have lots of good science to use, good science rarely leads to actionable, everlasting answers for everyone in the world. Society is always changing, and it’s not easy to offer simple, universal answers for an incredibly diverse populace.
For one, consistent answers require consistent questions, but there is an unfortunate discrepancy in the types of questions parents want answers to, and the ways science can provide these answers.
One of the most common parenting questions is “what’s the best way to get a child to sleep?,” but sleep is incredibly culturally and individually specific. It’s a simple question, but answering it at any large scale is incredibly complicated. Recommendations may change dramatically from place to place and time to time.
For instance, much of modern parenting advice on this topic revolves around getting a child to sleep on their own, but children have slept close to their parents for thousands of years. For most of human history, this would be a strange question. It’s still strange based on where you live. When told that German babies commonly sleep alone, a group of Nso mothers in Cameroon decried this arrangement as a form of maltreatment that would need immediate action, like sending a Nso woman to Germany to teach German mothers how to care for babies.
So what’s an authority to do? Often, they make educated guesses. In lieu of precise answers to specific questions, they extrapolate from what from what they do know to apply guidance to a wide range of questions that often don’t perfectly align with what experimenters have examined.
For instance, you may have heard your doctor recommend that, in order to encourage good sleep habits, you should ideally put your child down in their crib while they’re drowsy, but not fully asleep. The idea is that this will teach them to go to sleep on their own. But this specific recommendation is mostly based on opinion, not purely data.
“If you talk to specialists who see 12-, 15-, 18-month-olds with sleep problems … there is a huge overrepresentation of babies who never were given the chance to be put down on their own. Is that evidence? Well, sort of,” Dr. Joseph Hagan, an editor of the American Academy of Pediatrics’ (AAP) guidelines on sleep, was quoted in Slate.
Essentially, the AAP is trying to give parents answers to their questions, in lieu of having definitive proof. When science becomes clearer, those answers may change, because they weren’t ironclad to begin with. But parents don’t always realize the recommendations they get from doctors, which are based on guidelines from groups like the AAP, aren’t always set in stone.
Peanut butter is one famous recent example in which authorities seemed to abruptly change their recommendations in a relatively short span of time. In 2000, the AAP recommended that any children under age 3 avoid peanut butter, for fear of peanut allergies. Subsequent research suggested this was exactly the wrong direction – children who ate peanut butter more often actually had lower instances of peanut allergies – so the AAP reversed itself in 2008. Today, authorities state that introducing potential allergens like peanut butter can potentially be a good thing, but the science is continuing to evolve, and so will recommendations.
In other cases, recommendations change not because science has changed, but because of how parents reacted to the original recommendations, and how people in general are changing habits.
Take screen time. Before 2016, the AAP once offered straightforward limits: kids under age 2 should avoid screens of all kinds. Some parents liked that simplicity, but otherwise were confused. Did that apply to every type of screen? Even FaceTiming with a grandparent, or watching an educational show on a long car ride?
Now, the AAP has a more nuanced approach, acknowledging that hardly anyone seemed to be following the previous recommendations in our increasingly connected world.
“By moving away from the finger-wagging and diving deep into the shades of gray, the recommendations put more onus onto parents to decide what’s best for their families,” Laura Sanders summarized in the Growth Curve section of Science News.
That entire dynamic makes parenting recommendations an art as much as a science. And when you’re dealing with art, all kinds of preferences and biases come into play.
The AAP tends to be risk averse. For example, the AAP’s latest guidance on Safe Infant Sleeping Environments finally acknowledges that co-sleeping does happen, but implies that it only happens accidentally, and recommends that “if the parent falls asleep while feeding the infant in bed, the infant should be placed back on a separate sleep surface as soon as the parent awakens.”
A more nuanced look at this information would acknowledge that co-sleeping is a lot less dangerous in countries where everyone in the family sleeps on hard beds with light coverings or no coverings at all, and that when practiced safely, co-sleeping is unlikely to increase the risk of SIDS. Instead, the AAP’s “reduce risk at all costs” gets translated into advertising campaigns likening co-sleeping to putting your baby to bed with a cleaver.
But even if health authorities and scientists have nuanced recommendations, that all goes out the window in the hands of lawmakers and administrators. Elected officials and leaders of all kinds often awkwardly apply science in ways it wasn’t intended, leading parents to blame the science itself.
Take, for example, Dr. Carol Dweck’s work on Growth Mindset – the idea that you can improve your intelligence and skills through hard work, rather than accepting that these are fixed abilities. Schools in California are piloting a program to evaluate teachers and schools on their students’ growth mindset, despite Dr. Dweck’s insistence that this is not an appropriate use of the concept. One Master’s student describes how these programs are sometimes implemented: administrators describe the new strategies that teachers must use, with no opportunity for discussion, one week before classes were scheduled to begin, when teachers had already planned their work for the semester, with no support to integrate the new curriculum and the one teachers had already developed. All this, when the research points toward growth mindset being a useful tool for some children, but that implementing this on a large scale by teachers in classrooms is extremely problematic.
Understanding all the dynamics at work in these situations is often far too much for a normal parent to wrap their mind around.
What can parents do in the face of all this change?
So, science changes, society changes, authorities change the philosophy behind their recommendations, and dubious experts and uncareful decision-makers distort recommendations. How are parents supposed to keep track of all this?
Well, we won’t prevent all reversals of advice related to child psychology and parenting. There simply aren’t enough organizations willing to fund the type of long-term, randomized controlled trials it would take to be sure that an intervention “worked” for us to do this for every topic we want to understand. Even if we could, we wouldn’t want to wait the extra decade or two it would take to do these studies before trying to implement the results.
And in the time we were waiting, our culture would likely have shifted in that time anyway – much as we are now shifting away from the overparenting of children (which can cause higher levels of depression and less satisfaction with life in college students) and (slowly) toward an increased acceptance of risky play.
Your best bet are sources that tell you how they judge the quality of their sources, so you can judge for yourself whether you agree with these criteria. Both the Your Parenting Mojo podcast and Parentifact website tell you about small sample sizes, samples that are not representative of the general population, and implicit biases in the way questions are asked or instructions are framed that could impact the validity or applicability of the results in the work we cite. Not many sources will do this for you – mainly because it takes so long to go back to the original sources and check through them individually.
It’s also generally helpful to recognize that there are so many variables influencing a child’s outcomes that it is highly unlikely that any single one of them alone will have a meaningful impact – from predetermined genetic factors through micro-scale factors like the amount of money your family earns and whether your neighborhood is safe to macro-scale factors like national policies on maternity leave and free childcare to psychological interventions related to growth mindset, grit, and the ability to delay gratification – all of these have some impact, and I could probably find a psychologist to argue that each one of these factors is the most important of all of them. In reality, they all exert some influence, and some of them are out of your control.
That’s why the best place to be may be the middle ground between ignoring all expert advice and constantly bouncing back and forth between the latest sensationalist headlines. Becoming so jaded that you refuse to listen to those people working tirelessly to discover the truth will likely make life harder than it needs to be. Feverishly trying to keep track of every piece of advice, no matter the source, will also likely hurt more than help.
In between those extremes is understanding that good science, validated by multiple studies, is worth listening to, even though it might change in the future. It’s cultivating high-quality sources of information, but not even taking those at face value. It’s understanding that experts and authorities have agendas, which may be legitimate, but don’t necessarily align with every individual’s situation or philosophy.
Ultimately, parenting advice may seem to change extremely often, but the reality is parents don’t need to.
Also published on Medium.