June 3, 2013 | 20
Last week, the anchors at Fox News made headlines when they covered the recent Pew Research Center finding that 40% of all households in America have a female primary breadwinner. About 1/3 of these households consist of two-parent households where the mothers make more money than their husbands, and the remaining 2/3 consist of single mothers. Left completely in a tizzy after learning of this statistic, Fox assembled an entirely-male panel to mourn, in no uncertain terms, the downfall of society as we know it. This was quickly followed by another segment in which lead mouthpiece Erickson doubled down on his statements, and a third in which female anchor Megyn Kelly criticized the men for their backwards, offensive views. Many of the things said on these segments were hurtful to thousands of men, women, and children who were essentially being told that they are responsible for society crumbling into ruins. However, in addition to being offensive, many of the “facts” and much of the “science” on these videos were not quite correct. In a series of blog posts this week, I will be comparing what the Fox news anchors have claimed in these segments with empirical reality.
The Claim: In response to Megyn Kelly citing a 2010 study released by the American Psychological Association stating that there are no differences in positive outcomes between children whose mothers work and those whose mothers stay home, Erickson says he tends to “discredit” the data because the samples are largely “self-selective” and higher income. According to Erickson, “when you study higher-income families, you’re absolutely right, working mothers who are very high income, their children, there really isn’t a big difference…but when you go into the middle class, where a lot of these issues are bubbling up, when you have a Mom who’s working 12 hours a day and a Dad who’s working 12 hours a day, and they come home and they’re also trying to be good parents, you can’t have it all, and they’re making compromises.”
The Reality: If Erickson had read the 2010 paper cited by Kelly, he would have found that the actual data tell a completely different story. In fact, they tell the opposite story.1
As Kelly mentions, the paper being discussed is a meta-analysis spanning 69 studies including 128,738 children over 50 years. You can access the paper itself by clicking the link here.
To understand how this study gets its data, it’s important to understand exactly what a meta-analysis is. Imagine you were to sit down and write a literature review on a topic that interests you — say, the effects of moms working outside of the home on their children’s later development. If you wanted to look at this qualitatively, you could search a bunch of databases and find every single relevant study conducted over the past 50 years or so. You could then write up a nice report summarizing what each of these papers found, somewhat like a scientific book report.
But this could get a little complicated, especially if you started to wonder how you should treat each of the studies. For example, what if one study sampled 50,000 families in different states all across the country, and the second study only sampled 100 families in one specific college town? Should you treat both of these studies equally? Probably not. But how should you treat them unequally? Should you just write all of the details for every study in the report and allow people to form their own judgments? That might get tedious, especially if you have to do that for hundreds of samples.
So a meta-analysis is basically a quantitative method of addressing this problem in a systematic way. It’s something like combining this idea of a “scientific book report” together with the concept of a weighted average. It takes all of the effects found in every single study that you’ve gathered and averages them together so you get one, final “THIS IS YOUR EFFECT” measure of the relationship that you’re interested in studying — but it goes one step further and weights each effect differently before it gets entered into the average, based on how many people are in the sample and how large (or small) the variance in the sample happens to be. So, in our example above, if the first study found a correlation of .6 and the second study found a correlation of .2, an average of these two effects would give you an “overall correlation” of .4. But, because the first study was so much bigger, that correlation of .6 might get a little more weight in your average — so your “overall correlation” would take that into account and might be something more like .45 or .46. Now imagine extending this logic over dozens or hundreds of studies, weighting each one to account for differences in sample size, variance, etc. This is the bare bones of how a meta-analysis works. It’s like a literature review with numbers — or a weighted average of the effects in every single study that you found — that treats the effects that you get from better-powered studies more seriously.
So what do the samples involved in this particular meta-analysis look like? Are they all wealthy, Caucasian, two-parent families? Not quite. Although a little over half of the samples were categorized as “mixed SES,” 30% of the samples in these studies were primarily working/lower-middle class, and only about 15% of the samples were primarily middle/upper-class. Almost 20% of the samples came strictly from single-mother households, and about 20% of the samples consisted of families on welfare. With regards to racial diversity, approximately 20% of the samples were primarily Black or African-American, and approximately 10% of the samples were primarily Hispanic. So, while these numbers could be higher, it’s certainly not the case that these studies are only recruiting wealthy families — there’s a fair representation of economic and racial diversity.
Now we can see that there are plenty of lower-income families in this sample. So does Erickson have a point? Is it the case that the kids are doing OK in wealthy families, but struggling if they have working moms and they’re not as financially stable?
One of the real benefits of doing a meta-analysis is that you can actually look for important moderators of your effect of interest. So, let’s say you have 20 studies that look at the effects of maternal employment on child IQ in families that are on welfare, and 20 studies that look at the effects of maternal employment on child IQ in families that are super wealthy. Long story short, you can basically sort those studies into groups that you want to compare (i.e., “Welfare Families” vs. “Wealthy Families”) and then run some statistical tests that will let you know if the effects in those two groups are significantly different from each other. So, if maternal employment is really bad for one type of family but not so bad for another, this moderator analysis will let you know that information.
The authors of this meta-analysis — knowing full well that the effects of maternal employment on children might differ based on socioeconomic status, as Erickson insinuates — actually did compare the studies based on whether or not the majority of the families involved in the study were on welfare. If Erickson were correct, children from families on welfare (aka the poorest families) should be struggling the most, so you would expect to see lower levels of achievement and more behavioral problems if mothers in those families work while their children are young, whereas those from higher-income households should be showing no real effect.
However, when looking at samples where the families were on welfare, children whose mothers worked while they were very young (1-3 years old) actually performed significantly better on measures of overall achievement and had significantly higher IQs , although there were no differences when it came to performance on formal achievement tests. On the contrary, when looking at samples where the families were not on welfare, there were no differences in overall achievement or IQ between the children whose mothers worked and did not work during their early childhood years, although higher SES children whose mothers worked while they were young actually did slightly worse on formal achievement tests.
What if we look at whether or not the child is coming from a single-parent household? Same story. Children who lived with single mothers performed better on measures of overall achievement and IQ if these single moms worked while the kids were very young. Children who lived in two-parent households, on the other hand, showed no differences in overall achievement or IQ, but did worse on formal achievement tests if their mothers had worked.
And what about behavioral problems, like externalizing behaviors (aggression or impulsivity) or internalizing behaviors (depression or anxiety)? After all, if lower-income children whose parents work outside the home have higher IQs but also have higher rates of depression and anxiety, that’s still a problem, right?
Sure, it would be a problem — if that were the case. But it’s not. Once again, the pattern is the same. Children who lived with single mothers who had worked outside of the home while the kids were very young actually exhibited significantly lower rates of overall behavior problems, significantly lower rates of aggression and impulsivity, and marginally lower rates of depression and anxiety. Children from two-parent households showed no such difference in overall behavior problems, aggression, or impulsivity, though they also showed lower rates of depression and anxiety. So, across the board, when mothers worked outside of the home where their babies were very young, it didn’t matter if they were single mothers or members of a two-parent household. Looking across a wide variety of racial and socioeconomic groups, studies either found no relation between employment and behavioral problems, or they found that children whose mothers worked while they were young actually had fewer behavioral problems and better academic outcomes than their counterparts whose mothers stayed at home.
The data keep telling the same story, no matter how you look at it. According to the data (which, again, covers over 100,000 children across almost 70 different studies conducted over 50 years), having a mother who works outside of the home, if anything, might actually be MORE BENEFICIAL for children from lower-income families or single-mother households than having their mothers stay at home. For children from wealthier families, there is either no difference between the children of working and stay-at-home Moms, or the children of stay-at-home mothers fare a bit better.
The authors have a painfully simple and obvious explanation for this difference, which makes sense as soon as you read it. Erickson assumed that children from higher-income families will probably do fine in the world regardless of whether or not their mothers are employed, but children from lower-income families will suffer if both parents are working long hours and then coming home stressed. After all, as Erickson notes, you “can’t have it all,” and mothers who work are “making compromises.”
But the problem here is that Erickson is forgetting one very crucial thing — if there’s anything more stressful than being a single mother with a full-time job, it’s being a single mother without a full-time job. As the authors note, the “added financial security and health benefits that accompany [maternal] employment [in lower-income households]…improved food, clothing, and shelter because of increased income…[and] the psychological importance of having a role model for achievement and responsible behavior” are all really important things for vulnerable, lower-income children to have in their lives. If Mom is not outside the home working, sure, she is staying at home with her child — but if that family has little to no income coming in from other sources, that child is really going to suffer. Mom can only do so much to raise that kid well if he/she is starving or doesn’t have warm clothing in the winter. The added financial security from having a working mother more than makes up for the detriments of not having Mom around all day. Turns out, it’s more important to have food, clothing, and a roof over your head than it is for a two-year-old to avoid the perils of daycare. For wealthier families, on the other hand, Mom often returns back to work because she wants to, not necessarily because she has to. Although this is certainly not always the case, higher-income households likely have more financial security than their lower-income counterparts, regardless of Mom’s employment. Because of this, however, Mom returning back to work likely doesn’t raise the family income level so much that it makes a meaningful difference in the child’s quality of food, clothing, and shelter. It might mean that the child can get a few extra toys at Christmas, but it certainly doesn’t mean the difference between going to bed full or hungry. The increases in family income from the wealthier Mom working and the provision of a positive female role model in the household don’t always offset the challenges to bonding and attachment that maternal employment might also pose. For the lower-income Moms, on the other hand, that source of income is so crucially important, and it is so potentially devastating to a family if that income source is removed, that it certainly better for the kid’s future prospects if Mom goes back to work.
So, there don’t seem to be strong differences in children of stay-at-home and working mothers. When there are differences, they actually trend towards the children of working mothers having more positive academic and behavioral outcomes, and these positive benefits are especially noticeable in lower-income households, where the added financial security provided by Mom’s income does a whole lot to improve a child’s prospects.
1. This is actually a great example of why funding for the behavioral sciences is so critically important, and you can’t simply argue that it’s all just an exercise in proving “common sense.” When Erickson first said this, I had not read this paper and I’m somewhat embarrassed to say that I completely bought the idea that this effect was what they had found. I even wrote out a long argument in a comment that I posted on Facebook about why you would expect to see negative effects of maternal employment in lower SES (but not upper SES) households. When I read the actual paper, I felt incredibly silly and immediately understood why the effect could (and does) go the other way. However, let this be a lesson that hindsight bias is a cruel beast, and if you attempt to ridicule social science research by saying it’s simply “common sense,” you’d better make sure that you’re not actually claiming the opposite of what empirical studies have found.
APA Press Release: http://www.apa.org/news/press/releases/2010/10/working-mothers.aspx
Lucas-Thompson, R.G., Goldberg, W.A., & Prause, J. (2010). Maternal work early in the lives of children and its distal associations with achievement and behavior problems: A meta-analysis. Psychological Bulletin, 136 (6), 915-942 PMID: 20919797
Image Credit: Fair and Balanced slogan via Wikipedia and available under free use, as the purpose of this post is “for identification and critical commentary on the station ID or program and its contents.”