In their just-released study, “Maternal Employment and Child Cognitive Outcomes in the First Three Years of Life,” researchers Jeanne Brooks-Gunn, Wen-Jui Han, and Jane Waldfogel have more bad news for working mothers. The trio used data from the National Institute of Child Health and Human Development Study of Early Child Care to measure the impact on cognitive skills of maternal employment in a child’s first year. (The skills were evaluated by the Bracken School Readiness Test, which is administered when a child is 3.) The study found “negative effects” on children whose mothers worked 30 hours or more per week in the first nine months, “even when controlling for child-care quality, the quality of the home environment, and maternal sensitivity.” In other words, quality childcare and quality mothering notwithstanding, a mother who works fulltime in the first nine months of her child’s life is statistically likely to have a child, who at age 3, scores lower than children of non-working mothers on a cognitive test.
The New York Times headline, “Study Links Working Mothers to Slower Learning,” sparked a new wave of guilt and frustration in the ranks of working mothers, or at least those of the 25 million in the workforce who went back to jobs when their children were infants. Others were further exhausted by the new onslaught of statistics, bracing for the attendant criticism and finger wagging about their selfish priorities, even though many of these mothers have no choice when it comes working outside the home.
But when they are asked about the implications of the study, published in the academic journal “Child Development,” the authors point out that individual outcomes can vary widely. Some variables, such as the quality of the childcare or the attitudes of the mothers, had a significant influence on individual children. Overall, they emphasize that the study shouldn’t necessarily discourage mothers of young children from working, but encourage government policy that would improve the quality of childcare, change the length of family leave and include pay for those who take it, as well as promote job-sharing and flexible hours that would help working parents who struggle with work and caring for young children.
Dr. Jane Waldfogel, associate professor at Columbia University in the School of Social Work, and a research associate at the Centre for Analysis of Social Exclusion at the London School of Economics, spoke about the study from her office in London. Expressing her own occasional frustration with the lack of nuance and explication in news reports about the work, Waldfogel talked about the data used in the study and the implications of her research, as well as research trends in her field, and, in particular, the ways in which fathers get left out of the equation.
Can you begin by discussing the research from the National Institute of Child Health and Human Development Study of Early Child Care? As a researcher in the field, what do you find helpful about this study as opposed to earlier data sets? What kinds of problems do you see with the study? In particular, were there any questions you would have liked to study about working parents and child health and development that were precluded by the structure and setup of the NICHD-SECC study?
As you probably know, this study was a collaborative effort on the part of several teams of leading developmental psychologists from around the country, and it was specifically designed to study the effects of early childcare on child development. So, it’s got wonderful measures of childcare quality and also measures of parenting quality. It’s really head and shoulders above any other resource that we’ve ever had before to look at the effects of early parental employment or early childcare on children’s outcomes.
Of course, there are always more things that any individual researcher would want to have in the data set, and I’m sure there must have been some compromises that were made. When you have families included in a survey like this, there are limits as to how much time you can be in their homes or have them come into the laboratory, and so you can’t ask them everything that you’d want to ask. Decisions have to be made about what to include and where to place the most emphasis.
Were there things that I wish had been included? Sure. There are always more things that I would want. I wish they had done more to look at the quality of the fathers’ care of the child/children, in addition to the mother’s care. I wish that they had done more to gather information about the quality of the mothers’ jobs and how the mothers felt about their jobs. But that’s difficult to do. In the scheme of things, I’m about 99 percent satisfied with what they did.
How is childcare defined and how do different forms of care function in the NICHD-SECC study? Did the study allow you to differentiate between and compare the effects on development of different forms of care?
Yes. The researchers who put together this study kept track of the type of care the children were in at different points in time during their early years, and then at different points in time they visited the childcare arrangements. If a child was in childcare for more than 10 hours a week, they went out and visited that childcare arrangement.
They did an assessment of the quality of the arrangement using a rating scale that they developed that is very focused on the individual child and the individual child’s experience. Is the caregiver being responsive to that child? When the child is distressed, when the child is bored, when the child is engaged and doesn’t want to be interrupted, how sensitive and responsive is the caregiver being to that child? So it’s a terrific scale.
The reason they had to develop their own scale was that they wanted to be able to assess quality across different types of childcare arrangements. In the previous research that had been done on childcare, there was one scale used for individual providers, like a father, or a grandmother, or a babysitter, or a family day care provider, and there was a different scale used for centers, like day care centers or nursery schools. That’s why they had to develop their own.
In the studies that the NICHD group have done, researchers looked extensively at the long-term effects of the type of care that children are in and the long-term effects of the quality of care that children are in. We also looked at those in our study, although the main focus for us was the effects of early maternal employment.
So, childcare is defined as care by anyone other than the mother?
And is the mother’s care rated along the same scale?
No. There are two different measures. There’s a measure of childcare quality that’s used for non-maternal childcare, and then there’s a measure of the mother’s sensitivity, and that’s used for all mothers, working or not working. It’s a different assessment, that involves bringing the mother and the child into a lab setting, I think, and doing some assessment of her care of the child and there’s also an assessment of the home environment, so they are different measures.
Do your findings suggest that child cognitive development is negatively affected for the child of a working mother even if, for instance, her partner functions as a stay-at-home parent?
I don’t think we looked at that question directly. In other work, when we’ve had larger samples, we’ve tried to look at how the effect of maternal employment in that first year differs based on the type of childcare she used, including whether the childcare was father care. And I don’t think we found that father care is particularly protective.
But that data set came from an earlier study that we did with the children of the National Longitudinal Survey of Youth. Those were children born in the 1980s. For them, having a mother who worked full time in the first year and a father who provided care had negative effects on children’s cognitive outcomes. But fathers who were stay-at-home fathers back in ’82 and ’83 are probably different from fathers who are staying home now. So, I wouldn’t generalize from that to what might be the case today.
I think the world has changed a lot in the last 20 years. In ’82 and ’83 there was a pretty deep recession in the U.S., and so a number of the fathers who were staying home taking care of children were unemployed, and that’s different from a family where the partners have talked it over and have reached a decision that this is what they prefer to do and he’s home by choice.
It’s hard in the NICHD study, we have 900 kids, but you start splitting them up into groups and it gets kind of small.
Are there studies and analyses about paternal employment and child development? Have researchers considered the effect, for example, of fathers who work more than, say, 40 hours a week?
That’s a good question. I’ve definitely seen studies of fathers’ employment, about whether they are working or not, but I’m not sure that I have seen a study that looks at whether fathers are working more than 40 hours, and that’s a good question. I’m not sure that I have seen that.
Has research considered the importance of paternal care and development, looking, for example, at the relationship between paternal sensitivity and child cognition?
We didn’t look at that with the NICHD data. There is some information available on fathers’ sensitivity in the NICHD data set — not for all fathers but for some (some of the study sites collected this and some did not). We didn’t realize until recently that the data were there — now that we do know, we will take a look at this.
Why do you think that people aren’t paying attention to fathers?
That’s a very complicated question.
What are your thoughts?
That’s not really my realm. I don’t really have a thought to articulate about that.
Has research been conducted on the effect of maternity or paternity leave on child development?
It’s a question that I’m actually looking at currently. What’s the effect of maternity leave entitlements on child development? I have looked some at paternity leave. It’s very tough to estimate effects because most men take such short periods of leave that it’s like looking for a needle in a haystack. American men are now taking paternity leave in pretty large numbers, but they take on average only about a week or two, and so it’s very hard to find men on leave in a data set. It’s very hard to pick it up. And then it’s very hard to estimate effects of something that only lasts for a week or two.
What about in other countries where they have longer-term paternity leave?
Even in Sweden they’ve had trouble getting men to take longer periods of leave, although they’re working very hard on it and they’re making some progress. I haven’t seen any research that looks at effects on child outcomes. It’s funny, I think in other countries this is a no-brainer because they think it’s good for children and families, so why would you go out and study it? Is it better if the parents can share the leave or work part time? They’re interested in those questions, but the question of whether it is better for children if their fathers had the right to paternity leave, that’s just taken for granted.
In your discussion of your research, you write that “children who are in early child care typically have mothers who are employed, and vice versa.” But many working families utilize a shift-work organization for the parents’ labor, to be able to avoid or minimize traditional childcare — Francine Deutsch, for example, analyzes this phenomenon in “Halving It All.” Is there any research on whether this kind of care arrangement affects the negative correlation between maternal employment and child development? In other words, can we isolate whether it is the childcare or the maternal employment that is the causative agent in delayed cognitive development?
I’d answer that in two different ways. One is that in our study, we were specifically looking at the effects of early full-time maternal employment, maternal employment of 30 or more hours per week in the first year. And that’s where we saw the negative effects on children’s cognitive outcomes at age 3. In other studies, using the same data set, when researchers look at the effect of early childcare, they don’t see negative effects at age 3. So that would suggest that it’s the long hours of maternal employment, and not the use of childcare, that’s leading to the lower school readiness that we saw at age 3.
Regardless of the relationship between that employment and the actual childcare for the child? For example, if the mother works at night …
We didn’t look at whether the effects of maternal employment varied by whether it was shift work, night, evening, or other hours. The big share of the mothers in this study were working nonstandard hours. But one of the three authors — Wen-Jui Han — is actually now looking at that question, so sometime down the line she’ll have some results that she can share.
You indicate that women with more education returned to work earlier than women with less education. Yet you also found that mothers who worked more than 30 hours per week by the time their children were 9 months old were rated as providing what the NICHD-SECC defines as less-sensitive care than mothers who did not work in the first year. Can you tell us a little more about what the NICHD-SECC defines as sensitive mothering, parenting, and care giving? Can we draw the conclusion that you found more educated, working women to be less sensitive, which is to say less responsive, mothers?
No, that second statement is completely wrong. The first statement that you were making has to do with associations in the raw data. Women who go back earlier tend to be the more highly educated — the higher earners, the more professional women. But when we look at the effects of early maternal employment we’re holding all that constant. So we’re saying, if two women had the same education, the same marital status, the same depression score in the first year — they were the same in every way we could measure, but one of them had worked full time in the first year and one of them hadn’t — what’s the difference in their children’s school readiness? We’re comparing like to like.
And in terms of the difference in their sensitivity?
Yes, we found that, for women who worked full time in the first year, by the time their children were age 3 they were providing less sensitive care than women who hadn’t worked full time in that first year. And it makes me wonder if working full time in the first year is starting women down a pathway that results in them being more hassled, more stressed out when they’re with their child. That’s one possible interpretation.
The sensitivity measure is what I was describing to you before: Does a mother respond when her child is distressed, does she reach out to engage her child when the child is bored, does she also not interrupt the child when the child is happily playing, is she not overly intrusive and interrupting? It’s a measure of sensitivity.
In your findings, you discuss the fact that the lower cognitive development does not show up until the 36-month Bracken School Readiness test. You mention that earlier tests measure different cognitive competencies. How satisfied are you with the Bracken School Readiness test as a substantive measure of something we should really care about in terms of child development? What is the correlation between low scores on the Bracken test and longer-term cognitive development?
The Bracken measures things we should care about — how well a 3-year-old knows the things that they will need to know to be ready for school. And it is predictive of later school outcomes. So yes, I think it is a good measure.
You also mention a Swedish study that found that Swedish children who entered childcare earlier in the first year of life had better cognitive outcomes than those who entered care later. A study by Dr. Jay Belsky, based on the NICHD-SECC data set, suggested that children in childcare displayed better language skills and short-term memory. How do you think we should assess the seemingly contradictory nature of different studies on this issue?
Can I take the questions separately, because they’re completely different points? There have been a number of studies using the NICHD data that look at whether children who are in childcare have different development, and I think the most interesting finding that has come out of those studies is that the more time children spend in center-based care, the better prepared they are for school, and the better is their cognitive development. And the lead authors of those studies are people like Deborah Vandell and Kathleen McCartney.
I wouldn’t put Jay Belsky with those studies. He’s taken the lead on studies that have to do with behavioral outcomes, but he doesn’t particularly study the cognitive outcomes. So I don’t think there’s any contradiction between a finding that children whose mothers work early and full time in the first year have poorer cognitive outcomes, and the finding that children who spend more time in center-based care in their preschool years have better cognitive outcomes — because not all children whose mothers are working full time in the first year are in center-based care.
We’re talking about different populations. The study from Sweden is really interesting because it says that in a different context, having a mother work full time in the first year and being in childcare in the first year can be a positive thing. So that says something to me about the quality of childcare that’s available in Sweden as compared to the quality that’s available in the United States.
But in your study you found that the quality of childcare didn’t have an impact on development.
Sure it did. We found huge effects on the quality of childcare. The effects of quality of childcare and the quality of the mother’s care together are more important than the effects of whether she worked full time in the first year or not. If you moved a child from a mother with low quality care to higher quality care, and if you moved a child from low quality day care to higher quality day care, the child would gain about seven points on the Bracken.
But didn’t your study find that regardless of the fact that the child might have a sensitive mother and high quality childcare, that child would still have statistically lower cognitive development than a child whose mother didn’t go to work during the first nine months?
Right, on average, the effect was negative, right. But that’s an average, given the average quality of care the children were in.
What are the consequences as measured on the Bracken test for maternal employment of more than 30 hours per week after nine months?
There are no negative effects, and they may well be positive. Usually, if there’s any effect of employment after the first year, it’s positive, and that’s usually because the employment brings some good things, and also children are increasingly likely to be in center-based care as their mothers are working after the first year.
Center-based care is strongly associated with better cognitive outcomes and school readiness outcomes. I’m not seeing significant effects of employment after the first year. There’s certainly not anything negative after the first year. They’re positive, but they’re not significant.
So, just to clarify your statement: There are positive cognitive effects for children whose mothers work after the first year. Has there been research comparing cognitive effects for children whose mothers are working after nine months as compared to children whose mothers aren’t working?
Are you asking if children do better on cognitive tests at 3 or 4 if their mothers are working, as compared to if their mothers have not worked at all up to that point?
I think in general they tend to do better and sometimes those differences are significant and sometimes they’re not. Generally, employment after the first year is positive in terms of cognitive development. Our study is about full-time employment in the first year. It’s not about employment after the first year; it’s not about part-time employment. It’s really important that it not be generalized to maternal employment.
The New York Times originally headlined the story about your study: “Study Links Working Mothers to Slower Leaning.” Is it possible that the media has interpreted your work as applying to all maternal employment?
Right, the word “full-time” got left out and the words “first year.” I’ve thought some about what they could have done. I understand the length of headline issue, but that’s not what the headline should have been. But there’s been a fair amount of coverage about the policy implications, and people have talked about parental leave and about childcare quality.
Is there anything else you want to say about media coverage of this work?
I’ve tried to emphasize in talking with the media that I think parents make the best choices they can, taking into account a lot of factors. Many of those factors are things that we can’t measure in our data. So, I’d hate to see any individual parent second-guess a decision that they’ve made on the basis of our study or any other study.
We have to assume that parents are making the best decisions they can, taking a lot of things into account. But I also think we need to give parents a better set of choices in terms of paid parental leave, in terms of higher quality childcare, in terms of more options for part-time and flexible working. I think the media do have a tendency to cover the individual story and to draw out implications for individual parents and to not place as much emphasis on the policy implications. I think the policy implications are important.
The most direct implication for current policy is on the move to raise the working hours required for mothers on Temporary Assistance to Needy Families (TANIFF) from 20 hours a week to 40 hours a week for women who have preschool age children.
Right now states have the option to require women to go back to work as early as three months, but women with preschoolers only have to work 20 hours per week to satisfy the work requirement. And the Bush administration has been proposing to raise that to 40 hours per week. This research says that’s completely the wrong thing to do. And for children who really are the most vulnerable, these are children who are going to be at risk, you know on average they’re going to tend to be more at risk in terms of school readiness to start with, so this is just not what you want to be doing for them.
What do you think about the direction of research in this field generally?
I think we’re actually making progress. I think it’s good that we’re looking at a range of factors that influence child development, and that we’re not looking just at maternal employment or just at childcare or just at the quality of the home environment. But I think we don’t pay enough attention to fathers. I think they tend to be forgotten by most people doing the research and that includes us. I think it’s absolutely true that we haven’t paid enough attention to fathers.