MMO: The national dialog about what's best for children often gets mired in cultural debates about ideal family forms and whether young children are more likely to thrive when they have full-time maternal care. What's the advantage of using a research-based approach to assessing children's health and developmental needs, as you do in "What Children Need"?
Jane Waldfogel: In thinking about children and families, people have a tendency to draw on their own personal experiences and assumptions. But we also need to recognize that the world has changed since we adults were children, and that not all families are alike. If we are to make sound decisions about what children need and what we as a society should be doing to help meet children's needs when parents work, we need to be clear both about our values and about the research evidence.
To start with values, there are three core values that underpin our thinking about what's best for children. The first is the importance of respecting choice. Whatever policies we introduce, these policies should, to the extent possible, support families making their own choices about how their children are cared for. A second fundamental principle, and one that sometimes conflicts with choice, is the importance of promoting quality. We now know that the quality of children's care arrangements has a lasting impact on their growth and development. But currently, too many children and youth are in arrangements that are not of good quality. The third key principle is the importance of supporting employment. The work ethic is a widely shared American value, and work is a financial necessity for most parents. Moreover, women's employment is seen by many as key to gender equity and women's well-being.
Articulating these values, however, is not sufficient to tell us what's best for children. We need to know what the research shows -- about the effects of parental employment, or about the effects of preschool child care or after-school care. To take infant child care as an example, is it good or bad for children to enter non-parental child care at 3 months, as many children in the U.S. do? Saying that we value parental choice, quality care, and supporting employment does not provide the answer to this question. We need to know what the best evidence from research says about how children are affected by entering child care at that age.
I emphasize the point about best evidence because not all research is equally informative. Social scientists agree that where available the strongest evidence comes from controlled experiments, which randomly assign one group, the treatment group, to receive an intervention, and another group, the control group, to not receive it. If the samples are large enough and if the groups have been randomly assigned, then it is possible to measure the effect of an intervention by comparing the change in a given outcome for the treatment group to that for the control group. In the absence of a controlled experiment, the next best option is a "natural experiment," which mimics a laboratory experiment by randomly exposing one group to an intervention. For instance, one state or a few states (the treatment group) might implement a new program for infants, while other states (the control group) do not. If the two groups of states are otherwise comparable, then outcomes for the treatment and control groups can be compared, and the effect of the new program can be gauged.
Often, we lack either a laboratory or natural experiment, in which case we have to rely on observational studies. Such studies take advantage of naturally occurring variation in experiences across individuals and then attempt to measure the impact of those experiences holding all else equal. In the infant care example, we could identify families who used out-of-home child care for their infants and compare them to families who did not. If we could hold all else equal, and compare children who were identical except for the difference in their early child care experience, then we would be able to estimate the effect of early child care. In the real world, however, it is often impossible to hold all else equal. There may be many differences between children who did and did not attend infant child care, and researchers may not be able to control for all of them. For this reason, we have to be very cautious in drawing conclusions from observational studies and should place the most weight on studies that use rigorous methods to test whether the associations founds in observational data are likely to be causal. When studies use rigorous methods, and when several studies all point in the same direction, then we can have greater confidence in them.