If you take a bath on Friday, and get dirty on Saturday, this means the bath was ineffective. The preceding, clearly flawed, logic sums up the misguided hysteria over a recent HHS report showing that the positive effects of Head Start fade out by third grade. The early childhood field has grappled with the "fade-out effect" multiple times before. We seem to have no trouble understanding why people lose ground after weight loss, addiction recovery, or treatment for depression because we know intuitively that an intensive intervention represents an extraordinary divergence from the natural course of human difficulty. And so it is with children growing up in poverty.
But, you say, everyone knows that a bath's effects are supposed to be temporary. The very promise of Head Start is that it provides a long-lasting benefit to disadvantaged children. Plus, baths don't cost taxpayers 8 billion dollars every year. If Head Start can't protect children against the "dirt" they may encounter once they enter the public school system, then why continue to pay for it, right? In fact, it is the "de-fund immediately!" chorus that is engaging in magical thinking, and not the "self-congratulatory progressives," as Head Start supporters were dubbed by the Wall Street Journal.
Consider this: despite the fact that cognitive impacts were not sustained into third grade for the overall sample, children who received their Head Start treatment year when they were three years old and whose mothers were not depressed did sustain cognitive outcomes. Given the large size of this group, this equates to a full 34 percent of the overall three-year-old cohort not experiencing fade-out on these critical outcomes, five years after attending Head Start. This proportion should be considered remarkable given the fact that most of the control children still attended some type of formal early education program.
Compare this to an approximately 18 percent long-term weight loss for bariatric surgery -- widely believed to be the most effective treatment for obesity; an approximately 22 percent success rate inpatient drug rehabilitation treatment -- an intervention that cost taxpayers upwards of $9 billion way back in 2005; and the 14 percent success rate of magnetic brain stimulation therapy for depression -- called "an attractive therapy" by Scientific American. Yet, applying magical thinking unevenly to the domain of early education, headlines declare "Head Start: Still Useless" (Heritage Foundation), "Head Start's Sad and Costly Secret -- What Washington Doesn't Want You to Know" (Fox News); "The Head Start Ruse" (Tribune-Review); and "Head Start's Unfulfilled Promises" (Washington Examiner). On the other hand, kudos to the Media Matters piece entitled, "Media Cherry-Pick Facts to Falsely Label Head Start Program a Failure," which dispassionately lists multiple positive scientific findings from Head Start over the years, including those that extend into later life such as increased high school graduation rates and decreased crime.
Surely I too can be accused of cherry-picking one of the positive findings I described above, but someone had to, because as far as I can tell every commentator except for Media Matters stopped reading after the Executive Summary. The truth is that the findings are mixed: effects were positive post intervention (the standard to which most interventions are held) on a broad range of findings (a much higher standard than to which most interventions are held), but the control group mostly catches up after that. There were some lasting benefits into third grade (18 percent), lots of equivalences between Head Start and control group children (74 percent), and some negative effects too (8 percent where it appears that Head Start caused disadvantages.) The real truth is that the findings are nothing to jump for joy about, not because of public dings to anyone's pet program, but because the Head Start and control group children were doing equally poorly by the third grade, lagging behind their age-mates in reading and grade promotion rates. When treatment and control groups do equally well, that is evidence that a program is unnecessary; when they do equally poorly, that is evidence that 1) the program needs to be fixed or an alternative found, and 2) a greater degree of systemic coordination is needed, because no single approach can be a silver bullet, and comprehensive approaches are needed to solve comprehensive problems.
Why am I so stubbornly convinced that the solution is to fix rather than chop? Because I don't believe in magic, and because of research that indicates that higher resources and additional enrichments in elementary school lead to more long-lasting effects of early educational programs. Such an analysis hasn't been done yet for Head Start (and given that these data are available, it is frustrating that the fade-out results were released without looking into this obvious question), but it has been done for Early Head Start. A long-term follow-up of Early Head Start children through fifth grade showed, familiarly, that many of the broad impacts seen in early childhood had faded. However, children who had Early Head Start, plus formal care at ages three and four (which, in many cases was Head Start), plus a lower-poverty school at age five had superior outcomes on measures of cognitive performance seven years later in fifth grade to those students who had any two, any one, or none of these early educational experiences (and having any two of the experiences was also significantly better than having one or none.) This shows the critical importance of that ongoing force on the other end of the tug-of-war rope that gives children a fighting chance to combat the natural course of risk. If you throw disadvantaged children back into environments rife with their own risk factors and provide no further mitigating (let alone enriching) efforts, why would you expect any outcome other than a slide? Try to do something similar with obesity, addiction, or depression and then call their interventions ineffective, and you would be (or should be) laughed out of any room.
The biggest reason I will likely stay on the "fix it" side of this issue is that not a single critic seems to be able to identify specific explanatory flaws in the Head Start model, or any new ideas that aren't already being tried, with their own varying degrees of struggles and success. Casting aside the opinions of those who don't know any better, the most disappointing response to the Head Start findings came from someone who does: Russ Whitehurst. Currently at Brookings, Whitehurst is a highly accomplished early education scientist and the founding director of the Institute of Education Sciences. Whitehurst first makes the inaccurate comparison of the iconic Perry Preschool Project, a small-scale controlled study with a sample size of 123 to the federally scaled Head Start, the study of which had a nationally representative sample of nearly 5,000. Then, he makes a further unfair comparison to "sustained" effects found in an IES-led study of preschool curricula by writing that "the most effective programs at the end of the pre-K year continued to show effects on cognitive outcomes." He doesn't mention that 1) the sustained effects are measured only one year later; 2) they are only true for six of the 15 curricula evaluated and only on scattered outcomes; 3) these were all individual, small-scale studies; and 4) the sample was overall much lower risk than the Head Start population (e.g., 19 percent of mothers not completing high school vs. 40 percent in Head Start).
Whitehurst concludes by suggesting state control and parent choice in local early education centers as the alternative to Head Start. The problem with this logic is that Head Start quality is already well-documented as being superior to community child care on average, and mostly in the good or better range. For parental "shopping" to be a viable alternative to Head Start, Whitehurst would have to believe not only that broad-scale child care quality can be reached by transferring the money from one effort to the other, but also that without all the additional family supports provided by Head Start, attending a quality child care center would result in narrowing the achievement gap significantly into at least third grade. For what is the validity of an alternate proposal if it too would not be on the chopping block for failing to meet the same standard?
As for me, I don't have the answers either, but I take my hints from the following: Early Head Start's longitudinal impacts could be said to be somewhat stronger than Head Start's, and the three-year-old cohort within the Head Start study had somewhat more positive and sustained effects (20 percent) than the four-year-old cohort (15 percent). Furthermore, one of the strongest evidence bases for effects into adulthood of an early intervention is the Nurse-Family Partnership, which is a nurse home visiting program from pregnancy to age two. As part of the numerous statewide efforts to create coordinated systems of early childhood education and care, why not promote the co-location and full integration of programs that create a no-wrong door approach for families, starting with pregnancy? At a minimum, Early Head Start and Head Start should be integrated and co-located as a rule, and we have to continue to learn as much as possible from the Head Start Centers of Excellence, which are setting a high bar for quality early education in general.
As a society, we have larger issues to grapple with about why it should be that children, who by any measure were well-prepared for school, became unprepared for it shortly after entering. In the meantime, let's examine the obvious questions about the conditions under which positive early findings are sustained. Nothing else will be a greater detractor from the uneven application of the "inoculation expectation" to the realm of early education and care, and a greater contributor to finding ever more effective ways of narrowing the income-based achievement gap.