Woodhead, M. (1988). "When psychology informs public policy." American Psychologist 43(6): 443-454.
From the abstract: "The potential of early educational intervention to make along-term impact on the life chances of socially disadvantaged children has now been clearly demonstrated... The evidence has been... frequently cited in support of public policies on Head Start and early childhood programs in general... [L]ess attention has been paid to the sense in which long-term effects are embedded in a wider context of family, community, and school processes that affect children not only at the time of intervention but also throughout their childhood years..."
Jeremy Szteiter
April 13, 2008
Summary of key points, by section:
Main themes:
-Head Start as a successful model of early childhood education as a single reference point in support of future formation of public policy
-dissemination of knowledge based on research may be focus of researchers, but more problematic issue may be need of researchers to validate their claims and clarify contexts which contributed to the strength of their research findings
-strong data from the Consortium for Developmental Continuity (Con. for Longitudinal Studies) that supports benefit of early childhood intervention programs (particularly for socially disadvantaged)
-strong, consistent results of several researchers engaging in well-planned and well-designed studies, along with follow-up data
-study data has strengthened support for Head Start, etc.: cited in public debates about ECE, used in evidence for laws related to education for handicapped/preschoolers, used in "third world" countries as support for ed. changes there
"Wider Context of Early Intervention Research"
-CLS studies (in mid-60's, reported in late 70's/early 80') part of a longer-term push for ECE policy research, and many other projects in related areas show strong results
-CLS projects uniquely target 2-4 year olds, generally applicable, designed to be directly useful for policy
-possibly have been some flaws in methodology that detract from analysis of the CLS program effectiveness
-CLS projects seen as needing less rigorous experimental design needed to implement programs and therefore are more effective; also, their long-term attention has added to their credibility
"Political Status of Early Childhood Intervention"
-ECE (Early Childhood Education) has not always received direct and independent attention as a social science topic, so researchers have made extra effort to get it recognized
-ECE issues often subsumed under a broader polical issue, not recognized on their own (War on Poverty, etc.)
-Head Start initiated by Pres. Johnson in 1965
-easy to tie to policy because program addressed highly political issues such as poverty and unemployment
-given attention in media with dramatic headlines
-Head Start budget greatly increased after study finding, even when other public spending was held back in the early 80's
Central questions that need to be posed against the supposed success of the programs:
1. What causal processes could have produced effects over such a long period?
2. What significance does an understanding of possible causal processes have for policy?
3. How valid are the implications that have commonly been inferred from the data?
"Evidence of Long-Term Effects"
-program group - higher high school completion rates (related to higher expectations of parents and students)
-no direct effects on success in job market observed
-David Weikart's study - did show significant employment gains, and clear (though diminishing) effects on school achievement, lower levels of crime/arrests, teen pregnancy
--these real-life effects have been highly influential on policy
"A Promise for Social Policy"
-very limited academic review of the findings (challenge by Zigler)
-researchers reluctant to challenge CLS studies because they have political appearance of doing much for children in poverty, but widespread acceptance of findings leads to overly simplistic interpretation going forward for future policy
-may facilitate political expedience but may harm future decisions
-also suggests that this particular intervention is globally applicable as time goes forward (starts to forget the context of the original study)
-need to remember that some parts of scientific study may be difficult or impossible to reproduce accurately in future attempts at the program
-creates risk of full political backlash if program does not succeed as intended
"Constraints on Generalizability"
-specific contexts of the children and program design of the original study needs to be considered
-original children were "extremely disadvantaged", lowest social class, parents with limited education, majority of Black children
-various programs in the study were themselves varied (initial child age, length of program, etc.)
-success cannot be tied to a single strategy, no single curriculum model within the study determined to be superior
-study suggests possibility that a wide range of strategies for ECE might be effective, but ALL may not be effective
-study contexts included high involvement of parents, careful program designs, low child-to-adult ratios, etc.
"Models of Early Intervention Effectiveness"
-"direct effects" model - preschoolers are more effective in school because they are more capable and "smarter" through the experience of being in preschool
-cognitive dev. only 1 of 7 goals of Head Start
-difficult to tie long-term prog. effects - cog. ability higher through the next two years, but the effects still diminished during that time for the experimental group
-IQ not strong foundation for determining long-term effects (unless some latency occurs)
-key variables that offer evidence of long-term effects: "school competence" (retention of grade and referral to special education classes) => study group less likely to be held back or referred to special ed.
-CLS suggested that program caused teacher intervention which started a cycle of positive support and reinforced children's competence - became a self-fulfilling prophecy
-even though program group showed differences compared to control group, the program group still had relatively poor performance compared to broader school community - this effect could carry through even though cog. effects had diminished
-labeling of children also a self-fulfilling prophecy - children live up to standards set for them
-suggests that being held back/special ed. actually harmful by starting a long-term cycle of low achievement and low self-perception
-suggestion from British 11+ study - program group more likely to benefit because the program prepared children with respect to putting them in a perspective of achievement relative to another type of evaluation, which had a long-term effect on them in real terms
-this suggests that ECE intervention is more complex than realized
-preschool affected children but also parents,teachers in direct and indirect ways
-Bronfenbrenner's "ecological" framework - advocates social research in context
-methods and measures develop over time leaving longitudinal studies in a bind to maintain consistent measures, yet aknowledge a decreased validity.
-Often individual studies lack the proper requirements to be projected onto the general population, yet can show precise differences between specific groups.
-Meta studies, although general enough to be projected onto the general public, often lack the specific details and nuances of individual studies.
-Policy often requires much more than good evidence. In fact, sometimes the evidence does not appear to matter at all for the creation of policy.
"Policy Implications of Long-Term Effects"
-"transactional" model of the CLS study more complex,less certain than a direct model, which would directly relate change to policy
-transactional model long-term implications must consider other interactions that happen between parents and children, influence of culture, even specific experimental design differences between studies
-social sciences require many more replication experiments compared to hard sciences, which still may not "prove" a case
-must also consider that an absolute model should not simply be replaced with a culture-specific model in all cases, because some evidence may apply across cultures
-need to further examine the path of students after the study - "unseen" factors may dampen or amplify the path to long-term achievement (Latin America study - low number of teachers, poverty, limited resources can dampen benefits of preschool such that they start to become insignificant compared to these other factors)
-high-quality research of CLS study focused on child achievement but also might have provided evidence about the overall school structure and practices and helped to suggest strategies about how they could be improved as well
-original goal of Head Start included simply getting federal funds to local communities, and partially to counteract negative influences in homelife of children and counteract deficiencies in the school system (Head Start employed for the "worst reasons")
-CLS studies have taken place at a different time in history, so the reason for their relevance may be weakened by different/changing school climates
-reasons to believe that family attitudes about preschool were particularly sensitive at the time of the CLS studies: during period of racial desegregation, civil rights legislation, MLK, new hope for education/children, optimism of the country
-Zigler: argues that evidence has at times been misapplied in political terms to suggest blame for secondary school failure
-Zigler (1983): emergence of "new roles for scientists and experts as advocates"
-generally, overemphasis on "researched" study causes attention/recognition of "unresearched" areas (e.g. other context) to be lessed, but that does not mean those areas are not significant or crucial
-this might actually undermine role of ECE programs, since times when they fail to achieve goals may also be blamed solely on the program, ignoring other factors of culture, family and school-wide issues
Mary Beth LeFaivre
April 22, 2009
“When psychology informs public policy”
This article provides analysis of how social science, in particular early childhood education research, is translated to and utilized in policy. The article, published in 1991, raises a concern still prevalent in education research. What happens when policymakers and advocates use outcomes from research studies to push their own agenda? The author traces the history of Head Start, a federal program born out of the Great Society reforms that addressed poverty and racial inequalities. Head Start was created to provide early child development and health programs to low income families. Numerous evaluations have studied the effectiveness of this federally funded early intervention program over the years; a meta-analysis of 1,500 studies (Hubbell, 1983) including longitudinal comparison studies have measured the cognitive, health, and social development of children in Head Start programs, and “real life” variables over time, such as employment and juvenile delinquency.
Because these studies are not comparable in population or in rigor (of research design), it is difficult to conclude a causal link between the intervention and student outcomes. As policymakers have hailed the success of Head Start based on positive findings from these longitudinal studies, which are indeed scientifically based, federal funding has increased and sustained Head Start programs for more than 50 years. For example, the Consortium for Longitudinal Studies report showed in a comparison study in 1980, socially disadvantaged children who participated in Head Start programs had higher high school completion rates. Another study, Weikart’s Perry Preschool Project (Clement et al, 1984), showed positive effect of preschool on their levels of employment, juvenile delinquency and teen pregnancy. The authors of the Consortium projects note that children who participate in preschool have a good start in the school system, with stronger school competence, retention, and fewer referrals to special education as compared to control groups; these “transmission pathways” might have transformed short-term effects of the intervention into long-term outcomes.
The authors point out that ot’s important that preschool programs and the benefits they offer to disadvantaged children are not viewed as a magic bullet for improving achievement and social inequities in our society. In his analysis of these studies, Woodhead raises several concerns on the generalizability and validity of findings from the two seminal projects (Weikart’s and Consortium studies) that served as impetus for policy; their findings were promoted in the media and to the public as justification for public funds. While the studies show Head Start works, and it does make a difference in narrowing achievement gaps and social inequities, Woodhead notes that one cannot look at these studies in isolation. The growing acceptance in research is that of a situated learning or sociocultural or “ecological” framework, which takes into account the effect of the environment and social contexts that contribute to learning and social development. From this perspective, there are mediating variables – family, home, school, culture, and the child’s own self-concept – that actively interact with or mediate the short and long-term effects of the preschool intervention. Furthermore, the transmission pathways noted above will vary from culture to culture or school to school. Models that consider the complexity of these indirect variables, as opposed to a “direct effects model”, are called a transactional model. Translated to policy, this “cultural relativistic model” is not as powerful and absolute as a direct effects model, thus changing -- and challenging -- how one interprets social science research. Woodhead suggests that in social science, it is important to compare contextual conditions of studies, including the time in which it took place (in this case, the 1960s), and to replicate studies in order to generalize the potential for an intervention.
As policymakers use research such as Head Start evaluation research to inform social policy, it is critical they recognize the contextual variables that play a role in the success of early intervention programs. What are the community and social variables that interact with Head Start? How do these variables contribute to or detract from the long-term “real life” variables? As the author indicates, there could be other influences that are worthy of focus and funding for intervention, in addition to or instead of Head Start, that lead to healthy development of children in low SES communities. The point of this article is not to diminish the value of experimental social science research; it is a reminder that there is no simple cause-effect for early childhood intervention. Researchers and policymakers should consider the contextual factors and processes and not rely solely on the manipulated variable in experimental studies. Such multilevel analysis will help scientists identify the settings, contexts, and social variables that contribute to a broader goal of helping all children have access to healthy start in life.
