Preschool Matters

Preschool Matters Edward C. Melhuish Science 333, 299 (2011); DOI: 10.1126/science.1209459

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your colleagues, clients, or customers by clicking here. Permission to republish or repurpose articles or portions of articles can be obtained by following the guidelines here.

Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/content/333/6040/299.full.html A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/content/333/6040/299.full.html#related This article cites 16 articles, 4 of which can be accessed free: http://www.sciencemag.org/content/333/6040/299.full.html#ref-list-1 This article appears in the following subject collections: Education http://www.sciencemag.org/cgi/collection/education

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2011 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.

Downloaded from www.sciencemag.org on July 20, 2011

The following resources related to this article are available online at www.sciencemag.org (this infomation is current as of July 20, 2011 ):

PERSPECTIVES Making connections. A network with eight components (including two components with just a single node each). A rigorous mathematical proof given by Riordan and Warnke shows that as the network evolves by adding more links, the “giant” component that emerges at the phase transition does so continuously.

Further work is needed to understand why the numerical simulations were misleading. Furthermore, are there analogous results for other random network models, perhaps with consequences for applications to networks appearing in the real world? The new results and methods by Riordan and Warnke will further stimulate the study of the detailed behavior of various types of random network. References

1. R. Durrett, Random Graph Dynamics (Cambridge Univ. Press, Cambridge, 2007). 2. D. Achlioptas et al., Science 323, 1453 (2009). 3. O. Riordan, L. Warnke, Science 333, 322 (2011). 4. R. A. da Costa et al., Phys. Rev. Lett. 105, 255701 (2010). 5. A.-L. Barabási, R. Albert, Science 286, 509 (1999). 6. P. Erdös, A. Rényi, Publ. Math. Inst. Hung. Acad. Sci. 5, 17 (1960). 7. T. Bohman, A. Frieze, Random Structures Algorithms 19, 75 (2001). 8. J. Spencer, N. Wormald, COMBINATORICA 27, 587 (2007). 9. S. Janson, J. Spencer., http://arxiv.org/abs/1005.4494 (2011). 10. T. Bohman, Science 323, 1438 (2009). 10.1126/science.1208712

EDUCATION

Preschool Matters Edward C. Melhuish Evidence grows that starting preschool at age 3 or 4 can produce benefits decades later.

S

cholars have long debated the benefits of preschool, or prekindergarten, education. Several small-scale studies have documented that preschool contributes to better educational, occupational, and social outcomes for disadvantaged children over the long term (1) and is cost-effective (2). Large-scale, long-term studies, however, are unusual. On page 360 of this issue, Reynolds et al. (3) help fill that void. They report on a 25-year-long study, involving nearly 1400 former students, that examines the effect of the Child-Parent Center Education Program, a publicly funded preschool program in Chicago, Illinois, on subsequent educational achievement, socioeconomic status, health, and crime. Their results demonstrate consistent and enduring benefits for Birkbeck, University of London, London, WC1B 3RA UK, and Kyung-Hee University, Seoul, Korea. E-mail: e.melhuish@ bbk.ac.uk

children who began preschool at age 3 or 4 (compared with children who began kindergarten when older), and especially for males and children of high-school dropouts. In particular, by age 28, the former preschool students had higher educational levels, incomes, socioeconomic status, and rates of health insurance coverage—and lower rates of substance abuse and legal problems— than the kindergarten students. Most of the students in the Reynolds et al. study are from disadvantaged African-American families living in the inner city. How well do their findings apply to wider populations? The answer is that there is strong evidence that not only the disadvantaged benefit from a preschool education. In the United States, for example, general population studies of representative samples of people have found benefits for school readiness (4, 5); these benefits are greatest if preschool starts between 2 and 3 years of age (6).

www.sciencemag.org SCIENCE VOL 333 15 JULY 2011 Published by AAAS

299


specific applications. Others, such as the Achlioptas model, are designed to explore the phase transition or other properties in order to better understand random networks in general. In many cases, a phase transition of the same general type as for the Erdös and Rényi model has been found (see the figure). The Achlioptas model is usually presented as a kind of game: A network is grown by adding links, but each time you are given two candidates for the new link, each randomly chosen, and you have to choose one of them (not knowing the future links). Your task is to delay the emergence of a giant component as long as possible. Some strategies succeed in delaying the phase transition (7, 8). In the simplest case, the phase transition is still of the Erdös and Rényi type (9), but for another (better) strategy, numerical simulations suggested an explosive behavior (2). The intuition is that by preventing the phase transition for some time, “pressure” builds up, so that when the phase transition occurs at last, it becomes a violent explosion. Random networks are studied both by mathematicians and mathematical physicists, using partly the same mathematical methods but with one important difference. In mathematical physics it is acceptable to make reasonable approximations and simplifying assumptions. One might, for example, assume that certain quantities have limits as the number of nodes tends to infinity, based on the experience that it works well in other related problems. One might say that mathematical physicists are optimists who believe that natural systems are well-behaved, whereas mathematicians are pessimists who see every conceivable complication as a possibility until it is firmly ruled out by a solid proof. Consequently, a mathematical proof is pure logical reasoning that will stand forever, while a mathematical physics proof, like physics or other sciences, may have to be revised if new evidence shows that the underlying assumptions are unfounded. A third approach is to use numerical simulations. This is more like experiments in physics; a well-performed numerical simulation gives results that most likely are true, but there is no guarantee, and the results should

be seen as well-supported conjectures rather than proofs. The Achlioptas (2) process combines all three approaches, illustrating how they complement each other to increase our understanding. Numerical simulations were used, giving a surprising result that seemed well-founded, but nevertheless was only a conjecture (10). A mathematical physics proof was given by da Costa et al. (4) that the simulations were misleading and that the phase transition really is continuous, although with a very small critical exponent. That is, the size of the largest component increases rapidly, and is close to being discontinuous. Riordan and Warnke have given a fully rigorous mathematical proof, by different methods, and there can be no more doubt. Other things remain to be proven mathematically, in particular the value of (and even the existence of) the critical exponent found by da Costa et al., and it seems likely that new methods will have to be invented for that.

PERSPECTIVES Literacy Numeracy

Girls (versus boys) Early development problems Birth weight Family income Mother‘s education Father‘s education

Similar evidence comes from Socioeconomic status European nations (7, 8) and shows Home learning environment that the effects of preschool are High-quality preschool long term: Preschool was associated with increased qualifications, Primary school employment, and earnings up to age 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 33 (9). In France, universal free preEffect size in standard deviation units school (école maternelle) is available to children starting at age 3. During for background influences showed that, by the 1960s and 1970s, a large expansion of age 11, children who attended high-quality the program led to the proportion of 3-year- preschools significantly outperformed those olds enrolled in preschool increasing from who had not attended preschool on literacy 35% to 90%, and of 4-year-olds from 60% and numeracy tests and that low-quality preto 100%. State-collected data show that preschool had no beneficial effect (8). Indeed, school had a sizeable and persistent posi- the benefit deriving from 18 months of pretive effect on a child’s ability to succeed in school is similar to that gained from 6 years school and obtain higher wages in the labor of primary school (see the figure). market. Preschool also appeared to reduce In other research, the Organization for socioeconomic inequalities, as children Economic Cooperation and Development from less advantaged backgrounds bene- (OECD) examined educational attainment fited more than those from more advantaged data for 65 countries. It found that literacy backgrounds (10). Similarly, in Switzerland, at age 15 was strongly associated with prethe impact of preschool expansion was asso- school participation in countries where a ciated with improved intergenerational edu- large proportion of the population use precational mobility, with children from disad- schools, where preschool is available for vantaged backgrounds benefiting most (11). longer periods each year, and where there In Norway, which expanded preschool edu- were measures to maintain preschool qualcation for 3- to 6-year-olds during the 1970s, ity. They concluded that widening access researchers found that preschool participa- to preschool can improve performance and tion was associated with strong benefits for equity by reducing socioeconomic disparilater educational and job outcomes (12). ties, if extending coverage does not comproThere is also evidence of preschool’s mise quality (18). benefits from Asia, South America, and elseThese issues are relevant to preschool as where. In Bangladesh, preschool boosted an intervention strategy. Some analysts have primary school achievement (13), with sim- argued (19) that, in the United States, govilar results reported for 10 other countries ernment-funded preschool programs (e.g., (14). When Uruguay expanded its preschool child care centers, Head Start, and statesystem, studies comparing siblings with and funded prekindergarten) offer services that without preschool, and regions with varying are of “mediocre or worse” quality, that levels of preschool, revealed clear benefits children attending the average center may for children attending preschool through gain little cognitive boost, and that greater secondary school (15). Similar analyses in benefits could be gained by improving the Argentina found that 1 year of preschool quality of these programs. Others ( 20) was associated with primary school attain- maintain that publicly funded preschool in ment increases of 0.23 of a standard devia- the United States narrows the achievement tion (16). gap between poor and nonpoor groups by Many studies, including the work of as little as 5% because of the prevalence of Reynolds et al., suggest that high preschool low-quality programs and that preschool quality is critical to success. General popula- could be narrowing the gap by up to 50% if tion studies from the United States (17) and quality were improved. England (7, 8) indicate that it is higher-qualSuch debates have fueled increasing ity preschools that produce greater long-term interest in the universal provision of prebenefits. In England, a study that controlled school education as a means of advanc-

300

ing school readiness for children and their later attainment of social, economic, and occupational success (2, 21). Indeed, some argue that preschool is not only an intervention for disadvantaged groups and a means of advancing social welfare for all but also a critical contributor to the economic health of the nation (22). For example, Ben Bernanke, the chairman of the U.S. Federal Reserve System, has argued that “the payoffs of early childhood programs can be especially high” (23). Some countries, including China (24), appear to have adopted this perspective as they pursue focused efforts to provide preschool as widely as possible. 1. 2. 3. 4. 5. 6. 7. 8.

9. 10.

11. 12. 13. 14. 15. 16. 17. 18.

19.

20. 21. 22. 23. 24.

References

W. S. Barnett, J. Hum. Resour. 27, 279 (1992). J. J. Heckman, Science 312, 1900 (2006). A. J. Reynolds et al., Science 333, 360 (2011). W. T. Gormley et al., Science 320, 1723 (2008). K. A. Magnuson et al., Econ. Educ. Rev. 26, 33 (2007). S. Loeb, M. Bridges, D. Bassok, B. Fuller, R. W. Rumberger, Econ. Educ. Rev. 26, 52 (2007). E. C. Melhuish et al., Science 321, 1161 (2008). P. Sammons et al., Influences on Children’s Attainment and Progress in Key Stage 2: Cognitive Outcomes in Year 6 (Department for Children, Schools, and Families, London, 2008); http://eppe.ioe.ac.uk/eppe3-11/eppe3-11%20 pdfs/eppepapers/DfE-RR048.pdf. Accessed 6 June 2011. A. Goodman, A. B. Sianesi, Fiscal Studies 26, 513 (2005). C. Dumas, A. Lefranc, Early Schooling and Later Outcomes: Evidence from Pre-school Extension in France. Thema Working Paper 2010-07 (Université de Cergy Pontoise, France, 2010); http://thema.u-cergy.fr/IMG/ documents/2010-07.pdf. Accessed 6 June 2011. P. C. Bauer, R. T. Riphahn, Econ. Lett. 103, 87 (2009). T. Havnes, M. Mogstad, No Child Left Behind: Universal Child Care and Children’s Long-Run Outcomes. Discussion Papers No. 582. (Statistics Norway, Oslo, 2009). F. E. Aboud, Early Child. Res. Q. 21, 46 (2006). J. E. Montie et al., Early Child. Res. Q. 21, 313 (2006). S. Berlinski et al., J. Public Econ. 92, 1416 (2008). S. Berlinski et al., J. Public Econ. 93, 219 (2009). D. L. Vandell, J. Belsky, M. Burchinal, L. Steinberg, N. Vandergrift, Child Dev. 81, 737 (2010). OECD. PISA 2009 Results: Vols II and IV. (OECD, Paris, 2011); www.oecd.org/document/61/0,3343, en_2649_35 845621_46567613_1_1_1_1,00.html. Accessed 6 June 2011. R. Haskins, W. S. Barnett, Eds., Investing in Young Children: New Directions in Federal Preschool and Early Childhood Policy (Brookings Center on Children and Families & National Institute for Early Education Research, New Brunswick, 2010); http://nieer.org/pdf/Investing_in_ Young_Children.pdf. Accessed 6 June 2011. R. C. Pianta, W. S. Barnett, M. Burchinal, K. R. Thornburg, Psych. Sci. Pub. Int. 10, 49 (2009). E. Zigler, W. Gilliam, S. Jones, A Vision for Universal Preschool Education (Cambridge Univ. Press, New York, 2006); 10.1177/1529100610381908. M. McCain, J. F. Mustard, Early Years Study: Reversing the Real Brain Drain (Publications Ontario, Toronto, 1999). B. S. Bernanke, Challenges for State and Local Governments, Speech, New York, 2 March 2011. L. Shenglan, in Early Childhood Care and Education: International Perspectives on Policy and Research, E. C. Melhuish, K. Petrogiannis, Eds. (Routledge, London, 2006), pp. 151–166.

15 JULY 2011 VOL 333 SCIENCE www.sciencemag.org Published by AAAS

10.1126/science.1209459


Preschool’s benefits. In a study done in England (8), the effect of attending a high-quality preschool on a child’s literacy and numeracy at age 11 can equal or surpass that of other factors, including primary school quality and early developmental problems.