Factors in Persistence Versus Fadeout of Early Childhood Intervention Impacts

摘要 该项目试图了解成功的幼儿教育是否、对谁以及如何产生影响。 准备干预在儿童的整个发展过程中都是持续的。过去看似矛盾的发现 关于所谓的淡出激发了这个项目。我们的首要目标是评估两个早期阶段的长期影响 在北卡罗来纳州大规模实施了儿童干预措施。我们将评估治疗的影响 More at Four(MF)幼儿园前(Pre-k)干预计划和Key上的Smart Start(SS)倡议 青春期的健康、教育和幸福结果(例如，高中毕业、青少年怀孕)。 Mf是一项针对低收入儿童的高质量学前教育项目，以入学人数为衡量标准。 在社区一级，作为国家对15个县中每100个县的人均拨款 好几年了。这笔资金分配给了100多万名儿童。MF和准随机的逐步铺开 差异导致了独立于儿童的年份和县之间的资金差异很大 特点。社会保障局向各县提供资金，以改善儿童保育的质量和获得儿童保健、健康检查、 从出生到5岁的所有儿童的其他服务。SS是使用相同的逐步推出和 资金公式用于MF，但在不同的年份，在计划中提供独立的随机变量 进入。在测试长期影响时，我们还将测试影响大小是否随着家庭收入的变化而变化， 民族、双语学习状况等特点。我们的第二个目标将检验这一假设 影响效应的大小随孩子进入的学校环境而变化，称为持续环境 假设。我们假设，经历了入学准备干预并随后被安置的儿童 进入一个可持续的环境(例如，教室里有更高比例的同龄人愿意学习或 经历过早期干预，受过促进学习培训的课堂教师)会有 更持久的影响(缓和效应)。第三个目标是检验可持续环境假说 通过实验操作在积木(BB)中进行数学干预研究。BB Pre-k 数学课程侧重于计算和几何/空间推理。2005年，波士顿的42所学校和 有1,375名儿童的水牛被随机分配到三种情况之一：(1)BB；(2)BB加小学 学校教师培训，以支持早期影响的持续性；或(3)控制(照常进行学前教育)。我们将添加 衡量孩子随后的同伴、教室和学校环境，以及青少年的结果， 要测试BB计划在与后续的 可持续的环境。最终目标是整合研究结果以得出概括性的结论。 关于早期干预的影响在整个发展过程中持续或消失的情况，例如 无论是发展科学家还是教育政策制定者。

ABSTRACT This project seeks to understand whether, for whom, and how the effects of successful early childhood school readiness interventions are sustained across a child’s development. Past seemingly contradictory findings about so-called fadeout motivate this project. Our first aim will assess the long-run effects of two early childhood interventions implemented at scale in North Carolina. We will estimate the treatment impacts of the More at Four (MF) pre-kindergarten (pre-k) intervention program and the Smart Start (SS) Initiative on key health, education, and well-being outcomes during adolescence (e.g., high school graduation, teen pregnancy). MF is a high-quality pre-k program for low-income children, measured at the individual level by enrollment and at the community level as funding allocations per capita by the state to each of 100 counties in each of 15 years. This funding was allocated to over one million children. Gradual rollout of MF and quasi-random variations led to sizeable differences in funding across years and counties that are independent of child characteristics. SS provides funding to counties to improve quality and access to child care, health screenings, and other services for all children from birth to age 5. SS was implemented using the same gradual rollout and funding formula used for MF, but in different years, providing independent random variation in program access. As we test long-run impact, we will also test whether effect sizes vary as a function of family income, ethnicity, dual-language learning status, and other characteristics. Our second aim will test the hypothesis that the impact effect size varies with the school environment that a child enters, called the Sustaining Environment hypothesis. We hypothesize that children who experience a school readiness intervention and are then placed into a sustaining environment (e.g., classroom with higher proportion of peers who are ready to learn or have experienced the early intervention, classroom teacher who has been trained to promote learning) will have longer-lasting effects (a moderation effect). The third aim is to test the Sustaining Environment hypothesis through experimental manipulation in the Building Blocks (BB) mathematics intervention study. The BB pre-k math curriculum focuses on numeracy and geometric/spatial reasoning. In 2005, 42 schools in Boston and Buffalo with 1,375 children were randomly assigned to one of three conditions: (1) BB; (2) BB plus elementary school teacher training to support persistence of early impact; or (3) control (pre-k as usual). We will add measures of the child's subsequent peer, classroom, and school environments, as well as adolescent outcomes, to test the hypothesis that the BB program has longer-lasting effects when implemented with a subsequent sustaining environment. The final aim is to integrate findings across studies to reach generalizable conclusions about the circumstances in which early intervention has impacts that persist or fade across development, for both development scientists and education policy makers.