Healthy & ready to learn: examining the efficacy of an early approach to obesity prevention and school readiness.

Winter, Suzanne M. ; Sass, Daniel A.

The collision of the childhood obesity epidemic with pressure to achieve high academic standards is of serious concern in the United States. Growing numbers of low-income, minority children face double jeopardy as alarming obesity rates further widen existing achievement gaps. Health and education disparities persist when children enter kindergarten lacking fundamental school readiness skills and are also at risk of obesity. The goal of this study was to address serious gaps in research by examining the efficacy of an innovative program, Healthy & Ready to Learn, an early approach to obesity prevention and promotion of school readiness. The study targeted low-income, predominantly Latino preschoolers who are particularly at risk of health and educational disparities. The pretest-posttest, quasi-experimental study involved 405 children, ages 3 to 5 years, enrolled in four matched Head Start centers. To ensure rigorous assessment, the study used a battery of objective and validated instruments as direct measures of child outcomes. Using multilevel modeling, several linear growth models were conducted with participant's growth at Level-1 (i.e., time) and subject-level variables at Level-2 (i.e., treatment, gender, age, & body mass index classification) for each outcome variable of interest. Results revealed statistically significant improvements in growth (i.e., height), gross motor skills, physical activity levels, and receptive language development when comparing the treatment and control conditions. These promising results suggest that the Healthy & Ready to Learn program has potential as an early approach to improve children's health and, simultaneously, enhance their trajectory toward better academic performance.

Keywords: child, obesity, preschool, education, readiness

**********

The collision of the childhood obesity epidemic with pressure to achieve high academic standards is of serious concern in the United States. Growing numbers of low-income, minority children face double jeopardy as alarming obesity rates further widen existing achievement gaps. Health and educational disparities persist when children enter kindergarten lacking fundamental school readiness skills and are also at risk of obesity (Currie, 2005; Datar, Sturm, & Magnabosco, 2004; National Institute of Child Health and Human Development [NICHD], 2000). To address this burgeoning problem, a new paradigm for school readiness is emerging under the stewardship of leading national agencies.

NICHD, the Centers for Disease Control and Prevention (CDC), and other agencies have articulated a broadened perspective of school readiness in recognition that strategies to improve children's health also may increase their capacity for academic achievement. From this basic premise have sprung calls for interdisciplinary approaches to investigate school readiness and examine potential relationships between health and development from different perspectives. The major goal is to identify effective school readiness and health promotion strategies to improve trajectories for children at high risk of school failure (NICHD, 2009).

SCHOOL READINESS AND THE ACHIEVEMENT GAP

Achievement gaps begin early and tend to widen across time, especially for children who are particularly vulnerable. Minority children in poverty are disproportionately at risk of entering school ill-prepared to achieve academic success. It has been estimated that by the year 2030, one fourth of U.S. students will be Latino (NICHD, 2000). Latino children, the predominant group in the current study, are the largest and fastest-growing minority group in the United States. Almost 11% of children younger than age 5 years are Latinos, and the majority of these children are of Mexican descent (U.S. Census Bureau, 2009). Closing the gap may be difficult for Latinos and other low-income, minority children who start school lagging behind their peers in language, literacy, cognitive processing, and other fundamental school readiness skills necessary for academic success (Lee & Burkam, 2002; Mashburn, Justice, Downer, & Pianta, 2009). Lack of access to high-quality early education programs, communication barriers, and cultural differences also can have a negative impact on the school readiness of low-income, minority children (Collins & Ribeiro, 2004; Pianta, 2007).

Promoting school readiness is a powerful strategy for narrowing the achievement gaps among young children. Strong, corroborating evidence supports the importance of providing high-quality environments and enriching experiences to improve children's early learning and skill development, especially when children are economically disadvantaged (Mashburn, 2008; NICHD Early Child Care Research Network, 2005). Early language and literacy skills, specifically phonological awareness, print knowledge, and oral language, are strongly associated with comprehension and other reading skills critical for overall academic achievement (Lonigan, Schatschneider, & Westberg, 2008). As a result, language and literacy skills have garnered national attention throughout the United States as essential components of school readiness; consequently, these skills have become a major focus of quality early childhood programs (Mashburn et al., 2009).

Obesity and School Performance

Achievement gaps of children are compounded by alarming rates of obesity, which has struck hardest among low-income, minority children already at high risk for poor academic performance. One in seven economically disadvantaged preschoolers is obese when they enter school. Obesity is more prevalent in Latino preschoolers, at 18.5%, compared to 12.6% of White, preschool-age children (CDC, 2009c). Obesity, a serious public health risk, contributes to health and educational disparities among the growing population of Latino and other minority children in the United States. Research has reported a consistent association of childhood obesity with poor academic performance (Taras & Potts-Datema, 2005a, 2005b). This evidence suggests that obesity can jeopardize children's development, further widening the achievement gaps for minority children (CDC, 2008; NICHD, 2000).

Childhood obesity is a complex phenomenon with many contributing factors, including genetic predispositions, behavior, and environmental factors. Although genetic susceptibility may exist for some children, experts recommend that interventions focus on behavioral factors, such as eating and exercise habits, and also on creating supportive environments to sustain good health habits once these behaviors have been acquired. Parents and teachers are vital role models who can influence children's behaviors. Children may follow the example that adults set regarding exercise habits, food preferences, and time spent engaged in sedentary activity. Adults also influence the environments children experience, including food access and opportunities for physical activity (CDC, 2009b). The complexity of childhood obesity has led many experts to recommend comprehensive approaches aimed at improving children's social and physical environments. It also has been recommended that strategies be aimed toward affecting multiple layers of the environmental context influencing children, including homes, schools, and the community, to create optimal support for children's health and obesity prevention (Campbell & Hesketh, 2007).

Lack of Early Obesity Prevention Research

Although large strides have been made in understanding the language, cognitive, social, and behavioral skills that children need to be successful in school, serious gaps in the research bases exist regarding childhood obesity. It is becoming very clear that obesity contributes to health disparities and educational inequities, especially for young minority children and those in poverty. Yet serious gaps remain in the accumulated research bases regarding early approaches to obesity prevention. Prevention programs for children younger than age 5 are very limited, and few have been examined empirically to establish efficacy and to prove effectiveness (Summerbell, Waters, Edmunds, Brown, & Campbell, 2006). However, reliable data sources suggest that children are transitioning into overweight status at younger ages, and overweight is an absorbing state, meaning that once a person becomes overweight, reversing this trajectory is extremely difficult.

Without effective programs of early intervention, statistics indicate a trend toward obese children becoming obese adults (CDC, 2009c; Fletcher, 2007). Thus, experts have recommended a focus on prevention by establishing a healthy lifestyle of nutritious eating habits and physical activity early in a child's life (American Academy of Pediatrics, 2007). Although children are becoming obese at increasingly younger ages, translational research to identify effective strategies and prevention programs aimed at young children is exceedingly sparse. Additional research is needed to identify successful and sustainable interventions (Neuhauser et al., 2007; Timmons, Naylor, & Pfeiffer, 2007).

Purpose

School readiness is a complex, multidimensional concept in which children's health, development, and experience are interrelated. Concerns have been raised about the potential effects of rising childhood obesity rates on school performance. Of particular concern are increasing rates of overweight during early childhood years and insufficient physical activity among children throughout childhood (Dockett & Perry, 2009; NICHD, 2009).

The current study addresses the need to investigate early obesity prevention from a developmental perspective. Programs aimed at reversing obesity have not been highly successful, yet few studies have investigated strategies for early prevention. Consequently, the current study was conducted to help fill a serious gap in the research bases on childhood obesity. Moreover, the current research study is among the first to concurrently examine health and school readiness constructs embedded in a single program. To address these gaps, the purpose of the current study was to examine the efficacy of an innovative program called Healthy & Ready to Learn, an early approach to obesity prevention and the promotion of school readiness. The current study targeted low-income, predominantly Latino preschoolers who are particularly at risk for health and educational disparities.

Research Hypotheses

For the current study, factors related to health were designated as primary variables of interest, with variables related to school readiness considered of secondary interest. Related to health characteristics, we hypothesized that participants in the treatment group would demonstrate significantly less growth/change from pretest to posttest in body mass index (BMI) and weight compared to the control group. Although preschool-age children are expected to gain weight as they grow, we hypothesized the program would help children avoid excessive weight gain. Associated with healthy activity, we hypothesized that participants in the treatment group would have significantly greater growth/improvement from pretest to posttest in motor development, as measured by the Brigance Diagnostic Inventory of Early Development--II (Brigance; Glascoe, 2004) and System for Observing Fitness Instruction (McKenzie, Sallis, & Nader, 1991) when compared to the control group. It has been argued that motor development is an important health correlate that might be useful in tracking childhood obesity (Timmons et al., 2007). Motor skill development enables children to be physically active, a condition associated with lower obesity rates and greater academic success (Carlson et al., 2008; Trost, 2008). Moreover, it has been postulated that good motor skills may have a direct effect on reducing body fat (Reilly et al., 2006).

Given that the Healthy & Ready to Learn program also sought to improve children's school readiness, we investigated the program's impact on children's language development, a key school readiness indicator. It was hypothesized that participants in the treatment group would display significantly greater improvement in receptive language development from pretest to posttest than the control group. Receptive language was selected as the secondary variable because it is a robust indicator of cognitive and overall development. Language development also is considered a strong predictor of school readiness and later academic performance (Nelson, Nygren, Walker, & Panoscha, 2006).

METHOD

Sample

Data for the current intervention study were collected from four Head Start centers matched on the basis of geographical location, size of center, and demographic characteristics of families served. The centers were located within a 1-mile radius of each other in a high-poverty, low-income neighborhood in a large metropolitan city located in South Texas. The centers chosen served families that were similar in ethnicity, income, and level of parental education. Nearly two thirds of the families reported annual income averages below $20,000, and less than 5% of parents reported earning a college degree from a 4-year institution. Each Head Start center was administered by the same agency and used a common curriculum, teacher professional development, and parent training program. Teachers in these centers typically reported an education attainment of less than a bachelor's degree.

From the four matched Head Start centers, 405 children ages 3 to 5 years were sampled. Participants were predominantly (95%) Latino of Mexican American origin, with English often (67%) the preferred language spoken in homes. No statistically significant gender differences between males ([n.sub.T] = 115, [n.sub.C] = 96) and females ([n.sub.T] = 91, [n.sub.C] = 103) in the treatment (T) and control (C) groups were revealed, [chi square] (1, N = 405) = 2.333, p = .127. Age differences between the treatment ([M.sub.T] = 49.93 months, SD = 7.45) and control ([M.sub.C] = 50.11 months, SD = 7.20) groups were also not uncovered, t(1,403) = 0.244, p = .807. Results revealed relatively few differences between the two groups at pretest (see [[beta].sub.01] in Table 1) on each outcome variable of interest. The exceptions were on the Brigance, with the treatment group scoring significantly below the control group by .63 units on the Nonlocomotor and .97 units on the Locomotor.

Groups were also comparable in terms of BMI classification at pretest for the treatment and control: underweight (0.5% & 1.0%, respectively), healthy weight (59.3% & 60.2%, respectively), overweight (17.6% & 15.0%, respectively), and obese (19.6% & 18.9%, respectively). Although prevention studies typically focus on children whose weight is not yet problematic, the unprecedented prevalence of overweight among children (10%-15%) has led to new sampling recommendations. It has been suggested that samples include children who are already overweight and in need of effective treatment to prevent adult obesity (Dietz & Gortmaker, 2001). Consequently, we did not eliminate children from our sample who fell into overweight and obese categories, although participants' BMI category was considered in the statistical analyses.

Description of Intervention

A pilot study was conducted aimed toward testing Healthy & Ready to Learn, an integrated program of research-based strategies to prevent obesity and enhance the school readiness of preschool-age children. The program was previously developed by the first author's research team and includes strategies for obesity prevention based upon guidelines and recommendations of experts, such as those issued by the CDC and the Institute of Medicine (Campbell & Hesketh, 2007). The Healthy & Ready to Learn program includes school readiness strategies based upon recommended practices of professional organizations (National Association for the Education of Young Children & International Reading Association, 1998; NICHD Early Child Care Research Network, 2005). Pilot data were collected from 42 participants to evaluate treatment feasibility. Program modifcations were conducted during this time to address methodological and application issues. From this pilot study, the protocol was refined to examine the efficacy of the Healthy & Ready to Learn program.

Healthy & Ready to Learn is an innovative program that uses a multilevel approach recommended by the Institute of Medicine, directing obesity prevention strategies at three groups of key stakeholders: children, parents, and teachers (Kumanyika, Kraak, Liverman, & Meyers, 2007). This strategy was adopted because research suggests that obesity prevention programs using multilevel approaches are most effective, because they engage layers of society surrounding children, such as schools and homes, and thereby support individual behavioral change (Campbell & Hesketh, 2007). Multilevel approaches based upon ecological theory (Bronfenbrenner & Morris, 1998) also have been widely used by school readiness focused programs (Dockett & Perry, 2008, 2009).

The current intervention program seeks to attain changes in children's behavior rather than focusing solely on an educational approach. Focusing on acquisition of positive health behaviors and establishing healthy habits in children have been recommended for obesity prevention to promote lasting, sustainable change to help children avoid lifelong obesity (Cole, Waldrop, D'Auria, & Garner, 2006; Gibbons, 2007). Although changing the behaviors of parents and teachers may be an indirect benefit of the program, the major goals of the intervention are changing children's health behaviors and increasing children's engagement in activities associated with school readiness. A unique feature of the program is the alignment of the curriculum across home and preschool contexts to increase the potential for children to receive maximum support for acquiring health behaviors and school readiness skills. By implementing the Healthy & Ready to Learn program, parents and teachers promote health and school readiness and provide a more supportive home and school context for children (Gibbons, 2007).

Healthy & Ready to Learn is also a comprehensive and multifaceted program aimed at multiple behaviors, simultaneously. This type of approach has been recommended as an effective strategy for programs seeking to prevent childhood obesity (Krishnamoorthy, Hart, & Jelalian, 2006). The program is designed to be implemented in home and school, two contexts that are highly influential in the early development of children. The two major goals for the program are to (1) prevent childhood obesity and (2) improve children's school readiness. To achieve these goals, the program uses a set of evidence-based intervention strategies aimed at five specific behaviors that influence children's growth and development: (1) reduced sweetened beverage and sweets consumption, (2) increased fruit and vegetable consumption, (3) increased amount of physical activity, (4) decreased screen time for entertainment (including television watching, computer use, and video game play), and (5) increased engagement in school readiness activities. The program was implemented in the school and family contexts by providing child-focused activities and materials aimed at improving school readiness and preventing obesity in children, and by training parents and teachers to effectively implement these activities with children. The three major components of the Healthy & Ready to Learn program, and the procedures used to implement these components, are outlined below.

Child activities. The program aligned home and school by providing parallel curricula consisting of a set of children's literature and corresponding, complementary activities. The format of the activities was designed to scaffold the health and school readiness promotion actions of parents and teachers. The semistructured format gave simple instructions and provided vocabulary to use, as well as questions and statements to help the adults provide a high-quality learning experience for children. This format was adopted as a way of aiding treatment integrity across parents and teachers who might vary in education levels and knowledge of how to promote children's health and school readiness. The format was based upon social cognitive theory (Cole et al., 2006) and provided a framework for parents and teachers to help ensure their success in engaging children in the activities. Management strategies also were embedded in the activities to help parents and teachers use authoritative interaction techniques recommended for facilitating positive adult-child interactions (Baumrind, 1991). Research suggests that these techniques are influential not only in promoting social development, but also in affecting the development of children's positive health habits, especially those related to eating (Bante, Elliott, Harrod, & Haire-Joshu, 2008; Rhee, 2008).

Child activity packets included simple, illustrated instructions in English and Spanish, children's books, and all materials needed for parents and teachers to engage the child. Parents and teachers read selected children's books on a health-related theme and engaged children in enjoyable activities to help them learn to make smart health choices. The activities were designed to improve children's language and literacy skills, while encouraging their acquisition of good health habits. Other studies have suggested that using children's literature to introduce concepts and encouraging storybook reading in preschool and home settings are also effective for promoting school readiness (Stanulis & Manning, 2002). Joint storybook reading promotes active reciprocal engagement of parent-child dyads across cultural contexts and may enhance socioaffective and cognitive development, as well (Cameron & Pinto, 2009).

A key intervention component involved increasing children's engagement in moderate to vigorous physical activity to 60 minutes, daily. Recent evidence suggests physical activity is highly associated with fitness and cognitive benefits that improve children's chances for later academic success (Carlson et al., 2008; Trost, 2008). Teachers and parents are trained to implement curricular activities targeting specific gross motor skills and designed to encourage movement. Equipment, materials, and suggested guidance strategies are provided to facilitate children's participation in fun, play-based physical activities (some of them incorporating music).

Parent training. Rather than focusing solely on increasing the health knowledge of parents, the training emphasized motivating parents to engage in health promotion behaviors by implementing the child activities at home. Training also focused on demonstration and practice of the activities to help ensure parents' success in implementing the activities. Parents attended monthly training sessions provided in small groups that included demonstrations modeling how to implement child activities at home. Face-to-face training was delivered in English and Spanish using Healthy & Ready to Learn videos and PowerPoint presentations. Media-based training approaches have been reported to be successful in other studies (McGarvery et al., 2004). Promotoras (i.e., trained Latino neighborhood leaders) assisted in training and provided an important cultural interface between parents and researchers. Nutritious snacks and bilingual child activity packets (consisting of children's storybooks and physical activity equipment) provided incentive for parent participation. Topics included strategies for promoting school readiness, such as story reading techniques, as well as obesity prevention strategies, including the importance of increasing one's intake of fruits and vegetables, reducing the use of sweets as rewards, and finding ways to involve children in 60 minutes of daily physical activity.

Teacher training. Preschool teachers received approximately 20 hours of training through face-to-face and media-based instruction. The training increased awareness of risks associated with overweight and encouraged behaviors and practices to promote healthy lifestyles in children. The training met the needs of preschool teachers, who usually have limited knowledge about nutrition and physical activity and about how to promote good health habits in children. Obesity prevention topics included identifying alternatives to using sweets as rewards, strategies for encouraging healthy eating habits, and strategies to increase physically active play opportunities for children. School readiness topics included strategies for encouraging print recognition and phonemic awareness during storybook reading, extending children's vocabulary, and using questioning to stimulate conversational language. As incentives, the teachers received activity plans, children's storybooks, and physical activity equipment to keep for classroom use.

Research Design Protocol

Data were collected using a pretest-posttest, quasi-experimental design with matched sites from four Head Start programs. These sites were matched on geographical location, size of center, and family demographics to increase internal validity. From the four centers, two centers were randomly designated as treatment sites, and the remaining two served as the control group. During the study, the treatment and control groups followed their standard protocol: a program consisting of a general framework curriculum, teacher professional development, and parent education. As a supplemental booster program, the treatment group also received the 24-week Healthy & Ready to Learn program, which consisted of six modules intended to boost children's school readiness and improve children's health as strategies for preventing lifelong obesity. It is important to note that the standard program did not include a specific focus on obesity prevention, nor did it include explicit teacher and parent strategies to intensively boost health and school readiness, as was included in the content of the Healthy & Ready to Learn program. Moreover, the Healthy & Ready to Learn program aligned the child activities implemented in the school and at home so these activities would be complementary.

Treatment Integrity

Multiple strategies were used to enhance and assess treatment integrity to increase the current study's internal and external validity. Based on pilot work conducted examining the feasibility and acceptability of program implementation with the target population, the investigators selected methods to address treatment integrity for the current study that would be practical, acceptable to participants, and capable of yielding accurate results. Several fidelity methods were triangulated to address treatment integrity based on methods recommended as practical (Elliott & Busse, 1993). First, a manualized intervention method was used whereby teachers and parents were provided with explicit, written lesson plans for each activity that included step-by-step instructions and statements or questions to facilitate children's participation and learning. Second, research staff conducted intermittent integrity checks by monitoring visits and interviews with teachers and parents to check whether each program activity was completed and implemented according to the instructions provided. Weekly interviews with individual teachers were conducted in their classrooms. Monitoring interviews with parents were conducted monthly when the next set of home activities was distributed. During teacher and parent sessions, research staff went through each step of the upcoming activities to ensure understanding. Staff answered questions and demonstrated techniques for teachers and parents to use in implementing program activities. Third, to help assess treatment integrity, teachers and parents completed self-report evaluations, which included a rating scale for each activity and open-ended questions, so they could describe their success or barriers to implementation of program activities.

These methods provided ongoing monitoring of treatment fidelity and documentation that enabled the investigators to assess treatment integrity formatively throughout the current study. The frequency of treatment integrity monitoring allowed research staff to provide immediate, corrective feedback or reinforcement of techniques through modeling, to better ensure fidelity of implementation by parents and teachers. Such systematic and multiple methods of enhancing and assessing treatment integrity have been reported as successful in similar intervention studies (Lane, Beebe-Frankenberger, Lambros, & Pierson, 2001). To decrease the possibility of contamination of current study results at all four matched sites, investigators checked with the Head Start program managers to ensure that the curriculum and programs already in place were the same at all sites and different from the intervention. Initial instructions to the program managers also included a request to notify the investigators of any programmatic changes that occurred during the current study so that investigators could identify any sources of contamination that might pose a potential threat to the current study findings. Regardless, a limitation was the lack of statistical results that documented and estimated the percent of adherence to the program by the parents and teachers. Although teachers were monitored to assess treatment integrity, the parents were not evaluated, as it was not feasible and thought to be invasive.

Outcome Measures

To provide a rigorous test of the intervention program, we used a battery of objective and validated instruments to directly measure child outcomes, rather than using less precise report measures, such as parent and teacher surveys. For treatment and control conditions, data were collected at pretest and posttest by administrators trained on each instrument and following strict protocols for administration, as recommended by the authors of each instrument. Researchers were trained on methods of assessing young children in keeping with the protection of human participants. Researchers also were required to practice administration under supervision and in field settings to ensure their competency in administering each instrument prior to data collection. The following section describes each measure used in the current study.

Height, weight, and body mass index (BMI). To assess changes in children's physical growth characteristics, standardized protocols and calibrated equipment were used to measure children's weight in kilograms and height in centimeters, and these measurements were used to calculate each child's BMI. The BMI evaluates the ratio (BMI=weight [kg] / height [[m.sup.2]]) between weight and height to estimate a healthy body weight based on one's height.

In absence of a standard definition of excess body fat in children to which a child's BMI value might be compared, the CDC have constructed BMI-for-age growth charts to use in screening children. Categories are as follows: underweight (BMI < 5th percentile), normal (5th percentile < BMI < 85th percentile), overweight (85th percentile < BMI < 95th percentile), and obese (BMI = or >95th percentile) (CDC, 2009a; Ogden, Carroll, & Flegal, 2008). Prior to administration, researchers collecting these measurements received 3 hours of training on the protocols, including practice administrations. To ensure accuracy of the BMI measurement, a calibrated Tanita scale was used, and each child was measured twice. If there was a difference of .25 kilograms or greater for weight, a third measurement was obtained to ensure an accurate measure. The same practice occurred if a height measurement differed by more than .25 centimeters.

System for Observing Fitness Instruction (SOFIT). The SOFIT (McKenzie, Sallis, & Nader, 1991) is a validated tool that has been widely used in research studies to objectively assess children's activity levels during teacher-led physical education instruction. For the current study, SOFIT was used to measure the activity levels of children during guided physical activity sessions conducted at the preschool as part of the Healthy & Ready to Learn program. Unlike the other measures used in the current study, SOFIT data were only collected from a random sample of 131 participants, given its extensive time demands. This practice is in accordance with the protocol for the instrument established by the validation studies conducted by the authors of the instrument.

A random sample was selected based on the Table of Random Numbers, with provisions to ensure equal sample by gender. From the entire sample of interest, two randomly selected children were observed every 20 seconds (10-second observation/10-second recording interval), using momentary time sampling to assess the intensity of their physical activity. Preprogrammed audiotape recorders were used to ensure the accuracy of the time sampling intervals. Codes 1 to 4 described the observed student's body position (i.e., lying down, sitting, standing, walking), with a score of 5 (very active) identified when the student was expending more energy than during normal walking. These activity codes were validated by heart rate monitoring in studies of young children (McKenzie, Sallis, & Nader, 1991) and also were validated for use with predominantly Hispanic populations of children similar to the sample for the current study (Pope, Coleman, Gonzalez, & Health, 2000). Validity and reliability studies of this instrument have been conducted in more than 1,000 school sites across the United States (McKenzie, 2002).

To establish reliability prior to administration, researchers were trained on the protocol for the SOFIT assessment for a total of 6 hours. Three hours of the training was conducted in the research lab using established protocols and videotaped clips. The remaining 3 hours was spent in actual field observation practice. Seven pairs of raters established interrater reliability before administering the SOFIT in the current study. Raters were paired and administered the SOFIT to the same child, simultaneously. Interrater reliability rates ranged from 90% to 92% to establish an overall interrater reliability rate of 91%.

Brigance Diagnostic Inventory of Early Development--II. The Brigance (Brigance; Glascoe, 2004) is a widely used instrument for evaluating children's developmental and early academic skills. Changes in gross motor skill development were measured using this assessment, which derives a sampling of developmental and early academic skills and includes updated standardization and norms that reflect rapid changes in development. Age-appropriate Nonlocomotor and Locomotor total scores were utilized. The Nonlocomotor subscale measures motor abilities, such as standing and jumping, whereas the Locomotor subscale measures such movements as crawling, walking, and running. The Brigance has been extensively evaluated for reliability and validity (Glascoe, 2004). To enhance reliability prior to administration, researchers received 6 hours of training from a trainer who was qualified and experienced in administration of the Brigance. Whole-group instruction consisted of 3 hours with the trainer. Small-group training (two to three researchers at a time) was conducted for an additional 3 hours.

Before administering the assessment in field, a performance-based assessment was conducted by the trainer in which individual researchers administered the Brigance to the trainer. The performance-based assessment was repeated, if necessary, until the researcher was able to administer the Brigance accurately, according to the specifications of the qualified trainer. The trainer supervised the administration of the test in the field sites for the current study to ensure adherence to standardized test administration protocols. Given that simplicity of scoring the Brigance (the participant can either perform the task or not), no interrater reliability was obtained. However, multiple researchers typically were available to direct and score the objectives to ensure agreement.

Peabody Picture Vocabulary Test III (PPVT-III). The PPVT-III (Dunn & Dunn, 1997) requires participants to point to one of four pictures that best represents the meaning of a verbally presented stimulus word. This standardized measure of receptive vocabulary is moderately correlated to cognitive abilities and is significantly correlated with children's Verbal Scale and Full Scale scores on Wechsler Preschool and Primary Scale of Intelligence--Revised (Wechsler, 1989). This measure also has strong internal consistency reliability, with coefficients consistently larger than .90 for this age group (Dunn & Dunn). To measure receptive vocabulary, the current study used the standardized scores ([mu] = 100, [sigma] = 15) of children's verbal intelligence, given that they are adjusted for age and provide commonly used standards and norms.

As was done with SOFIT and Brigance, a qualified trainer experienced in the administration of the PPVT conducted 6 hours of training on the proper administration to increase reliability. Three hours of training consisted of whole-group instruction with the trainer and another 3 hours of training in small groups. To ensure adherence to standardized administration procedures, each researcher was required to pass a performance-based assessment demonstrating competence in administering the instrument to the qualified trainer. For the current study, the trainer supervised the researchers in assessing children at the study sites to further ensure proper administration.

Missing Data

The impact of missing data is well-known and extremely problematic with longitudinal data (Newman, 2003), as participants are often missing data for numerous reasons. Missing data is also a problem when comparing statistical models, as adding variables to a model with missing data creates different sample sizes, and thus not perfectly nested models. Traditional missing data methods (e.g., listwise deletion, pairwise deletion, etc.) often produce biased parameter estimates and reduced statistical power, whereas modern procedures (e.g., multiple imputations and maximum likelihood estimation procedures) are robust to missing data when these data are missing completely at random or missing at random (Enders & Bandalos, 2001). For the reasons above, missing data were treated using multiple imputations (MI) via the Markov Chain Monte Carlo method within SAS (SAS, Version 9.1.3, Littell, Milliken, Stroup, Wolfinger, & Schabenberger, 2006). Regardless of the percent of missing data, missingness was assumed to be missing at random (see Little & Rubin, 2002), given that missing data were dependent upon other variables in the dataset. For the current study, 100 datasets were generated using 300 burn-in iterations before the first imputation in each chain and 600 iterations between imputations.

Multilevel model analyses were conducted on each of the 100 complete datasets, each with 405 participants. Imputations were only required for the outcome variables, as data were never missing for one's treatment condition, age, or gender. Each participant had at least one of the two time points for each outcome variable of interest. The percent of missing data over time for the outcome variables varied from 23% to 35% (M = 27%, SD = 3). SOFIT data, which were only collected from a random sample of 131 (32% of the 405) participants, had missing data on one time point for about 17% (n = 22) of cases. Although, statistically, one could impute missing data for the entire sample (n = 405), the authors selected a more conservative approach by only imputing missing data for those participants with SOFIT data.

Analytic Approach

Using SAS PROC MIXED, several two-level multilevel linear growth analyses were conducted with participant's growth rates at Level-1 (i.e., time) and subject-level variables at Level2 (i.e., treatment, gender, age, & BMI classification) for each outcome variable of interest. For each analysis conducted, two separate models were fit sequentially to explore the impact of adding additional predictor variables. Results revealed that only treatment status significantly predicted student growth rates. Neither gender, age, nor BMI classification predicted or moderated these outcomes. Therefore, only treatment status was used as a student-level variable.

Model 1, an unconditional linear growth model, was fit first to examine the amount of dependency in the outcome variables and establish baseline statistics related to changes in subject growth (i.e., slopes) and starting points (i.e., intercept or initial growth between the pretest and posttest). At Level-l, this unconditional model can be written as

[[??].sub.ij] = [[pi].sub.oj] + [[pi].sub.1j][(Time).sub.ij] + [r.sub.ij], (1)

which corresponds to the linear growth model. For this model, [pi] represents the Level-1 (within subject) variation for subject i(i = 1,2 ..., 405) over j(j = 1 to 2) time points. Level-2, which evaluates variation in the linear growth model unrelated to any subject-level covariates, substitutes [[pi].sub.0j] = [[beta].sub.00] + [u.sub.0j] and [[pi].sub.1j] = [[beta].sub.10] + [u.sub.1j] into Equation 1. For this Level-2 model, [[beta].sub.00] represents the intercept (or initial status) and [[beta].sub.10] represents the average slope (or average growth curve) across all participants.

Model 2 evaluated whether students' treatment status significantly predicted their initial status and growth rates on each outcome variable. This Level-2 model is an extension of the Level-2 unconditional model and is written as

[[pi].sub.0j] = [[beta].sub.00] + [[beta].sub.01] (treatment) + [u.sub.0j] (2)

and

[[pi].sub.1j] = [[beta].sub.10] + [[beta].sub.11] (treatment) + [u.sub.1j] (3)

As seen in Equation 2, the treatment status variable is used to predict that group's initial status or intercept, whereas in Equation 3, the treatment status variable predicts participant growth rates. Combining the above equations produces the following equation:

[[??].sub.ij] = [[beta].sub.00] + [[beta].sub.10][(Time).sub.ij] + [[beta].sub.01][(Treatment).sub.ij] + [[beta].sub.11] (Time)[(Treatment).sub.ij] + [u.sub.0j] + [u.sub.1j][(Time).sub.ij] + [r.sub.ij]. (4)

Collectively with this model (see Equation 4), [[beta].sub.00] is the average intercept (or starting point) across participants, [[beta].sub.10] is the average slope (or change) across all participants, [[beta].sub.01] is the difference between the treatment and control group at pretest, and [[beta].sub.11] is the difference in growth rates from pretest to posttest between the treatment and control group. For [[beta].sub.11], positive numbers reflect larger growth rates for the treatment group than the control group.

Effect Size Calculations

Effect sizes for each [[beta].sub.11] effect were calculated to determine the practical significance of the growth curve differences between the treatment and control group. Effect sizes were calculated using the following equation: d= [[beta].sub.11]/SD, where SD is the standard deviation at pretest. More detail regarding these multilevel model effect size measures can be obtained from Feingold (2009) and Raudenbush & Liu (2001). The effect size standards followed that of Cohen (1988): small, d = .20; medium, d = .50; and large, d = .80.

The variance-covariance components were also evaluated to (1) determine the amount of variation in the intercepts and slopes and (2) determine whether this variation can be explained by treatment condition status. To determine whether treatment status explains these variance components, the intraclass correlation coefficients (ICC or [??]) were calculated.

RESULTS

A two-level multilevel linear growth model was conducted for each of the seven outcome variables of interest (see Table 1). The average intercept ([[beta].sub.00]) and slope ([[beta].sub.10]) across participants were of less interest, as it estimates the average initial status (or pretest score) across groups and the overall average amount of change from pretest to posttest, respectively. For example, the PPVT results indicate that the sample mean PPVT score was 79.84 ([mu] = 100, [sigma] = 15) and that the scores increased by 3.39 units from pretest to posttest. The parameter estimate [[beta].sub.01] reflects differences between the treatment and control group at pretest. As seen in Table 1, the treatment and control group differed significantly from each other at pretest on Brigance motor skill development measures (Nonlocomotor & Locomotor), with the treatment group scoring lower than the control group.

The parameter estimates ([[beta].sub.11]) of primary interest tested whether the treatment group experienced significantly more growth (or change) compared to the control group from pretest to posttest. The only outcome variable tested of interest, but not hypothesized, was height, as this variable was primarily used to assess whether the groups differed at pretest and is connected to BMI (see Height, Weight, and BMI in the Outcome Measures section). Analyses indicated that

participants in the treatment group experienced a statistically significant increase in height, compared to the control group, from pretest to posttest with participants. In fact, the treatment group grew almost a half-centimeter ([[beta].sub.11] = .48, p = .0019) more than the control group. Treatment status also accounted for 13% of the explainable variation in student growth rates as measured by the ICC ([??]), but treatment status did not explain any of variation for initial status.

Given that the intervention was implemented to increase healthy eating and exercise, significantly lower weight and BMI scores were expected for the treatment group. Multilevel model analyses indicated, however, that the growth curves (changes from pretest to posttest) for the treatment and control group did not significantly differ for either weight or BMI. Moreover, the effect sizes were very small (d [less than or equal to] 1.051). The ICCs were also less than 1% when explaining the intercepts and slopes. This conclusion was supported with a chi-square analysis, as participants classified based on their BMI scores (underweight or health weight vs. overweight or obese) did not differ significantly between treatment and control group at posttest, [chi square](1, N = 405) = 0.120,p = .729.

The treatment group experienced significantly more growth from pretest to posttest in gross motor skills when compared to the control group on the Brigance Nonlocomotor and Locomotor scores (see Table 1). This suggests that the treatment group's strict physical activity plan did have a detectable difference on student motor skills. In fact, growth rates for the treatment group were about one unit ([[beta].sub.11] = 1.15 & 1.02) greater than the control group, with effect sizes of approximately .30 for both outcome variables. Although not statistically significant at [alpha] = .05 (p = .081) due to the smaller sample size (n = 131 vs. 405) and less statistical power, a medium effect size (d = .51) for differences in growth curves between the treatment and control group was found on the SOFIT measure of physical activity. This suggests that the growth curves were almost one-half standard deviation larger for the treatment group than the control group. The ICCs for the intercepts and slopes were very small ([??]'s [less than or equal to] 1%) on both Brigance measures. For the SOFIT, treatment status did account for a notable portion of explainable variation in growth rates ([??] = 10%), but not initial status ([??] [approximately equal to] 0%).

Using the PPVT standardized scores, participants in the treatment group experienced marginally statistical (p = .059) and practically (d = .18) significant growth from pretest to posttest when compared to the control group. These analyses imply that not only did implementing a treatment designed to improve physical health not impair the students' receptive language, but the treatment actually had a slightly positive impact. These results are probable given that learning components and other curricular activities were integrated into the program. Based on the ICCs, treatment status accounted for less than 1% of the variability in the initial status and growth curves. Collectively, these results suggest that though the treatment significantly increased gross motor skills, this increase may have indirectly influenced students' receptive language development, as measured by the PPVT.

DISCUSSION

The findings of the current study support the view that the Healthy & Ready to Learn intervention has potential as an early approach to prevent obesity and enhance school readiness. Collectively, multilevel model analyses revealed an immediate impact on children's gross motor development, physical activity levels, receptive language and height, with less influence on BMI and weight. These findings are significant for several reasons. Limited research has empirically evaluated early approaches to obesity prevention for preschool children ages 3 to 5 (Campbell & Hesketh, 2007), the age range targeted in the current study. Experts widely agree it is critical for young children to develop good eating and exercise habits so that these behaviors will persist into adulthood (American Academy of Pediatrics, 2007). The current study revealed promising results using a relatively short intervention with a low-income, predominantly Latino sample, historically prone to adult obesity and poor academic performance (NICHD, 2000).

Recognizing that improvements in growth and physical development are associated with prevention of obesity, the current study sought to help children avoid excessive weight gain and improve gross motor development. To achieve these outcomes, the intervention promoted healthy eating and physical activity behaviors at home and preschool. One very positive finding was the statistically significant improvement in locomotor and nonlocomotor gross motor skills of children in the treatment group when compared to the control group from pretest to posttest. Despite starting behind, the treatment group's motor development surpassed that of the control group by the posttest. This is a critical finding, because good motor skill development is considered an important health correlate in children and is associated with higher levels of physical activity. From this standpoint, good motor development is thought have direct effects on body fat content in the long term, decreasing the likelihood of obesity (Reilly et al., 2006; Timmons et al., 2007).

Contradictory to our hypotheses, neither BMI nor weight analyses revealed statistically significant differences in growth, curved from pretest to posttest, between children in the treatment and control groups. Other obesity intervention studies involving children also have reported a lack of immediate effects on BMI (Fitzgibbon et al., 2005, 2006; Reilly et al., 2006). Results are mixed regarding whether follow-up studies might reveal the emergence of positive effects on children's BMI 1 or 2 years after intervention. For example, a physical activity intervention study conducted with predominantly Black children did reveal positive effects in the 1 st and 2nd years of follow-up. In contrast, a replication of the current study with Latino children failed to produce long-term results (Fitzgibbon et al., 2005, 2006). Consequently, future examination of the Healthy & Ready to Learn treatment should include follow-up data collection to determine whether long-term effects to children's BMI emerge, or to decide what other treatment components can be modified to increase the treatment effect on BMI and weight. Moreover, it is possible that a program longer than 24 weeks is needed to demonstrate improvements in BMI. It is possible that BMI measurements were not sensitive to physiological differences that might have been present, which we discuss further in the Strengths and Limitations section of this article.

Although not statistically significant, the SOFIT assessment findings revealed a positive trend toward increased engagement in moderate to vigorous physical activity for children in the Healthy & Ready to Learn program. Perhaps most important, the SOFIT produced the largest effect size and arguably the most promising findings. High activity levels are related to improved gross motors skills; consequently, the investigators expected to find higher activity levels for children in the treatment group. Activity patterns of preschoolers, however, are not well understood, and research has yet to reveal optimal levels of physical activity for children in this age range (Timmons et al., 2007). Moreover, many environmental factors influence children's physical activity. For example, the type of environment, equipment, and teaching style can influence levels of physical activity children exhibited (Bower et al., 2008). Future studies could provide insight regarding the influence of other variables, and longitudinal studies might reveal whether activity levels increase with age for children who participated in the Healthy & Ready to Learn program.

Related to school readiness, there is reason to be cautiously optimistic regarding the impact of the Healthy & Ready to Learn program. Strong associations between physical development, physical activity, and academic achievement have been reported (e.g., Carlson et al., 2008). Consequently, evidence of improvements in physical growth (i.e., height) and gross motor development reported in the current study appear to support the view that the intervention produced benefits that might positively affect school readiness and later school performance. The investigators evaluated children's receptive language using the PPVT because receptive language is related to reading comprehension and is highly predictive of later academic success (Nelson et al., 2006). In the current study, the PPVT scores yielded marginally significant gains from pretest to posttest at the .06 level for the treatment group when compared to the control group. Although positive, these findings suggest that in future studies, collecting follow-up data as children enter kindergarten and 1st grade may be critical for determining whether there are any long-term benefits to children's school readiness that result in better school performance. Economically disadvantaged children, such as those in the current study sample, often begin school lagging behind their peers because they have experienced less interactive reading and early language stimulation at home (Senechal & LeFevre, 2002). This finding was replicated here, as our sample's PPVT scores (M = 79.84) were significantly below the average population scores. Thus, any positive effects related to school readiness and obesity prevention are certainly significant and suggest, at the very least, that the Healthy & Ready to Learn program did not impede school readiness development.

Although not stated as one of the primary or secondary hypotheses, differences in growth curved from pretest to posttest on height also were tested between the control and treatment groups. Anthropometric data revealed greater height gains among children in the treatment group compared to the control group from pretest to posttest. Height is an important marker of childhood growth; consequently, this finding is very encouraging. Height is primarily attributed to proper childhood nutrition and is associated with lower risks of cardiovascular disease and diabetes, conditions often associated with obesity (Davey Smith et al., 2000; Lawlor, Ebrahim, & Smith, 2002). Conversely, malnutrition can impede growth, psychomotor skills, and cognitive development affecting school readiness and academic performance. Malnutrition can also result in obesity due to large intake of low nutrient, high-calorie foods, which is a common dietary practice among Latino children from low-income households and other economically disadvantaged groups of children in the United States (Mier et al., 2007; Pollitt, 2000).

The Healthy & Ready to Learn program's parent and teacher training emphasized the importance of good nutrition; for example, various fruits and vegetables were introduced in child activities. Children from age 2 years to puberty have the potential to grow in height at an average rate of approximately 21/2 inches (6 centimeters) per year (CDC, 2010). Thus, the 24-week intervention had the potential to improve children's nutrition and eating habits, which might have resulted in a positive impact on growth (height) and development. Of course, additional research is needed to substantiate this finding and provide biological markers justifying this increased growth (Barcelo-Batllori & Gomis, 2009).

Strengths and Limitations

Several strengths and limitations of the current study are worthy of consideration. A major strength of the current study was the use of a battery of direct measurements of child outcomes with standardized procedures, and the use of validated instruments, including assessments, that provided normative data, such as the PPVT and Brigance. This strength counters criticism of studies that rely on proxy-report questionnaires, commonly used to obtain child data from parents and teachers (Oliver, Schofield, & Kolt, 2007). However, because the current study favored a direct measurement approach, it is unknown how closely participants followed healthy eating habits at home and whether parents encouraged their children to remain active and avoid sedentary behavior. Another clear strength of the current study was that it broke ground in advancing study of early approaches to obesity prevention. Further, a comprehensive, multilevel approach for treatment was used, as has been widely recommended for preventing obesity and enhancing school readiness.

Limitations related to instrumentation include reported inaccuracies in BMI measurement of children (Ellis, Treuth, Wong, & Abrams, 2000). In absence of a standard definition of adiposity (excess body fat) in children, it is a common practice in clinical and research settings to screen children for overweight by comparing their BMI to the CDC's BMI-for-age growth charts (Ogden et al., 2008). However, limitations to the use of BMI as a measurement in growing children have been reported. A large-scale USDA Children's Nutrition Research Center study (Ellis, Abrams, & Wong, 1999), with a sample of 979 children, reported that 16% of children with normal range BMI had unhealthy fat levels when verified by more accurate testing. One fourth of children with at-risk BMI levels were found to be at normal levels. Although it is practical to use BMI, it is important to note that ethnicity, physical activity, age, and other factors may influence accuracy of children's BMI (Ellis et al., 2000).

It is also unknown what program components (e.g., intervention with students, teachers, or parents) are the most effective in promoting healthy eating, exercise, and school readiness skills. Future studies will be necessary to isolate and test individual program components. For example, it is important to determine the most effective length of treatment. Extending the treatment length will help to determine whether additional exposure to the intervention will increase the overall effect sizes. The current study only revealed small to moderate effect sizes for the 24-week intervention, which appears insufficient to produce large effects. However, small effect sizes are commonly reported in early childhood research studies conducted in naturalistic settings (NICHD Early Child Care Research Network, 2002). Moreover, effects of small magnitude during early development and skill acquisition can have positive effects in the long term (Mashburn et al., 2009).

Determining optimal treatment time and the effect of time on effect sizes is especially applicable to weight and BMI. Similarly, the intercept and slope ICCs were nearly always small, regardless of the outcome variable. Thus, although the growth curves (i.e., slopes) often differed between the treatment and control groups, the explainable variability in these growth curves was often not reduced noticeably, and variables that may explain this variability (e.g., teacher and parent variables) were not explored. Similarly, these results suggest that further research is needed to explain why some children experience changes whereas others do not.

Another limitation is that participants were not randomly assigned to treatment conditions. Although specific criteria were used to achieve a closely matched sample and relatively few differences existed at pretest, other student variables may have influenced the results that would have been eliminated with random sampling. Variables related to teacher and site effects also were not tested, due to an inadequate sample size at each level, which could potentially explain variability in the ICCs. Perhaps most importantly, although teachers and parents participated in the treatment, these effects were not evaluated. Therefore, it is feasible that certain teachers or parents were more encouraging or effective, thus contributing to more positive outcomes and explained variability in student growth curves. The final limitation of the current study was that the sample was predominantly low-income, Latino children in a limited age range. Thus, it is unknown whether these results will generalize to other student demographics. From the authors' perspective, this limitation is curtailed by the fact that disadvantaged Latino children are often most at risk for obesity and are also at serious risk of experiencing developmental lags that can impede their acquisition of school readiness skills and abilities.

IMPLICATIONS AND FUTURE RESEARCH

The current study advanced research toward filling serious gaps on two critical fronts: the fight to prevent obesity and efforts to improve children's school readiness. The findings of the current study suggest that the Healthy & Ready to Learn program has the potential to deliver positive results. Although confirmatory studies are needed, the current study provided evidence that it is feasible for preschools to use the Healthy & Ready to Learn program as a supplemental curriculum to increase physical activity, while also possibly reducing obesity and increasing school readiness. Current study findings lend support to the use of comprehensive, multilevel approaches with research-based obesity prevention and school-readiness promotion strategies supplementing existing curricula, teacher professional development, and parent training. Infusing stronger support for children's acquisition of positive health behaviors across multiple layers of influence

in home and school contexts appears beneficial. Aligning the curriculum across school and home provides continuity for children and also may provide benefits to their language development and school readiness. The success of the Healthy & Ready to Learn intervention also suggests that typical curricula and training of teachers to meet standards for Head Start programs can be improved with research-based content, specific teaching strategies, and mentoring support.

Several policy implications also can be drawn from the current study findings. From a broad perspective, public policies are needed to prioritize efforts toward early prevention of obesity to coincide with promotion of school readiness during the preschool years. Policies that provide greater emphasis on providing more intensive booster programs, such as Healthy & Ready to Learn, to preschools serving low-income, minority children are needed to narrow achievement gaps and reduce health disparities. Local policies are needed to provide resources and support aimed at overcoming obstacles that teachers and parents may encounter when implementing effective programs in community preschools and homes.

Future confirmatory research will be necessary to fully establish the efficacy and effectiveness of the Healthy & Ready to Learn program. However, the 24-week program did have an immediate and positive impact on height, gross motor skills, and physical activity levels, which are important factors associated with healthy weight in children. The program also provided a boost to children's receptive language, a developmental area that is critical to school readiness. However, longitudinal study is needed to determine whether long-term benefits to language development and school performance emerge over time. Future research studies also will be necessary to isolate the school and home components of the program and to determine the contribution of each to the child outcomes. Although the focus of the current study was to examine the effectiveness of the intervention in changing child outcomes, future studies should examine the extent to which behaviors of individual teachers and parents change as a result of the intervention. Moreover, the development of additional instrumentation--specifically, direct assessments for use in the home setting--will permit more accurate evaluation of treatment integrity and overall results. Future longitudinal research is also needed to determine whether physiological differences in treatment and control group children, as measured by the BMI, emerge over time.

In conclusion, the promising findings of the current study suggest the Healthy & Ready to Learn program is worthy of replication and further testing through group-randomized trials involving diverse samples. Follow-up data are also needed to pinpoint long-range effects that might result from participation. Nevertheless, the current study provides good preliminary evidence that the Healthy & Ready to Learn program has the potential to provide a critical boost to children's health and early development in key areas, thereby improving their chances to avoid obesity and achieve academic success.

DOI: 10.1080/02568543.2011.580211

Submitted November 2, 2009; accepted June 13, 2010.

REFERENCES

American Academy of Pediatrics. (2007, June 8). Expert committee releases recommendations to fight childhood and adolescent obesity. Retrieved from www.aap.org/advocacy/releases/june07obesity.htm

Barite, H., Elliott, M., Harrod, A., & Haire-Joshu, D. (2008). The use of inappropriate feeding practices by rural parents and their effect on preschoolers' fruit and vegetable preferences and intake. Journal of Nutrition Education and Behavior, 40(1), 28-33.

Barcelo-Batllori, S., & Gomis, R. (2009). Proteomics in obesity research. Proteomics Clinical Applications, 3, 263-278.

Baumrind, D. (1991). The influence of parenting style on adolescent competence and substance use. Journal of Early Adolescence, 11(1), 56-95.

Bower, J. K., Hales, D. P., Tate, D. E, Rubin, D. A., Benjamin, S. E., & Ward, D. S. (2008). The childcare environment and children's physical activity. American Journal of Preventive Medicine, 34(1), 23-29.

Brigance, A. H. (2004). Brigance Diagnostic Inventory of Early Development-H (IED-II). North Billerica, MA: Curriculum Associates.

Bronfenbrenner, U., & Morris, P. A. (1998). The ecology of developmental processes. In R. M. Lerner (Ed.), Handbook of child psychology: Theoretical models of human development (5th ed., Vol. 1, pp. 993-1028). New York, NY: Wiley.

Cameron, C. A., & Pinto, G. (2009). A day in the life: Secure interludes with joint book reading. Journal of Research in Childhood Education, 23(4), 437-449.

Campbell, K. J., & Hesketh, K. D. (2007). Strategies which aim to positively impact on weight, physical activity, diet and sedentary behaviours in children from zero to five years. A systematic review of the literature. Obesity Reviews: An Official Journal of the International Association for the Study of Obesity, 8(4), 327-338.

Carlson, S. A., Fulton, J. E., Lee, S. M., Maynard, M., Brown, D. R., Kohl, H. W., & Dietz, W. H. (2008). Physical education and academic achievement in elementary school: Data from the Early Childhood Longitudinal Study. American Journal of Public Health, 98(4), 721-727.

Centers for Disease Control and Prevention. (2008). Contributing factors. In Overweight and obesity. Retrieved from www.cdc.gov/nccdphp/dnpa/obesity/contributing_factors.htm

Centers for Disease Control and Prevention. (2009a). About BMI for children and teens. Retrieved from www.cdc.gov/healthyweight/assessing/bmi/childrens_BMI/about_childrens_B MI.html

Centers for Disease Control and Prevention. (2009b). Childhood overweight and obesity: Contributing factors. Retrieved from www.cdc.gov/obesity/childhood/causes.html

Centers for Disease Control and Prevention. (2009c). Obesity prevalence among low-income, preschool-aged children United States, 1998-2008. Morbidity and Mortality Weekly Report, 58(28), 769-773.

Centers for Disease Control and Prevention. (2010). CDC growth charts: United States. Retrieved from www.cdc.gov/growthcharts/Default.htm

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

Cole, K., Waldrop, J., D'Auria, J., & Garner, H. (2006). An integrative research review: Effective school-based childhood overweight interventions. Journal for Specialists in Pediatric Nursing, 11(3), 166-177.

Collins, R., & Ribeiro, R. (2004). Toward an early care and education agenda for Hispanic children. Early Childhood Research and Practice, 6(2). Retrieved from http://ecrp.uiuc.edu/v6n2/collins.html

Currie, J. (2005). Health disparities and gaps in school readiness. Future of Children, 15(1), 117-138.

Datar, A., Sturm, R., & Magnabosco, J. L. (2004). Childhood overweight and academic performance: National study of kindergartners and first-graders. Obesity Research, 2004(12), 58-68.

Davey Smith, G., Hart, C., Upton, M., Hole, D., Gillis, C., Watt, G., & Hawthorne, V. (2000). Height and risk of death among men and women: Aetiological implications of associations with cardiorespiratory disease and cancer mortality. Journal of Epidemiology & Community Health, 54(2), 97-103.

Dietz, W. H., & Gortmaker, S. L. (2001). Preventing obesity in children and adolescents. Annual Review of Public Health, 22(1), 337-353.

Dockett, S., & Perry, B. (2008). Starting school: A community endeavor. Childhood Education, 84(5), 274-280.

Dockett, S., & Perry, B. (2009). Readiness for school: A relational construct. Australasian Journal of Early Childhood, 34(1 ), 20-26.

Dunn, L., & Dunn, D. (1997). Peabody Picture Vocabulary Test-Ill. Circle Pines, MN: American Guidance Service.

Elliott, S. N., & Busse, R. (1993). Effective treatment with behavioral consultation. In J. Zins, T. Kratochwill, & S. Elliott (Eds.), Handbook of consultation service for children (pp. 179-203). San Francisco, CA: Jossey-Bass.

Ellis, K. J., Abrams, S. A., & Wong, W. W. (1999). Monitoring childhood obesity: Assessment of the weight/height index. American Journal of Epidemiology, 150(9), 939-946.

Ellis, K. J., Treuth, M., Wong, W., & Abrams, S. (2000). Body mass index: Does it accurately screen children for increased adiposity? Obesity Research, 8, 38S.

Enders, C. K., & Bandalos, D. L. (2001). The relative performance of full information maximum likelihood estimation for missing data in structural equation models. Structural Equation Modeling, 8(3), 430-457.

Feingold, A. (2009). Effect sizes for growth-modeling analysis for controlled clinical trials in the same metric as for classical analysis. Psychological Methods, 14(1), 43-53.

Fitzgibbon, M. L., Stolley, M. R., Schiffer, L., Van, H. L., Kaufer-Christoffel, K., & Dyer, A. (2005). Two-year follow-up results for Hip-Hop to Health Jr.: A randomized controlled trial for overweight prevention in preschool minority children. Journal of Pediatrics, 146, 618-625.

Fitzgibbon, M. L., Stolley, M. R., Schiffer, L., Van, H. L., Kaufer-Christoffel, K., & Dyer, A. (2006). Hip-Hop to Health Jr. for Latino preschool children. Obesity, 14, 1616-1625.

Fletcher, J. M. (2007, May). Predictors and consequences of early childhood overweight. Paper presented at the Early Childhood Longitudinal Study, Birth Cohort (ECLS-B) First Release Conference: Development from Birth through Age Two, Bethesda, MD.

Gibbons, K. (2007). Interventions for childhood overweight and obesity: A place for parenting skills. Nutrition & Dietetics, 64(3), 142-143.

Glascoe, F. P. (2004). Brigance Diagnostic Inventory of Early Development-H (IED-II) standardization and validation manual. North Billerica, MA: Curriculum Associates.

Krishnamoorthy, J. S., Hart, C., & Jelalian, E. (2006). The epidemic of childhood obesity: Review of research and implications for public policy. Social Policy Report, 19(H), 1-19.

Kumanyika, S. K., Kraak, V. I., Liverman, C. T., & Meyers, L. D. (2007). Institute of Medicine evaluation framework for assessing progress in preventing childhood obesity. The Journal of the Federation of American Societies for Experimental Biology, 21 (5), A161.

Lane, K. L., Beebe-Frankenberger, M. E., Lambros, K. M., & Pierson, M. (2001). Designing effective interventions for children at-risk for antisocial behavior: An integrated model of components necessary for making valid inferences. Psychology in the Schools, 38(4), 365-379.

Lawlor, D. A., Ebrahim, S., & Smith, G. D. (2002). Association between components of adult height and type II diabetes and insulin resistance: British Women's Heart and Health Study. Diabetologia, 45, 1097-1106.

Lee, V., & Burkam, D. (2002). Inequality at the starting gate: Social background differences in achievement as children begin school. Washington, DC: Economic Policy Institute.

Littell, R. C., Milliken, G. A., Stroup, W. W., Wolfinger, R. D., & Schabenberger, O. (2006). SAS for mixed models. Cary, NC: SAS Institute.

Little, R. J. A., & Rubin, D. B. (2002). Statistical analysis with missing data (2nd ed.). New York, NY: John Wiley.

Lonigan, C. J., Schatschneider, C., & Westberg, L. (2008). Identification of children's skills and abilities linked to later outcomes in reading writing, and spelling. In National Institute for Literacy (Ed.), Developing early literacy: Report of the National Literacy Panel (pp. 55-106). Washington, DC: National Institute for Literacy.

Mashburn, A. J. (2008). Quality of social and physical environments in preschools and children's development of academic, language, and literacy skills. Applied Developmental Science, 3(12), 113-127.

Mashburn, A. J., Justice, L. M., Downer, J. T., & Pianta, R. C. (2009). Peer effects on children's language achievement during pre-kindergarten. Child Development, 80(3), 686-702.

McGarvery, E., Keller, A., Forrester, M., Williams, E., Seward, D., & Suttle, D. E. (2004). Feasibility and benefits of a parent-focused preschool child obesity intervention. American Journal of Public Health, 94(9), 1490-1495.

McKenzie, T. L. (2002). SOFIT: System for Observing Fitness Instruction Time. San Diego, CA: San Diego State University.

McKenzie, T. L., Sallis, J. F., & Nader, P. R. (1991). SOFIT: System for Observing Fitness Instruction Time. Journal of Teaching in Physical Education, 11, 195-205.

Mier, N., Piziak, V., Kjar, D., Castillo Ruiz, O., Velazquez, G., Alfaro, M. E., & Ramirez, J. A. (2007). Nutrition provided to Mexican-American preschool children on the Texas-Mexico border. Journal of the American Dietetic Association, 107(2), 311-315.

National Association for the Education of Young Children & International Reading Association. (1998). Joint position statement by NAEYC and the IRA/Learning to read and write: Developmentally appropriate practices for young children. Young Children, 53(4), 30-46.

National Institute of Child Health and Human Development. (2000). Health disparities: Bridging the gap. Retrieved from www.nichd.nih.gov/publications/pubs/upload/health_disparities.pdf.

National Institute of Child Health and Human Development. (2009). Child Development and Behavior Branch (CDBB), NIHCD, Report to the NACHHD Council. Washington, DC: Author.

National Institute of Child Health and Human Development Early Child Care Research Network. (2002). Early child care and children's development prior to school entry: Results from the NICHD study of early child care. American Educational Research Journal, 39(1), 133-164.

National Institute of Child Health and Human Development Early Child Care Research Network. (2005). Pathways to reading: The role of oral language in learning to read. Developmental Psychology, 41(2), 428-442.

Nelson, H. D., Nygren, P., Walker, M., & Panoscha, R. (2006). Screening for speech and language delay in preschool children: Systematic evidence review for the US Preventive Services Task Force. Pediatrics, 117(2), 298-319.

Neuhauser, L., Constantine, W. L., Constantine, N.A., Sokal-Gutierrez, K., Obarski, S. K., Clayton, L., ... Syme, S. L. (2007). Promoting prenatal and early childhood health: Evaluation of a statewide materials-based intervention for parents. American Journal of Public Health, 97(10), 1813-1819.

Newman, D. A. (2003). Longitudinal modeling with randomly and systematically missing data: A simulation of ad hoc, maximum likelihood, and multiple imputation techniques. Organizational Research Methods, 6, 328-362.

Ogden, C. L., Carroll, M. D., & Flegal, K. M. (2008). High body mass index for age among US children and adolescents, 2003-2006. Journal of the American Medical Association, 299(20), 2401-2405.

Oliver, M., Schofield, G. M., & Kolt, G. S. (2007). Physical activity in preschoolers: Understanding prevalence and measurement issues. Sports Medicine (Auckland, N.Z.), 37(12), 1045-1070.

Pianta, R. C. (2007). Preschool is school, sometimes. Education Next, 7(1), 44-49.

Pollitt, E. (2000). Developmental sequel from early nutritional deficiencies: Conclusive and probability judgements. Journal of Nutrition, 130(2), 350S-353S.

Pope, R. E, Coleman, K. J., Gonzalez, B. E, & Health, E. M. (2000). Validating SOFIT during physical education in a predominantly Hispanic community. Medicine and Science in Sports and Exercise, 32(5 Suppl), 328.

Raudenbush, S. W., & Liu, X. (2001). Effect sizes for growth-modeling analysis for controlled clinical trials in the same metric as for classical analysis. Psychological Methods, 6, 387-401.

Reilly, J. J., Kelly, L., Montgomery, C., Williamson, A., McColl, J. H., Conte, R. L., ... Grant, S. (2006). Physical activity to prevent obesity in young children: Cluster randomised controlled trial. British Medical Journal, 333, 1041-1043.

Rhee, K. E. (2008). Childhood overweight and the relationship between parent behaviors, parenting style, and family functioning. The Annals of the American Academy of Political and Social Science, 615(11), 12-37.

Senechal, M., & LeFevre, J. (2002). Parental involvement in the development of children's reading skill: A five-year longitudinal study. Child Development, 73, 445-460.

Stanulis, R. N., & Manning, B. H. (2002). The teacher's role in creating a positive verbal and nonverbal environment in the early childhood classroom. Early Childhood Education Journal, 30(1), 3-8.

Summerbell, C. D., Waters, E., Edmunds, L. D., Brown, T., & Campbell, K. J. (2006). Interventions for preventing obesity in children. Cochrane Library, 3, CD001871

Taras, H., & Potts-Datema, W. (2005a). Chronic health conditions and student performance at school. Journal of School Health, 75(7), 255-266.

Taras, H., & Potts-Datema, W. (2005b). Obesity and student performance at school. Journal of School Health, 75(8), 291-295.

Timmons, B. W., Naylor, P., & Pfeiffer, K. A. (2007). Physical activity for preschool children--how much and how? Canadian Journal of Public Health, 98(Suppl 2), S122-S134.

Trost, S. G. (2008). Physical education, physical activity, and academic performance in youth. Chronicle of Kinesiology and Physical Education in Higher Education, 19(3), 33-40.

U.S. Census Bureau. (2009). The Hispanic origin population in the United States: 2007 and 2008. Retrieved from www.census.gov/Press-Release/www/releases/archives/ hispanic_origin_population/014335.html

Wechsler, D. (1989). Wechsler Preschool and Primary Scale of Intelligence--Revised. San Antonio, TX: The Psychological Corporation.

Suzanne M. Winter

Child and Adolescent Policy Research Institute, The University of Texas at San Antonio,

San Antonio, Texas

Daniel A. Sass

The University of Texas at San Antonio, San Antonio, Texas

This research was supported by the Baptist Health Foundation of San Antonio and The Max and Minnie Tomerlin Voelcker Fund.

Address correspondence to Suzanne M. Winter, Child and Adolescent Policy Research Institute, The University of Texas at San Antonio, 501 West Durango Boulevard, San Antonio, TX 78207-4415. E-mail: suzanne.winter@utsa.edu

TABLE 1
Parameter Estimates and Effect Sizes for Each Multilevel Model

                                   [[beta].sub.00]   [B.sub.10]

Height                                 100.87           3.52
Weight                                  17.33           1.42
Body mass index                         16.92           0.15
Brigance (Nonlocomotor total            14.70           1.02
  score)
Brigance (Locomotor total score)        29.80           0.95
SOFIT (Overall score)                   42.04          -0.78
PPVT (Standardized score)               79.84           3.39

                                   [[beta].sub.01]   [[beta].sub.11]

Height                                   0.51             0.48 **
Weight                                   0.02             0.17
Body mass index                         -0.10            -0.06
Brigance (Nonlocomotor total            -0.63 *           1.15 **
  score)
Brigance (Locomotor total score)        -0.97 **          1.02 *
SOFIT (Overall score)                    0.63             3.42
PPVT (Standardized score)               -0.81             2.62

                                   [ES.sub.[beta]11]

Height                                    0.09
Weight                                    0.05
Body mass index                          -0.03
Brigance (Nonlocomotor total              0.34
  score)
Brigance (Locomotor total score)          0.30
SOFIT (Overall score)                     0.51
PPVT (Standardized score)                 0.18

Note. Brigance = Brigance Diagnostic Inventory of Early
Development/II; SOFIT = System for Observing Fitness Instruction;
PPVT = Peabody Picture Vocabulary Test III. [[beta].sub.00],
[[beta].sub.10], [[beta].sub.01], and [[beta].sub.11] are the
estimated interception, time effect, treatment effect, and time
by treatment effect. Recall the time by treatment effect
([[beta].sub.11]) is of primary interest, as it tests whether the
treatment condition (treatment vs. control) differs from pretest
to posttest. [ES.sub.[beta]11] represents the overall effect (i.e.,
effect size) associated with [[beta].sub.11]. The parameter estimates
[[beta].sub.01] and [[beta].sub.11] marked with an * and ** were
statistically significant at 0.05 and 0.007 (.05/7),
respectively. Also note that [[beta].sub.11] on the PPVT
(p = .059) and SOFIT (p = .081) were marginally significant.