The effects of school physical education grants on obesity, fitness, and academic achievement

Preventive Medicine 78 (2015) 44–51

Contents lists available at ScienceDirect

Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed

The effects of school physical education grants on obesity, fitness, and academic achievement Paul T. von Hippel ⁎, W. Kyle Bradbury Center for Health and Social Policy, LBJ School of Public Affairs, University of Texas, Austin, TX, USA

a r t i c l e

i n f o

Available online 8 July 2015 Keywords: Obesity Overweight Motor activity Policy

a b s t r a c t Objective. Foundations and governments fund a number of programs that provide grants to improve school physical education or other forms of school-based physical activity. The effects of these grant programs are unknown. We evaluate the effects of Texas Fitness Now, a program in which the state of Texas granted $37 million to improve physical education in high-poverty middle schools over the 4 school years from 2007–08 to 2010–11. The stated goals of Texas Fitness Now were to reduce obesity, increase fitness, and raise academic achievement. Method. We summarize how Texas Fitness Now funds were spent and estimate the impact of Texas Fitness Now using a fixed-effects longitudinal model that exploits changes in schools' eligibility over time. Changes in eligibility occurred when eligibility expanded to new schools after year 2 and when the program was terminated after year 4. Results. Most Texas Fitness Now funds were spent on sports and fitness equipment. Smaller amounts were spent on anti-obesity curricula. Texas Fitness Now improved strength and flexibility, especially among girls, but it did not improve BMI or academic achievement, and it had mixed effects on aerobic capacity. The fitness benefits were not lost in the year after the program ended, perhaps because schools kept the equipment that they had bought during their years of eligibility. Conclusion. The results of Texas Fitness Now were typical for an intervention that relied almost exclusively on physical activity. Programs that improve BMI as well as fitness tend to have a more fully developed nutrition component. © 2015 Elsevier Inc. All rights reserved.

Introduction The effect on obesity of physical education (PE) and other schoolbased physical activity (PA) is a controversial subject. While medical authorities including the American Heart Association, the Institute of Medicine, and the American Academy of Pediatrics have endorsed using PE and PA to reduce child obesity (Institute of Medicine, 2013; Pate et al., 2006; American Academy of Pediatrics et al., 2006), a recent review in the New England Journal of Medicine classified the idea that conventional PE classes can reduce obesity as a “myth” (Casazza et al., 2013, 2014). Despite endorsing school PA to prevent child obesity, the American Academy of Pediatrics has described interventions that rely on PA alone as “somewhat disappointing to date” (American Academy of Pediatrics et al., 2006). A Cochrane review of child obesity prevention found that out of 11 PA-only interventions, only 1 significantly reduced body mass index (BMI) gains in girls, and none significantly reduced BMI gains in boys (Waters et al., 2013). The average effect of the 11 ⁎ Corresponding author at: LBJ School of Public Affairs, Sid Richardson Hall 3.251, University of Texas, Austin, 2315 Red River, Box Y, Austin, TX 78712. E-mail address: [email protected] (P.T. von Hippel).

http://dx.doi.org/10.1016/j.ypmed.2015.06.011 0091-7435/© 2015 Elsevier Inc. All rights reserved.

interventions was significantly better than zero, but the Cochrane review cautioned that the average might be inflated by a publication bias against small studies with disappointing results (Waters et al., 2013). The two largest PA-only interventions had standardized effect sizes of 0.00, suggesting that null results were not due to a lack of statistical power. Interventions that combined PA with dietary changes had better results, reducing BMI growth significantly in 10 trials out of 27 (Waters et al., 2013). The effects of PA interventions on outcomes other than BMI are somewhat more encouraging. A meta-analysis of PA interventions for obese children found that although PA's “effects on body weight and central obesity [were] inconclusive,” on average, PA interventions did improve body composition by reducing percent body fat (Atlantis et al., 2006). PA can also increase strength and cardiovascular fitness (Sallis et al., 1997), producing “fat but fit” (Eisenmann, 2007) children whose blood chemistry and risk of metabolic syndrome are better than those of similarly obese children who are less fit (Eisenmann, 2007; Sung et al., 2002). In addition, PA can increase flexibility, which is impaired in many obese individuals (Gilleard and Smith, 2006). The effect of PA on academic achievement is also controversial. A recent systematic review concluded that “participation in [PA] is positively related to academic performance” (Singh et al., 2012), but a reanalysis

P.T. von Hippel, W.K. Bradbury / Preventive Medicine 78 (2015) 44–51

of the same data concluded that “physical activity is not related to performance at school” (Hattie and Clinton, 2012). It is true that PA improves cognition, particularly executive function (Tomporowski et al., 2008), and fitter, lighter students tend to have higher grades, test scores, and graduation rates (Crosnoe and Muller, 2004; Crosnoe, 2007; Van Dusen et al., 2011). However, the correlation between fitness and academic achievement is hard to interpret causally, in part because both fitness and achievement are correlated with socioeconomic status (von Hippel and Lynch, 2014). Compared to correlational studies, prospective studies of PA interventions suggest much smaller benefits for academic achievement, and some PA interventions have not improved achievement at all (Tomporowski et al., 2008).

PA grant programs and Texas Fitness Now While much of the literature has focused on specific PA interventions, it is just as important to understand the effects of broader PA policies. At the level of state government, the most common policy is to mandate the amount of time that children must spend in PE. PE time mandates can reduce obesity in elementary school, but only one-third of elementary schools comply with their states' PE time mandates (Cawley et al., 2013). PE time mandates have not reduced obesity in high school (Cawley et al., 2007). In this article, we evaluate a different policy, namely grant programs which provide schools with funds to improve PE or other aspects of school PA. Unlike PE time mandates, PA grant programs are voluntary and offer schools autonomy as well as financial incentives for participation. As a result, PA grant programs may attract fuller, more enthusiastic participation than unfunded PE time mandates. In addition, schools often use PA grant programs to defray the costs associated with particular PA interventions or curricula (SPARK, 2009a; Office of Safe and Drug-Free Schools, 2010). In this article, we evaluate a 4-year, $37 million PA grant program called Texas Fitness Now (TFN), which while active was the secondlargest government PA grant program in the US, second only to the federal government's Carol M. White Physical Education Program (PEP) (U.S. Department of Education, 2013). TFN was authorized in 2007 by the Texas legislature and ran for 4 years, from the 2007–08 school year through the 2010–11 school year, before being terminated in the recession-crimped budget passed in 2011. The goals of TFN, according to the legislation authorizing it, were to “[reduce] childhood obesity and Type II diabetes in school districts that have proportionately high numbers of economically disadvantaged [ED] students” (Legislative Reference Library, 2007), where ED means that the student qualified for public assistance such as school meal subsidies. Further goals, according to the Texas Comptroller and the Texas Education Agency (TEA), were to “decrease body fat, increase strength and endurance…prevent exercise-related injuries” (Combs and Texas Education Agency, 2007), and “improve academic achievement” (Combs and Texas Education Agency, 2010), since “through increased fitness, students' cognitive ability will improve” (Combs and Texas Education Agency, 2007) In years 1–2, TFN eligibility was limited to schools whose students were at least 75% ED. In years 3–4, eligibility expanded to schools whose students were at least 60% ED. In all years, eligibility was limited to schools serving 6th, 7th, or 8th graders, except for schools in the disciplinary and juvenile justice systems. The total budget for TFN was $10 million per year in years 1–2 and $8.5 million per year in years 3–4—a total of $37 million over 4 years. In years 1 and 2, eligible schools could apply to TEA for annual grants of $1,500 plus $32 for each 6th–8th grader. In years 3 and 4, schools that had not previously participated could apply for annual grants of $1,500 plus $28 per 6th–8th grader, while schools that had previously participated could receive a continuing annual grant of $1,000 plus $11 per 6th–8th grader.

45

TFN grants may seem small compared to other health or education spending, but they are very large compared to other spending on PE. The median discretionary PE budget for US middle schools in 2009 was just $900 (National Association for Sport and Physical Education, 2009), while the average TFN grant in 2009 was $15,500 per school (Texas Education Agency, 2011). TFN grants were large enough to pay for evidence-based obesity interventions; for example, the Planet Health intervention costs $14 per student (Cawley, 2007)—less than half of the initial TFN allocation. In a TEA survey, at least 93% of schools reported that TFN grants were adequate to achieve the program's goals (Texas Education Agency, 2011). Schools had flexibility in applying for TFN grants, but applications had to include detailed budgets devoted 75% to PA and 25% to nutrition and had to outline plans to improve obesity and other fitness measures by 5%. In addition, TFN grant recipients had to agree to more stringent requirements than other Texas middle schools (Combs and Texas Education Agency, 2010; Combs, 2008). TFN grant recipients had to require 6 semesters of PE, while other middle schools could require just 4. TFN grant recipients had to assess fitness at both the start and end of the school year, while other middle schools only assessed fitness at the end. TFN grants were limited to new initiatives that would “supplement…and not supplant” existing PE programs (Combs and Texas Education Agency, 2010). A TEA survey found that compliance with grant conditions rose by about 10% between years 1 and 2. By year 2, 90% of TFN middle schools required 6 semesters of PE, and 93% were testing fitness twice per year (Texas Education Agency, 2011). Methods Data We assembled publicly available annual data for all 4 years of TFN and the first year after termination. For each school and year, the data summarized demographics including the percentage of each school's students who were ED, as well as indicators for whether each school was eligible for TFN, and whether it participated. Because of TFN's eligibility requirements, we limited the data to middle schools that were not in the disciplinary or juvenile justice systems and that only enrolled students in 6th, 7th, or 8th grade. We also obtained all the TFN grant proposals approved in year 3, which was the year with the highest level of TFN participation. These proposals contained budgets which broke TFN spending into standard categories. Some proposals named items that the district would purchase for participating schools, as well as specific anti-obesity curricula which we coded using keywords. Dependent variables: Fitness and achievement Starting in 2007–08 (year 1 of our study), Texas required all middle schools to administer the FitnessGram assessment (Human Kinetics, 2013) every spring. Using FitnessGram, schools had to assess at least one measure of body composition, one measure of aerobic capacity, one measure of flexibility, and three measures of strength. To assess body composition, 89% of schools chose BMI (the alternative was skinfold thickness). To assess strength, 91–92% of schools chose pushups, curlups, and trunk lifts. To assess flexibility, 68% of schools chose the shoulder stretch and 46% chose the sit and reach (some chose both). To assess aerobic capacity, 47% of schools chose the mile run and 62% chose the progressive aerobic cardiovascular endurance run (PACER), “a paced, 20-meter shuttle run that increases in intensity as time progresses” (Texas Education Agency, 2012). On each component, the FitnessGram classifies students as being inside or outside a “healthy fitness zone” (HFZ) whose definition varies by gender and age. The HFZ for aerobic capacity is associated with high maximal oxygen consumption (VO2max) and low risk of metabolic syndrome (Welk et al., 2011a; Cureton and Warren, 1990). The HFZ for BMI is associated with having low percent body fat (Laurson et al., 2011), which in turn is associated with reduced cardiovascular risk factors including systolic blood pressure, HDL cholesterol, triglycerides, C-reactive protein, insulin, and fasting glucose (Going et al., 2011). The HFZs for strength and flexibility have not changed since 1992, but for BMI and aerobic capacity, the boundaries of the HFZ were revised in 2011 (year 4 of our study), and the area outside the HFZ was broken into subcategories indicating “some risk” and “high risk” (Welk et al., 2011b).

46

P.T. von Hippel, W.K. Bradbury / Preventive Medicine 78 (2015) 44–51

Texas middle schools were also required to administer the Texas Assessment of Knowledge and Skills (TAKS), which includes tests of reading and math. TAKS scores exceeding a moderate threshold were classified by TEA as “proficient,” and scores exceeding a higher threshold are classified as “commended.” According to TEA, the meaning of proficiency was stable from year to year, but in 2009, a large increase in proficient 8th graders led some observers to suggest that TEA had lowered the 8th grade standard (Mellon, 2010). TAKS was given in years 1–4 before being replaced by a different exam in year 5. For privacy reasons, TEA did not publish individual students' TAKS scores and did not collect individual students' FitnessGram results. Instead, every year, TEA summarized, for each grade and gender within each school, the number and percentage of students who scored at the “proficient” or “commended” level in TAKS of reading or mathematics, as well as the number and percentage whose FitnessGram component scores were in the HFZ or, where applicable, the some risk and high-risk categories. School-level percentages are adequate to evaluate the impact of TFN, because TFN was a school-level rather than individual-level intervention. While power would have been greater with individual scores, the power of the study was substantial because of its large sample size.

Table 1 Texas middle schools in the study.

Model

may be different in years 1–3 and 4–5. In years 1–3, we are estimating what was gained when TFN expanded to new schools in year 3. In years 4-5, we are estimating what was lost when TFN ended in year 5. Alternative models and estimation methods are given in the online Appendix.

We identified TFN's effects by exploiting changes in program eligibility over time. When TFN expanded in year 3, many previously ineligible schools became eligible, and when TFN ended in year 5, all schools became ineligible. These changes in eligibility were exogenous. That is, for a given school, eligibility depended on the year and the percentage of students who were ED, but net of those variables eligibility did not depend on the school's fitness or achievement outcomes. In each year, 88–95% of eligible schools participated in TFN. This is a high participation rate, but it is not 100%, and is possible that some schools chose to participate in an endogenous fashion that depended on their fitness and achievement levels net of their eligibility status. To estimate the effect of participation, we used the exogenous variable of eligibility as an instrument for the potentially endogenous variable of participation. Our statistical model, which we estimated using the xtivreg command in Stata 13, was a fixed-effects linear regression that used eligibility as an instrument for participation. The model is a system of two equations or stages: participateit ¼ α 0i þ α 1 eligibleit þ α 2 year t þ α 3 EDit þ α 4 year t  EDit þ uit ð1Þ P ijkt ¼ β 0i þ β 1 participateit þ β 2 year t þ β3 EDit þ β 4 year t  EDit þ eit

ð2Þ

In Eq. (2), the dependent variable Pit is the percentage of students, in school i and year t, who were in some category of fitness or achievement—for example, the percentage of students who were in the HFZ for BMI. The key independent variable is participateit, which is a dummy variable indicating whether school i participated in TFN in year t. In Eq. (1), participateit is modeled as a function of eligibleit which is a dummy variable indicating whether school i was eligible for TFN in year t. α0i and β0i are school-specific fixed effects, which control for time-invariant confounders. yeart is a vector of dummy variables that distinguish the 5 years of the study and control for average time trends in participation and in the average value of Pit. It is possible that the time trends are different in the high-ED schools that were eligible for TFN than in the low-ED schools that were not. To control for this possibility, the model includes an interaction between yeart and EDit, which is the percentage of school i's students who were ED in year t. In addition to the yeart × EDit interaction, the model also includes EDit on its own, although this is not really necessary since the model already has a time-invariant fixed effect, and EDit is almost time-invariant. (The serial correlation between EDit values from consecutive years is .98.) The final terms are random residuals uit and eit which are nonnormal, heteroskedastic, and serially correlated. Because of the complex residual structure, we estimated the model with bootstrapped standard errors. We fit the model separately to each grade and gender. We also fit a model to all grades together. In the all-grade models, grade indicators interacted with every term in the model, and there was a separate fixed effect for each grade within each school. We fit the model separately to years 1–3 and years 4–5. There were two reasons for this. First, for BMI and aerobic capacity the definition of the HFZ changed between years 3 and 4, and we did not want this change of definition to confound our estimate of the program effect. Second, the estimated effect

%

Schools Students Asian Black Hispanic Native American White ED Female Participated in TFN Eligible for TFN Participation among eligible schools

2008

2009

2010

2011

2012

1150 772,559 3 13 44 0.4 40 55 50 24 26 93

1161 790,687 3 13 45 0.4 39 56 50 24 26 95

1225 839,137 3 13 46 0.4 38 59 50 40 44 91

1216 832,984 3 12 47 0.5 36 59 50 38 43 88

1156 817,465 3 12 48 0.5 35 60 49 0 0

Note. These statistics are limited to schools that reported FitnessGram results to the state.

Results Table 1 describes the middle schools in the study. This may be the largest study ever conducted of a PE intervention, with 772,559 students and 1,150 schools in year 1 alone. The percentage of schools that were eligible for TFN rose from 26% to 44% when the eligibility threshold was lowered in year 3, then fell to 0% when TFN was terminated in year 5. Among eligible schools, the participation rate was at least 88% in every year. Table 2 summarizes the budgets of TFN grant proposals that were submitted and approved by TEA in year 3. 71% of TFN funds were spent on supplies and materials, with 35% going to fitness equipment and 19% going to sports equipment. An additional 12% was spent on capital outlay, which often included additional sports and fitness equipment. These large equipment purchases are consistent with a TEA survey in which 95% of TFN grant recipients reported that they used program funds to buy “traditional PE equipment (e.g., balls, rackets)” (Texas Education Agency, 2011). One survey respondent wrote: “We have turned our old field house into a top notch fitness center. We have weights, a TV for exercise videos, step benches, fitness bars, medicine balls and much more. It's our mini 24 Hour Fitness!” (Texas Education Agency, 2011). Another wrote: using “heart rate monitors purchased with the funds, [s]tudents collected data and compiled a personal fitness plan with attached goals for improvement” (Texas

Table 2 TFN expenditures in year 3 (2009–10). Category

Amount ($)

% of total

Payroll Professional services Supplies and materials Fitness equipment Sports equipment Coordinated school health programs Nutrition education equipment School health advisory council Technology General supplies Other Capital outlays Other Total

642,278 299,359 6,153,797 2,981,330 1,635,336 387,630 566,876 28,365 511,750 40,251 2,250 1,024,042 409,874 8,616,147

7 3 71 35 19 4 7 0 6 0 0 12 5 100

P.T. von Hippel, W.K. Bradbury / Preventive Medicine 78 (2015) 44–51

Education Agency, 2011). Large equipment purchases are not unique to Texas Fitness Now. In the federal PEP program, for example, one-third of 2010 grants to school districts were spent on equipment (Taylor and Jones, 2012). Contrary to program guidelines, nutrition accounted for far less than 25% of TFN spending. To understand this, we called TEA grant officers who told us that the nutrition requirement was informally waived in year 2 after some districts reported difficulties identifying nutritionrelated purchases. About 40% of district applications reported using specific fitness and anti-obesity curricula. The most popular choices were the Coordinated Approach to Child Health (Webber et al., 1996; Coleman et al., 2005) (CATCH, 17% of districts); Sports Play and Active Recreation for Kids (SPARK, 9% of districts), usually paired with Healthy and Wise (SPARK, 2009b) (8%); and the Bienestar/NEEMA coordinated school health program (3%), which targets Hispanics and military families (Clearinghouse for Military Family Readiness, 2012). Figs. 1 and 2 graph trends in the percentage of 8th grade boys and girls who were in the HFZ for 8 different FitnessGram components, as well as the percentage who were proficient or commended in reading or mathematics. Trends are graphed separately for the very high-ED schools that were eligible for TFN in years 1–4, the high-ED schools that were eligible in years 3–4, and the low-ED schools that were never eligible. As expected (Van Dusen et al., 2011; von Hippel and Lynch, 2014), the higher ED schools had lower percentages of children in the HFZ, and lower percentages who were proficient or commended. This is one of the reasons that high-ED schools were targeted by TFN. It is clear from Figs. 1 and 2 that TFN did not improve BMI. The percentage of students whose BMI was in the HFZ did not rise when

47

schools were eligible for TFN, and did not fall when TFN was terminated. In fact, the percentage did not materially change in eligible or ineligible schools in any year, except for the artificial change that occurred when the HFZ threshold changed in 2011. Trends for some other outcomes look more hopeful. Most outcomes improved over the years of TFN, and for some outcomes, the improvement was greater among eligible schools than among ineligible schools. For two outcomes, the improvements among eligible schools stopped in year 5 after TFN was terminated. But not all outcomes display these patterns. Table 3 uses our fixed-effects model to estimate the effect of TFN participation in years 1–3. The estimates show that TFN participation had no significant effect on BMI or academic achievement. The lack of significance cannot be blamed on lack of power. The sample is very large, the point estimates are close to zero, and the confidence intervals, in the all-grade model, are just 1–2 percentage points wide. Although TFN did not improve BMI or achievement, it did have significant benefits for fitness, especially among girls. Among girls, in the all-grade model, TFN participation increased by 2–6 percentage points, the percentage of girls who were in the HFZ for all 3 measures of strength (pushups, curlups, and trunk lifts) and one measure of flexibility (the sit and reach but not the shoulder stretch). TFN also increased the percentage of girls who were in the HFZ on one measure of aerobic capacity (the PACER), but it decreased the percentage in the HFZ on another measure (the mile run). Boys also benefited from TFN participation, but their benefits were limited to two FitnessGram components: pushups and the PACER. Table 4 displays the estimated effects of TFN in years 4–5; that is, it estimates what benefits were lost when TFN was discontinued in year

Fig. 1. Trends in the percentage of 8th grade boys who were in the Healthy Fitness Zone (HFZ) for 8 different FitnessGram components, and the percentage who were commended or proficient in reading or math. Note: in 2011, the HFZ threshold changed for BMI, the mile run, and the PACER.

48

P.T. von Hippel, W.K. Bradbury / Preventive Medicine 78 (2015) 44–51

Fig. 2. Trends for 8th grade girls.

5. The estimates suggest that little was lost. For both boys and girls, the all-grade estimates are small and nonsignificant for every outcome. Among the single-grade estimates, a few are significant, but the fraction of estimates with p b .05 is just 5%—exactly the fraction that we would expect by chance if there no true effects (Hsu, 1996). Notice that the estimates for year 4–5 include the high-risk categories that were introduced for some FitnessGram measures in year 4. However, they do not include results for the TAKS math and reading tests, since those tests were discontinued in year 5. The online Appendix gives results from alternative models.

Discussion TFN significantly improved fitness, and the fitness benefits were greater for girls than they were for boys. This gender difference is not unusual; several PA interventions (Sallis et al., 1997; Gortmaker et al., 1999; Datar and Sturm, 2004), though not all (Cawley et al., 2013), have shown greater benefits for girls than for boys. Girls may be more sensitive to school PA in because they are less physically active out of school (Nader et al., 2008). The fitness benefits of TFN did not reverse in the year after the program ended. A possible reason for this is that most TFN funds were spent on sports and fitness equipment, which schools got to keep after TFN ended. It is also possible that TFN changed students' bodies and behaviors in ways that outlasted the program. Although TFN improved fitness, it did not achieve its primary goals of reducing obesity and increasing academic achievement. This pattern of results is not uncommon for a school intervention that relied almost entirely on PA. As discussed in the introduction, PA-only interventions

improve fitness more often than they reduce obesity or improve academic achievement. The question arises whether TFN could also have reduced obesity if it were planned differently. Although results cannot be guaranteed, perhaps TFN would have had a better chance of reducing obesity if the program design had followed three principles of “evidence-based policy” (Haskins and Margolis, 2014; Campbell, 1969). The first principle is that all grants should be competitive. TFN was a noncompetitive formula grant program, in which eligible districts were practically guaranteed to receive funding if they applied for it. Although formula grants are not necessarily ineffective, their effects can be limited because districts lack the knowledge and incentives to align their proposals with program goals. More effective use can be made of funds if districts have to compete for them, with the competition judged by an expert panel applying criteria that are aligned with program goals such as obesity reduction. A second principle is that, while some funding can be set aside for innovation, the bulk of funding should be reserved for interventions that have previously shown rigorous evidence of efficacy. In some federal and state programs (Haskins and Margolis, 2014; Osborne et al., 2015), an expert committee draws up a menu of evidence-based interventions, and grant applicants choose from that menu. Had the state drawn up a menu of interventions that have rigorous evidence of reducing obesity, it would have discovered, as discussed in the Introduction, that such interventions rarely rely on PA alone (Waters et al., 2013). Instead, evidence-based obesity interventions tend to have a welldeveloped nutrition component, either alone or combined with PA (Waters et al., 2013). A final principle of evidence-based policy is ongoing evaluation. Even interventions that have previously demonstrated effectiveness

Table 3 Effect of TFN participation in years 1–3. All grades

6th grade

7th grade

8th grade

%

Girls

Boys

Girls

Boys

Girls

Boys

Girls

Boys

BMI

HFZ

Mile run

HFZ

PACER

HFZ

Pushup

HFZ

Curlup

HFZ

Trunk lift

HFZ

Sit and reach

HFZ

Shoulder stretch

HFZ

Math

Proficient

−0.4 (−1.5, 0.7) −4.1⁎ (−8.1, −0.2) 4.4⁎⁎⁎ (1.8, 7.1) 3.7⁎⁎⁎ (2.1, 5.4) 2.0⁎⁎ (0.5, 3.5) 1.9⁎⁎ (0.5, 3.3) 5.9⁎⁎ (2.0, 9.8) −0.8 (−2.9, 1.4) −0.5 (−1.3, 0.4) −0.1 (−0.9, 0.8) 0.5+ (−0.1, 1.1) −0.1 (−1.2, 0.9)

−0.4 (−1.5, 0.6) −0.7 (−4.8, 3.5) 3.0⁎ (0.5, 5.4) 3.5⁎⁎⁎ (1.8, 5.2) 1.2 (−0.5, 2.9) −0.0 (−2.0, 1.9) 2.2 (−1.0, 5.5) −0.0 (−1.5, 1.4) −0.4 (−1.6, 0.9) −0.5 (−1.3, 0.3) 0.1 (−0.5, 0.6) −0.2 (−1.4, 0.9)

1.0 (−0.7, 2.6) −2.5 (−12.7, 7.7) 4.3 (−1.3, 10.0) 3.9⁎ (0.9, 6.8) −0.3 (−3.5, 2.9) 3.7+ (−0.5, 7.9) 6.8⁎ (0.5, 13.0) 0.5 (−2.2, 3.3) −1.2 (−3.4, 0.9) −0.2 (−2.1, 1.7) 0.1 (−0.9, 1.0) −0.5 (−2.5, 1.5)

−0.3 (−2.6, 1.9) −2.1 (−13.1, 9.0) 1.6 (−2.7, 5.9) 3.2+ (−0.1, 6.5) 1.1 (−2.7, 4.8) 0.5 (−3.2, 4.1) 4.8 (−1.2, 10.7) 1.9 (−1.3, 5.1) 0.4 (−2.0, 2.7) 1.0 (−1.5, 3.5) 0.1 (−1.4, 1.6) 0.9 (−1.2, 3.1)

0.1 (−1.8, 1.9) −3.9 (−9.8, 2.0) 4.9⁎ (0.8, 8.9) 3.8⁎ (0.4, 7.1) 4.0⁎⁎ (1.6, 6.5) 2.8 (−1.0, 6.6) 8.1⁎ (1.5, 14.6) −0.1 (−3.6, 3.4) 0.3 (−1.7, 2.3) −0.2 (−1.3, 0.9) 1.0 (−0.3, 2.3) 0.3 (−0.9, 1.5)

−0.9 (−2.4, 0.7) −2.1 (−8.7, 4.5) 3.0+ (−0.3, 6.4) 3.7⁎ (0.7, 6.6) 0.7 (−1.9, 3.3) 0.1 (−4.0, 4.2) −0.1 (−6.5, 6.2) −0.2 (−2.4, 2.0) −0.1 (−1.9, 1.7) −0.7 (−1.8, 0.3) 0.1 (−1.4, 1.5) −0.5 (−1.9, 0.8)

−2.0+ (−4.0, 0.1) −5.4⁎ (−10.8, −0.1) 4.0⁎ (0.6, 7.5) 3.6⁎ (0.6, 6.5) 1.8 (−0.6, 4.3) −0.4 (−5.0, 4.2) 2.9 (−3.7, 9.5) −2.4 (−5.7, 0.9) −0.7 (−2.2, 0.9) 0.1 (−1.2, 1.4) 0.4 (−0.4, 1.2) −0.3 (−1.6, 0.9)

−0.1 (−1.8, 1.6) 1.8 (−5.8, 9.4) 4.0⁎ (0.1, 7.9) 3.5⁎ (0.4, 6.6) 1.9 (−0.8, 4.6) −0.6 (−3.8, 2.7) 2.6 (−3.0, 8.1) −1.4 (−4.1, 1.3) −1.1 (−2.6, 0.3) −1.3 (−2.9, 0.3) 0.0 (−0.9, 0.9) −0.8 (−2.4, 0.8)

Commended Reading

Proficient Commended

P.T. von Hippel, W.K. Bradbury / Preventive Medicine 78 (2015) 44–51

Outcome

Estimate (95% CI). HFZ = Healthy Fitness Zone. + p b .10. ⁎ p b .05. ⁎⁎ p b .01. ⁎⁎⁎ p b .001 (two tailed).

49

50

P.T. von Hippel, W.K. Bradbury / Preventive Medicine 78 (2015) 44–51

Table 4 Effect of TFN participation in years 4–5. All grades

6th grade

7th grade

8th grade

Outcome

%

Girls

Boys

Girls

Boys

Girls

Boys

Girls

Boys

BMI

HFZ

0.7 (−1.0, 2.3) 0.9 (−2.2, 4.1) −0.8 (−2.9, 1.2) −1.5 (−3.4, 0.4) 1.0 (−0.4, 2.4) 1.2 (−0.6, 3.1) −0.2 (−2.2, 1.8) 1.6 (−0.5, 3.6) −1.1 (−5.0, 2.8) −1.0 (−3.5, 1.6) −1.0 (−2.6, 0.6)

0.9 (−0.6, 2.4) 1.1 (−2.5, 4.6) −1.7 (−4.8, 1.5) −0.4 (−3.7, 2.9) 0.7 (−1.2, 2.7) 1.2 (−0.9, 3.2) 1.1 (−0.9, 3.1) 0.5 (−1.2, 2.1) 1.8 (−2.5, 6.2) 0.2 (−1.7, 2.0) −1.6 (−3.7, 0.5)

−0.7 (−3.7, 2.3) 2.2 (−2.2, 6.7) −2.0 (−6.1, 2.1) −0.7 (−4.6, 3.2) 1.3 (−1.2, 3.8) 3.8+ (−0.5, 8.1) 2.0 (−2.1, 6.2) 1.4 (−3.3, 6.1) 1.0 (−7.0, 9.0) 0.9 (−3.4, 5.1) 1.1 (−1.7, 3.9)

0.6 (−2.1, 3.2) −3.5 (−8.5, 1.5) −1.3 (−5.5, 2.9) 0.3 (−4.2, 4.8) 0.7 (−2.7, 4.0) 1.4 (−4.3, 7.2) 0.3 (−4.0, 4.5) 3.8+ (−0.5, 8.0) 3.0 (−5.8, 11.8) 0.8 (−2.5, 4.2) −1.6 (−4.6, 1.5)

0.7 (−2.4, 3.8) 1.9 (−4.2, 8.0) −1.2 (−5.8, 3.3) −2.0 (−5.6, 1.6) 0.6 (−1.7, 2.9) 0.2 (−3.0, 3.4) −0.7 (−3.7, 2.4) 1.2 (−1.6, 4.0) −5.8 (−13.8, 2.3) −0.8 (−4.0, 2.5) −0.4 (−3.0, 2.2)

2.1 (−0.7, 4.8) 0.3 (−5.9, 6.6) 0.0 (−5.1, 5.2) −1.3 (−5.9, 3.3) 1.5 (−1.5, 4.5) 3.2⁎ (0.1, 6.2) 2.1 (−1.3, 5.5) −1.2 (−3.7, 1.3) −0.4 (−6.5, 5.7) 0.8 (−2.3, 3.9) −2.5⁎ (−4.9, −0.2)

1.8 (−0.4, 4.0) −1.0 (−6.2, 4.2) 0.4 (−3.5, 4.4) −1.8 (−5.2, 1.6) 1.2 (−1.5, 3.9) 0.3⁎⁎ (−3.1, 3.6) −1.5 (−4.8, 1.8) 2.1 (−1.3, 5.5) 1.9 (−4.8, 8.6) −2.6 (−6.7, 1.5) −3.2⁎ (−5.6, −0.7)

0.1 (−3.1, 3.3) 5.1 (−2.0, 12.3) −3.6 (−9.1, 1.8) −0.0 (−5.4, 5.3) −0.1 (−3.3, 3.1) −1.0⁎⁎⁎ (−5.0, 2.9) 0.7 (−3.0, 4.5) −0.4 (−3.5, 2.7) 3.2 (−3.9, 10.2) −1.0 (−4.5, 2.6) −0.7 (−4.6, 3.2)

High risk Mile run

HFZ High risk

PACER

HFZ High risk

Pushup

HFZ

Curlup

HFZ

Trunk lift

HFZ

Sit and reach

HFZ

Shoulder stretch

HFZ

Estimate (95% CI). HFZ = Healthy Fitness Zone. + p b .10. ⁎ p b .05. ⁎⁎ p b .01. ⁎⁎⁎ p b .001 (two tailed).

should be evaluated, since the fact that an intervention has succeeded once does not guarantee that success will be repeated if the intervention is implemented in a different time and place. By the time this article was written, the state of Texas had already authorized, expanded, and canceled TFN without clear evidence regarding its effectiveness. Even this evaluation was not funded by the state, but by a private foundation interested in the promise of PA programs in general. Evaluation of TFN would have had more value to Texas while the program was still active. Yet the lessons of this evaluation will be valuable to other organizations seeking cost-effective ways to reduce obesity, increase fitness, or raise academic achievement. Conflict of interest The authors declare that they have no conflict of interest. Acknowledgments This research was funded in part by a grant from the St. David's Foundation. We thank Roxanne Garza, Amy Suntoke, and Margaret Tollefson for their contributions to a pilot analysis of the first year's data. Appendix A. Supplementary analyses Can be found online at http://dx.doi.org/10.1016/j.ypmed.2015.06. 011. References American Academy of Pediatrics, Council on Sports Medicine and Fitness, Council on School Health, 2006. Active healthy living: prevention of childhood obesity through increased physical activity. Pediatrics 117 (5), 1834–1842. http://dx.doi.org/10. 1542/peds.2006-0472. Atlantis, E., Barnes, E.H., Singh, M.A.F., 2006. Efficacy of exercise for treating overweight in children and adolescents: a systematic review. Int. J. Obes. 30 (7), 1027–1040. http:// dx.doi.org/10.1038/sj.ijo.0803286. Campbell, D.T., 1969. Reforms as experiments. Am. Psychol. 24 (4), 409–429. http://dx. doi.org/10.1037/h0027982. Casazza, K., Fontaine, K.R., Astrup, A., et al., 2013. Myths, presumptions, and facts about obesity. N. Engl. J. Med. 368 (5), 446–454. http://dx.doi.org/10.1056/NEJMsa1208051. Casazza, K., Brown, A., Astrup, A., et al., 2014. Weighing the evidence of common beliefs in obesity research. Crit. Rev. Food Sci. Nutr. http://dx.doi.org/10.1080/10408398.2014. 922044 (0(ja):00-00).

Cawley, J., 2007. The cost-effectiveness of programs to prevent or reduce obesity: the state of the literature and a future research agenda. Arch. Pediatr. Adolesc. Med. 161 (6), 611–614. http://dx.doi.org/10.1001/archpedi.161.6.611. Cawley, J., Meyerhoefer, C., Newhouse, D., 2007. The impact of state physical education requirements on youth physical activity and overweight. Health Econ. 16 (12), 1287–1301. http://dx.doi.org/10.1002/hec.1218. Cawley, J., Frisvold, D., Meyerhoefer, C., 2013. The impact of physical education on obesity among elementary school children. J. Health Econ. 32 (4), 743–755. http://dx.doi.org/ 10.1016/j.jhealeco.2013.04.006. Clearinghouse for Military Family Readiness, 2012. Bienestar/NEEMA Coordinated School Health Program. Clgh Mil Fam Readiness (http://www.militaryfamilies.psu.edu/ programs/bienestarneema-coordinated-school-health-program. Accessed January 29, 2014). Coleman, K.J., Tiller, C.L., Sanchez, J., et al., 2005. Prevention of the epidemic increase in child risk of overweight in low-income schools: the El Paso coordinated approach to child health. Arch. Pediatr. Adolesc. Med. 159 (3), 217–224. http://dx.doi.org/10. 1001/archpedi.159.3.217. Combs, S., 2008. Gaining Costs, Losing Time - The Obesity Crisis in Texas. http://www. window.state.tx.us/specialrpt/obesitycost/index.php (Accessed January 21, 2014). Combs, S., Texas Education Agency, 2007. Program Guidelines, Texas Fitness Now Program: Middle School Physical Education and Fitness Programs. Texas Education Agency, Austin, TX. Combs, S., Texas Education Agency, 2010. Texas Fitness Now Program Guidelines (Austin, TX). Crosnoe, R., 2007. Gender, obesity, and education. Sociol. Educ. 80 (3), 241–260. http://dx. doi.org/10.1177/003804070708000303. Crosnoe, R., Muller, C., 2004. Body mass index, academic achievement, and school context: examining the educational experiences of adolescents at risk of obesity. J. Health Soc. Behav. 45 (4), 393–407. http://dx.doi.org/10.1177/002214650404500403. Cureton, K.J., Warren, G.L., 1990. Criterion-referenced standards for youth health-related fitness tests: a tutorial. Res. Q. Exerc. Sport 61 (1), 7–19. Datar, A., Sturm, R., 2004. Physical education in elementary school and body mass index: evidence from the early childhood longitudinal study. Am. J. Public Health 94 (9), 1501–1506. Eisenmann, J.C., 2007. Aerobic fitness, fatness and the metabolic syndrome in children and adolescents. Acta Paediatr. 96 (12), 1723–1729. http://dx.doi.org/10.1111/j. 1651-2227.2007.00534.x. Gilleard, W., Smith, T., 2006. Effect of obesity on posture and hip joint moments during a standing task, and trunk forward flexion motion. Int. J. Obes. 31 (2), 267–271. http:// dx.doi.org/10.1038/sj.ijo.0803430. Going, S.B., Lohman, T.G., Cussler, E.C., Williams, D.P., Morrison, J.A., Horn, P.S., 2011. Percent body fat and chronic disease risk factors in U.S. children and youth. Am. J. Prev. Med. 41 (4, Supplement 2), S77–S86. http://dx.doi.org/10.1016/j.amepre. 2011.07.006. Gortmaker, S.L., Peterson, K., Wiecha, J., et al., 1999. Reducing obesity via a school-based interdisciplinary intervention among youth: Planet Health. Arch. Pediatr. Adolesc. Med. 153 (4), 409–418. http://dx.doi.org/10.1001/archpedi.153.4.409. Haskins, R., Margolis, G., 2014. Show Me the Evidence: Obama's Fight for Rigor and Results in Social Policy. Brookings Institution Press, Washington, D.C. Hattie, J., Clinton, J., 2012. Physical activity is not related to performance at school. Arch. Pediatr. Adolesc. Med. 166 (7), 678–679. http://dx.doi.org/10.1001/archpediatrics. 2012.334.

P.T. von Hippel, W.K. Bradbury / Preventive Medicine 78 (2015) 44–51 Hsu, J., 1996. Multiple Comparisons: Theory and Methods. 1st ed. Chapman and Hall/CRC. Human Kinetics, 2013. Fitnessgram: Activity and fitness assessment, reporting, and tracking. http://www.fitnessgram.net/home/ (Accessed May 9, 2013). Institute of Medicine, 2013. Educating the Student Body: Taking Physical Activity and Physical Education to School. http://www.iom.edu/Reports/2013/Educating-theStudent-Body-Taking-Physical-Activity-and-Physical-Education-to-School.aspx (Accessed October 15, 2013). Laurson, K.R., Eisenmann, J.C., Welk, G.J., 2011. Body mass index standards based on agreement with health-related body fat. Am. J. Prev. Med. 41 (4, Supplement 2), S100–S105. http://dx.doi.org/10.1016/j.amepre.2011.07.004. Legislative Reference Library, 2007. HB 1, 80th Regular Session. http://www.lrl.state.tx.us/ legis/ApproBillSections.cfm?billID=60463&from= (Accessed January 21, 2014). Mellon E. A lowered bar raises doubts over TAKS standards. Houston Chronicle. http:// www.chron.com/news/houston-texas/article/A-lowered-bar-raises-doubts-overTAKS-standards-1712503.php. Published June 8, 2010. Accessed April 7, 2015. Nader, P.R., Bradley, R.H., Houts, R.M., McRitchie, S.L., O'Brien, M., 2008. Moderate-tovigorous physical activity from ages 9 to 15 years. JAMA 300 (3), 295–305. http:// dx.doi.org/10.1001/jama.300.3.295. National Association for Sport and Physical Education, 2009. Physical Education Trends in Our Nation's Schools: A Survey of Practicing K-12 Physical Education Teachers. Roslow Research Group, Port Washington, NY. Office of Safe and Drug-Free Schools, 2010. Carol M. White Physical Education Program CFDA #84.215 F: Information and Application Procedures for Fiscal Year 2010. U.S. Department of Education, Washington, DC. Osborne, C., Bobbitt, K., Boelter, J., Kwang, T., 2015. Early lessons learned from building local early childhood comprehensive systems in Texas. J. Appl. Res. Child Informing Policy Child Risk 5 (1) (in press). Pate, R.R., Davis, M.G., Robinson, T.N., Stone, E.J., McKenzie, T.L., Young, J.C., 2006. Promoting Physical activity in children and youth a leadership role for schools: a scientific statement from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism (Physical Activity Committee) in collaboration with the Councils on Cardiovascular Disease in the Young and Cardiovascular Nursing. Circulation 114 (11), 1214–1224. http://dx.doi.org/10.1161/CIRCULATIONAHA.106.177052. Sallis, J.F., McKenzie, T.L., Alcaraz, J.E., Kolody, B., Faucette, N., Hovell, M.F., 1997. The effects of a 2-year physical education program (SPARK) on physical activity and fitness in elementary school students. Sports, play and active recreation for kids. Am. J. Public Health 87 (8), 1328–1334. http://dx.doi.org/10.2105/AJPH.87.8.1328. Singh, A., Uijtdewilligen, L., Twisk, J.R., van Mechelen, W., Chinapaw, M.M., 2012. Physical activity and performance at school: a systematic review of the literature including a methodological quality assessment. Arch. Pediatr. Adolesc. Med. 166 (1), 49–55. http://dx.doi.org/10.1001/archpediatrics.2011.716.

51

SPARK, 2009a. Grant Finder Tool. http://www.sparkpe.org/grants/grantfundingresources/ (Accessed November 16, 2014). SPARK, 2009b. SPARK and Healthy & Wise. http://www.sparkpe.org/coordinated-schoolhealth/tx-coordinated-school-health/texas-csh/ (Accessed January 29, 2014). Sung, R.Y.T., Yu, C.W., Chang, S.K.Y., Mo, S.W., Woo, K.S., Lam, C.W.K., 2002. Effects of dietary intervention and strength training on blood lipid level in obese children. Arch. Dis. Child. 86 (6), 407–410. http://dx.doi.org/10.1136/adc.86.6.407. Taylor, T., Jones, W., 2012. Evaluation Brief: The Carol M. White Physical Education Program Project Implementation During Year 1 of the Grant. American Institutes for Research. Texas Education Agency, 2011. Evaluation of the Texas Fitness Now Grant Program: 2007-08 to 2009-10 School Years. http://www.tea.state.tx.us/index4.aspx?id= 2930&menu_id=949 (Accessed May 9, 2013). Texas Education Agency, 2012. FitnessGram Procedures Manual. http://www.tea.state.tx. us/FitnessGram_Manual.html (Accessed January 27, 2014). Tomporowski, P.D., Davis, C.L., Miller, P.H., Naglieri, J.A., 2008. Exercise and children's intelligence, cognition, and academic achievement. Educ. Psychol. Rev. 20 (2), 111–131. http://dx.doi.org/10.1007/s10648-007-9057-0. U.S. Department of Education, 2013. Carol M. White Physical Education Program. http:// www2.ed.gov/programs/whitephysed/index.html (Accessed July 14, 2015). Van Dusen, D.P., Kelder, S.H., Kohl 3rd, H.W., Ranjit, N., Perry, C.L., 2011. Associations of physical fitness and academic performance among schoolchildren. J. Sch. Health 81 (12), 733–740. http://dx.doi.org/10.1111/j.1746-1561.2011.00652.x. von Hippel, P.T., Lynch, J.L., 2014. Why are educated adults slim—Causation or selection? Soc. Sci. Med. 105 (3). http://dx.doi.org/10.1016/j.socscimed.2014.01.004. Waters, E., de Silva-Sanigorski, A., Burford, B.J., et al., 2013. Interventions for preventing obesity in children. Cochrane Database of Systematic Reviews. John Wiley & Sons, Ltd (http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD001871.pub3/abstract. Accessed October 15, 2013). Webber, L.S., Osganian, S.K., Feldman, H.A., et al., 1996. Cardiovascular risk factors among children after a 2 1/2-year intervention—The CATCH Study. Prev. Med. 25 (4), 432–441. http://dx.doi.org/10.1006/pmed.1996.0075. Welk, G.J., Laurson, K.R., Eisenmann, J.C., Cureton, K.J., 2011a. Development of youth aerobic-capacity standards using receiver operating characteristic curves. Am. J. Prev. Med. 41 (4, Supplement 2), S111–S116. http://dx.doi.org/10.1016/j.amepre. 2011.07.007. Welk, G.J., Going, S.B., Morrow Jr., J.R., Meredith, M.D., 2011b. Development of new criterion-referenced fitness standards in the FITNESSGRAM® program: rationale and conceptual overview. Am. J. Prev. Med. 41 (4 Suppl. 2), S63–S67. http://dx.doi. org/10.1016/j.amepre.2011.07.012.