School Garden Prevalence Before and After the Implementation of the Healthy Hunger-Free Kids Act

Abstract

BACKGROUND:

The Healthy Hunger-Free Kids Act (HHFKA) of 2010 supported implementation of school gardens for promoting fruit and vegetable consumption. We examined school gardens’ prevalence over time by school-level factors during the period before and after the implementation of HHFKA.

METHODS:

Using data from the New Jersey Child Health Study conducted in four low-income New Jersey cities, prevalence of school gardens among K-12 schools (n=148) was assessed between school year 2010–11 to 2017–18. Multivariable analysis estimated changes in garden prevalence over time adjusting for school-level factors.

RESULTS:

Overall, the sample included 97 elementary and 51 middle/high schools. Multivariable logistic regression showed that compared to 2010–11 (19%), a higher proportion of schools reported having a garden in 2013–14 (32%, p=0.025). Over the entire study period, schools with majority Hispanic student enrollment had approximately half the odds of having a garden compared to schools with majority Black students (p=0.036).

CONCLUSION:

School garden prevalence increased in the year immediately following the implementation of the HHFKA but this increase was not sustained over time. Future research should investigate the reasons for this decline and potential disparities by race/ethnicity.

According to the Centers for Disease Control and Prevention, in the USA in 2017–2018, 19.3% of children between 2 and 19 years of age had obesity, with a higher prevalence among school-age children: the obesity rate was 21.2% for 12–19 year-old, 20.3% for 6–11 year-old, and 13.4% for 2–5 year-old.¹ Schools are an ideal setting for implementing obesity prevention efforts, as almost 60 million children spend a considerable amount of time in early care and school settings each day,² creating opportunities for establishing foundations for good health.

Recent studies and systematic reviews suggest that school gardening programs can improve fruit and vegetable (FV) consumption,^3–7 which is typically well below the daily recommended amount in youth diets.⁸ Exposure to school garden programs has shown to help elementary school-age children (1) become more familiar with FV,^5,9–12 (2) increase their preference for^5,12,13 and willingness to taste them,⁹ as well as (3) eat more FV.^4,5,10,13 While the literature for school gardens among high school students is scarce, a pilot study conducted in New York found that when school-garden-produced vegetables were in the salad bar, the percentage of students selecting salad increased by up to 10%.¹⁴ Further, for high school students, the sole exposure to prior farm experience or having a home garden created more willingness to try new FV, compared to students without a garden or previous farm experience.¹⁵ These findings suggest that sustained exposure to gardening through school years can be beneficial in changing dietary consumption patterns as children grow.

Every five years, the Child Nutrition Reauthorization Act reviews and authorizes funding for all federal child nutrition and school meal programs, providing the opportunity to continually improve these programs.¹⁶ As part of the 2010 Child Nutrition Reauthorization Act, the United States Department of Agriculture enacted the Healthy Hunger-Free Kids Act (HHFKA).¹⁷ The HHFKA included recommendations for increasing the amount and variety of FV in students’ diets through various strategies. One of these strategies included the allocation of earmarked funding for state and local establishments to the USDA Farm to School Grant Program, which came into effect in 2012. These funds provided the necessary support to plan, develop, and implement school gardens.^17,18 To apply to the Farm to School Grant Program, eligible entities must submit detailed applications to the Request for Applications for the fiscal year, considering logistics such as technical assistance and institutional sustainability. To guide the application process, the USDA website offers a manual alongside a database of other resources and previous examples.¹⁹

Few studies to date have examined the uptake of school gardens in schools. The Centers for Disease Control and Prevention’s School Health Policies and Practices Study reported that in a nationally representative sample of USA schools in 2014, 19.2% of elementary schools had a school garden, compared to 24.4% of middle schools and 22.3% of high schools.²⁰ Turner and colleagues, in 2016, published the only study that examines the prevalence of gardens in public elementary schools in the US over time and found that school gardens almost tripled from 2006–07 (11.9%) to 2013–14 (31.2%).²¹ The authors did not examine prevalence of school gardens in middle and high schools. Recent changes in the prevalence of school gardens in K-12 schools, primarily since the implementation of HHFKA, have not been investigated. This longitudinal study, spanning over eight years, aims to (1) assess school garden prevalence in K-12 public schools in four school districts in New Jersey and (2) examine differences in trends by school-level factors. We hypothesize that more schools will report having school gardens after the implementation of the HHFKA and that, contrary to School Health Policies and Practices Study findings,²⁰ prevalence of school gardens over time will be higher in elementary schools because parental involvement is higher at the elementary school level,²² and the literature agrees on the importance of stakeholders participation in garden prevalence.²³

METHODS

Participants

Survey data were collected over an eight-year period from school year (SY) 2010–11 to SY 2017–18 from 148 K-12 public schools located in four New Jersey cities: Camden, Newark, New Brunswick, and Trenton. Responses were provided by school nurses with assistance from school food service and physical activity personnel. All public schools located in the four study cities were invited to participate in the survey if they offered any grades between K and 12.

Instrumentation

School surveys were conducted as part of the New Jersey Child Health Study,²⁴ a longitudinal study examining whether and how changes in the food and physical activity environments in schools and communities influence children’s health behaviors and weight outcomes. Information on school food offerings, physical activity opportunities, and health-promoting policies were collected at three different time points: School Year (SY) 2012–13, SY 2015–16, and SY 2017–18. For each survey administration, questions pertained to information regarding the current and preceding SY(s). Overall, data were collected for 8 consecutive SYs, from 2010–11 to 2017–18. The survey questions were adapted from previous research^25–28 to measure elements of the food and physical activity environments in schools that can affect students’ obesogenic behaviors. Additional details on the current survey instrument are available elsewhere.²⁹

School level (elementary vs middle/high), racial/ethnic composition of the student body, percentage of students eligible for free and reduced-price meals (FRPM), and total enrollment, were obtained for each SY from the National Center for Education Statistics common core data repository.³⁰ Schools were grouped into 3 mutually exclusive race/ethnicity categories: majority Black (>50% of students are Black), majority Hispanic (>50% of students are Hispanic), and majority White/mixed (>50% of students are White or school does not have a majority racial/ethnic group). The percentage of students eligible for FRPM was used to control for school-level income. In multivariable analyses, schools were grouped into three categories based on the tertiles of the distribution.

Procedure

Surveys were distributed to school nurses, who were asked to consult with school food service personnel and physical education staff for answering questions specific to those domains. The survey had three sections addressing (1) food environment, (2) physical activity environment, and (3) school-level practices and program participation. Under the school-level practices portion of the survey, respondents were asked about the presence of a school garden in their schools: “In which of the following school years did your school have a fruit and/or vegetable garden in which students could participate?”.

To ensure a high response rate, a $50 incentive was offered to the respondents at the completion of the survey. Across all four districts, the average response rate was 92%. Data collected after SY 2012–13, the year HHFKA policies came into effect, were considered post-HHFKA. Respondents had the option to respond using a paper format or an online version via Qualtrics (Provo, UT).

Data Analysis

All analyses were conducted in Stata 15 (StataCorp LLC, College Station, TX). Each school included in the analysis contributed data for between 2 and 8 time points (average 6.3). Missing data were imputed using a chained equation procedure.^31,32 Bivariate Chi-square tests and t-tests were used to examine whether there were differences in the control variables (i.e., school level, racial/ethnic composition, students’ eligibility for FRPM, city, and total school enrollment) across schools with and without school garden.

Multivariable logistic regression analysis was used to model the prevalence of school gardens between SY 2010–11 and SY 2017–18, adjusting for all covariates. The margins postestimation command was used to obtain the adjusted school garden prevalence for each year. Time first entered as a categorical variable (with 2010 as the reference category) and then as a continuous variable to estimate the linear trend over the 8 years. The main model was then expanded by adding an interaction term between time and school-level variables (one at a time) to examine whether the prevalence of school garden over the years was different by school level, school race, and students’ FRPM eligibility. Statistical significance was determined at p<0.05.

RESULTS

Demographic characteristics of the study sample are described in Table 1. From SY 2010–11 to 2017–18, a total of 927 school-year observations was available from 148 schools (ranging from 108 to 127 schools in any given year). The number of schools surveyed varied each year as schools close and open based on district needs. Newark, the largest of the four school districts, included most schools in the sample across each year (46% to 51%), followed by Camden (17% to 22%), Trenton (16% to 21%), and New Brunswick (11% to 12%). Elementary schools made up approximately two-thirds of all schools each year. Over 90% of schools had either majority Black or majority Hispanic student enrollment throughout the study period, with few White majority/mixed-race schools (between 2.7% and 6.7%). Across the entire study period, the average students’ FRPM eligibility rate was 69.6%, 84.5%, and 92.9% for the three tertiles (low, medium, and high, respectively) of the FRPM eligibility distribution.

Table 1.

Descriptive Characteristics for All Schools in the Sample Between SY 2010–11 and SY 2017–18

School Level Factors	School Year
	2010–11	2011–12	2012–13	2013–14	2014–15	2015–16	2016–17	2017–18
Number of Schools	127	127	127	110	110	110	108	108
Race/Ethnicity of Enrolled Students (%)
Majority Black	48.5	50.5	50.6	46.3	45.0	47.0	42.9	41.1
Majority Hispanic	47.3	44.6	44.5	47.1	49.5	50.3	54.2	56.0
Majority White and mixed	4.2	4.9	4.9	6.7	5.5	2.7	2.8	2.8
Eligible for Free and Reduced-Price Meals (%)	80.6	84.6	90.1	88.0	85.1	74.6	76.2	76.8
School Level (%)
Elementary	67.7	67.7	67.7	67.3	67.3	67.3	67.6	67.6
Middle and High	32.3	32.3	32.3	32.7	32.7	32.7	32.4	32.4
Total enrollment (mean)	530	546	598	572	579	605	645	675
Study City (%)
Camden	21.3	21.3	21.3	21.8	21.8	21.8	17.6	17.6
New Brunswick	11.0	11.0	11.0	11.8	11.8	11.8	12.0	12.0
Newark	51.2	51.2	51.2	46.4	46.4	46.4	49.1	49.1
Trenton	16.5	16.5	16.5	20.0	20.0	20.0	21.3	21.3

Table 2 shows results from the bivariate analysis. There were significant differences in prevalence of school gardens by (1) racial/ethnic composition of schools (p<0.001), (2) students’ eligibility for FRPM (p=0.013), (3) total enrollment (p=0.016), and (4) city (p<0.001). No differences in garden prevalence were observed between elementary and middle/high schools (p=0.519).

Table 2.

Bivariate Analysis Showing Prevalence of School Garden by School Level Factors Across All Years Combined.

School Level Factor	Schools without Garden (%)	Schools with Garden (%)	p-value
	(n = 668)	(n = 259)
Race/Ethnicity of Enrolled Students (%)			<0.001
Majority Black	64.8	35.2
Majority Hispanic	79.8	20.2
Majority White and mixed	64.0	36.0
Students’ FRPM eligibility (%)	81.6	83.9	0.022
School Level (%)			0.519
Elementary	72.8	27.2
Middle/High	70.6	29.4
Total enrollment (mean)	619	553	0.023
Study City (%)			<0.001
Camden	64.3	35.7
New Brunswick	93.3	6.7
Newark	73.9	26.1
Trenton	63.1	36.9

P-values are calculated through chi-square tests for categorical variables and t-tests for continuous variables.

Table 3 shows results from the multivariable logistic regression model predicting prevalence of school gardens for each year, adjusting for school-level factors. Compared to baseline (SY 2010–11), K-12 schools had twice the odds of having a school garden in SY 2013–14 (OR=2.03, p=0.025), while other years were not significantly different from the base year. Overall, there was no observable linear trend (p=0.570). Compared to majority Black schools, majority Hispanic schools had half the odds of having a school garden (OR=0.54, p=0.036), while no associations were observed by school level or students’ FRPM eligibility status.

Table 3.

Multivariate Analysis Predicting School Garden Participation Across School Level Characteristics for SY 2010–11 To SY 2017–18

School Level Factors	Odds Ratio	95% Confidence Interval	p-value
Year
2010	Ref
2011	1.54	(0.95, 2.51)	0.078
2012	1.47	(0.89, 2.43)	0.132
2013	2.04	(1.09, 3.80)	0.025
2014	1.65	(0.91, 3.01)	0.101
2015	1.43	(0.77, 2.66)	0.262
2016	1.69	(0.88, 3.25)	0.118
2017	1.30	(0.64, 2.66)	0.471
School Level
Elementary school	Ref
Middle or High school	1.00	(0.52, 1.93)	0.991
Students’ FRPM eligibility
Lowest eligibility category	Ref
Middle eligibility category	1.09	(0.68, 1.75)	0.729
Highest eligibility category	1.43	(0.85, 2.40)	0.172
School Majority Race
Majority Black	Ref
Majority Hispanic	0.55	(0.31, 0.96)	0.036
Majority White and Mixed	1.21	(0.38, 3.89)	0.750

Odds ratios from logistic regression model predicting participation in school garden programs, adjusting by school level, school year, proportion of students eligible for free and reduced-price meals, race/ethnicity, and city.

Using estimated prevalences from the model reported in Table 3, Figure 1 presents yearly adjusted prevalence of school gardens for the full sample. While an overall increase in garden prevalence was observed in the earlier years after the implementation of the HHFKA, going from 18.9% in SY 2010–11 to 32.2% in SY 2013–14, this initial increase was not sustained, with the estimated prevalence slowly declining over the following 4 years.

Figure 1. Adjusted Prevalence of School Gardens Across All School Levels for Each School Year Between 2010 to 2017.

Note: Adjusted for year, city, students’ FRPM eligibility, school level, school majority race/ethnicity, and enrollment. P for trend = 0.570

Figures 2–4 were generated from the interaction models. They report the yearly estimated prevalence of school garden by school-level characteristics. Specifically, Figure 2 shows that the garden prevalence was remarkably similar for elementary and middle/high schools over the study period. The initial increase in garden prevalence post-HHFKA observed in the full sample was seen in both elementary and middle schools. There was a divergence only in SY 2017–18 due to a steeper decline among middle/high schools. There was no significant time trend for either elementary or middle/high schools.

Figure 2. Adjusted Prevalence of School Gardens by School Level for Each School Year Between 2010 to 2017.

Note: Adjusted for year, city, students’ FRPM eligibility, school majority race/ethnicity, and enrollment. Elementary school p for trend = 0.451; middle/high school p for trend = 0.926

Figure 4. Adjusted Prevalence of School Gardens by Students’ FRPM Eligibility for Each School Year Between 2010 to 2017.

Note: Adjusted for year, city, school level, school majority race/ethnicity, and enrollment. FRPM lowest eligibility category p for trend = 0.647; FRPM middle eligibility category p for trend = 0.743; FRPM highest eligibility category p for trend = 0.222

Trends in garden prevalence by school racial/ethnic composition are shown in Figure 3. For most years, majority Black schools had a higher prevalence of school gardens than majority Hispanic schools. However, despite some year-to-year variations, there was no linear trend in garden prevalence for either of the two groups. Similarly, Figure 4 shows that the prevalence of school gardens did not follow any specific pattern when schools were grouped by tertile of students’ FRPM eligibility.

Figure 3. Adjusted Prevalence of School Gardens by School Majority Race/Ethnicity for Each School Year Between 2010 to 2017.

Note: Adjusted for year, city, students’ FRPM eligibility, school level, and enrollment.

School majority Black p for trend = 0.888; school majority Hispanic p for trend = 0.430.

The trend for majority White and Mixed schools was omitted from this figure because of the small sample size.

DISCUSSION

This study examined school garden prevalence in K-12 public schools in four New Jersey cities across an 8-year span, which included periods before and after HHFKA implementation. HHFKA mandated funds through the reauthorization of the 2010 Child Nutrition Act for Farm to School and school garden programs.¹⁷

The main finding of this study is that after HHFKA enactment in SY 2010–11, school garden prevalence tended to increase until SY 2013–14. However, this increment was not sustained over time. Turner and colleagues²¹ examined elementary school garden prevalence changes in a nationally representative sample and found a steady increase from SY 2006–07 to SY 2013–14. The study findings are consistent for the overlapping study years, but in the current study the observed upward trend did not persist after SY 2013–14. A possible explanation is that federal funds for gardening programs did not succeed at sustaining an increase in school gardens within the study cities. Only $5 million was allocated to the Farm to School program nationally.¹⁷ It is quite likely that this modest amount of funding is not adequate to meet the demand to start and maintain gardens over time. Based on the United States Department of Agriculture data, in fiscal year 2013, only 71 garden projects were funded in the United States.³³ Although there has been tremendous interest in school gardens, inadequate skills and knowledge among school personnel have been considered a barrier to having a school garden.³⁴ While funds from HHFKA include provisions for technical assistance,¹⁷ which can provide the necessary training for school garden program personnel,²³ further research is needed to evaluate if this type of assistance adequately meets the needs of schools in diverse communities. In addition, there is ample evidence showing that community stakeholders are critical for creating, maintaining, and sustaining a successful garden program.^34–39 These stakeholders are also crucial for securing funding for a school garden program.^35,36,39 Sustaining high levels of engagement among unpaid staff or community members often depends on individuals who can bring together stakeholders and resources. In addition to federal laws, state and local laws can also play an important role. A study published in 2012 found that schools with Farm to School Programs were more common in states with Farm to School Program specific laws.⁴⁰ However, a more recent study published in 2017 found no similar association.⁴¹ Yet, it is noteworthy that in addition to the federal support through the HHFKA, New Jersey enacted the “New Jersey Farm to School and School Garden Fund,” in 2014.⁴² This new legislation created an opportunity to use community tax dollars to support Farm to School Programs and school garden grants. Despite this, we found no significant increases in school garden prevalence in SYs following 2014. This could show how lack of awareness of grant availability can play a role regardless of local and federal financial support. An example of that is reported by Hazzard and colleagues,³⁹ who compared schools applying for the California Instructional School Garden Program with those that did not, and found that almost two-thirds of non-applicants were unaware of the grant. Another constraint is lack of availability of teachers’ and school staff members’ time, which makes it difficult for them to take on additional (often unpaid) tasks, such as taking care of a school garden on a regular basis.^{34,37,39,43,44} Further, in the context of the current study, considering that New Jersey has cold winters, there are only two short seasons (Fall and Spring) where students could work in the garden, potentially affecting the incentive of having a garden.

Additional results from this study show that school gardens were less likely to be found in schools serving majority Hispanic students than majority Black students. Turner et al.,²¹ found that school garden prevalence was higher in non-white majority schools compared to white majority schools (22.0% and 17.1%). Future research should examine the effect of the HHFKA school garden grant and school garden prevalence through a racial/ethnic equity lens.

This study did not find any differences on garden prevalence by students’ FRPM eligibility. This result also differs from Turner and colleagues,²¹ who found garden programs were less common at schools with lower-socioeconomic status. This could be attributed to very high FRPM participation rates in our study sample, where over the study period, 69.2% of schools had over 80% of students eligible for FRPM.

This study fills an important gap in the literature by providing school garden prevalence rates before and after the implementation of the HHFKA. In addition, this study provides data on the prevalence and trend of school gardens in elementary and middle/high schools from SY 2010–11 to SY 2017–18. The study results confirmed the hypothesis that school garden prevalence increased the years following the HHFKA implementation; however, this increment was not sustained. In addition, the hypothesis about elementary schools having higher school garden prevalence was not supported.

Limitations

The presence of school garden was determined through a single survey question, and no additional information on size or scope of the garden was collected. This constraint limited the ability to study prevalence based on specific gardening program characteristics. Because schools close and open based on district needs, some schools changed locations, which might have affected the presence of garden. In addition, the current study did not include data on district wellness policies, state and local regulations, school district priorities, or how much time students spend working on the garden—all factors that are likely to affect school garden prevalence and sustainability. We also did not have information about schools being awarded the HHFKA grant. Lastly, the sample for this study was limited to high minority communities; as a result, the current findings are only generalizable to other similar school districts.

Conclusions

School garden prevalence tended to increase in the years immediately following the HHFKA implementation but was not maintained over time. There was no linear trend in the prevalence of school garden over the study period for the full sample, by school level, by majority race/ethnicity school, nor by students’ FRPM eligibility. Significant racial/ethnic disparities in school garden prevalence were observed that warrant further exploration in future studies.

IMPLICATIONS FOR SCHOOL HEALTH.

Recognizing the range of health benefits that school gardens offer to students is crucial to providing opportunities and incentives for the establishment and maintenance of school gardens. Policymakers should increase the money allocated to promote garden sustainability. These resources could support technical assistance for teachers and in-charge personnel. In addition, collaboration with school districts at the local level and the Department of Education at the state level could play a significant role due to their greater knowledge about communities’ needs, therefore enhancing the program’s outcomes. Furthermore, it is recommended to increase community outreach by making grant application information more accessible for schools and school districts and connecting with neighborhood committees, nonprofit organizations working with schools, and other stakeholders. Finally, the USDA should make its database of resources easier to navigate, especially regarding technical garden care and integration of gardens into school curriculum, thus facilitating better use of resources for teachers/school staff, garden volunteers, and other stakeholders.