We collected data from January to May 2011 before implementation of SB 1413 and the Act. We selected public schools from the National Center for Education Statistics’ Common Core of Data (CCD) (13). The 1,313 standard California public schools (excluding private, K-8 or K-12, special educational, vocational, and alternative schools) were from California’s Bay Area region (Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma counties). To understand water provision in schools of varying type and location, we stratified eligible schools by CCD urban-centric locale and school type. We collapsed locales (large, midsize, or small cities; large, midsize, or small suburbs; fringe, distant, or remote towns; and fringe, distant, or remote rural areas) into 4 categories: city, suburb, town, and rural area. In these categories, we further stratified schools by primary, middle, and high school to generate 12 separate sampling strata (eg, city high, city middle, rural elementary). Using a random number generator, we generated a random list of all eligible California schools for each of the 12 strata.

We developed a telephone survey and observational tool to assess provision of drinking water and policies and practices related to school drinking water on the basis of a literature review, a previous qualitative study conducted by the first author (A.I.P.) on access to school drinking water (14), and feedback from study collaborators. To improve the content validity of the survey and observational tool, we used a snowball sampling approach to identify 15 experts with content knowledge of drinking water access and policies and practices related to drinking water access in schools. We used these experts’ comments to revise the survey and observational tool, which we then pilot tested with 10 staff from ineligible schools (ie, located outside of the Bay Area or nonstandard public schools).

The observational tool assessed the following domains: free drinking water access and quality, student drinking water intake in FSAs, and bottled water and other beverages available for purchase. We also reviewed school documents to assess school drinking water policies.

Two observers used the tool to code access to free drinking water, characterized as the school locations where drinking water was available (ie, FSAs) and the type of water source available (eg, fountains, water dispensers). Observers also documented drinking water quality by coding the temperature (1 = very cold, 2 = cold, 3 = room temperature, 4 = warm), clarity (1 = clear, 2 = cloudy, 3 = yellow, 4 = brown), and flow strength (1 = high, 2 = medium, 3 = low, 4 = none) of free drinking water in school FSAs. Observers also characterized the cleanliness of the drinking water sources (1 = very clean, 2 = clean, 3 = unclean, 4 = very unclean).

Observers assessed student intake of water in FSAs by counting the number of students who drank free drinking water in the FSA during all lunch periods on an observation day. Although students may eat lunch in locations throughout the campus, it is impractical to ensure that drinking water is accessible in all of these locations. For these reasons, we defined an FSA as an area within 100 feet of where reimbursable meals were served. The water source observed was the one closest to the FSA. We estimated the percentage of students observed drinking water at lunch by dividing the number of students observed drinking water at lunch by the daily attendance.

Observers also recorded the type, location, and cost of beverages provided with the school lunch and available for purchase. After the visit, we obtained wellness policies from each participating school’s district for analyses of language regarding drinking water provision.

The school administrator survey measured drinking water availability (ie, type, location, and source), drinking water policies and practices (eg, quality and fountain accessibility or maintenance), and barriers (eg, cost, policies and contracts, student behavior) to school drinking water access. We obtained school sociodemographic data from the Education Data Partnership (15).

We mailed an invitation letter to principals of eligible schools. A research assistant (RA) then contacted study participants to confirm receipt of materials, answer questions, and schedule an interview time. RAs contacted potential study participants until they declined. If a school declined, we sampled the next randomly chosen school from the study stratum. To obtain our goal sample size of 24 schools, we had to contact 44 schools from the overall sampling frame of randomly selected California schools. We obtained consent from survey respondents (eg, school principals, school administrators) before collecting data.

RAs or the first author (A.I.P.) conducted all surveys. We audio recorded surveys, which lasted 10 to 20 minutes. After conducting the surveys, RAs scheduled an observational visit. Each RA had extensive training using the observational protocol before data collection; 2 RAs and the principal investigator (A.I.P.) visited schools until consistency in coding was achieved (κ > 0.80) for all variables. Thereafter, 2 RAs visited each school and simultaneously coded all observational data on paper checklists. Kappa statistics ranged from 0.88 to 1.0 for observational variables, indicating excellent interobserver reliability.

We provided study participants with $50 gift cards for participation. University of California, San Francisco’s Committee on Human Research approved the study.

We analyzed data using Stata version 11 (StataCorp LP, College Station, Texas). We used descriptive statistics (frequencies, percentages, and means) to summarize school characteristics (ie, school population, racial/ethnic breakdown, percentage of students eligible for free and reduced-price meals through the National School Lunch Program [NSLP]), and main study outcomes (eg, drinking water availability in FSAs and students’ water intake). We used t tests and χ² tests to assess the differences between study participants and nonparticipants.

Of the 14 schools that offered free drinking water in FSAs, 10 offered water via fountains, while the remaining schools had an alternative delivery system (Table 2). Beverages most commonly offered with the NSLP included 1%/skim unflavored and flavored milk; 100% juice and bottled water were offered in some schools. Only 1 elementary school offered competitive beverages or beverages for purchase in the FSA. In secondary schools, the most commonly available competitive beverages were bottled water and sports drinks. The mean price of bottled water in FSAs was $0.92 per bottle. Despite legislation (Senate Bill 965) that since 2009 has prohibited the sale of SSBs other than sports drinks in California public schools, 1 school offered “slushies,” a frozen flavored beverage.

At schools with free water in FSAs, only 4% of the 11,226 students in daily attendance were observed drinking free water at lunch. The percentage was highest in elementary schools, followed by middle and then high schools (Figure 1). In the schools that did provide water in FSAs, most provided water of good overall quality; the mean water temperature was cold (1.9), the mean water clarity was clear (1), mean water flow was of medium strength (1.8), and mean cleanliness of the water delivery system was very clean to clean (1.6). Only 1 school dispensed water that was not cold or very cold and only 2 schools dispensed water of low or no strength. Although all water delivery systems in school FSAs were described as clean, 2 fountains contained gum and 1 fountain contained a small amount of dirt.

We photographed alternative water delivery systems in school FSAs (Figure 2). Only 4 schools, all secondary, offered water through a nonfountain source. One high school purchased a dispenser that dispensed filtered hot and cold tap water. According to the school, the dispenser was installed because no other functional drinking water source was available in the school. A school administrator estimated that the unit, paid for by the nonprofit organization that operates the charter school, costs $200 to $400. Filter changes required every few months were estimated as $20 per change. Students were expected to bring their own container to access water from the dispenser. At this same school, 1 student was selling a plastic foldable bisphenol A (BPA)–free water bottle at the school as a fundraiser ($4 per bottle).

Another high school installed a hydration station with a mural backsplash displaying promotional messages to encourage water intake. The hydration station was installed by the food service director, a champion of health and nutrition. The cost of the station, estimated at $2,000, was paid for through the food services budget, which often covers ancillary costs (eg, trays, napkins) associated with meal service. Students at this school, similar to the school with a dispenser, were expected to bring their own containers for drinking water due to concerns about waste (eg, cups) not being discarded appropriately in trash receptacles.

Half (12 of 24) of school administrators had heard about the Act or SB 1413. In our study, 14 schools had water in the FSA before implementation of SB 1413 and the Act. Of these 14 schools, 3 reported that they began providing water in the FSA as a result of hearing about SB 1413. Administrators most commonly cited more pressing academic concerns (n = 19) and the cost of new water programs (n = 16) as barriers to implementation of water requirements.

As a result of hearing about SB 1413, food services departments in 2 middle schools filled water dispensers with ice water and provided water with disposable cups at lunch. At both of these schools, dispensers were on hand at the school, and food services absorbed the cost of cups. Per estimates, water coolers cost approximately $30 each and cups are $0.02 to $0.05 each for 4- to 6-ounce foam cups and Dixie-brand cups.

Among wellness policies for the 20 school districts represented in our study (4 schools were in duplicate districts), 11 included language related to drinking water (Table 3), but none specifically mentioned that free drinking water should be available in FSAs. Three district wellness policies contained water-related language in more than 1 thematic area (eg, water allowable as a competitive beverage, provision of water with snacks). Only half of schools that offered water via a nonfountain source had water-related language in their wellness policy.