>School Lunch Program Enrollement

School Lunch Program Enrollement


Public assistance comes in various forms and indicates local or regional poverty rates. One of the best measures of this is enrollment of children in free and reduced cost meals programs. These data are available for individual schools. Rates of enrollment are also available by ethnicity.
Category: 
Social Condition
Objective: 
Economic Condition

QT: Indicators

What is it?

Enrollment of children in school lunch programs is considered to be a sensitive measure of poverty at the sub-municipal scale. Children between 6 and 17-years-old are eligible if family income is less than the federal poverty level. Data are available for every school that is participating in the federal program, including schools in the study area, for the last 20 years.

Why is it important?

Poverty and income inequities are correlated with reduced life expectancy (Singh and Siahpush, 2006), child well-being (Pickett and Wilkinson, 2007), and academic performance (Caldas and Bankston, 1997). Enrollment in school lunch programs is an extensive (data available for every school), but fairly general indicator of poverty. We can answer questions related to rate of poverty for individual schools (K-12) and change in this rate over at least the last 20 years. Because rates of enrollment are available for each school, correlations can be drawn between this poverty indicator and other municipal or subwatershed condition and trends in condition.

Target or Desired Condition

Community economic conditions can affect opportunities and sense of welfare for children and adults. Absent a state or local policy that defines an acceptable level of poverty, this analysis rates 0% enrollment in a school lunch program as a good target score (100) and 100% school lunch program enrollment as a poor score (0). A linear function was used to calculate score, where Score = 100% - % children enrolled.

What can influence or stress condition?
Poverty is caused by a variety of factors, including employment availability, legacy of poverty, regional economy, and skills for employment. In this region, historic mining, logging, and contemporary agriculture provide much of the land-based income. Over the last few decades, influx of retirees and ex-urban migration has led to changing demographics, including income. Global and statewide economic trends are likely to influence community economic condition. Communities that derive their economic well-being from productivity that is not controlled or commodified by global markets may be less negatively impacted by economic declines.

Data Sources

California Department of Education; USGS Geo-Names Database (http://gis.ca.gov).

Data Transformations

Data were manually assembled from downloadable files. For school years where the year was given by “88/89” or similar, a new column was created and actual year-dates manually entered corresponding to the end of the school year (e.g., “1989”). Only percent of students receiving “free meals,” as opposed to reduced-price, were calculated and used to be consistent over the whole time-span.

Condition Analyses

The percentage of students receiving free meals was extracted from the California Department of Education database for 2008, the last year with complete data. The location of the school was determined using the USGS Geo-Names database in ArcGIS. Each school was attributed with a subwatershed (e.g., North Fork Feather). For each subwatershed, mean, standard deviation, minimum, and maximum percentage of children receiving free lunches were calculated.

Trends Analyses

Random numbers were assigned to each school in the watershed using the Microsoft Excel random number generator. The 20% of schools with the lowest random numbers were chosen for trends analysis, resulting in trends for 20 schools being analyzed. The Mann-Kendall trends analysis was used, using the methods described in Section 4.3.

What did we find out/How are we doing?

The economic condition score in the last year for which there was complete data (2008) varied by subwatershed (Figure 1 and Table 1), ranging from a score of 32 (Lower Feather subwatershed) to a score of 75 (Upper Bear subwatershed). At the scale of individual schools, enrollment rates varied from 0% to 100%.

The high score for the Deer Creek Watershed suggests that communities in that watershed have lower levels of poverty than other subwatersheds. The low score for the Lower Feather and Middle Yuba watersheds suggests that communities in these watersheds have higher levels of poverty.

Trends Analysis

School lunch enrollment rates at 20 randomly-selected schools were analyzed over the 20-year time period of data (1989 to 2009), excluding 2004 and 2005 data because of rates of enrollment that exceeded 100% for these years (Figure 2). For 11 schools there were statistically-significant increases in enrollment over this time, meaning that community economic conditions were worsening. The trends varied from 0.4% to 2% increases in enrollment per year. For only one school (Nuestro Elementary School in Sutter County) was there a statistically-significant decrease in enrollment (1.5%/year). The remaining schools showed no change in enrollment. An increase in enrollment corresponds to decreasing condition scores.

Temporal and spatial resolution

Because there are data for every school and every year, spatial and temporal resolution for this indicator is moderate. Annual enrollments are reported for each school participating in the program. Therefore, the assessment can be updated annually, unless monthly data were to be collected from individual schools.

All subwatersheds had at least one school, but there was a wide difference in number of schools between the Lower Feather (51 schools) and the North Yuba (1 school), which affects the calculation and meaning of the average score for each subwatershed.

How sure are we about our findings (Things to keep in mind)

The data used for this indicator are based on school enrollment figures, reported to the California Department of Education. They are complete for 1989 to 2008, with potential problems with data for certain schools in 2004 and 2005 (e.g., enrollment rates >100%). The precision of these data is likely very high. The calculated average score for each subwatershed reflects the average condition for that area, the 95% confidence interval and minimum and maximum scores reflect the variation around the averages, which can be fairly large (Table 2). One data gap in this analysis is that not all schools or parts of schools enrolled in the Free and Reduced Priced Meal Program have location information that can be used in GIS.

Enrollment is based on a family being below federal poverty level. This means that the metric is not particularly sensitive to geography-specific cost of living variation (Curran et al., 2006; Heflin et al., 2009), which is a limitation in its use. There may also be an effect of peer-pressure on children’s desire for enrollment.

Overall our confidence is high in the precision of the indicator, moderate and variable about how well the average value for each subwatershed reflects conditions, and moderate to high for how well the indicator reflects community economic well-being.

Sub-watershed
Area Code Score Trend Confidence
North Fork Feather NFF 52 Downward
East Branch North Fork Feather EBNFF 49
Middle Fork Feather MFF 54 Downward
Lower Feather LF 34 Downward
North Yuba NY 64 Downward
Middle Yuba MY 32 Downward
South Yuba SY 40
Lower Yuba LY 35 Downward
Upper Bear UB 70 Downward
Lower Bear LB 61