Joe Pacha and Norm Durflinger are both with the Educational Administration and Foundations Department at Illinois State University. Sherrilyn Billger is with ISU’s Department of Economics and Frank Beck is with ISU’s Department of Sociology.
I n the 1940’s, Illinois had more than 11,000 school districts, most of which were one-room schools. Since then, district consolidations and school closures have whittled the number to 866, as of July 1, 2010.
Little research exists about the predictors and outcomes of school closure. Research attempting to do so used only part of the measures for a limited time. If in-depth analyses of the causes and consequences of school closure have been studied, it has been on a case-by-case basis.
As Illinois State University researchers, we worked together three years gathering data from 1972 to 2005 for a study of Illinois school closure to answer the following questions:
• What are the demographic, economic and educational causes of school closure?
• What trends lead up to a closure decision and which are most important?
• What are the demographic, economic and educational impacts resulting from school closure? Are these effects immediate or do they manifest over time?
• Under what circumstances does the closure of a school bring about demographic, economic and educational benefits for a county, district or community?
We wanted to understand the relative size and importance of these forces, expecting that multiple factors are important and at work in these cases. We also wanted to understand the issues in order to help school boards and administrators better understand what is involved as they make difficult decisions concerning their schools.
We believe that our analysis confirms that school closure is not tied to the two most often cited issues — money and enrollment. Other issues and factors come into play in the decision making process and are important for both legislators and school leaders to understand.
The September/October issue of The Illinois School Board Journal looked at closure of elementary schools. This issue looks at junior high schools, which in the context of reporting information is interchangeable with a designation of “middle school.”
Predictors over time
Graphs 1-4 on the next page represent the significant predictors of junior high school closure over time: the education fund; per pupil operating expenditures; enrollment; and equalized assessed valuation (EAV). These graphs represent a 10-year history of each of these four variables just prior to the closure of junior high schools. It is important to remember that this is an aggregate of all the closed schools. The bottom line of the graph represents “time” and starts at 10 years and ends with closure at zero.
So what do these charts mean? What do they tell us about closing or not closing schools?
Graph 1, the education fund, shows a steady decrease over the nine years prior to the time of the school closure. This is predictable since there is a similar decline in the EAV shown in Graph 4, and that would also cause a decline in the ability of the school to raise funds. Couple this with Graph 3, enrollment, and it becomes apparent that declining enrollment and an inability to raise funds to make the school function widens to a point where the community is unable to provide the support needed.
Graph 2, per pupil operating expenditures, shows a steady increase throughout the 10-year period. What is striking is that as the enrollment falls, the cost to operate the school continues to increase at a steady rate. This gap between enrollment and costs to operate demonstrates one of the major factors in school closure.
Graph 3, school enrollment, is very telling and demonstrates why this is one of the main predictors of school closure. When declining enrollment is combined with increasing operating expenditures per pupil and decreasing education fund expenditures, the formula for school closure is high. Conversely, when enrollment increases, operating expenditures per pupil decrease and the education fund stabilizes, so the health of the junior high becomes more stable.
Graph 4, EAV, demonstrates a radically changing value that ends up lower over the 10-year period beforeschool closure. When compared to schools that remained open, this amount is significantly lower and is highly connected to the inability of the school to raise funds in the same manner as the open schools.
Additional predictors
Not all predictors can be shown over time and to the degree that they affect the closure overall. Analysis of the data suggests 27 predictors shown in Table 1 and Table 2 (page 17). These tables provide an overview of the predictors and their values in helping understand their relationship to junior high school closure.
Table 1 addresses the “educational/school” predictors of closure. The second addresses “community” predictors.
To read the tables, use the following formula: “Increasing (insert the variable name) (insert the column designation) the likelihood of closure.” The first variable would read: “Increasing enrollment is neutral (has no affect) toward the likelihood of closure” and the third variable would read, “Increasing expenditures per pupil decreases the likelihood of school closure.”
Let’s look at each of these variables to see what they mean and to better understand them in the context of the whole.
First, several variables have no influence on junior high closure whether they increase or decrease. These variables have been designated on the tables as “neutral.”
Five variables increase the likelihood of closure: percent of African American students; poverty rate; vacancy rate; percent involved with agriculture; percent with high school diplomas; and percent of residents with graduate degrees.
• An increase in the percent of African American students significantly increases the likelihood of school closure. To understand why an increase in the percentage of African American students in the building is a strong factor in the likelihood of junior high school closure is a phenomenon that definitely deserves more study. Statistics alone do not help us understand the reasons behind the data. This cannot be equated with an increase in the total number of students, which decreases the likelihood of closure in an elementary school but is neutral at this level. This was a neutral factor at the elementary level.
• An increase in the poverty rate increases the likelihood of school closure. Increases in the poverty rate in a school signals several factors that are hard for schools to overcome and often require extra resources: students who are in more need of extra help, students who have less resources to draw on and students who have fewer “life experiences.” Making up for these factors requires additional resources from the school. However, this is in direct contradiction with the findings for elementary schools, which showed a decrease in the likelihood of closure with an increase in the poverty rate.
• An increase in the community home vacancy rate increases the likelihood of school closure. The fewer people living in a community directly relates to other factors that play on school closure. Vacancy rate is a community measure that indicates the vitality of the community. While the vacancy rate was neutral for elementary schools, at the middle school/junior high level, it is a more significant factor.
• An increase in the percentage of workers in agriculture increases the likelihood of school closure. While this variable reacts oppositely for elementary schools, it needs more study to understand why this variable would react this way. While we speculated that community stability might play into this factor at the elementary level, it plays the opposite at the middle school/junior high level.
• An increase in the population with high school diplomas and graduate degrees increases the likelihood of school closure. Why would these two seemingly positive variables work against keeping a school open? The answer may be found in the extremes these two might exemplify. Communities with higher percentages of the population with graduate degrees often demand greater student achievement and services for the students. If those services are not provided, these same community people will look for “solutions” or alternatives to public schools.
Factors you might influence
Only three variables, if increased, decrease the likelihood of closure. Of these three, two appear to be under the control of school leadership. Therefore, knowing and understanding them is important, because influencing them is within the realm of the school leadership’s responsibility.
• An increase in the expenditures per pupil decreases the likelihood of closure. While this may seem counterintuitive, the fact remains that open schools spend more per pupil than closed schools did, and by a fairly significant amount. Therefore, this variable is a means of being able to address an issue of school closure.
• An increase in teacher experience decreases the likelihood of closure. This may also seem to be counter to what would be thought true. If you raise teacher experience this also raises the cost of the teachers and costs the district more. This variable is very much connected to the one above because an increase in teacher costs would increase the expenditures per pupil. Increased teacher experience is often related to increased student achievement and better student learning.
• An increase in the percentage of immigrants in the community decreases the likelihood of closure. This community variable could be explained in that immigrant families are very education minded and are determined to make sure their children attend school and do well. This would be helpful in a school that needs students to be students.
What is it that school board members and administrators must do when the above findings are similar to their own schools? Understanding which variables drive the predictability of closure and its effect can help the board and community know which areas to work on by recognizing which factors might help the most. Having that knowledge is far better than guessing at what to do. The closure of a school is a complex issue and all the variables must be reviewed to understand the complete picture.
What does it mean?
Breaking the conclusions into three parts will make it possible to better understand the overall ramifications of the power of this study. First, when comparing closed and open junior high schools without taking into account other similarities or differences, the following are important:
• Schools that closed had greater percentages of African American students than those remaining open.
• Schools that closed had a higher home vacancy rate in their community, were in more rural places, and had a higher percentage of people with high school degrees and graduate degrees.
• Schools that remained open had higher expenditures per pupil, had greater teacher experience, and higher percentages of immigrants in the community.
Second, when looking at closed junior highs alone, the following findings are important:
• A downward trend in enrollment precedes closure by four to eight years.
• A downward trend in the education fund seems to precede closure by two to five years.
• A pronounced upward trend in per pupil operating expenditures precedes closure decisions.
Finally, when comparing open and closed junior highs that are similar on all other characteristics measured, the findings are:
• Larger expenditures per pupil decrease the probability of closure.
• The economic health of communities (e.g. poverty, household income and unemployment) is inconsistently (and not strongly related) to the probability of closure.
• Larger immigrant populations in communities are linked to lower closure probabilities.
• Greater percentages of high school education levels and graduate degrees in the local population increase the likelihood of closure.
• The greatest school level predictor of closure is percentage of African American students in the school (and needs further study).
• When all school and community level effects are compared within levels of “ruralness,” schools in the smallest and most isolated of rural counties (i.e. counties with no single community larger than 2,500 residents) have the highest probability of closure. This is compared to schools in other rural counties with larger communities all the way up to and including schools in metropolitan counties. [This was not demonstrated in the tables and graphs above.]
When school board members and administrators compare the above findings to their own school(s), they may find one or two characteristics of a school that closed. Does that mean that their school is on a track that will lead to closure?
Probably not. Careful analysis of all of the predictors must be made, not just one or two. However, one or two factors can be a warning sign of things to come. Taking a careful look at all of the factors together can help give school leaders the answers they need.
Part I: Elementary Schools (September/October)
Part III: High Schools (January/February)
For more information
If individual schools would like a quick determination of the probability of closure, please contact the authors. By providing some basic data this can be done in a short amount of time. In-depth analysis of the probability of school closure will take additional time and research but can be accomplished for a nominal fee to cover expenses.
For additional information concerning individual school closure predicted probabilities contact:
• Frank Beck — or 309-438-7770
• Sherrilyn Billger — or 309-438-8720
• Norm Durflinger — or 309-438-8989
• Joseph Pacha — or 309-438-8575
References:
Alan J. DeYoung, The Life and Death of a Rural American High School: Farewell Little Kanawha, New York: Garland, 1995
David R. Reynonds, There Goes the Neighborhood: Rural School Consolidation at the Grass Roots in Early Twentieth-Century Iowa, Iowa City, Iowa: University of Iowa Press, 1999