The US national graduation rate is 84.1%. However, it is estimated that a “90 percent high school graduation rate would likely create more than 65,000 new jobs and boost gross domestic product by $11.5 billion annually.”[1] Certainly a worthy goal and one that would compound for each graduating class.

Currently only two states, Iowa and New Jersey have a 90% high school graduation rate.[2] The big question for this research became: could a model be developed to predict graduation rates and could the model’s variables be changed to predict a 90% graduation rate for individual high schools.

The Data

The Department of Education graduation data provides Adjusted Cohort Graduation Rates (ACGR) for 23,090 high schools. The four year ACGR is the number of students who graduate in four years or less with a regular high school diploma divided by the number of students who formed the cohort for that graduating class, starting in 9th grade. However, to protect student privacy, if there are less than 5 students in the cohort, no data is provided and for cohorts less than 200, a range is given. The smaller the cohort the wider the range will be. When a range is given, this research assumed the average number of the range.

The Office for Civil Rights (OCR) data has 1,836 variables including bullying, teacher-student ratio, certification, absenteeism, sports participation, and more divided by gender, ethnicity, LEP, and 504 subsets.

The Model

95% Confidence Intervals of several variables that influence US high school graduation rates.

Of the many different variables, 12 were chosen to measure their correlation with graduation rates. On the chart, the 95% confidence intervals are plotted. If the 95% confidence interval includes the zero dotted line, one can assume with 95% confidence that the variable has zero influence on graduation rates. Three variables included zero, teacher/student ratio, harassment/bully allegation ratio, biology class enrollment. Three other variables were very close to zero, but did not include it, teacher absenteeism, in-school suspension rate, and money per student spent on activities. Of the remaining six, this research choose to focus on three variables, chronic absenteeism and AP class participation most strongly influenced graduation and sports participation also seems to be an important positive influence on graduation rates.

 

 

 

Model Prediction

If a school principal or district is to use the model, it must be a simple model. Complicated models are hard to remember and focus on, simple models can be implemented. Rather than use all six variables, this research focused on three that had significant influence. The model shows the result below with several related metrics. 

In order to test the model, three high schools were chosen that had less than 90% graduation. John F. Kennedy High School was chosen as it one of the lower performing high schools in Montgomery County, Maryland, a wealthy suburb of Washington, DC. H.D. Woodson High School in Washington, DC has previously been identified as a high school with an improving graduation rate in the District of Columbia Public Schools system, which has historically low graduation rates. South Lakes High School is a graduation rate similar to the Virginia state rate. The three high schools also have different enrollments, John F. Kennedy High School has about 1,500 students; H.D. Woodson has about 775 students and South Lakes High School has about 2,400 students.

Predicted vs. Actual Graduation Rates for three high schools, South Lakes High School, H.D. Woodson High School, and John F. Kennedy High School.

The chart shows the predicted graduation rate using the model and the actual graduation rate as reported by the Department of Education. The last three columns show if a 90% prediction rate was possible and how close it could be to 90%.

 

 

 

 

 

 

 

 

90% Prediction

Changes Necessary to Achieve 90% Graduation . Estimates for three high schools

This chart shows what a specific school would have to change to achieve a predicted 90% graduation rate. All of them would have to increase of students that participate in a high school sport by 250 students. They would all have to increase the number of students taking AP classes. All the schools would have to make a greater effort to reduce chronic absenteeism among their student populations. However, the good news is that all these changes are possible and within the range what could be expected over the next three years, if school administrators and leaders devoted the time, money and energy to the effort.

 

 

 

Links to the data:

https://www2.ed.gov/about/inits/ed/edfacts/data-files/acgr-sch-sy2013-14.csv

https://www2.ed.gov/about/inits/ed/edfacts/data-files/acgr-sch-sy2015-16.csv

https://www2.ed.gov/about/offices/list/ocr/docs/crdc-2013-14.html

https://www2.ed.gov/about/offices/list/ocr/docs/crdc-2015-16.html

Notes:

[1] http://gradnation.americaspromise.org/about

[2] http://gradnation.americaspromise.org/