KIPP LA Schools (KIPP LA) is a nonprofit organization that operates free, open-enrollment, public charter schools focused on preparing students in under-resourced communities in Los Angeles for success in school, college, and life. With seven elementary and six middle schools in East and South Los Angeles, KIPP LA serves nearly 5,000 students or “KIPPsters” and supports nearly 1100 alumni in reaching high school and college with their KIPP Through College program.
When a child enters a KIPP LA elementary or middle school, the students, parents, and teachers make a promise to help the student successfully “climb the mountain to and through college.” But how can teachers and parents know if a student is indeed on the right path and help guide them accordingly if they’re not?
To help this volume of students make it to college, their six-person data team supports its schools, teachers, and regional leaders with data management and analytics. They wondered, could they use their demographic, academic, behavioral and socio-economic data from students in their elementary and middle schools to predict a student’s level of college readiness and provide more customized support and interventions?
At the 2015 Teradata Cares DataDive, Data Ambassadors William High, Farrukh Ali and Surbhi led a team of volunteers to do exploratory data analysis and probabilistic modeling to see if they could help KIPP LA take the first steps towards understanding what factors impact a student’s level of college readiness.
Over two days, the teams looked to answer, do middle school student records contain indicators of college acceptance? Were there any correlations between college entrance and test scores with past test scores, grades, or attendance that could help explain admission success? And ultimately, could the team help KIPP design a data-driven approach to support the next generation of college students?
To better understand the data itself, the team first created data dictionaries, defining the various fields within the data and their relationships to one another, then plotted histograms to understand the distribution of the data. Next, the team identified key variables for each question they were trying to answer and began modeling the data using a variety of techniques like linear regression to understand the relationship between variables, k-means clustering to reveal any natural groupings within the data, and two types of predictive models to see which variables might be predictive of college acceptance.
Of the many analyses the team did, they did find that the eighth grade Measurement of Academic Progress or MAP score, a standardized test used in the U.S. to assess student progress, was indicative of college acceptance. As shown in the correlational analysis above, the higher a student’s MAP test scores in middle school reading and mathematics, the more options for college the student has. This is potentially an important signal for parents and teachers as it occurs four years before the important college decision is made so there is opportunity to intervene.
KIPP LA is exploring further analysis of how achievement and standardized test data impacts college readiness of its students. They would ultimately like to build a data-informed curriculum that features teacher-facing dashboards and student-facing report cards to inform individuals on their progress and are exploring doing a longer-term DataCorps project to do just this.
It’s clear that the path of “climbing the mountain to and through college” requires paying attention early on and looking for signs like middle school reading and mathematics MAP scores. When assessing the probability a student will not only enter a good college but complete their studies, it’s important to consider other factors including effectiveness of study habits for college level coursework, whether the student has to work (and how many hours), and social readiness. This understanding will help KIPP LA build the knowledge and tools for teachers, parents and students to ensure that KIPPsters go to and through college.