Estimating the Impact of Alabama's First Class Pre-K on K-12 Teacher Turnover

Methods

Summary

Complimentary data sets will be blended using Alteryx. These include comprehensive collections containing student demographics, discipline, attendance, and academic performance, and student-teacher matches including courses, with teacher demographics, salaries, credentials, education, and experience.

Challenges

I currently have data from 2009-2019 for some sets, but I await additional data from the post-pandemic period (2020-2022). Previous data requests have been slow. Meeting the proposed timeline targets and milestones will depend on having the required data in a timely manner.

At a technical level, I will not have a student-teacher-incident match, so any such connection would be inferential.  Also, some students have had multiple incidents, with different infractions and an assortment of dispositions. This will require a pivot of the data to pair with demographic data for statistical analysis.

Randomly selected exact matching will be used to make statistical comparison sets between student peer groups. Comparisons will be fully blocked with randomly selected peers on race, ethnicity, gender, age, grade, and context (within 5% of school poverty, percent black).