Analytical injustice league: understanding statistical manipulation of student retention data using modification methods of missing values
Date
2021
Authors
Long, Sarah E., author
Gloeckner, Gene, advisor
Anderson, Sharon, committee member
Folkestad, James, committee member
Eakman, Aaron, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Missing values that fail to be appropriately accounted for may lead to reduced statistical power, biased estimators, reduced representativeness of the sample, and incorrect interpretations and conclusions (Gorelick, 2006). The current study provided an ontological perspective of data manipulation by explaining how statistical results can fundamentally change depending on specific data modification methods. This has consequential implications, specifically in higher education, that depend on quantifiable methodologies to substantiate practices through evidence based policy making (Gillborn et al., 2018; Sindhi et al., 2019). The results of the current study exposed how examining patterns of data missingness can have critical implications on student retention initiatives including intervention programs, identification of high-risk students, and funding opportunities for support programs. It is imperative for both data scientists and data stakeholders to be critically aware of what data they collect, report, and utilize from the variable selection to statistical methodologies.
Description
Rights Access
Subject
first-year retention
quantitative methodology
missing data
critical race theory