AE Rodriguez
March 17, 2022
More Specifically: the Admissions Intake Process
To discuss dysfunction in the intake processes. Intake processes (an element of supply chains) have come under scrutiny lately and will probably remain an issue in the future near and not so near.
To examine machine learning processes: their uses, applications and limitations in business operations. Here I focus on OneClass algorithms and Propensity Score Matching.
Possible alternatives: Semi-Supervised Learning, Misclassification
Between 2025 and 2029, the number of college-going students will decline by 15%—that’s over 400,000 fewer students in a span of four years, an average loss of 100,000 students per year.
Afia Tasneem, EAB.com (Feb 2019)
After a 4-year slide due to Trump administration immigration policies, enrollment of new international students fell dramatically in fall 2020 because of COVID-19, according to the Institute of International Education.
Andrea Widener, C&EN (Nov 2020)
Source: Forbes
- Since their inception a century ago, standardized tests have been instruments of racism and a biased system.
- Students of color, particularly those from low-income families, have suffered the most from high-stakes testing in U.S. public schools.
- Today, a movement is growing across the country to resist testing abuse and overuse, and to promote authentic assessment.
National Education Association,
By: John Rosales and Tim Walker
Published: 03/20/2021
RIP SAT? Harvard makes standardized tests optional through 2026
Harvard announced it won’t require SAT or ACT scores through 2026. If other schools follow suit, admissions offices will be ‘looking at applications they would not have seen previously,’ says Jeff Selingo, author of ‘Who Gets in and Why’
Source: CNN, SMerconish, Dec 2021
Here’s the general logistic regression model:
Prob{Y=1|X}=11+exp(−Xβ)
The X represents our predictors. The β represents weights or coefficients for our predictors.
Our three predictors, age, experience, and gpa
then Xβ=β1*age+β2*experience + β3*gpa
In order to reduce the selection bias, Propensity Score Adjustment can be used in this case for
reweighting.