Improving graduation rate estimates using regularly updated Markov chains

08/30/2020 ∙ by Shahab Boumi, et al. ∙ 0

American universities use a rolling six-year graduation rate (SYGR) to calculate statistics regarding their students final educational outcomes (graduate or not graduate). Meanwhile application of absorbing Markov chains (AMC) is commonly used to estimate graduation rates in research settings. In both cases a frequentest approach is used by counting the number of students who finished their program within six years for the standard SYGR method; in the case of Markov chains a frequentest approach is used to compute the associated transition matrix. Both approaches have significant limitations related to sensitivity when applied to small sample sizes or sub-populations at a university. In this paper, we use sensitivity analysis to compare the performance of the standard SYGR method, and absorbing Markov chains. We also propose and evaluate the use of a regularly updating Markov chain in which the transition matrix is updated year to year. Results indicate that the regularly updating Markov chain approach reduces the estimation variation by 50 especially for population with small sample sizes.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.