Using Embeddedness Theory to Understand and Promote Persistence in STEM Majors

Student retention in science, technology, engineering, and mathematics (STEM) majors is a national concern. Research typically emphasizes attrition in STEM fields by examining factors driving students to switch career paths. This study used embeddedness theory to instead focus on factors that better anchor students within their majors. Focus groups were conducted with 21 junior and senior students majoring in STEM to contextualize the 3 tenets of embeddedness: fit, links, and sacrifice. Results supported embeddedness theory as a viable framework for understanding and promoting STEM student retention because fit, links, and sacrifice were all cited as factors that enriched the student experience. Gender differences suggested a heightened sense of pride for women as compared with men who pursue STEM degrees. Results inform ways to contextualize quantitative measures of STEM embeddedness. To facilitate retention, career counselors, instructors, and others may assist students to frame challenge as normal among persisting STEM upperclassmen.     

Response surface methodology in quality improvement

Recent work has lead to development of robust parameter design which alleviates criticisms of the Taguchi approach. A dual response surface approach is discussed and illustrated.     

CONCEPTS,THEORY, AND TECHNIQUES

The beta distribution is becoming more widely used in business applications and in economic model-building. A simple way to estimate the parameters of the univariate distribution and the general multivariate distribution would be useful but there appears to be no simple way of accomplishing this task. For example, the method of maximum likelihood does not yield tractable results. The method of moments presented in this paper is a practical way to obtain parameter estimates for the beta distribution. Explicit estimators are derived for the univariate and bivariate case and estimators are then inferred for the n-variate case. Empirical tests suggest that the method of moments may be used to establish useful estimates for these parameters. Also, estimates obtained by the method of moments may be used to develop bounds on the parameters so that computational techniques can be utilized with other estimation methods possessing more desirable properties.