Using Logistic Regression to Examine College Course Delivery Effectiveness

Conference Registration Edit Registration WebCT Homepage

Thursday, 21 July 2005
This presentation is part of 4: Poster Session
Using Logistic Regression to Examine College Course Delivery Effectiveness

Description: Logistic regression models and six years (fall 1998 – spring 2004) of enrollments (n = 263,717) were used to examine the likelihood of enrollment completion for courses delivered via the internet and in-person. Independent variables included gender, ethnicity, age, GPA, cumulative credit hours, transferred credit hours, and year of the enrollment. The methods and results should be of interest to those who want to evaluate the effectiveness of alternative modes of course delivery.
Presentation Format: Poster
Topic: Measuring outcomes: Student achievement and course success
Target Audience: Academic Technology Directors, Course Designers, E-learning Managers, Faculty and Other Instructors, Institutional Research and Assessment Staff, K-12 Educational Staff
Appropriate Audience Level: Beginning or new users of WebCT, Experienced WebCT users
WebCT Version:
Abstract Text: Higher education faces scarce resources, increasing enrollments, alternative course delivery methods proposed as the solution, relative effectiveness of alternative course delivery methods debated, and the need for appropriate comparison methods. Concern has been expressed that studies of have excluded dropouts, tilting the outcome findings toward those who are successful. The thesis of this study is that enrollment completion can be used as an indicator of institutional effectiveness and used to compare delivery methods.
SPSS's binary logistic regression was used to model completion using six years (fall 1998 – spring 2004) of enrollment data (n = 263,717). Results indicate that enrollments in sections delivered in-person were more likely to be completed (Odds ratio 2.5), likelihood of completing an enrollment decreased over the six years (Odds ratio 0.88) for in-person delivered and only slightly (if at all) in internet (Odds ratio 0.98, p-value 0.062), females were more likely than males to complete enrollments (Odds ratio 1.16) and even more likely than males for internet enrollments (Odds ratio 1.3), and students with higher GPAs were more likely to complete enrollments (Odds 3.8 times higher for a 1 point higher GPA) for either delivery method.
The methods and results should be of interest to those who want to evaluate the effectiveness of alternative modes of course delivery.

See more of Poster Session
See more of The 7th Annual WebCT User Conference

Description:	Logistic regression models and six years (fall 1998 – spring 2004) of enrollments (n = 263,717) were used to examine the likelihood of enrollment completion for courses delivered via the internet and in-person. Independent variables included gender, ethnicity, age, GPA, cumulative credit hours, transferred credit hours, and year of the enrollment. The methods and results should be of interest to those who want to evaluate the effectiveness of alternative modes of course delivery.
Presentation Format:	Poster
Topic:	Measuring outcomes: Student achievement and course success
Target Audience:	Academic Technology Directors, Course Designers, E-learning Managers, Faculty and Other Instructors, Institutional Research and Assessment Staff, K-12 Educational Staff
Appropriate Audience Level:	Beginning or new users of WebCT, Experienced WebCT users
WebCT Version:
Abstract Text:	Higher education faces scarce resources, increasing enrollments, alternative course delivery methods proposed as the solution, relative effectiveness of alternative course delivery methods debated, and the need for appropriate comparison methods. Concern has been expressed that studies of have excluded dropouts, tilting the outcome findings toward those who are successful. The thesis of this study is that enrollment completion can be used as an indicator of institutional effectiveness and used to compare delivery methods. SPSS's binary logistic regression was used to model completion using six years (fall 1998 – spring 2004) of enrollment data (n = 263,717). Results indicate that enrollments in sections delivered in-person were more likely to be completed (Odds ratio 2.5), likelihood of completing an enrollment decreased over the six years (Odds ratio 0.88) for in-person delivered and only slightly (if at all) in internet (Odds ratio 0.98, p-value 0.062), females were more likely than males to complete enrollments (Odds ratio 1.16) and even more likely than males for internet enrollments (Odds ratio 1.3), and students with higher GPAs were more likely to complete enrollments (Odds 3.8 times higher for a 1 point higher GPA) for either delivery method. The methods and results should be of interest to those who want to evaluate the effectiveness of alternative modes of course delivery.