Previous research indicates first term GPA is one of the best predictors of first year retention (Zhang, 2006; Luo, Williams, & Vieweg, 2007; Hosch, 2008; Cambra & Stanley, 2008; and Colorado State University, 2013). The next logical question is what factors are the best predictors of first term GPA? Knowing what factors best predict first term GPA might provide valuable insights and predictive analytics for improving first term GPA and consequential first year retention outcomes.
Data, Variables and Statistical Analysis
To gain an understanding of variables that are associated with first term GPA I exploit a dataset from a midwest community college. The variables of the analysis are summarized in the following table. Information for all variables were obtained for 10,218 students.
The variable, accel, requires more elaboration. It captures high school students who have earned college credit. Programs that provide this opportunity are known as dual enrollment, concurrent enrollment, Tech Prep, and STEM.
Multiple regression was chosen as the statistical tool of choice as the dependent variable, first term GPA (ftgpa), is a continuous variable and the predictors are a mixture of categorical and continuous variables. Multiple regression provides an opportunity to assess the association of each predictor on ftgpa.
Multiple Regression Output and Interpretation
The following represents the output from regressing ftgpa on the seven predictors.
First, we observe the full model with the seven predictors is statistically significant (Prob F > 0.0000). However, gender is not statistically significant so we will drop the variable ‘female’ and recast the regression with six predictors, resulting in the following output.
The full model as well as the six predictors are statistically significant. The model accounts for 32% of the variance in first term GPA. The regression coefficients may be interpreted in the following manner:
- White students have a .10 first term GPA advantage when compared to non-whites, all other variables set to their means.
- Accelerated students achieve a .45 first term GPA advantage over non-accelerated students, ceteris paribus.
- First generation students exhibit a -.56 (lower) first term GPA compared to students with backgrounds from families with previous experience in higher education, controlling for all other covariates.
- A one unit increase in fall term credit hours is associated with a .08 increase in first term GPA, holding all other predictors to their means. Thus an additional 3 credit course would be expected to yield a .24 increase in first term GPA, controlling for all other independent variables.
- A one unit increase in high school GPA is associated with a .69 increase in first term GPA, holding constant all other variables.
- Each additional unit increase in high school graduation year decreases students first term GPA by a factor of .06. (Older students achieve a higher first term GPA than more recent high school graduates.)
It’s useful to observe the association of predictors with first term GPA at specified levels of the predictors, controlling for the influence of the other covariates. The following graphs provide that visual opportunity.
I. Predicting First Term GPA by Race and High School GPA
II. Predicting First Term GPA by Year of High School Graduation
IV. Predicting First Term GPA by First Generation and High School GPA
V. Predicting First Term GPA by First Term Credits and Race
With data in hand and the completed multiple regression analysis it is possible to predict each student’s first-term GPA. For example, the following table is the outcome associated with asking my statistical software program (Stata) to list the first five students with predicted first term GPAs less than 2.0.
All five are white, none of them have been accelerated, three are first generation college students, all are part-time students (based on their credit hours), all have high school GPAs under 2.0 and all are older than traditional aged college students.
With predictive analytics community colleges can exploit this strategic information to design intervention programs which improve mission fulfillment, student outcomes and the financial position of the college.
Implications for Open Door Community Colleges
Open door community colleges admit individuals who have a high school diploma or GED certificate. ( I hasten to add that I have years of data indicating students who graduate from community colleges and then transfer to a four-year college do as well or better than cohorts who entered the university as freshmen.)
What can community colleges do to improve first term GPA with the expectation that such improvements yield positive retention outcomes? How can they accomplish this objective without compromising their missions?
- Use “predictive analytics” to identify high-risk students and sub-groups unlikely to achieve a 2.0 GPA during the first term.
- Provide financial support for long-term intervention programs. In short, institutionalize strategic initiatives to improve first term GPA and consequential retention.
- Use the above information for “targeted”student advising.
- Develop initiatives which promote credit momentum. Educate students on the cost implications for not graduating on‐time. Recall, an additional 3 credit course is associated on average with a .24 increase in first term GPA, holding all other predictors to their means. This strategy appears to be especially productive for accelerated students. See graph VII.
- Improve recruitment and admissions of accelerated students. (Recall accelerated students have a .45 first term GPA advantage over non-accelerated students.)
- Increase the number of students participating in accelerated programs (dual enrollment, concurrent enrollment, Tech Prep, and STEM.)
- Improve recruitment, admissions and financial aid for older students. (Older students achieve a higher first term GPA than more recent high school graduates.)
- While community colleges are open-door institutions there’s certainly no mission conflict in recruiting students with higher GPAs in high school.
- Consider offering a “summer bridge” program for at-risk students.
- Provide staff development opportunities for the “freshman year experience.”
- “Measure it to improve it.” Adopt CQI strategies and use the information to improve first term GPAs.
- Establish a “First-Term GPA & Retention Committee,” reporting directly to the President’s Council. Establish clear and measurable yearly objectives, which involve the buy-in of all stakeholders. Annual progress reports, including both sub-group and aggregate analyses, should be routinely available for all faculty and staff, but most importantly, acted upon.
- Never forget the power of faculty/staff reaching out to students.
- Celebrate success — with students! If we desire higher first term GPAs and consequential retention outcomes then we need to make those goals explicit and provide appropriate recognition, rewards and status on the basis of achieving prescribed goals.
Cambra, Ron & Stanley, John. (2008). Access to Success:Leading Indicators Workgroup. The University of Hawaiʻi at Mānoa.
Colorado State University. (2013). Early Indicators of Student Progress and Success. Office of Institutional Research.
Hosch, Braden J. (2008). Institutional and Student Characteristics that Predict Graduation and Retention Rates. Paper presented at the North East Association for Institutional Research Annual Meeting, Providence, Rhode Island.
Luo, M., Williams, J. E., Vieweg, B. (2008). Transitioning transfer students: Interactive factors that influence first-year retention. College and University, 83(2), 8-19.
Zhang, Biao. (2006). Factors Affecting 1st Year Transfer Retention at a Midwest University.