Preliminary research suggests that certain student behaviors can be early signals of success or struggle in college and beyond. For example, early enrollment and completion of required courses might indicate favorable prospects, while late registration and switching majors can foreshadow trouble ahead, particularly for first-generation students, students of color and students from low-income backgrounds.

Some states have developed student databases with information from high school through college and into the labor market. But restricted and siloed access to these databases makes it difficult to coordinate research and generalize early indicators of student success across institutions, systems and regions of the country. Without a common set of early indicators, colleges lack insight to prioritize support services and evaluate the impact of their interventions.


Ascendium and the Bill & Melinda Gates Foundation are funding a networked community of researchers in four states, led by Dr. Paul Attewell at City University of New York (CUNY) Graduate Center. With a combined $955,000 in grants, researchers from New York, Texas, Virginia and Illinois will use data mining techniques to analyze educational data across multiple systems and connect it to state-level data about employment and income.

Using information available at entry to college plus the student's performance in the first college semester, the researchers will develop models for accurately predicting students' likelihood of retention, degree completion and long-term earning potential. A common set of early indicators of student success will help identify which behaviors might lead some students to fall behind, and the earnings consequences of taking different routes through college.

The research consortium will share knowledge gained from this two-year project with other researchers and practitioners, such as student success and institutional research staff. Providing colleges with a set of common early indicators will allow them to prioritize students at greatest risk of dropping out for academic support services, which can help increase retention and completion rates.

Equally valuable to the colleges will be the ability to assess the effectiveness of their interventions early on, without waiting years to see if students drop out or graduate. Examining whether interventions move the needle on early indicators of student success would provide rapid feedback and allow for adjustments to be made while students are still in school.


During the project’s first year, the consortium analyzed state-level datasets from high schools, colleges and the labor market. Researchers found that nine times out of ten, early indicators correctly predicted which students are at greatest risk of not enrolling in a second year of college and which are least likely to graduate.

Researchers also identified patterns that challenge prevailing assumptions in the postsecondary field. They found that:

  • Students who have “some college, but no degree” earn significantly higher wages than their “no-college” counterparts.
  • Undergraduates who attend college part time earn at least as much after graduation as those who attend full time.
  • Transfer students who start at a two-year college earn about 14% less than students who complete their entire bachelor’s degree at a four-year institution.

Through presentations and convenings, researchers are sharing these findings with the field and exploring opportunities for deeper research.


Contact Senior Program Officer Sue Cui at