RETENTION

The question of students “at risk” is a multifaceted one. Nearly 50% of community college freshmen enrolled this year will not return next. Four-year colleges and universities don’t fare much better—approximately 27% are non-returns the second year. The results of this loss are staggering to the student, the college or university, and often to the society in general. These data are taken by most with a sense of surprise or loss, but little is done to avoid this calamity.

  1. The Validity of Prediction.
  1. To determine through empirical measurement the percentage of students “at risk,” any instrument used must be validated to prove its ability to predict return and non-return.In six recent studies using the CSFI, results predicted non-return the second year at an extremely high level. While this research was not designed to match students, each of these studies shows the reported efficacy of the CSFI in predicting non-return the second year.
  2. Grade point average is an important element of “high risk” and in predicting “non-return.” In another recent study by Santa Barbara City College the CSFI scales were compared from all samples of students (gender, ethnicity, etc.) in a multiple regression analysis to predict GPA. A correlation of r.=.36 was obtained, which means that the eight CSFI scales accounted for 13% of the variance of GPA>. This is very similar to the predicted value of high school GPA, which shows a r.=.40 correlation with college GPA. When the CSFI scales are combined with high school grades, a correlation of r.=.50 is obtained, accounting for 25% of the variance in college GPA. A modest but statistically significant increase.
  3. Rochester Community College designed a student development course emphasizing the eight factors of the College Success Factors Index. Students in the program were matched with others who were not given the intervention class, and after several semesters, there was 11% more retention with those students who took the course designed around the CSFI.
  1. The Identification of Specific Students.

Unlike the Registrar’s aggregated predictive data, CSFI data are able to identify “at risk” students by name, so individual interventions are possible. In each of the studies above “at risk” students were identified. The Cypress College study, for example, identified 93students who were at high risk, and at the University of Colorado 24 students were identified at high risk (CSFI scores of greater than 220).

  1. Individual student Factor Identification.

We are able to identify which of the “at risk” factors exist for those students identified. The results have indicated that a watchline score for each of the factors of the CSFI exists. Further, we found that with approximately two standard deviations below the watchline, on each of the factors of the CSFI, students were at extremely high risk.

Factorial scores may also be aggregated for administrators and faculty to better understand risk factors in general, and to design curriculum interventions.

  1. Students of Probation. A large number of students are on probation. Many factors may lead to probation; however, several important factors have been correlated with GPA in the College Success Factors Index. With the College Success Factors Index program, we are able to identify specific characteristics of lack of success, and then suggest areas of interventions necessary for the addressed probationary students. The program has several steps that include the following:
  2. Assess the probationary students and develop a profile of the CSFI characteristics of probation students.
  3. Assess a random sample of successful students utilizing the CSFI and develop a similar profile. Matching of samples are desirable.
  4. Compare the two profiles and determine areas of difference. For example, there could be differences in the student’s control, precision, or time scores as indicated by the CSFI in a recent study at Cypress College.
  5. Develop interventions around these significant vectors and instruct the probationary student through group sessions, college success classes, or seminar delivery systems.
  6. Post tests of the CSFI probationary students who have been given the interventions to see if improvements have indeed been made.
  7. Communicate these important success factors and program results to faculty, students, and feeder high schools.

Percent of Probation Students Below the Watchline on the CSFI Midwestern University 2004

In a study conducted by the research and Freshman experience coordinatorsat a Midwestern university, it was found that 207 students who were on PROBATION scored, on the average for eight variables, two to three standard deviations below the mean on the CSFI.

The specific percentages were the following:

These differences show the continuing need to pinpoint and improve the stated fundamental criteria for probationary students. Most of these percentages were statistically significant.

Retention Studies Early Alert Predictions

While retention from one year to another for college students is based on many facets, a statistical significant relationship was found as a part of our Early Alert research efforts. We found that the scores of beginning freshmen on the CSFI could partially predict non-return the second year. Our research included six community colleges, research universities, state universities and private colleges.

Early Alert Comparisons (2000 - 2007)

In one mid-western university, we studied 215 freshmen and compared their CSFI, and found a resulting Chi square of 5.46 between expected and observed second year return data. This chi square proved to be significant at the .02 level.

In the following study of students on probation from a recent study at Cypress College, Control, Precision and Time scores showed major differences, as did total scores. These differences occurred between the 3.5+ G.P.A. students and the 1.99 and below G.P.A. students who were on probation.