Leaving General College: Interests and Issues Related to Student Departure

Cathrine Wambach, Amy Mayer, Jennifer Hatfield and Jennifer Franko

General College Office for Research and Evaluation

University of Minnesota

download MSWord file

Abstract

Many of the freshmen who begin postsecondary education at the University of Minnesota General College (GC) leave before they complete degrees. The purpose of this study was to describe the patterns of achievement, academic interests, reasons for leaving, and plans after leaving for a cohort of GC leavers. Advising files, the U of M student database, and the General College Student Inventory (GCSI) were used as sources of information. Half of the leavers were found to have academic performance issues. Leavers expressed a wide variety of academic interests on the GCSI, the most common being business and social sciences. There was some evidence to suggest that students with interests in physical science and computer science were more likely to leave, which warrants future investigation. Information in advising files suggested that about 25% of the leavers intended to transfer to another college. Issues that may have been related to leaving that were mentioned most frequently in the files were lack of motivation, physical and mental health concerns, financial problems, and conflicts between family or cultural expectations and the demands of college.

Introduction

The problem of student attrition from colleges and universities has been well studied. In 1975, Tinto published his enduring Student Integration Model of attrition. This model included measures of a student’s initial goal for their education and their institutional commitments. More recently, Cabrera, Nora, and Castaneda (1993) combined Tinto’s conception of student persistence with work done by others (e.g. Bean, 1985) resulting in a model that included external factors that shape student intentions and preferences. Cabrera et al. found that social integration, financial attitudes, and academic integration have the strongest effects on retention in their model.

Government agencies have also dedicated time and resources to the problem of attrition.  A recent United States Department of Education, Office of Educational Research and Improvement, National Center for Education Statistics (NCES) report studied the topic in depth using 1996-1998 NCES data (Bradburn & Carroll, 2002). These data included self-reports on reasons for leaving. Nine reasons were listed in the report, which can be grouped into four categories, listed here in order of frequency cited: finance/work, family/home, academic, and personal.

The problem of student attrition is perhaps most closely felt at the college or university level. Individual colleges and universities have conducted their own studies and written reports in an attempt to understand the problem. Recent examples can be found from across the country, including the University of California-Santa Cruz, West Virginia State College, and the University of Minnesota-Twin Cities. These studies have tended to find results similar to the national studies. A study conducted at the University of California-Santa Cruz (Nelson & Davies, n.d.) found the following four types of reasons for leaving to be the most prevalent: financial, personal, academic, and social. A report from the West Virginia State College (Leaver Attrition Survey: Spring 1996, n.d.) found the following four reasons for leaving to be the most important: conflict between work and/or home and school, problems registering for classes, finances, and health.

Retention has come to the forefront of institutional concerns at the University of Minnesota because of persistent low graduation rates. With a six-year graduation rate of 50%, the University of Minnesota has the lowest rate in the Big Ten. A recent report of the Graduation and Retention Subcommittee of the Council of Undergraduate Deans (Improving our Graduation Rates, 2001) examined a broad range of information on leavers and reached several conclusions including:

1.      The U of M graduation rate is below its predicted rate based on pre-college academic achievement of its freshman admits.

2.      The U of M has lower retention of students in the junior and senior years compared to other similar institutions.

3.      Leaving might be influenced by lack of institutional expectations and incentives to graduate.

The report identified three types of leavers: underperformers, disenchanted, and drop-ins.  The underperformers left with low GPAs and often were suspended. The disenchanted had problems with course availability, inadequate advising, and an unfriendly environment. About 40% of the dropouts were disenchanted sophomores who tended to transfer to another college. The drop-ins were not dissatisfied with the university, but did not make it the focus of their lives. The drop-ins were more likely to be involved in jobs and pay their own college expenses. 

An analysis of all University of Minnesota students enrolled in Spring 2000, (Matross & Huesman, 2001) found that 13.6% did not return during the 2000-2001 school year. Of those, 36% had a registration hold, for either academic or financial reasons, that prevented them from registering. Seniors and freshman had the highest departure rates (21%) and juniors the lowest (8%). American Indian (30%), African American (20%), and Chicano Latino (21%) students had higher departure rates than Asian and White students (13%).

Matross and Huesman (2001) found that the University of Minnesota General College (GC) retention rate for the cohort they identified was 81%. Since GC admits students who are not usually qualified for admission to a research university, and admits significant numbers of students of color, the expected retention rate for the college is lower than that for other freshman admitting colleges. Other unique characteristics of the GC population might also raise the possibility that GC leavers are different from non-GC leavers. Matross and Huesman found that about half of GC leavers from the cohort they studied left for involuntary reasons including academic and financial holds.  Research by Schmitz and Andreozzi (1990) suggests that this might be an underestimation of the role of poor academic performance in students’ decisions to leave GC. They found that leavers attributed their academic problems to work commitments, personal problems, health, lack of readiness for college, the size of the university, and dissatisfaction with their GC experience.

In a recent survey of students who matriculated in GC and then transferred to a degree granting college, Wambach, Hatfield, and Merabella (2001) asked students to identify barriers to eventual graduation. Though 85% of students surveyed expressed complete confidence in their likely graduation, many students surveyed also identified access to courses and programs as potential barriers. This study suggested that some students might leave the University of Minnesota because desired majors and professional programs are not available at this institution. 

The problems encountered by GC students that might be related to their leaving the University have been documented for small samples of leavers (Schmitz & Andreozzi, 1990; Beach, Lundell & Jung, 2002). However, no recent study of a large cohort of GC leavers has been conducted. The study described in this report explores the following questions about GC leavers:

1.      How does academic performance in college impact leaving?

2.      Are leavers’ academic interests related to patterns of leaving?

3.      What do leavers say about their reasons for leaving?

4.      Do leavers plan to continue their education at another institution?

Method

The survey is the most common way to collect information about student leavers. However, this method is problematic because it is difficult to reach students after they leave the university due to the loss of easily available up-to-date contact information. The initial strategy for this study was to design a survey for enrolled students that would reach students who did not intend to enroll again. A copy of the survey is in Appendix A. The survey was designed to first identify students who did not intend to return. These students were asked a number of questions about their occupational goals and reasons for leaving. The survey was distributed on-line to 1475 General College students at the end of spring semester 2002. It yielded an overall response rate of 241 (16.3%). The response rate for the group of students who had not yet registered for the next semester was only 5.4%. Since the response rate for the population of interest was so low, we decided to use a different research method to address the research questions. 

An archival study based on a content analysis of leaver files was designed. All students in the General College are assigned to an advisor who maintains a file for her or him. From our prior experience with student files, we expected them to contain: an entrance questionnaire including questions about academic plans, academic alerts, mid-semester grade reports, registration planning forms, transfer plans, and advising notes.

The population studied was the segment of all GC students enrolled in the fall of 2000 who did not enroll during the next academic year (fall 2001/spring 2002). Of the 1,909 students enrolled in fall 2000, 448 (23%) met the criteria for inclusion in the study. The U of M student records data base was used to gather information about these students’ gender, ethnicity, ACT composite test score at entrance, last term of enrollment, and credits completed and cumulative GPA at the end of Fall 2000.

We attempted to obtain advisor files for 448 students of interest and were able to retrieve 403 (90.0%). All 403 files were coded by one of four raters: two raters coded approximately one third of the files each (n of 138 and 133), and two raters coded the remainder of the files (n of 70 and 62). Each file was read twice by the same rater. During the first reading, notes were made on leavers’ occupational goals, academic problems, and/or observations made by advisors. The raters also recorded reasons for leaving and information about transfer to another college. The notes were reviewed and codes developed for academic/vocational interests, reasons for leaving, presence of academic alerts, and plans after leaving. The files were then re-read and coded using these categories.

Coding

The academic/vocational interests identified by leavers were organized into 15 categories listed in Table 1. When coding the files, these categories were not used in a mutually exclusive manner; multiple academic/vocational interests could be assigned per student. The categories were selected to represent the broad range of leaver interests while allowing identification of leavers with interests in majors that require admission to highly competitive programs. The highly competitive majors include design, business, health sciences, and computer sciences. Interests expressed by students both at the beginning of their academic career and any time after their first term were recorded, resulting in two sets of academic/vocational interests: incoming and post-matriculation.

Reasons for leaving were organized into 13 categories listed in Table 2. When coding the files, these categories were not used in a mutually exclusive manner; multiple reasons for leaving could be assigned per student. The categories were suggested by prior research but also include categories not often seen in retention research such as disability issues and incongruities/resistances. The disability issues category includes perceived lack of accommodation for physical or learning disabilities. The incongruities/resistances category includes cases where there appeared to be advisor frustration regarding a student’s unwillingness to follow advice, inconsistencies between a student’s expressed plans and his or her enrollment records, or lack of compliance with institutional processes such as required program planning meetings.

Academic feedback consists of academic alerts and mid-semester reports that were sent from faculty to advisors either out of concern for student performance or as part of regular reporting of grades. In aggregate, they were categorized in a mutually exclusive manner as: none on file, on file with problems reported, or on file with no problems reported. On file with problems reported means that at least one instructor indicated that a leaver was excessively absent from class, missing assignments, or not making satisfactory progress.

Planned actions after leaving GC were organized into six mutually exclusive categories listed in Table 3. For 66% of leavers this information was not available in the file. When it did exist, it was contained in advising notes and correspondence from students. 

Reliability Analysis

In order to determine whether the files were coded consistently by the four raters, a reliability analysis was conducted using files from a random sample of 12 leavers. Files were available for all selected students except one. Each of the four raters coded all 11 files. As an index of inter-rater reliability, the agreement between raters regarding the codes assigned to these files was assessed. For the coding schemes that were not mutually exclusive (i.e. academic/vocational interests and reasons for leaving) rater agreement was assessed as follows. First, we identified all instances where an incoming academic/vocational interest category was selected by any one of the four raters. Next, the raters were divided into all possible pairs of two, resulting in six unique pairs of raters. Then, for each of the six pairs of raters, the number of times that the pair agreed in their indication of an incoming academic/vocational interest was counted. This number was then divided by the number of distinct categories selected across both raters in the pair. For example, if rater one selected the interest categories “business” and “computer science” and rater two selected “business” and “social science” then their agreement for the category would be 33% since they agreed on only one interest (business) of the three distinct interests indicated (i.e. business, computer science, and social science). In this manner a percent agreement statistic was calculated for each of the six pairs of raters, resulting in six percent agreement statistics. The average of these six statistics was taken as an index of inter-rater reliability for incoming academic/vocational interest coding. This procedure was repeated for post-matriculation academic/vocational interests and reasons for leaving in order to obtain indices of inter-rater reliability for these coding schemes as well. In order to assess inter-rater reliability for the mutually exclusive coding schemes (academic feedback and plans after leaving), the proportion of all possible rater pairs where coding agreement occurred was calculated for each coding scheme. The results are summarized in Table 4. The analysis suggests that there was more agreement in the coding of academic/vocational interests than in coding reasons for leaving, which often involved making an inference from information in the file. 

Results

Description of leavers

Approximately 70% of the students admitted to GC over the past 5 years have been white, whereas 65% of the leavers were white. Likewise, 50% of students admitted to GC have been male, whereas 52% of leavers were male. This suggests that students of color and males are somewhat over-represented among leavers. See Table 5 for descriptive statistics.

How does college academic performance impact leaving?

Among the 403 leavers, the average cumulative GPA in their last semester was 1.89, a C- average. Of these, 206 (51%) had GPAs lower than a 2.0 (see Table 6) making them ineligible for transfer to most degree granting U of M colleges. In addition, 294 leavers (73%) had academic feedback suggesting academic problems (see Table 8). Only 55 leavers (13.6%) had mid-semester reports that did not identify academic problems. Another 54 leavers (13.4%) had no mid-semester reviews or alerts in their files.

Are leavers’ academic interests related to patterns of leaving?       

When analyzing leavers’ academic interests, two groups of students were excluded. Both groups had registered for classes fall of 2000, but never attended. One group (n=21) had completed credits before fall semester 2000, the other had not (n=16).  Information about length of enrollment by incoming and post-matriculation academic/vocational interests can be found in Tables 7 and 8 below. The data suggest that business was the most frequently identified academic/vocational interest both at entry (n=94) and after matriculation (n=41). Social science was the second most frequently identified interest both at entry (n = 88) and after matriculation (n=41). Among leavers with an interest in business, 50% left after one or two terms. Among leavers with an interest in social science, 42% left after one or two terms. The tendency for about half of any group to leave by the end of two terms was apparent in most of the other interest groups, the exceptions being design (35%), biology (36%), computer science (37%), and undecided (37%).  

Much less information was available on leavers’ interests after their initial enrollment.  Table 8 summarizes this information and its relation to length of enrollment. The data suggest that students with interests in design, biology, or humanities are more likely to leave after 5 or more terms. Forty percent of leavers interested in design, 50% of those who were interested in biology, and 44.4% of those who were interested in humanities left the university some time after five terms of enrollment.

Because the above analyses did not include a comparative group of persisting students, no conclusions could be made regarding the relationship between academic/vocational interests and persistence. Therefore, in order to fully examine the relationship between interests and persistence, analyses were conducted on another set of data, which included incoming academic/vocational interests for all new freshmen entering the General College fall 1999-2001. These data were drawn from students’ (N=2714) responses to the General College Student Inventory (GCSI), the same instrument from which the leavers’ entry interests were extracted. Interests were coded by one rater in the same manner as previously described.

Figure 1 below shows 1-year retention rates for students who did and did not indicate each interest upon entry. No interest appears to be markedly related to one-year retention. The largest effects are seen for students interested in the humanities and architecture. Students who indicated an interest in humanities showed one-year retention rates about five percentage points below those who did not indicate this interest. Students who indicated an interest in architecture showed one-year retention rates about five percentage points above those who did not indicate this interest.

Figure 1. Incoming Academic/Vocational Interests and Rates of Retention to Second Fall Term: 1999-2001 Cohorts

Although the data in Figure 1 address the relationship between incoming interests and short-term persistence, it does not provide information regarding whether students with some incoming interests are more likely to drop/stop out later in their academic career. The charts in Figure 2 below illustrate the relationship between each incoming interest and persistence to the 2^nd, 3^rd, and 4^th fall terms for the fall 1999 cohort and persistence to the 2^nd and 3^rd term for the fall 2000 cohort.

Figure 2. Incoming Academic/Vocational Interests: Rates of Retention for Students from Fall 1999 and 2000 Cohorts Indicating Each Interest.

Note. Ind. = interest was indicated on GCSI; NI = interest was not indicated on GCSI. Fall 1999 cohorts were tracked to their 4^th fall term and fall 2000 cohort was tracked to their 3^rd fall term (e.g. fall 2002).

These charts illustrate some relationship between entry interest and longer-term retention within a cohort, but they also demonstrate the unpredictability of such relationships. For example, in the 1999 cohort at all retention points students expressing an interest in architecture, fine arts, communications, business, social science, health science, and/or biological science are retained at about the same rate (within four percentage points) as those not expressing the interest. However, this same pattern is not evident for students from the fall 2000 cohort who indicated interests in fine arts or design. The pattern also deviates from that of the fall 1999 cohort for students in the fall 2000 cohort with interests in business, social science, health science, and/or biological sciences. Over both cohorts, only the students with an interest in communications appear to be retained at the same rate as those not expressing the interest. Also replicated across the two cohorts was the trend for students interested in hard science or computer science to show increasingly lower retention rates over time relative to students without the interest at entry.

What reasons do GC leavers give for leaving college?

The most commonly described factor related to leaving was lack of motivation, which was indicated in 18.7% of the files. Health (16.1%), financial problems (15.7%) and family or cultural conflicts (12.6%) were also indicated in the files.  See Table 9. 

Factor Analysis

Rotated Principal Components analysis was used to determine whether reasons for leaving clustered into broader categories, or typologies.  Five factors were identified. The reasons for leaving with the highest loading on each factor, along with the factor loading values, are displayed in Table 10. A description of the five factors and the percentage of variance accounted for by each are listed in Table 11. The first factor combined family and cultural conflicts, financial problems, and multiple problems in a way that suggested that the costs of staying in school were too high for the student. The second factor, which combined performance problems with low motivation and the need for time-off from school, suggested a lack of purpose for and commitment to college. The third factor, which combined health problems with the incongruities and resistances category, suggested stress related physical and mental health problems. The other two factors were related to students’ perceptions that the U of M was not meeting their needs either for educational or other reasons.