Preparation for transfer to four-year colleges and universities is one of the primary missions of the community colleges. It would thus seem that a standard and comparable method for calculating transfer rates would exist and that all community college would have this rate readily available. This is not the case, however, for a number of reasons having to do both with technical data problems and the wide variety in the missions of community colleges and the characteristics of students they enroll.

Data problems

Counting the number of students who transfer to a four-year college or university is not something which a community college can do from its own student records. That information depends on a report back in some form from the receiving institution. The only long-term and reasonably comprehensive record of transfer numbers has been maintained by the California Post-Secondary Education Commission (CPEC) which reports the number of new transfers each fall received by all UC and CSU institutions by community college last attended. Such figures are available for Fall 1978 through Fall 1999.

Though probably fairly consistent, these fall transfer, California public-institution-only figures must be considered minimal counts. Substantial numbers of students transfer from community colleges to the public four-year colleges and universities in the spring semesters. Since the 1986-87 academic year, CPEC has reported full-year transfers as well as the fall only, but this shortens the time series considerably. Inspection of the full-year, public institution figures in comparison to the fall numbers further suggests that the full-year series may be less reliable.

CPEC also makes some attempt to compile fall semester figures for transfers to private colleges and universities in California from the community colleges. This depends entirely on voluntary cooperation on the part of the private institutions, and it is clear from examination of this series that the degree of such cooperation has varied substantially from year to year. No attempt is made by CPEC to collect full-year transfer figures from the private schools.

Completely missing from the CPEC totals are transfers from California community colleges to either public or private institutions out of state. All indications are that this is a significant, and probably growing, number particularly as the California public institutions have become relatively more expensive in recent years.

The fall semester, public-institution-only figures, provided by CPEC, when compared to enrollment changes in the community colleges, are thus probably the best indicator of any change in the rate of transfer. This comparison has been done in the accompanying figure in which both the number of fall transfers from the Los Angeles Community Colleges since 1978 reported by CPEC have been plotted along with the number of first-time students two years prior to the transfer semester. The chart shows that though there has been considerable variation in the relationship of the two lines from year to year, ther is no evidence of any long-term decline in the rate of transfer. Change in the absolute number of transferring students appears to be relate solely to enrollment levels.

The Denominator Problem

The second area of difficulty in constructing a transfer rate is frequently referred to as the denominator problem and stems from the disparate missions of community colleges an the variety of purposes for which students attend these institutions. Simply put, if we could know the true number of transferring students, what denominator would we use to convert this figure into a percentage or rate?

Should the denominator be total enrollment? On its face this is clearly not appropriate though it is a frequent assumption by those wishing to comment on supposedly low community college transfer rates. Only some fraction of all students are available for transfer in any given semester since most are still in the pipeline of preparation. If all students entered the institution only in the fall term, if all were full time, if all were in parallel curricula which prepared them for transfer in two years, and if transfer occurred only at the end of that period, we might discount total enrollment to 50% and use that as the denominator.

This is, of course, not the reality of community college students with some entering and others moving on in every term. Most community college students are part-time and take much longe than the presumptive two-year period. Many students have substantial insufficiencies in their preparatio for college-level work at entrance, necessitating quite different periods of time to prepare for transfer Finally, some students are eligible to transfer to a four-year institution after only a short enrollment in a community college and do so at quite idiosyncratic intervals.

A cohort approach in which students are tracked from entrance into the communit college to transfer to a four-year institution makes the most logical sense. A transfer rate should be calculated as the percentage of an entering group of students who ever transfer to an advanced institution. This introduces several additional data collection problems, however, which are even more difficult than those noted above in the description of the CPEC process.

For the cohort approach to work, individual student enrollments in community colleges must be matched to those in the receiving institutions. In California, such an approach has been instituted in order to comply with Federal Student Right-to-Know legislation requiring that students be informed of the probabilities of success at a given institution Unfortunately this system began only with those students entering in Fall 1994 and who transferred by Fall 1997, and with individual colleg figures available only beginning with the Fall 1996 entering cohort. Further, the published reports count completers of community college degrees or certificates and exclude them from the transfer figures. Thus, no long-term or comparative analysis of transfer rates is possible with that data.

The cohort approach requires a significant time lag for the compilation of a transfer rate. The Student Right-to-Know process uses a lag of 150% of "normal time," or three years. This is possible because it will report a rate for full-time students only. But, this window may be too narrow and result in not counting a significant number of transfers by students who attend part-time. A five-year or longer window would be required if part-time students were included. That would be far too long for the Student Right-to-Know purpose of holding institutions accountable for their performance. Yet the restriction to full-time students means that any indicators derived from this system will reflect at most the enrollment of only a third of the students in typical community college.

A final part of the denominator problem is that not all community college students enroll with any intent to transfer. Some enroll to pursue specific vocational programs, which can be completed at the community college level and do not lead to transfer. Some enroll with equally specific, but short-term vocational, goals such as the acquisition of computer skills or perhaps knowledge from som single, job-related class. Some number enroll for avocational or continued self-development reasons and take classes in art, music, literature, the social sciences, etc., but have no reason to pursue a transfer curriculum. Some, particularly a portion of those who continue straight from high school, may have no clearly formulated goal and are using the community college as a resource and perhaps a refuge for clarifying educational and career plans.

A community college transfer rate ought to measure the performance of the institution in preparing those students who are, or become, transfer bound to actually transfer. Institutional performance in vocational preparation, skill acquisition, goal clarification and general community service and enlightenment should be measured by different indicators. All of these are worthy missions for the community colleges. Individual institutions may have quite different mixes of these purposes depending on their local communities and settings.

However, true student intent or goal is a very difficult thing to discern in a systematic, statistical basis given the diversity of our students and the multiple purposes for which they may enroll. Given the current state of community college data systems and programming structure, efforts to refine the transfer rate so that it more truly reflects the success of an institution in that portion of its work actually directed toward transfer, will probably produce figures which are unreliable for assessing change from year to year in an individual school or appraising the performance of one institution as compared to another.

A Los Angeles Community Colleges Transfer Rate

I proceed here in two steps. The previous chart in which transfers to in-state public institution were plotted along side of enrollments of first-time students is probably the best way at this point of assessing whether any change in transfer performance is taking place. In that chart I have used a three-year lag between community college entrance and transfer. This is not a cohort tracking process, however. The approximately 2,500 Los Angeles Community College District students reported as transferring in Fall 1999, for example, did not all begin in the Fall of 1996. Some began earlier, some later. I have followed the logic of the Student Right-to-Know cohort tracking process which uses a three-year window between entrance and transfer. Experimentation, however, revealed that both two-year lag and four-year lags produced essentially the same fit between the transfer and enrollment curves and thus that period was chosen.

Short-term separations of the two curves should not be over emphasized given the crudeness of the data and the approach. The rapid enrollment growth to 1981, and then even more rapid decline through 1985, is probably much more responsible for the imbalances of the two curves during that period than any fundamental change in institutional performance. Enrollment growth after 1985 was disproportionately of older students less likely to be transfer bound, producing the appearance of a lesser rate of transfer success. Likewise, the apparent gain in transfer since 1994 may be primarily an artifact of some increase in younger, more transfer oriented students among the total entrants, rather than an indicator of institutional improvement. I would repeat the previous general observation--th most systematic data available does not indicate any significant change in the performance of the transfer function by the Los Angeles Community Colleges over the two decades which can be observed.

This first step, though the more systematic and reliable, only produces information which allows us to say that we appear to be doing no worse nor no better than we have done in the past in preparing students for transfer. It does not produce information which tells us whether we are doing "well enough." If we were to construct a transfer rate from this approach, it would be just over 13%. That would be unconscionably low if it were the "true" rate.

This "true" rate can only be approximated and is probably completely unreliable as an indicator of change from year to year or of differences in performance among institutions. In this second step, I proceed as follows. The total number of students who transfer must first be estimated. As indicated previously, the CPEC figures are reasonably complete only for transfers to California public institutions. In the best year of reporting of transfers to private institutions in California, 1987, an additional 533 students, 25.4% of the number transferring to public institutions, were counted as originating in one of the Los Angeles Community Colleges. That is a substantial number and can not be ignored in the appraisal of the performance of the transfer function.

In addition, as previously noted, the CPEC data collection process makes no attempt to count transfers to out-of-state institutions. I have made the assumption that one-half th number of students who transfer to California private institutions go out of state to either public or private schools. This is an arbitrary but not unreasonable assumption given the popularly of public institutions in Washington, Oregon and Arizona for California students, as well as other public and private schools across the country. Using this assumption, the number of transfers should be increased by another 12.7%.

Total transfers for the Los Angeles Community Colleges can thus be estimated by multiplying the number of California public institution transfers by 138%. For 1999 this means that the "true" transfer number should be somewhat over 3,400.

To produce a "best estimate" of the transfer rate it is also necessary to refine the base or cohort available to transfer. Only 27% of all entering students in 1996 (including transfers from other colleges) indicated that they planned to transfer to a four-year institution. The "true" base of entering students in Fall 1996 desiring to transfer may thus be as low as 8,200 students. The resulting rate of transfer in 1999 would be 42%.

Further adjustments to the base and consequent increases in the transfer rate would be appropriate. The base should probably be discounted by some proportion representing the number of students who arrive at the community college doors with transfer aspirations but with reading, math and other skills substantially below college level and who are not able to progress through the remedial program Institutional performance for these students should be measured by indicators focusing on the basic skills curriculum alone. Likewise some discount to the base should be applied to represent those students whose motivation, family or work situation prevent them from persisting long enough, perhaps beyond one semester, so that the institution can have any impact and be held accountable for the student's success.

At this point such discounts can not be estimated. Their noting, however, means that an estimated transfer rate of 42% undoubtedly is too low. The true rate, if it could be determined, would seem to put the Los Angeles Community Colleges well within any acceptable range of performance of the transfer function.