Introduction

Interest in improving the retention of enrolled engineering undergraduate students has been ongoing for decades. The various efforts at engineering colleges in US universities benefit both the students and the universities.

Retention is a complex issue and efforts to improve retention are typically tailored to meet the needs of a specific university. Analysis of the loss of engineering students must consider a variety of issues including interdepartmental transfers, transfers to other colleges at a university, transfers to another university, abandonment of interest in university education and transfers into engineering from within the university or from other universities. Even mismatches between the demands of engineering programs and the credentials of enrolled students are important. Thus, the retention efforts of any engineering program must be program specific and so too must be measures of the effectiveness of these efforts.

Retention inquiries to Engineering Trends occur often. Due to complexities such as those indicated above and the lack of availability of measures of success of individual engineering colleges, we have been at a loss to respond in a substantive (quantitative) manner, much to the dissatisfaction of the inquirers. Many of the inquiries stem from the commonly accepted premise that the success rate of a student in an engineering education program is low. Engineering Trends continues to hear the half-century old mis-quotation that starts with "look left and look right, only one of you will make it". This statement remains as a detrimental issue for all engineering educators and efforts to erase it would be worthwhile.

From this complex setting, Engineering Trends set out to derive from statistical data a measure of retention that applied to engineering education in the US, as a whole. The analysis to be presented here in no way applies to any individual engineering college. Rather, the information applies to the composite behavior of all engineering programs in the US.

Analysis

A very reasonable starting point is that undergraduate full-time engineering enrollments should be linked, with a lag time, to the awarding of bachelor's degrees. With current undergraduate part-time enrollments of about 8% and the time to attain a degree in this manner being large, the contribution of part-time enrollments to the analysis was not included.

Data for full-time undergraduate enrollments in Fall 2004 were distributed as follows: 27% first year, 21% second year, 21% third year and 31% fourth year. The decrease of 6% from the first to the second year indicates the only overall loss, as might be expected. The large fourth-year fraction is due to the inclusion of students extending their studies over more than four years and five-year programs such as cooperative education programs.

Enrollments for first-year undergraduate engineering students are not reported by a few universities to the organizations that conduct annual surveys. Thus, the "effective" first year fraction should be a bit higher than that indicated.

Data

The data shown in the graph below indicate the expected correlation between full-time enrollments and degrees. The enrollment data are for the Fall of the academic year indicated. Degree maxima and minima follow enrollment maxima and minima by about two or, occasionally, three years.

The enrollment and degree data were analyzed via comparisons of the degrees at a maximum or minimum with the total full-time enrollments at the preceding maximum or minimum. It was assumed that, for the most part, the students associated with a maximum or minimum were those who were awarded degrees in the subsequent degree maximum or minimum. These degree/enrollment ratios are presented in the graph below for the last half century. The academic year on the graph refers to the degree data.

Except for the AY1949-50 ratio, all data for maxima and minima adhere to a common line. The AY1949-50 data reflect the return to campus of veterans from World War II. Many of these students had short enrollments since they were engaged in university study prior to their service in the military.

The graph above shows that over a period of a half century, the degree/enrollment ratio increased from about 0.17 to 0.20. This increase can be interpreted as an increase in degrees for the total number of enrolled students from 17% to 20% over the period that was analyzed.

Even if these data show that retention has improved, the question remains as to the significance of the increase in magnitude. An increase of 3% after many decades of effort may at first appear to be insignificant. If the measure of retention is the ratio of degrees to full-time enrollment in prior years as defined in this study, what is the best ratio that engineering education can achieve? The answer is unclear. However, the ratio for "perfect" retention must be near 0.25. Thus, the increase from 0.17 to 0.20 is a significant advance and further efforts by engineering colleges should yield continued successes.

Conclusions

Enrollment-to-degree ratios for undergraduate engineering education in the US have been slowly increasing for over a period of five decades. Efforts to improve retention, therefore, can be judged to be successful. It appears that there is room for further improvement and retention programs should continue to be stressed.

Acknowledgments

The undergraduate enrollment and degree data used in this study originated from the annual surveys of the Engineering Workforce Commission of the American Association of Engineering Societies. Engineering Trends acknowledges the efforts of this organization in providing credible data and expresses its gratitude for their services to the engineering profession. Persons seeking further information about their surveys and the availability of survey data should visit the AAES Web site (www.aaes.org).

Footnote

Engineering Trends data are compiled mainly from information submitted by universities to the annual surveys of EWC and ASEE. On the very rare occasions where errors in data appear, Engineering Trends corrects the error, if possible, or deletes the data if the error is large enough to alter significantly the trend of the university or the US total.