These are challenging times in the consideration of the technological growth of the United States. Comparisons of the technological strength of the US to that of other countries have been receiving increased attention. Amid this interest, there has been concern about the fact that bachelor's, master's and doctoral degree data for engineering programs resulting from surveys of various organizations are not always in agreement.

This report compares and analyzes the bachelor's, master's and doctoral degrees awarded in the US reported over the past four decades as published by the American Society of Engineering Education (ASEE), the Engineering Workforce Commission (EWC) of the American Association of Engineering Societies (AAES) and the National Science Foundation (NSF).

Bachelor's Degree Data

Bachelor's degree data reported by ASEE, EWC and NSF are shown in the graph below.

NSF data are from NSF Report 04-311 ("Science and Engineering Degrees, 1966-2001", March 2004) and from the Science and Engineering Indicators 2006 (S&EI 2006) that are currently posted on the NSF Web site. (Bachelor's degree data from S&EI reports (2000, 2002, 2004 and 2006) and NSF Report 02-327 (July 2002) are coincident with those shown for Report 04-311.) Clearly, all three sets of bachelor's degree data (ASEE, EWC and NSF) are in agreement from about AY1975-76 through AY1990-91.

EWC data from AY1967-68 (their first survey) through the mid-1970s are in reasonable agreement with NSF data except for one year (AY1972-73). ASEE surveys began with AY1966-67, but focused mainly on graduate program data at that time. Their first bachelor's degree data appeared for AY1970-71. Presumably, it required several years for all of their survey respondents to provide the undergraduate information that was requested.

ASEE and EWC data are in general agreement over the period from the mid-1970s through about AY2001-02. (ASEE survey data for AY1993-94 through AY1998-99 were excluded from the present study due to a decline in the number of university participants in the surveys during this period.) For the last three years, ASEE surveys also requested that engineering colleges provide data for computer science programs at their universities which were in colleges other than the engineering college. National totals for bachelor's degrees reported by ASEE continue to include only those computer science degrees awarded within engineering colleges; EWC national totals include all engineering college degrees plus degrees awarded in computer science programs accredited by ABET (Accrediting Board for Engineering and Technology) "outside" of the engineering college. For AY2004-05, ASEE reported that approximately 3,300 bachelor's degrees in computer science were awarded by sixty-seven universities "outside" of the engineering college that was surveyed. The average difference in the total number of engineering bachelor's degrees reported by ASEE and EWC over the past three years is about 3,200. Thus, it appears that this difference is due mainly to whether computer science degrees awarded by non-engineering colleges are included in the total number of degrees.

The data from NSF surveys from AY1990-91 depart from ASEE and EWC data and the departure has increased annually. From AY1995-96 through AY1997-98, annual NSF data are, on the average, about 2,400 lower than EWC data; from AY1999-00 through AY2001-02, the average difference is about 6,000. For the same two three-year periods, computer science degrees derived from the EWC surveys were annually about 3,100 and 5,800, respectively. ASEE data for the period AY1999-00 through AY2001-02 were annually about 5,600 higher than NSF data and ASEE reports an annual average of 6,000 computer science degrees during the three-year period. Thus, it appears that the NSF data for bachelor's degrees in engineering do not include computer science degrees awarded either within or "outside" of engineering colleges.

In summary, ASEE engineering bachelor's degree totals include all degrees awarded by engineering colleges. EWC totals include all engineering college degrees plus ABET-accredited computer science degrees "outside" of engineering colleges in universities surveyed. Bachelor's degrees reported by NSF as "engineering" appear to exclude all computer science degrees. It is uncertain as to whether other degrees awarded by engineering colleges that do not specify "engineering" in the degree name are excluded as well by NSF.

The delay of several years in the reporting degree data by NSF is difficult to overlook. Accurate predictions of trends require up-to-date information. Engineering bachelor's degree data for AY2004-05 have already been reported by both ASEE and EWC. The most recent bachelor's degree total reported by NSF is for AY2001-02 (S&EI 2006).

Master's Degree Data

Master's degree data reported by ASEE, EWC and NSF are shown in the graph below. Clearly, the ASEE and EWC data are in agreement over the four decades shown. Thus, the "computer science effect" noted previously for bachelor's degrees reported by ASEE and EWC does not seem to apply to master's degree data. ASEE data for AY2004-05 indicate about 2,000 master's degrees in computer science (sixty-seven universities) were awarded by colleges other than the engineering college that was surveyed. The number of universities currently reporting "outside" engineering computer science master's degrees in EWC surveys is probably small.

NSF data for the period AY1999-00 through AY2001-02 indicate an annual average of about 5,300 fewer master's degrees than EWC during this period. This difference is greater than the average number of computer science master's degrees for computer science (about 3,900) derived from EWC survey data. Master's degrees reported by ASEE during the same three-year period were about 4,400 annually; they report an annual average of only about 3,500 computer science degrees. Thus, the "computer science effect" appears to explain only a fraction the difference between NSF master's degree data and data reported by ASEE and EWC.

The comments made previously about the excessive delays in the reporting of data on bachelor's degrees by NSF also apply to the master's degree data in the graph above. The most recent NSF data indicate a constant number of degrees awarded (about 26,000) since the mid-1990s. The ASEE and EWC data indicate that a substantial growth trend is underway with a current output of about 41,000 master's degrees annually. New annual records are shown in the graph above for the last three years. Clearly, the missing years in the NSF data (AY2002-03 through AY2004-05) obscure the ongoing situation.

It is of additional concern that master's degrees reported by NSF are not year-to-year consistent. The graph above indicates that master's degrees reported by NSF in S&EI 2002 are the same as those in NSF Report 04-311 (also Report 02-327), but were shifted forward to later years. Master's degree data for AY1976-77 through AY1980-81 reported in S&EI 2002 are the same as those reported in Report 04-311 two years earlier. Those in S&EI 2002 for AY1984-85 through AY1988-89 were shifted forward by four years; that in S&EI 2002 for AY1989-90 was shifted by three years; that in S&EI 2002 for AY1990-91 was shifted forward by two years; that for S&EI 2002 for AY1991-92 was shifted forward by one year. The basis for these "lateral" shifts is unknown. However, the credibility of NSF data is certainly of concern because of these shifts as well as by the lower numbers of degrees reported since the mid-1980s.

(The NSF degree data shown in the graph above were obtained from Appendix Table 02-22 in the NSF Science and Engineering Indicators 2002 report (NSB 02-1). A discerning Engineering Trends member pointed out that Appendix Table 02-23 in NSB 02-1 contained master's degree data that were consistent with those presented in NSF report 04-311 (but inconsistent with those in Appendix Table 02-22). Engineering Trends reported the NSF Appendix Table 02-22 data without checking to determine whether these data were the same as those in Appendix Table 02-23. In the future, Engineering Trends will make every effort to check for internal consistency in reports of engineering education data where data are presented in multiple tables and graphs.)

Doctoral Degree Data

The graph below compares doctoral degree data reported by ASEE and EWC to those reported by NSF (Reports 04-311 and 06-308 ("Science and Engineering Doctorate Awards: 2004", March 2006) and S&EI 2006). The data from Report 06-308 coincide with those of Report 04-311 through AY2000-01 and provide additional data for AY2001-02 through AY2003-04.

The ASEE and EWC data are in agreement over the four-decade period shown and indicate a substantial increase in doctoral degrees since the late 1990s. Doctoral degree data reported by both ASEE and EWC indicate a new annual record being set in AY2004-05. Full-time enrollment data reported by both ASEE and EWC for fall 2005 indicate that doctoral degree growth (with new records) will continue for four or five more years.

NSF doctoral degree data in Report 04-308 and S&EI 2006 show no indication of this doctoral degree growth trend. The most recent NSF doctoral data (Report 06-308) indicate a degree growth rate from AY2001-02 through AY2003-04 that is essentially the same as the rates shown by the ASEE and EWC data for the three-year period. However, for AY2001-02 through AY2003-04, the NSF data are about 750 doctoral degrees lower (annual average) than the average of the ASEE and EWC data. The number of doctoral degrees in computer science derived from EWC data for AY2001-02 through AY2003-04 is about 400. For ASEE data during this same three-year period, a total of about 520 computer science doctoral degrees was awarded (annual average). Thus, the difference in degrees between NSF and both ASEE and EWC data is greater than the number of computer science degrees.

It is of additional concern that doctoral degrees, in addition to master's degrees, reported by NSF are not year-to-year consistent. The graph below indicates that doctoral degrees reported by NSF in S&EI 2006 are the same as those in Report 04-311 (and Report 02-327), but have been shifted to later years. Doctoral degree data for AY1982-83 through AY2002-03 reported in S&EI 2006 are exactly the same as those reported in NSF Report 04-311 (and S&EI 2000 through 2004) years earlier.

Those in S&EI 2006 for AY2001-02 and AY2002-03 are the same as those in AY1999-00 and AY2000-01 in Report 04-311 (shifted forward by two years); that for AY2000-01 is essentially the same as AY1996-97 (shifted forward by four years); that for AY1999-00 is the same as AY1994-95 (shifted forward by five years); those for AY1982-83 through AY1998-99 are the same as AY1976-77 through AY1992-93 (shifted forward by six years). Thus, S&EI 2006 provides only prior data shifted forward by two to six years. Trend predictions based on the doctoral degree data in this report will be flawed.

(The NSF degree data shown in the graph above were obtained from Appendix Table 02-30 in the NSF Science and Engineering Indicators 2006 report (NSB 06-1). A discerning Engineering Trends member pointed out that Appendix Tables 02-31 and 02-32 in NSB 06-1 contained doctoral degree data that were consistent with those presented in NSF report 04-311 (but inconsistent with those in Appendix Table 02-30). Engineering Trends reported the NSF Appendix Table 02-30 data without checking to determine whether these data were the same as those in Appendix Tables 02-31 and 02-32. In the future, Engineering Trends will make every effort to check for internal consistency in reports of engineering education data where data are presented in multiple tables and graphs.)

NSF data in Report 06-308 are important in providing more recent data (through AY2003-04) and in indicating the onset of the period of growth in doctoral degrees in engineering. It is disconcerting, however, that the data in the report do not agree with the S&EI 2006 data for AY2001-02 and AY2002-03. Both sets of data became available on the NSF Web site at about the same time.

Summary

The study described in this report was aimed at providing an insight to the differences in the national totals of engineering bachelor's, master's and doctoral degrees resulting from the annual surveys of the American Society for Engineering Education (ASEE), the Engineering Workforce Commission (EWC) and the National Science Foundation (NSF). Significant understanding of the differences has emerged, but some issues remain for future examination.

The degree data from the three surveying organizations can be summarized as follows:

a. ASEE national totals are for the total number of degrees awarded by US engineering colleges.

b. EWC national totals are for degrees awarded by US engineering colleges plus ABET-accredited computer science degrees "outside" of the engineering college being surveyed. (The differences between ASEE and EWC data are observed only in bachelor's degrees, are minor and affect only the data reported in the last three years.)

c. NSF national totals for bachelor's degrees appear to be those awarded by US engineering colleges, with computer science degrees excluded. NSF data for master's and doctoral degrees appear to exclude somewhat more than just computer science degrees. (NSF reports a separate US national total for computer science degrees awarded by all universities in the US.) The reports for bachelor's and master's degrees by NSF lag those of ASEE and EWC by three years, making trend assessments based only on NSF degree data unreliable.

National totals for master's degrees reported by ASE and EWC are in agreement for the last four decades. The three-year lag in the reporting of NSF data has obscured the last three years of new annual records for master's degrees shown in the ASEE and EWC data.

National totals for doctoral degrees reported by ASEE and EWC are in agreement for the last four decades. These data show significant growth in the number of doctoral degrees awarded in the last few years and the record number of doctoral degrees awarded in AY2004-05. The latest NSF data reported for doctoral degrees (through AY2003-04) indicate an onset of growth at the same rate as that shown by the ASEE and EWC data.

No "computer science effect" is evident in comparisons of master's and doctoral degrees reported by ASEE and EWC. Master's and doctoral degree data reported by NSF are lower than those reported by ASEE and EWC. The differences are measurably larger than the master's and doctoral computer science degrees derived from the ASEE and EWC survey data.

It is troubling that various tables containing engineering master's and doctoral degrees (in NSF Special Reports and in Science and Engineering Indicators Reports available on the NSF Web site) are inconsistent. Data presented in separate reports can be found to have different values for the same year; occasionally, the differences are large. For example, Science and Engineering Reports contain some master's (2002) and doctoral (2006) degrees that were the same numerically as data in prior reports, but were shifted to later years. Some of these shifts were as large as six years.

Acknowledgments

Some of the data used in this study originated from the annual surveys of the American Society for Engineering Education and the Engineering Workforce Commission of the American Association for Engineering Societies. Persons seeking further information about their surveys and the availability of survey data should visit the ASEE (www.asee.org) and the EWC (www.aaes.org) Web sites.

Degree data reported by the National Science Foundation that appear in this report are from NSF Reports 02-327 (July 2002), 04-311 (March 2004) and 06-308 (March 2006) and from the semi-annual Science and Engineering Indicators reports to the National Science Board dated 2000, 2002, 2004 and 2006. All of the NSF Reports and the Science and Engineering Indicators reports are available on the National Science Foundation Web site (www.nsf.gov). 

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.