Extrapolation of Taulbee Survey Data

Each year, the Computing Reserach Association sponsors the Taulbee Survey of Ph.D. schools. While this survey covers many topics, one focus involves the plans of new Ph.D.s in computer science and computer engineering. While the survey includes various categories for positions, this page focuses on possible conclusions for schools whose highest degree is at the Masters or Bachelors level. While the Taulbee Survey publishes results on the basis of surveys completed, what follows also gives one possible extrapolation for all schools and new Ph.D.s.


Taulbee Survey -- Raw Data

1996-1997 1997-1998 1998-1999 1999-2000 2000-2001
Ph.D. Departments Surveyed 162 182 203 214 215
Departments Responding 135 (83%) 144 (77%) 156 (77%) 173 (81%) 173 (80%)
   
Academic Positions
   Ph.D.-Granting Schools 237 274 293 273 294
       Tenure-Track 111 140 172 158 157
       Research 49 59 47 52 54
       Post Docs 55 55 45 40 56
       Instructors 22 20 29 23 27
   Other (Masters' and Bachelors') Schools 42 31 27 29 32
   
Industry/Self Employed 437 409 402 365 377
Government 27 30 19 21 13
Abroad 43 41 23 26 31
Unknown/Unemployed 108 148 88 160 186
   
Totals
    Known Employment 786 785 764 714 747
    Overall 894 933 852 881 933


Taulbee Survey -- Extrapolations to 100%

1996-1997 1997-1998 1998-1999 1999-2000 2000-2001
Academic Positions
   Ph.D.-Granting Schools 300 380 425 405 450
       Tenure-Track 140 195 240 235 240
       Research 60 80 70 75 85
       Post Docs 70 75 70 60 85
       Instructors 30 30 45 35 40
   Other (Masters' and Bachelors') Schools 60 45 35 45 50
   
Industry/Self Employed 530 570 565 540 575
Government 35 40 20 30 20
Abroad 55 55 35 40 50
Unknown 20 20 20 20 20
   
Totals
    Known 980 1080 1080 1060 1145
    Overall 1000 1100 1100 1080 1165

Notes

  1. All extrapolations are rounded to the nearest five and should be regarded as rough estimates -- at best.

  2. These extrapolations are based on extending the proportions from the raw data to include the schools which did not respond. Also, the proportions of various known categories were used to place most of those listed as "Unknown" in the original surveys.

  3. Extrapolations for 1996-1997 largely reflect presentations made at the 1998 Snowbird Confernece.

  4. Extrapolations for 1997-1998, 1998-1999, 1999-2000, and 2000-2001 reflect my own computations, based primarily on extending maintaining proportions of known categories to the larger populations.

  5. While the number of Ph.D.s is shown to increase from 1996-1997 to 1997-1998, this number can be explained completely by an expanded number of schools surveyed and a corresponding increase in surveys returned. With this adjustment for reporting, it is not clear that any more Ph.D.s were graduated in 1997-1998.

  6. While the raw data shows an increase in the number of Ph.D.s from 1998-1999 to 1999-2000, this increase is more than explained by an increase in the number of schools surveyed and then reporting. The number of schools surveyed increased 5.4%, from 203 to 214; and the number of schools responding jumped from 156 to 173, a 10.9% increase.

    If the percentage increase responding were applied to the number of new Ph.D.s produced in 1998-1999, the expected number of reported new Ph.D.s for 1999-2000 would be 944. However, only 874 new Ph.D.s were reported -- a significant decrease from what one would have expected from the increase in reporting.

  7. The increase in the number of new Ph.D.s from 1999-2000 to 2000-2001 seems real. The number of Ph.D.-granting schools remained about constant during this time (an increase of only 1 or 0.5%), while the number of new Ph.D.s increased by about 8.1%.

  8. Looking at somewhat different data over a longer term, the authors of the 1999-2000 Taulbee Survey note that that the number of CS/CS Ph.D.s awarded in 2000 "is the lowest number in more than 10 years, and it reverses the trend of moderate increases for the past several years. (Computing Research News, March 2001, page 5.)"


This document is available on the World Wide Web as

    http://www.cs.grinnell.edu/~walker/dept/taulbee-97-01.html


created April 11, 2000
last revised November 12, 2002

Henry M. Walker (walker@cs.grinnell.edu)