Academic journal article
By Loftsgaarden, Don O.; Watkins, Ann E.
The American Statistician , Vol. 52, No. 4
1. THE FALL 1995 CBMS SURVEY
A brief overview of the fall 1995 CBMS Survey is given here as background for this article. Two extensive questionnaires were used to gather the data. One was used for mathematics programs at two-year colleges and the second was used for departments in the mathematical sciences at four-year colleges and universities.
The two-year college population consisted of the 1,023 schools with mathematics programs which were divided into 10 strata on the basis of control (public or private) and total enrollment of the school. The questionnaire was mailed to a stratified random sample of 250 two-year college mathematics programs and the overall response rate was 65%. No "for profit" two-year schools were included in the CBMS survey.
The second population consisted of 1,396 four-year colleges and universities which were divided into 20 strata on the basis of control (public or private), highest degree offered by the mathematics department (bachelor's, master's, or doctor's), and total enrollment of the school. The questionnaire was mailed to a stratified random sample of 349 four-year college and university mathematics departments and the overall response rate was 66%. In both the two-year and four-year samples, the strata for the schools with the smallest enrollments were sampled less heavily than the strata for the schools with the largest enrollments. The optimum allocation of the sample among the strata in each population was made assuming the sampling cost was the same in each stratum and using the enrollment for each school and the number of schools in each stratum.
In addition, in the four-year college population, the questionnaire was mailed to any statistics departments at the 349 schools in the sample. In particular, the statistics departments included in the CBMS survey were those teaching undergraduate statistics courses. These statistics departments were classified as bachelor's, master's, or doctor's according to the highest degree offered by the mathematics department at that school. The number of statistics departments in each stratum in the sample was used to project the total number of statistics departments in that stratum in the population. These numbers of statistics departments were in turn used to make projections of all results for statistics departments. Projections for statistics departments were made separately and results kept separate from those for mathematics departments. There were 50 statistics departments in the sample and 35 responded to the survey.
The 1990 CBMS Survey estimated that there were 60 separate statistics departments at the four-year colleges and universities in the population for the survey. For the 1995 CBMS Survey the estimated number of separate statistics departments was 71. Some of these were new since 1990 and the rest missed in 1990. This increase in the number of separate statistics departments undoubtedly accounts for some, but not all, of the increase in enrollments in statistics departments, in the number of faculty in statistics departments, and in the bachelor's degrees in statistics found by the 1995 CBMS Survey.
Counting the number of statistics departments that teach undergraduate courses is not an easy task. The CBMS survey sampled departments that were separate from the mathematics departments, that taught undergraduate courses in statistics, and that may have offered a bachelor's degree in statistics. The best list to start with in finding such departments is ASA's Schools Offering Degrees in Statistics in the U.S. and Canada Including Departments With Statistics Concentrations, 1995 Edition. This is a self-reported list and may not be complete. On the other hand, the list includes statistics programs within mathematics departments, biostatistics departments, and so on. Here are two examples. The statistics department at one university has only a graduate program and does not teach undergraduate statistics courses or give a bachelor's degree in statistics. …