Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 THE RELATIONSHIP BETWEEN THE FIVE VECTORS OF VARIABLES AND TTD AND ITS COMPONENTS: A COMPARISON OF ZERO-ORDER CORRELATIONS This chapter examines the zero-order correlations between a number of the independent variables used in the model and l-! l) and its component parts. The analysis discusses why the data are broken down by field and describes zero- order correlations between select variables in the five vectors in the model- family background, student attributes, tuition and financial aid, institutional environment, and market (economic and social) forces and TTD and its components. It also provides correlations among the several salary and employment variables themselves and analyzes the amount of variation in 1 TIP explained by each vector. The Importance of Disaggregation by Field - Existing studies either addressed issues related to ~-1L) aggregated over all fields (Wilson, 1965) or controlled for field differences using a set of dummy variables (Abed) and Benkin, 1987~. The former approach ignores the possibility that a given independent variable (e.g., whether the student has an undergraduate degree in the same field) may have a different effect in some fields than in others, while the latter makes the rather stringent assumption that a one-unit change in an independent variable has the same effect on TTD for a student in chemistry, for example, as it does for a student in the biosciences. A number of studies of student aspirations and persistence suggest both assumptions are wrong (Biglan, 1973; Girves and Wemmerus, 1988; Thistlethwaite, 1962~. And economic research suggests market conditions differ among scientific and engineering fields (Berger, 1988; Freeman, 1971~. Failure to recognize that differences among field exist can give rise to models that give inaccurate explanations of why TTD changes. The following sections provide field-specific data on the variables that the opportunity-cost analysis and the literature suggest have an effect on AID. The zero-order correlations are suggestive, since the actual relationship between an independent variable and the dependent variable is captured by a model that 45
tests their effect, holding all other things equal. These correlations highlight the differences among fields and help to explicate the interrelationship of the variables in each vector with TTD and its components. They also make it possible to examine the relationship between TTD and variables for which data are not available for sufficiently long periods of time. Unless otherwise noted, all data are for the 1967-1986 period. All correlations are between the selected variable and TTD and its components. A single asterisk (*) denotes the correlation is significant at 1-percent level or greater. A double asterisk (**) denotes a significance level of 5 percent. Changes in Family Background Characteristics Of particular interest are the percentage of graduates in each doctorate cohort who are married (Table 4.1), the average number of dependents of doctorates in each cohort (Table 4.2), the percentage who are black (Table 4.3), the percentage who are Hispanic (Table 4.4), and the percentage who are women (Table 4.5~. Changes in Student Attributes The variables of interest are average age of the cohort at the start of the doctoral program (Table 4.6), percentage in the cohort who attended a highly selective undergraduate school (Table 4.7), and percentage of the cohort with an undergraduate degree in the same field as their doctorate (Table 4.~. For the reasons discussed in Chapter 3, a thorough analysis of TTD should employ a measure either of student ability, such as undergraduate or graduate cumulative grade-point average,l2 or of achievement level, using scores from the SAT, ACT, or ORE. Unfortunately, the DRF does not contain data either on student grades or on predoctorate test scores. To develop a "proxy" measure of the skills that a given cohort possesses, we used a variable equal to the percentage of new doctorates in each cohort who attended a selective undergraduate institution, where the average incoming 1973-74 freshmen earned a combined SAT verbal and math score of 1,300 or higher. The assumption is that the larger the percentage of students from institutions of this type, the larger the overall ability level of the students in a given cohort. 12 Student grades pose a technical problem when they are aggregated at the cohort level because the ordinal scales used to grade students at different institutions are not additive. 46
TABLE 4.1: Correlation for Percent Married, 1967-1986 Field l lo RTD TPGE TNEU Chemistry -0.46** -0.67* -0.37 0.80* Physics/Astronomy -0.68* -0.82* -0.59* 0.57* Ea~/Atmospheric/Marine Sciences -0.44* * -0. 89* -0.08 0. 84* Mathematics/Computer Sciences -0. 80* -0. 87* -0.74* -0.29 Engineering -0.69* -0.80* -0. 11 0.05 Agricultural Sciences -0.07* -0.86* -0.63* 0.68* Biosciences -0.62* -0.92* -0.60* 0.78* Health Sciences -0. 66* -0.80* -0.48* * -0.41 Psychology -0.82* -0.85* -0.85* -0.38 Economics -0.66* -0.91* -0.04 0.68* Social Sciences -0.77* -0.93* -0.45** 0.65* NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.2: Correlation for Average Number of Dependents, 1967-1986 Field ITD RTD PAGE LIEU Chemistry -0.66* -0.84* -0.42 0.76* Physics/Astronomy -0.83* -0.91 * -0.64* 0.46* * Earth/Atmospheric/Marine Sciences -0.48** -0.90* -0.03 0 . 8 5 * Mathematics/Computer Sciences -0.84* -0.91* -0.74* -0.36 Engineering -0.85* -0.90* -0. 15 -0. 11 Agricultural Sciences 0 .04 -0.88* 0.72* O .7 8* Biosciences -0.50** -0.88* -0.46** 0.85* Health Sciences -0.5 8 * -0.74* -0.48* * -0.29 Psychology -0.74* -0.81 * -0.78* -0.24 Economics -0.47** -0.96* 0.18 0.85* Social Sciences -0.71 * -0.92* -0.37 0.75* NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 47
TABLE 43: Correlation for Percent Black, 1974-1986 Field 1-ID RTD TPGE INEU Chemistry 0.09 0.008 -0.20 0.33 Physics/Astronomy 0.05 0.49 0.24 -0.48: Earth/Atmospheric/Marine Sciences -0.007 0.25 -0.05 -0.59 Mathematics/Computer Sciences -0.20 -0.19 -0.27 -0.03 Engineering 0.25 0.47 -0.37 0.03 Agricultural Sciences 0.16 0.33 0.01 -0.13 Biosciences -0.09 -0.15 -0. 15 0.3 3 Health Sciences 0.01 0.06 -0.15 0.08 Psychology 0.57* 0.62** 0.53 0.46 Economics 0.60* 0.74* 0.55** -0.93* Social Sciences 0.70* 0.75* 0.61** 0-.10 NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.4: Correlation for Percent Hispanic, 1974-1986 Field l lV RTD TPGE TNEI J Chemistry 0.16 0.54 -0.14 -0.33 Physics/Astronomy 0.07 0.25 0.53** -0.40 Earth/Atmospheric/Marine Sciences 0.22 0.34 -0.03 -0.32 Mathematics/Computer Sciences 0.79* 0.73 * 0.73 * 0.53 Engineering 0.05 0.30 -0.49 -0.41 Agricultural Sciences 0.22 0.60** -0.26 -0.22 Biosciences 0.84* 0.86* 0.87* -0.17 Health Sciences 0.66* 0.62** 0.71* 0.52 Psychology 0.93 * 0.93 * 0.92* 0.91 Economics 0.68* 0.75* 0.41 -0.53 Social Sciences 0.89* 0.93* 0.83* 0.31 NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 48
TABLE 4.5: Correlation for Percent Female, 1967-1986 Field l lV RID TPGE INAJ Chemistry 0.63* 0.78* 0.48** -0.70* Physics/Astronomy 0.68* 0.82* 0.69* -0.58* Earth/Atmospheric/Marine Sciences 0.59* 0.92* 0.06 -0.76* Mathematics/Computer Sciences 0.93 * 0.97* 0.86* 0.46** Engineering 0.65* 0.72* -0.07 0.04 Agricultural Sciences 0.08 0.87* -0.67* -0.67* Biosciences 0.70* 0.96* 0.68* -0.72* Health Sciences 0.82* 0.88* 0.71* 0.58* Psychology 0.91* 0.94* 0.93* 0.48** Economics 0.63* 0.91* 0.03 -0.70* Social Sciences 0.86* 0.96* 0.56* -0.57* NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.6: Correlation for Average Age at Start of Doctoral Program, 1967- 1986. Field TIN RTD TPGE luau Chemistry 0.85* 0.89* 0.73* -0.51* Physics/Astronomy 0.62* 0.61* 0.78* 0.23 Earth/Atrnospheric/hIarine Sciences 0.31 -Q.16 0.49* * 0.49 * * Mathematics/Computer Sciences 0.93* 0.88* 0.91 * 0.65* Engineering 0.29 0.08 0.83 * 0.42 Agricultural Sciences 0.44** -0.45** 0.75* 0.70* Biosciences 0.90* 0.77* 0.91 * -0.11 Health Sciences 0.76* 0.61* 0.88* 0.65* Psychology 0.99* 0.96* 0.98* 0.76* Economics 0.72* -0.04 0.76* 0.42 Social Sciences 0.89* 0.61* 0.93* 0.19 . NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 49
Table 4.7: Correlation for Percent from Selective Undergraduate Schools, 1967 1986 Field l lV RID TPGE TNAJ Chemistry -0.61* -0.62* -0.40 0.27 Physics/Astronomy -0.55 * -0.43 -0.44* * -0.31 Earth/Atmospheric/~Iarine Sciences -0.30 -0.13 -0.29 -0.11 Mathematics/Computer Sciences 0.01 -0.003 0.11 -0.13 Engineering -0.25 -0.22 -0.37 -0.02 Agricultural Sciences -0.32 0.60* -0.56* -0.80* Biosciences 0.30 0.19 0.37 0.01 Health Sciences -0.06 -0.13 -0.30 0.15 Psychology -0.39 -0.47** -0.42 0.02 Economics 0.11 0.11 0.08 -0.11 Social Sciences -0.70* -0.90* -0.34 0.68* NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.8: Correlation for Percent with Undergraduate Degree in Doctoral Field, 1974-1986 Field l lV RID TPGE TNEU Chemistry -0.002 -0.06 0.07 0.22 Physics/Astronomy 0.76* 0.84* 0.45** -0.26 Earth/Atmospheric/Marine Sciences -0.29 0.67* -0.05 0.68* Mathematics/Computer Sciences -0.82* -0.80* -0.83 * -0.40 Engineering -0.80* -0.89* -0.07 -0.02 Agricultural Sciences 0.28 -0.70* 0.65 * 0.85* Biosciences 0.52* 0.89* 0.50* -0.85* Health Sciences 0.77 0.83* 0.66* 0.57* Psychology 0.07 0.16 0.08 -0.16 Economics 0.24 0.85* -0.27 -0.85* Social Sciences 0.57* 0.79* 0.23 -0.69* NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 50
Changes in Tuition and Financial Aid The variables of interest here are average tuition and fees paid in a given year (Table 4.9~; percentage of students with federal support (Table 4.10), private foundation support (Table 4.11), research assistantships (Table 4.12), or teaching assistantships (Table 4.13~; and percentage of students who relied on their own earnings as their gnmary means of support (Table 4.14~. Changes in Institutional Environment This subsection examines the relationship between a select number of aggregate measures of institutional environment and MID. These are the percentage of students with a baccalaureate from a foreign institution (Table 4.15, p. 54), the ratio of full-iime equivalent faculty to doctorate recipients (Table 4.16, p. 55), the ratio of the dollar value of government R&D TABLE 4.9: Correlation for Average Tuition Paid Field llV RID IPGE TNEU Chemistry 0.07 Physics/Astronomy -0.007 EarthJAtmospheric/Marine Sciences -0.05 Mathematics/Computer Sciences -0.35 Engineering -0. 14 Agricultural Sciences 0.28 Biosciences -0.44* * Health Sciences -0.26 Psychology -0.45** Economics -0.54* * Social Sciences 0.42 -0.05 -0.16 -0.30 -0.3 1 -0.15 -0.17 -0.39 -0.53 -0.40 -0.27 -0.33 0.05 0.12 -0.05 -0.36 0.27 0.17 -0.46 0.16 -0.46 -0.18 -0.54 0.33 0.38 0.40 -0.17 -0.10 0.36 0.09 0.21 -0.52* -0.06 0.02 NOTES: (1) These are zero-order correlation coefficients. (2) Tuition lagged three years. Weights were used to aggregate public end private institutions. Since national averages are not available for graduate tuition and fees, our analysis assumes that undergraduate tuition is a good proxy variable. The assumption is that undergraduate and graduate tuitions are highly correlated and that increases in the former are accompanied by similar increases in the latter. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 51
TABLE 4.10: Correlation for Percent with Primary Support from Federal Government, 1967- 1986 Field l lV RTD TPGE EMU Chemistry -0.52** -0.70* -0.44 0.76* Physics/Astronomy -0.51 * -0.67* -0.51 * 0.64* Earth/Atmospheric/Marine Sciences -0.49* * -0.83 * -0.04 0.74* Mathematics/Computer Sciences -0.87* -0.91 * -0.81 * -0.38 Engineering -0.61* -0.69* 0.14 -0.04 Agricultural Sciences -0.10 -0.59* 0.47* * 0.39 Biosciences -0.69* -0.94* -0.67* 0.72* Health Sciences -0.74* -0.86* -0.48** -0.52** Psychology -0.94* -0.94* -0.95* -0.60* Economics -0.91 * -0.60* -0.47* * 0.20 Social Sciences -0.88 * -0.67 * -0.87 * -0.07 - NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.11: Correlation for Percent with Primary Support from Private Foundations, 1967- 1986 Field l lV RTD TPGE TNUJ Chemistry -0.14 -0.18 0.06 0.05 Physics/Astronomy -0.58* -0.56* -0.21 -0.07 Earth/Atmospheric/Marine Sciences 0.01 -0.26 0.18 0.40 Mathematics/Computer Sciences -0.06 -0.23 0.03 0.26 Engineering -0.26 -0.38 -0.33 0.31 Agricultural Sciences 0.43 0.30 0.02 0.09 Biosciences 0.61* 0.18 0.59* 0.51** Health Sciences -0.16 -0.24 -0.40 0.10 Psychology -0.56* -0.65 * -0.62* -0.07 Economics -0.02 -0.78* 0.43 0.86* Social Sciences -0.21 -0.61* 0.24 0.90* NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 52
TABLE 4.12: Correlation for Percent with Primary Support from Research Assistantships, 1967-1986 . . Field TI,D RTD TPGE TNEU Chemistry 0.63* 0.76* 0.46** -0.65* Physics/Astronomy 0.60* 0.73* 0.56* -0.54* Ear~/Atmospheric/Marine Sciences 0.33 0.82* -0.09 -0.88* Mathematics/Computer Sciences 0.72* 0.75 * 0.72* 0.23 Engineering 0.79* 0.80* 0.09 -0.06 Agricultural Sciences -0.01 0.85 * - 0.66 * -0.77 * Biosciences 0.77* 0.96* 0.74* -0.63* Health Sciences 0.80* 0.91* 0.62* 0.58* Psychology -0.64* 0.66 * -0.73 * -0.29 Economics 0.41 0.67* -0.09 -0.59* Social Sciences 0.71* 0.77* 0.43 -0.39 NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.13: Correlation for Percent with Primary Support from Teaching Assistantships, 1967-1986 Field l lo RTD IPGE TNE]J Chemistry 0.68* 0.72* 0.46** -0.38 Physics/Astronomy 0.87* 0.88* 0.48** -0.03 Earth/Atmospheric/Marine Sciences -0 .54* -0 . 61 * -0 . 24 0 .40 Mathematics/Computer Sciences 0.45 0.56* 0.31 O.1S Engineering 0.66* 0.69* 0.43 0.09 Agricultural Sciences -0.02 0.65* -0.46** -0.61* Biosciences 0.1 7 0. 64* 0.10 -O.8S* Health Sciences 0.20 0.14 0.39 0.07 Psychology 0.04 0.10 0.04 -0.18 Economics 0.24 0.86* -0.29 -0.86* Social Sciences 0.61* 0.81* 0.24 -0.65 NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 53
TABLE 4.14: Correlation for Percent with Primary Support from Own Earnings, 1977-1986 Field l lV RTD TPGE TNEU Chemistry 0.63** 0.77* 0.11 0.25 Physics/Astronomy 0.52 -0.16 0.47 0.62 Earth/Atmospheric/Marine Sciences 0.71** 0.81* 0.40 0.14 Mathematics/Computer Sciences 0.81* 0.69** 0.72** 0.79* Engineering -0.32 -0.32 -0.33 -0.08 Agricultural Sciences 0.66* * 0.31 0.49 0.77* Biosciences 0.79* 0.73** 0.78* 0.63** Health Sciences 0.92* 0.89* 0.93* 0.85* Psychology 0.84* 0.85* 0.81 * 0.78* Economics 0.66** 0.79* 0.05 -0.46 Social Sciences 0.91* 0.92* 0.84* 0.86* NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.15: Correlation for Percent with B accalaureate from Foreign Institutions, 1967-1986 Field l lV RTD TPGE TNEU Chemistry 0.42 0.34 0.14 0.19 Physics/Astronomy 0.32 0.17 0.07 0.53* Earth/Atmospheric/hIarine Sciences 0.22 -0.13 0.40 0.41 Mathematics/Computer Sciences 0.90* 0.90* 0.88* 0.50* Engineering 0.76* 0.74* 0.70* 0.16 Agricultural Sciences 0.21 -0.29 0.32 0.41 Biosciences -0.52** -0.69* -0.53* 0.51* Health Sciences -0.46** -0.45 * * -0.18 -0.48* * Psychology -0.28 -0.27 -0.34 -0.09 Economics 0.68* 0.53* 0.35 -0.29 Social Sciences -0.32 -0.42 0.19 0.38 NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 54
TABLE 4.16: Correlation for Number of Full-Time Equivalent Faculty l Field l ill) RID TPGE LIEU Chemistry 0.68* 0.85* 0.37 -0.65* Physics/Astronomy 0.88* 0.95* 0.65* -0.33 Eard~lAtmosphericIMarine Sciences 0 .3 8 0 . 82* -0.09 -0 . 81 * Mathematics/Computer Sciences 0.80* 0.90* 0.69* 0.33 Engineering 0.79* 0.93* 0.19 -0.14 Agricultural Sciences -0.06 0.83* -0.66* -0.79* Biosciences 0.28 0.75* 0.25 -0.91* Health Sciences 0.48 0.60* 0.49 0.19 Psychology 0.60* 0.69* 0.64* 0.04* Economics 0.22 0.90* -0.37 -0.92* Social Sciences 0.52** 0.83* 0.11 -0.85* NOTES: (1) These are zero-order correlation coefficients. (2) Period for AD is 1967-1986; FACULTY, a crude proxy for the number of mentors available to doctorate students, is lagged, 1964-1983. *Denotes correlation is statistically significant at 1 percent level or greater. **Denotes correlation is significant at 5 percent level or greater. TABLE 4.17: Correlation for Government R&D Spending Field l lo RID TPGE TNEU Chemistry 0.54* 0.70* 0.47* -0.72* Physics/Astronomy 0.59* 0.73 * 0. 64* -0.61 * Earth/Atmospheric/Marine Sciences 0.63* 0.92* 0.07 -0.72* Mathematics/Computer Sciences 0.92* 0.91* 0.90* 0.49** Engineering 0. 62* 0.71 * -0.14 0.01 Agricultural Sciences 0.18 0.86* -0.56* -0.60* Biosciences 0.80* 0.98* 0.79* -0.63* Health Sciences 0.89* 0.95* 0.67* 0.70* Psychology 0.97 * 0.98 * 0.97 * 0.63 * Economics 0.75* 0.89* 0.14 -0.61* Social Sciences 0.95 0.96* 0.71* -0.40 NOTES: (1) These are zero-order correlation coefficients. (2) Period for AD is 1967-1986 and for R&D is 1964-1983. *Denotes correlation is statistically significant at 1 percent level or greater. **Denotes correlation is significant at 5 percent level or greater. 55
expenditures to doctorate recipients (Table 4.17), the percentage of doctorate recipients who received an undergraduate degree from a Research I school as identified by the Carnegie Classification (Table 4.18), the percentage of students who received an undergraduate degree from a "top 40" school as identified by the NRC's Assessment of Research-Doctorate Programs in the United States (Table 4.19), the percentage of students who received a graduate degree from a Carnegie- classified Research I or Research II school (Table 4.20), and the percentage of students who received a graduate degree from a "top 40" school (Table 4.21~. Changes in Market Forces Salary Variables An exhaustive review of salary data revealed differences in the quantity and quality of various sources (Tables 4.22-4.25~. Only seven data files were used; others were excluded either because their academic field classifications were incompatible with those in this study or because the time spans of data collection were inadequate.l3 13- The sources for data on salary were the following: the American Institute of Physics, baccalaureate salary data beginning in 1965 for physics and astronomy [three missing years of data (1964, 1966, and 1967) were generated using an instrumental variable based on the Endicott Report data for physics]; Battelle Columbus Laboratones' data series that begins in 1968 for baccalaureate and doctorate salaries in engineering, chemistry., and physics (BCL's data series for life sciences was considered too aggregated for use in the model, but the data are shown in the correlation table with SDR salary for biological scientists); College Placement Council data on salary offers to baccalaureates, starting in 1964 for chemistry and math (excluding computer sciences); Endicott Report data on baccalaureate salary starting in 1964 for chemistry, math, engineering, and the combined field of economics and finance; the Survey of Doctorate Recipients, the only source of doctorate salary data for all 11 fields (such data have been collected on a biennial basis since 1973; however, an instrumental variable was created based on Bureau of Labor Statistics' Weekly. Earnings data to provide even-year data and to project salaries back to 1964~; and baccalaureate salary data from the National Survey of Hospital and Medical School Salaries starting in 1964 for staff nurses, used as a proxy for health sciences. 56
TABLE 4.18: Correlation for Percent with B accalaureate from Category I Research University, 1967-1986 Field al lV RTD TPGE TNEU Chemistry -0.34 -0.37 0.26 -0.14 Physics/Astronomy 0.26 0.42 0.31 -0.58* Ear~/Atmospheric/Marine Sciences -0.34 -0.22 -0.22 -0.03 Mathematics/Computer Sciences 0.31 0.30 -0.38 0.02 Engineering -O. 14 -0.03 -0.3 1 -O. 19 Agricultural Sciences 0.17 -0.35 0.26 0.50** Biosciences 0.58* 0.82* 0.58* -0.68* Health Sciences 0.04 -0.08 -0.10 0.14 Psychology -0.49* * -0.5 0* * -0.48 * * -0.35 Economics 0.45** -0.27 0.51** 0.52** Social Sciences -0.40 -O. 67* -0.10 0.72* efficients. Category I Research University is taken from the Carnegie Classification of Colleges arid Universities. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.19: Correlation for Percent with Baccalaureate from "Top 40" School, 1967-1986 Field l lV RTD URGE 1NAJ Chemistry -O. 63 * -0.75 * -0.13 0.3 6 Physics/Astronomy 0.01 0.12 0.06 -0.40 Earth/Atmospheric/Marine Sciences 0.04 0.12 0.06 -0.22 Mathematics/Computer Sciences -O. 10 -O.OB -0.08 -O. 16 Engineering -0.79* -0.78* -0.59* -0.15 Agricultural Sciences -0.39 0.06 -0.25 -0.35 Biosciences 0.54* O. 81 * 0.54* -0.71 Health Sciences 0.03 -0.05 -0.18 0.19 Psychology -0.55 * -O. 61 * -0.57* -0.20 Economics 0.30 -0.1 1 0.21 0.24 Social Sciences -0.41 -O. 65 * -0.12 O. 65 NOTES: These are zero-order correlation coefficients. "Top 40" refers to those schools so identified in the NRC's Assessment of Research Doctorate Programs in the United States. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 57
TABLE 4.20: Correlation for Percent with Graduate Degree from Category I or Category II Research School, 1967-1986 Field l lV RTD TPGE TNEU Chemistry 0.03 0.20 0.17 -0.64* PhysicstAstronomy 0.13 0.27 0.25 -0.59* Earth/Atmospheric/Marine Sciences -0;46** -0.79* -0.02 0.66* Mathematics/Computer Sciences -0.17 -0.13 -0.13 -0.38 Engineering -0.14 -0.06 -0.53 * -0.08 Agricultural Sciences -0.17 -0.47** 0.18 0.29 Biosciences 0.003 -0.44** 0.04 0.74* Health Sciences - 0.15 - 0.33 - 0.23 0.06 Psychology -0.79* -0.85* -0.82* -0.26 Economics 0.17 -0.63* 0.67* 0.77* Social Sciences -0.68* -0.87* -0.38 0.67* NOTES: These are zero-order correlation coefficients. Category I Research University is taken from the Carnegie Classification of Colleges and Universities. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.21: Correlation for Percent with Graduate Degree from "Top 40" School, 1967-1986 Field ~ lV RTD TPGE TNEU Chemistry -0.16 0.01 0.02 -0.58* Physics/Astronomy 0.17 0.32 0.25 -0.58* EarthtAtmospheric/Marine Sciences 0.005 0.14 -0.29 -0.20 Mathematics/Computer Sciences 0.08 0.10 0.12 -0.19 Engineering -0.46** -0.47* * -0.50* * -0.02 Agricultural Sciences -0.28 -0.82* 0.26 0.53* Biosciences 0.05 -0.29 0.05 0.56* Health Sciences -0.08 -0.28 -0.17 0.16 Psychology -0.80* -0.87$ -0.84* -0.29 Economics 0.35 -0.40 0.72* 0.59* Social Sciences -0.45* -0.73* -0.09 0.77* NOTES: These are zero-order correlation coefficients. "Top 40" refers to those schools so identified in the NRC's Assessment of Research Doctorate Programs in the United States. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 58
\ TABLE 4.22: Correlation for Average Salary of Recent Doctorate Recipients Field TID RTD TPGE EMU Chemistry 0.70* 0.79* 0~40 -0.50** Physics/Astronomy 0.67* 0.62* 0.25 0.23 Ear~/Atmospheric/hIarine Sciences -0.57* -0.52** -O.28 0.19 Mathematics/Computer Sciences 0.49** 0.59* 0.35 0.24 Engineering 0.78 * 0. 84* 0 .40 0. 02 Agricultural Sciences -0.20 -0.86* 0.57* 0.57* Biosciences -0.73* -0.96* -0.73* 0.70* Health Sciences -0.80* -0.82* -0.60* -0.70* Psychology -0.96* -0.97* -0.97* -0. 60* Economics -0.68* -0.37 -0.53* 0.05 Social Sciences -0.89* -0.97* -0.63* 0.52** NOTES: (1) These are zero-order correlation coefficients. (2) Specifically, SDR salary is regressed on weekly eamings, and the coefficients from this regression are used to estimate salaries in the missing years; SDRSAL is lagged three years. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.23: Correlation Between SALRAT1 and TTD and Its Components Field llV RTD TPGE TNEU Chemistry -0.69* -0.87* -0.43 0.7 1 * Physics/Astronomy -0.59* -0.75* -0.52* 0.54* Ear~/Atmospheric/Marine Sciences -0.007 0 .10 -0.002 -0 .13 Mathematics/Computer Sciences -0. 87* -0.95 * -0.78* -0.3 7 Engineering -0.75 * -0. 85 * 0.08 0.04 Agricultural Sciences 0.18 0. 66* -0.37* -0.47* * Biosciences 0.60* 0.91* 0.57* -0.77* Health Sciences 0.38 0.34 0.42 0.32 Psychology 0.76* 0.76* 0. 80* 0.44** Economics 0.41 0.56* 0.20 -0.39 Social Sciences 0.23 0.29 0.19 -0.25 NOTES: (1) These are zero-order correlation coefficients. (2) The years prior to 1973 are projected; SALRAT1 is lagged three years. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 59
TABLE 4.24: Correlation for Salary Ratio of Doctorates 10 Years After Degree Field l lV RID TPGE LIEU Chemistry Physics/Astronomy Ear~/Atmosphenc/Marine Sciences Mathematics/Computer Sciences 0.56 Engineenng 0.55 Agncultural Sciences Biosciences Health Sciences Psychology Economics Social Sciences 0.63 0.62 -0.51 0.72* 0.57* 0.63* 0.63* 0.5 1** 0.34 -0.56* 0.14 0.42 0.33 0.13 0.01 -0.6 1 * -0.56* -0.26 0.21 0.07 -0.01 -0.01 NOTES: (1) These are zero-order correlation coefficients. (2) A comparison is made to the baccalaureate rather than the master's salary because of the larger number of observations in the former category; SALRAT10 is lagged Tree years. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. Employment Indicator Variables The percentage of doctorate recipients seeking postgraduate employment is a reliable indicator of job market opportunity. Job opportunities, in turn, are likely to affect TTD. Data on job-seeking behavior are easy to obtain and reasonably reliable (Tables 4.2~4.30~. However, because such data are collected at the time the doctoral candidate is completing the degree, they may understate employment prospects, because finding a job after graduation takes time. Data on job-seeking activity have been used in studies by Freeman (1971~. The Stock Variable The zero-order correlations between TTI) and its components and the number of doctorates in the United States divided by the U.S. population are shown in Table 4.31. 60
to Go 1 to or a) o v, o c~ .5 v, a: c~ ~ - v, 3 m o - C. o ·- V) e' E~ - 0 ~ :~ 1 ~ _ _ <: _ od m V, ~ 0 ~ 03 <; o 0 ~ ~ ~, V, C~ ~n ~ S 0 ~ v, c~ ~ m~ V, ;^ 0d U. m ;^ Ct _ ~ _ 0 ~ _ ._ * * * * ~ O ~oo oo ~oo O 0 1 1 0 1 o 1 11 1 * * * ~ ~oo oo ~n oo ·· · ~ o0 1 1 o 1 o 1 1 1 1 * o~ ~o · ~ 1 0 1 1 1 1 1 o 1 1 1 oo ~oo O C~ · · · ~ o 1 1 Ol o 1 1 1 1 o 1 * * C~ · - 0 1 1 0 1 1 1 1 1 1 1 . _ V, ^~- ~ E = ti 5 '~ w E =0 ~ = m :: 61 ~ ._ e~ ~ 3 ·- 2 ._ _ _ ~ _ o . ~ K _ _l e. m~ ~.O u: 'c, - 4 . _ 8 .o ._ ~ c~ ~ O ~ ~ ~ _ O ~ ·_ ~ O-0 . ~ ~ r- ca cn c~ ._ O.~ - g ~ ~ ~ _ ,~,0 ~D ~ Om ~ ~D o - _ - - - ~ ~ ·~ oo~ · - o ,'t u, ~ - ~ _ - c, e) ~ · ~ - ~ - - - ~ ~ ~ o o ~ ~ ~c- c -c- Z.3 ~* *
TABLE 4.26: Correlation for Percent Seeking Postgraduate Employment Field llV RTD TPGE TN~ Chemistry 0.62* Physics/Astronomy 0.7S* Earth/Atmospheric/Marine Sciences -0.02 Mathematics/Computer Sciences 0.30 Engineering 0.65 * Agricultural Sciences 0.20 Biosciences 0.24 Health Sciences 0.51 * Psychology 0.74* Economics 0.19 Social Sciences 0.84* 0.58* 0.60* 0.42 0.41 0.62* 0.80* 0.61* 0.55* 0.78* 0.75* 0.95* 0.44* 0.36 -0.12 0.12 0.55* -0.39* 0.19 0.59* 0.78* -0.33 0.54* -0.17 0.44** -0.60* 0.11 0.24 -0.60* -0.71* 0.27 0.30 -0.72* -0.58* NOTES: (1) These are zero-order correlation coefficients. (2) SEEK variable is lagged three years. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. TABLE 4.27: Correlation for Percent with Definite Employment or Postdoctoral Appointment Field l lV RTD TPOE ~J Chemistry -0.64* Physics/Astronomy -0.75 * Earth/Atmospheric/hIarine Sciences -0.02 Mathematics/Computer Sciences -0.31 Engineering -0.63 * Agricultural Sciences -0.20 Biosciences -0.1 ~ Health Sciences -0.62* Psychology -0.71 * Economics -0.09 Social Sciences -0.82* -0.61* -0.61* -0.44** -0.43 -0.60* -0.77* -0.54* -0.67* -0.75* -0.66* -0.95* -0.45** -0.38 0.15 -0.14 -0.53* 0.37 -0.11 -0.62* -0.75 * 0.33 -0.52* 0.19 -0.37 0.60* -0.09 -0.24 0.56* 0.68* -0.36 -0.26 0.66* 0.60* NOTES: (1) These are zero-order correlation coefficients. (2) DEFIN variable is lagged three years. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 62
TABLE 4.28: Correlation for Overall U.S. Unemployment Rate Field T1D RTD TPOE TNAJ Chemistry 0.55* 0.61* 0.61* -0.55* Physics/Astronomy 0.61 * 0.64* 0.63 * -0.28 Earth/Atmospheric/hIarine Sciences 0.39 0.73* 0.15 -0.72* Mathematics/Computer Sciences 0.81* 0.77* 0.73* 0.57* Engineering 0.54* 0.63* -0.15 0.06 Agricultural Sciences 0.18 0.71 * -0.37 -0.51 * Biosciences 0.71* 0.82* 0.69* -0.47* Health Sciences 0.72* 0.81* 0.50** 0.53* Psychology 0.78* 0.78* 0.82* 0.52* Economics 0.61* 0.71* 0.07 -0.44** Social Sciences 0.77* 0.78* 0.63* -0.34* NOTES: (1) These are zero-order correlation coefficients. variable is lagged three years. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. (2) Unemployment TABLE 4.29: Correlation for Unemployment Rate of College-Educated Population Field l lo RID TPGE EMU Chemistry 0.58 * 0.63 * 0.63 * -0.57* Physics/Astronomy 0.74* 0.73* 0.68* -0.15 Earth/Atmospheric/bIarine Sciences 0.38 0.70* 0.16 -0.69* Mathematics/Computer Sciences 0.77* 0.78* 0.67* 0.48** Engineering 0.68* 0.74* 0.02 0.14 Agricultural Sciences 0.14 0.73* -0.39 -0.58* Biosciences 0.58* 0.78* 0.54* -0.55* Health Sciences 0.63* 0.73* 0.47** 0.42 Psychology 0.70* 0.71* 0.75* 0.39 Economics 0.44** 0.72* -0.07 -0.55* Social Sciences 0.67* 0.76* 0.48** -0.47** NOTES: (1) These are zero-order correlation coefficients. rate is lagged three years. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 63 (2) Unemployment
TABLE 4~30: Correlation Between Percent Seeking Postgraduate Employment and the Other Market Variables Between SEEK and . Field I3NEMP* UNEMP4YR** Chemistry 0.43 0.65 Physics/Astronomy 0.41 0.60 Ear~/Atmospheric/Marine Sciences 0.55 0.71 Mathematics/Computer Sciences 0.36 0.56 Engineering 0.40 0.5 8 Agricultural Sciences 0.72 0.85 Biosciences 0.64 0.80 Health Sciences 0.69 0.82 Psychology 0.81 0.88 Economics 0.70 0.82 Social Sciences 0.88 0.88 *UNEMP = Overall unemployment rate for the U.S. labor force (obtained from He Bureau of Labor Statistics) **UNEMP4YR = Unemployment rate for persons with four or more years of college (obtained from the Bureau of Labor Statistics) TABLE 4.31: Correlation for Per Capita Number of Doctorates in the United States, 1967-1986 Field l lV RTD TPOE TOFU Chemistry -0.71* -0.82* -0.49* * 0. 62* Physics/Astronomy -0. 87* -0.93 * -0.63* 0.3 0 Earth/Atmospheric/Marine Sciences -0 .47 * -0 . 89 * -0.010 0 . 84 * Mathematics/Computer Sciences -0. 8 7 * -0. 91* -0.7 8 * -0.44* Engineering -0.83* -0.91* -0.16* -0.06 Agricultural Sciences -0.05* -0.70 0.36 0.63* Biosciences -0.59* -0.87* -0.54** 0.70* Health Sciences -0.38 -0.38 -0.11 -0.39 Psychology -0.91* -0.89* -0.91 * -0.68* Economics -0.46** -0.91* -0.1 7 0.79 * Social Sciences -0.85* -0.84* -0.67* 0.3 6 NOTE: These are zero-order correlation coefficients. *Denotes correlation is statistically significant at 1% level or greater. **Denotes correlation is significant at 5% level or greater. 64
The Zero-Order Correlation of the Vectors The correlations between I-l ~ and all of the variables in each vector for which data are available for 1967-1986 are in Table 4.32. Regression analysis was used to derive an adjusted R2 for each vector on the assumption that this is Me only vector that affects l-! ~ (no one model consistently has the highest Rat. The F vector (family background charactenstics) explains most of the adjusted variation in TTD in math, health sciences, and social sciences. The I vector (individual attributes) explains most of the variation in chemistry, engineering, and psychology. Variations in two fields agricultural sciences and biosciences are best explained by the TLFA vector (tuition and financial aid). Finally, the O vector (organizational factors) explains most of IlD's adjusted variation in the remaining three fields: P&A; EAM; and economics. Remarkably, the E vector (economic variables) was not able to predict a larger amount of the variation than other vectors in any fields. TABLE 432: Amount of Adjusted Variation in TTD Explained by Each of the Five Vectors Vector Field F I TLFA O E Chemistry 0.73* 0.76* 0.75* 0.71* 0.48 Physics/Astronomy 0.80* 0.77* 0.81* 0.84* 0.76* Ear~/Atmospheric/Marine Sciences 0.33** 0.29** 0.44* 0.66* 0.18 Mathematics/Computer Sciences 0.96* 0.86* 0.93* 0.95* 0.89* Engineering 0.71 * 0.85* 0.57* 0.78* 0.63 * Agricultural Sciences 0.11 0.09 0.34** 0.31 -0 Biosciences 0.82* 0.82* 0.94* 0.84* 0.50 Health Sciences 0.91* 0.75* 0.71* 0.83* 0.41* Psychology 0.97* 0.98* 0.94* 0.95* 0.59* Economics 0.76* 0.64* 0.88* 0.89* 0.17 Social Sciences 0.98* 0.91* 0.95* 0.92* 0.67* NOTE: F = Family Background (MARRIED, DEPEND, TEMP, WOMEN); I = AGE, SAMEFLD, SELECT; TLFA = TUITION, SUPPED, SUPPRIV, SUPTA, SUPRA; 0 = FORBACC, BTOP40, BCARN1ST, PTOP40, PCARN1ST, FACULTY, R&D; E = SALRAT1, UNEMP4YR, SEEK. Acronyms are defined in Appendix B. pp. 175- 177. *Denotes correlation statistically significant at 1% level or greater. **Denotes correlation significant at 5% level or greater. 65
TABLE 4.33: Number of Fields in Which Each Variable Had a Statistically Significant Correlation with AD or RTD TTD RTD Negative Positive Negative Positive Correlation Correlation Correlation Correlation Family Background Characteristics MARRED 10 0 11 0 DEPEND 10 0 11 0 BLACK 0 3 0 3 HISPANIC 0 6 0 7 WORN O 1 0 0 1 1 Student Attributes AGE 0 9 1 7 SELECI 3 0 4 0 SAMEFLD 2 3 3 6 Tuition and Financial Aid TUllION ~ 3 0 1 0 SUPPED 1 0 0 1 1 0 SUPPRIV 2 1 4 0 SUPRA 1 7 0 1 1 SUPTA 1 4 1 8 SUPOWN O 9 0 8 Institutional Environment FORBACC 2 3 2 3 FACULTY 0 8 0 11 R&D 0 9 0 11 BCARN1ST 1 2 2 1 BTOP40 3 1 4 1 PCARN1ST 3 0 6 0 PTOP40 3 0 4 0 Salary Variables SDRSAL 7 3 6 4 SALRAT1 4 2 4 4 Employment Conditions SEEK 0 6 0 10 D1EFIN 6 0 10 0 UNEMP O 9 . 0 11 UNEMP4YR 0 9 0 1 1 PERPOP Stock Variable 10 0 9 0 NOTES: (1 ) These are zero-order correlation coefficients. (2) Statistical significance is at the .05 level. (3) Acronyms are defined in Appendix B. pp. 175- 177. 66
Summary Table 4.33 shows the number of fields with which each of the independent variables had a statistically significant correlation (p = .05) to time to the doctorate. The table is limited to zero-order correlations with T1 D and its component RTD, since other components did not appear to increase TTD. The greatest correlation to I1D was for variables indicating marital status, dependents, gender, and federal financial support. These correlations were apparent in 10 fields for 11D and in all 11 fields for RTD. As predicted by the opportunity-cost analysis, married members of the cohort and cohort members with dependents had a negative correlation to DID. Those with federal support also showed a negative correlation to TTD, which was not predicted. Female gender was positively correlated to 1-1~. Other variables that were strongly and positively correlated with RTD in all fields were research assistantships, number of full-time faculty, level of federal R&D support, the overall unemployment rate, and the unemployment rate for college graduates. The signs were not always as predicted; for example, the relation between the unemployment variables and RTD was expected to be negative but turned out positive. Zero-order correlations must be approached with some caution. While they are useful for demonstrating an association between ICED and/or its component parts and the variables posited by the literature and by opportunity- cost hypotheses, the nature of He relationship is speculative. In addition, some of"the independent variables are time-dependent and may move up or down together over long periods. 67
