NCES 2013-008
January 2013

Table A-30.  Estimated equations and model statistics for degrees conferred, by degree level and sex
Dependent variable Equation1     R2 Breusch-Godfrey Serial Correlation LM test statistic2 Time period
Associate's degrees, men DASSOCM =   1,877 + 96DUGFT2M + 57DUGFT2ML2     0.66 1.46 (0.482) 1980–81 to 2009–10
(2.3)   (6.4)   (3.0)

Associate's degrees, women DLOGASSOCW =   # + 0.7DLOGUGFT2WS3 + .3MA(1)     0.58 0.14 (0.932) 1980–81 to 2009–10
(4.5)   (1.9)

Bachelor's degrees, men DBACHM =   1153 + 59DUGFT4M + 144DUGFT4ML2     0.81 2.43 (0.297) 1980–81 to 2009–10
(1.10)   (2.9)   (5.9)

Bachelor's degrees, women DBACHW =   4845 + 57DUGFT4W + 118DUGFT4WL2     0.69 0.67 (0.716) 1980–81 to 2009–10
(2.3)   (2.3)   (3.5)

Master's degrees, men PCHMASTM =   # + 0.6PCHPBFTM + 0.5PCHPBFTML1     0.67 0.83 (0.661) 1980–81 to 2009–10
(3.57)   (3.15)

Master's degrees, women PCHMASTW =   # + 0.4PCHPBFTW + 0.6AR(1)     0.57 3.83 (0.148) 1980–81 to 2009–10
(2.2)   (4.03)

Doctor's degrees, men DDOCM =   -362 + 51DPBFTML1 + 59DPBFTML2     0.43 1.7 (0.413) 1980–81 to 2009–10
(-1.5)   (2.1)   (2.4)

Doctor's degrees, women DDOCW =   562 + 29DPBFTWL1 + 39DPBFTWL2     0.46 0.53 (0.70) 1980–81 to 2009–10
(1.8)   (1.6)   (2.1)
 # Rounds to zero. † Not applicable. 1AR(1) indicates that the model was estimated to account for first-order autocorrelation. To estimate the model, it was first transformed into a nonlinear model and then the coefficients were estimated simultaneously by applying a Marquardt nonlinear least squares algorithm to the transformed equation. MA(1) indicates that the model was estimated to incorporate moving average of the residual into model fit. For a general discussion of the problem of autocorrelation, and the method used to forecast in the presence of autocorrelation, see Judge, G., Hill, W., Griffiths, R., Lutkepohl, H., and Lee, T. (1985). The Theory and Practice of Econometrics. New York: John Wiley and Sons, pp. 315–318. Numbers in parentheses are t-statistics. 2 The number in parentheses is the probability of the Chi-Square associated with the Breusch-Godfrey Serial Correlation LM Test. A p value greater that 5% implies that we do not reject the null hypothesis of no autocorrelation at the 5 percent significance level for a two-tailed test or 10 % significance level for a one-tailed test, (i.e., there is no autocorrelation present). For an explanation of the Breusch-Godfrey Serial Correlation LM test statistic, see Greene, W. (2000). Econometric Analysis. New Jersey: Prentice-Hall. NOTE: R2 is the coefficient of determination. DASSOCM = First difference of associate's degrees awarded to men. DLOGASSOCW = First difference of the log of associate's degrees awarded to women. DBACHM = First difference of bachelor's degrees awarded to men. DBACHW = First difference of bachelor's degrees awarded to women. PCHMASTM = Percentage change in master's degrees awarded to men. PCHMASTW = Percentage change in master's degrees awarded to women. DDOCM = First difference of doctor's degrees awarded to men. DDOCW = First difference of doctor's degrees awarded to women. DUGFT2M = First difference of full-time male undergraduate enrollment in 2-year institutions. DUGFT2ML2 = First difference of full-time male undergraduate enrollment in 2-year institutions, lagged two periods. DLOGUGFT2WS3= First difference of the sum of the full-time female undergraduate enrollment in 2-year institutions over the present year and the previous 2 years. DUGFT4M = First difference of full-time male undergraduate enrollment in 4-year institutions. DUGFT4ML2 = First difference of full-time male undergraduate enrollment in 4-year institutions, lagged two periods. DUGFT4W = First difference of full-time female undergraduate enrollment in 4-year institutions. DUGFT4WL2 = First difference of full-time female undergraduate enrollment in 4-year institutions, lagged two periods. PCHPBFTM = Percentage change in full-time male postbaccalaureate enrollment. PCHPBFTML1 = Percentage change in full-time male postbaccalaureate enrollment lagged one year. DFPBTML2 = First difference of full-time male postbaccalaureate enrollment lagged two years. DFPBTWL1 = First difference of full-time female postbaccalaureate enrollment lagged one year. DFPBTWL2 = First difference of full-time female postbaccalaureate enrollment lagged two years. SOURCE: U.S. Department of Education, National Center for Education Statistics, Degrees Conferred Model, 1980–81 through 2009–10. (This table was prepared March 2012.)