Tuesday, June 10, 2008

Making the Grade

As it turns out, getting good grades in college is not a tricky empirical question. There is a simple formula:

  1. Be smart;
  2. Work hard.

Studies by James Michaels and Terance Miethe (1989, Social Forces) and William Rau and Ann Durand (2000, Sociology of Education) show a convincing relationship between time spent studying and working on classwork and grades. They also show that things like class rank/high school GPA/SAT scores/other preexisting characteristics have a strong effect.

One thing I found interesting is this: verbal SAT scores (satv) may have a stronger impact than math SAT (satm). Check it out (data analyzed using STATA):

Variable Mean Std. Dev. Min Max
gpas 2.797741 .368208 1.787 3.555
satm 584.1852 42.62034 484 730
satv 573.4815 38.59501 507 668
satc 1157.685 67.05358 1008 1330
------------------------------------------------------------------------------
. reg gpas class satm ecbu [aw=n], beta
-------------+------------------------------ F( 3, 50) = 12.71
-------------+------------------------------ Adj R-squared = 0.3986
------------------------------------------------------------------------------
gpas Coef. Std. Err. t P>t Beta
------+----------------------------------------------------------------
class -.125275 .0250027 -5.01 0.000 -.5486035
satm .001759 .0010401 1.69 0.097 .1894219
ecbu -.1364445 .0820951 -1.66 0.103 -.1812849
_cons 253.4883 50.38878 5.03 0.000
------------------------------------------------------------------------------

. reg gpas class satv ecbu [aw=n], beta
-------------+------------------------------ F( 3, 50) = 16.97
-------------+------------------------------ Adj R-squared = 0.4749
------------------------------------------------------------------------------
gpas Coef. Std. Err. t P>t Beta
------+----------------------------------------------------------------
class -.1028173 .0248065 -4.14 0.000 -.4502567
satv .0033683 .0010379 3.25 0.002 .3962096
ecbu -.0308692 .085759 -0.36 0.720 -.041014
_cons 207.4469 50.09281 4.14 0.000
------------------------------------------------------------------------------
. reg gpas class satm satv ecbu [aw=n], beta
-------------+------------------------------ F( 4, 49) = 12.98
-------------+------------------------------ Adj R-squared = 0.4748
------------------------------------------------------------------------------
gpas Coef. Std. Err. t P>t Beta
------+----------------------------------------------------------------
class -.0999089 .0249782 -4.00 0.000 -.4375205
satm .0010047 .0010068 1.00 0.323 .1081956
satv .0030886 .0010751 2.87 0.006 .3632994
ecbu -.0250653 .0859596 -0.29 0.772 -.0333027
_cons 201.1831 50.48664 3.98 0.000

Based on this, a one-standard-deviation increase in verbal SAT (about 38.6 points) increased these students' semester GPAs by about 0.13 points; math SAT was insignificant. Your choice of major does not seem to have a substantial impact, except perhaps on your salary when you're finished.

No comments:

Post a Comment