A Teacher’s Own Academic Ability

As Dale Ballou of Vanderbilt University and Michael Podgursky of the University of Missouri have noted, the weight of the available education research heavily supports the contention that academically able teachers tend to have higher-performing students.[16] Citing ample research evidence, Richard Murnane and Jennifer Steele of the Harvard University Graduate School of Education write, "One teacher characteristic that is somewhat helpful in predicting student outcomes is academic ability, as measured by verbal aptitude scores, ACT scores, or undergraduate college selectivity."[17]

Analyzing student and district data in Alabama, Ronald Ferguson and Helen Ladd measured the contribution that various education inputs make to student achievement. They found that smarter teachers — that is, those who had higher scores on the ACT when they applied to college — were more successful at increasing their students’ reading and math scores, with the impact stronger in reading than in math. They wrote, "The 0.10 coefficient for reading implies that a difference of one standard deviation in the distribution of teacher test scores would generate a difference of 0.10 standard deviations in the distribution of student test scores."[*]

The teaching profession, however, draws disproportionately from the lower end of the distribution of academic ability.[†] Robin Henke, Xianglei Chen and Sonya Geis of the federal government’s National Center for Education Statistics, for instance, analyzed data from the Baccalaureate and Beyond Longitudinal Data Survey.[18] This dataset includes information on the characteristics of individuals who had prepared to teach and who had entered the teaching profession in 1993 and on those same individuals’ teaching career decisions in 1997. Henke and colleagues showed that college graduates who became teachers were more likely than other college graduates to have scored in the bottom quartile on their college entrance exams.

Writing for the Education Policy Center at Michigan State University, Stanford University education researcher Susanna Loeb and Michelle Reininger, now of Northwestern University, confirmed this finding by studying teachers who graduated from the State University of New York.[19] Loeb and Reininger state that these data "show that elementary and secondary school teachers are more likely to have scored at the lower end of the distribution of SAT scores than non-teachers." As in the study of Henke and colleagues, Loeb and Reininger qualified their findings by showing that teachers of high school math and science are likely to have scored well on their college entrance exams compared to teachers of other subjects and even to nonteachers. To support their own research, Loeb and Reininger also cited Eric Hanushek and Richard Pace[20] and Emeliana Vegas and colleagues[21], who had similar findings.[**]

The teaching profession also tends to lose those with greater academic ability. Loeb and Reininger pointed out, "[T]eachers with higher test scores are more likely to transfer or quit teaching, leaving behind their lower scoring colleagues." This point is reinforced by Henke and colleagues,[22] who found that 32.1 percent of teachers who had scored in the top quartile on their college entrance exams quit teaching, compared to 16.1 percent of teachers who had scored in the bottom quartile.[23]

One trend affecting the teaching work force is that fewer of the most academically able women are entering the field than in prior generations. Using five longitudinal data sets spanning 35 years, Sean Corcoran, William Evans and Robert Schwab found, "Whereas close to 20 percent of females in the top decile in 1964 chose teaching as a profession (teaching was the most frequently reported occupation among this group in 1964), only 3.7 percent of top decile females were teaching in 1992."[24] These researchers also found a decline in the average teacher’s performance on standardized math and verbal tests relative to other female high school graduates. They attributed these declines in teacher quality to enhanced job opportunities for women in other competitive fields. They found, "Top scoring women in our 1992 cohort were much more likely to be working as computer specialists (5.9 percent), accountants (6.0 percent), or managers (15.1 percent).Top decile females were almost as likely to be lawyers (3.2 percent) as teachers."[25]

Analyzing this trend, Caroline Hoxby, now of Stanford University, and Andrew Leigh of The Australian National University found a different explanation. Hoxby and Leigh discovered that two factors related to teacher pay have contributed to the loss of more academically able women from the teaching work force. First, from 1963 to 2000, they showed that the wages for women gradually approached those for men across the range of nonteaching professions. Then they demonstrated that wages for teachers with higher academic aptitude fell relative to the average wage for all teachers and that wages for teachers with lower academic aptitude rose relative to the average for all teachers. Their conclusion is informative: "[W]e cannot expect high-performing college graduates to continue to enter teaching if that is the one profession in which pay is decoupled from performance."[26]

Since academically able teachers tend to be successful in promoting student achievement, it is important to find ways to reverse this trend and encourage them to enter and remain in the teaching profession.

[*] Ronald Ferguson and Helen F. Ladd, “How and Why Money Matters: An Analysis of Alabama Schools,” in Holding Schools Accountable, ed. Helen F. Ladd (Brookings Institution, 1996), p. 277. They continue: “Alternatively, the effect of teacher test scores can be compared with the estimated effects of the socioeconomic characteristics of the community. For example, the estimated coefficients imply that it would take an increase of 25 percentage points in the percentage of college-educated adults (which is equivalent to slightly less than a two-standard deviation change) to achieve the same gain in reading test scores that could be obtained by substituting teachers with test scores one standard deviation higher than those of the school’s current teachers” (p. 278).

[†] Goldhaber, “The Mystery of Good Teaching: Surveying the Evidence on Student Achievement and Teachers’ Characteristics.” In another paper, Goldhaber observes: “Not surprisingly, the non-teacher labor market again rewards ability at a much higher rate than the teacher labor market, with the teacher labor market actually giving a slight premium to those with the lowest SAT scores in 2004.” See Dan D. Goldhaber, “Teacher Pay Reforms” (Center for American Progress, 2006), 8, www.americanprogress.org/issues/2006/12/pdf/teacher_pay_report.pdf (accessed May 22, 2008).

[**] “... Dan Goldhaber and Albert Liu show that in a sample of recent college graduates, those who report considering a career as a teacher have SAT scores that are, on average, about 40 points lower than those who do not, and college graduates who become teachers have SAT scores that are more than 50 points lower than those who enter a different occupation. … The differential in aptitude (measured, for instance, by SAT scores) between teachers and non-teachers is certainly a concern given empirical evidence that higher-aptitude teachers tend to be more effective in the classroom” (Goldhaber, “Teacher Pay Reforms,” 9-10).