School officials are becoming more concerned with their
inability to attract applicants into the teaching profession, and how to entice
and retain talented, and highly qualified teachers. The education sector used
to have a large labor pool of women college graduates is now in competition
with a variety of professions – and the best and highly skilled are perceived
to unlikely become a teacher.
Aggravation over the quality of the teaching forces comes
from findings that suggest that the particular measures of teacher quality
(such as verbal and mathematical skills) are robustly linked to student output.
The strength of this result is in great opposition to the argument over the
significance of other inputs into the production of education, such as
per-student spending, and class size.
Although the hypothesis that labor market desegregation has
influenced the teacher quality seems to be broadly acknowledged, there is
remarkably little evidence that measures the degree to which it is true.
Several cross-sectional studies have pointed that university
graduates becoming teachers in the 1970’s and the 1980’s did not match up
satisfactorily to their counterparts. Furthermore due to lack of information,
less is known on how such relationship changed over time.
Currently, (to the authors’ knowledge) there are no
available data that:
1. Collect
data before and after the change of the labor market that started in the 1960’s
2. Contain
variables that can be deemed as reliable teacher quality measures
3. Recognize
practicing teachers with a high degree of certainty
This is with the possible exception of the set of National
Longitudinal Surveys (1968), which Richard J. Murnane et. al. (1991), and
Marigee P. Bacolod (2003) used to examine the changing teachers attributes.
Both found a reduction in the share of college graduates having high AFQT[1]
scores deciding to teach.
I. Data and
Methodology
Five longitudinal surveys of high school pupils across more
than four decades:
1. Wisconsin
Longitudinal Study (WLS) for the class of 1957
2. Project
Talent for the classes of 1960-1964
3. National
Longitudinal Study of the High School Class of 1972 (NLS-72)
4. High
School and Beyond (HSB) for the class of 1982
5. National
Education Longitudinal Study of 1988 (NELS) for the class of 1992
These surveys include results from a questionnaire given
during the senior year; all require involvement in a series of aptitude tests,
and all accomplish several follow-ups after graduation. By including
standardized test scores for all students, graduates can be placed in cohort
skill distribution and the tendency of women (or men) having high relative
scores to go into teaching over time can be assessed.
Advantages of large longitudinal surveys of high-school
graduates:
1. Follow-up
surveys allow identification of those individual who actually became teachers
2. They
provide large samples from a fairly stable population (of high school
graduates) over this period.
Imperfections of large longitudinal surveys of high-school
graduates:
1. Only
snapshots of the five groups over this period are provided
2. Not a
nationally representative sample of high school graduates, for instance the
oldest data set, (WLS), consists only of non-Hispanic whites in Wisconsin.
3. The
data are exposed to sample attrition (like all longitudinal surveys). (However,
participation in the follow-up survey was high).
The skill measure is a centile ranking based on a pupils’
placement in the high-school graduates distribution on the math and verbal
parts of the exam taken during the senior year – and the correlation between
the skill measure and the score is quite high.
While the limitations of the measure are accepted, previous
literature seems to support the notion that given observable attributes, a
teacher’s math and verbal aptitude is perceived to be the most significant for
student performance.
All women high school graduates (with an available test
score, and participated in a follow-up survey) from each group were selected.
Using self-reported occupation (which is reported despite of individual’s
labor-force status), teacher and non-teachers are distinguished from the data
sets. The broad description is useful in this context, if women likely to have
more time out of the labor force self-select into occupation like teaching.[2]
II. Results
Predictably, the centile ranking of the mean female teacher
in the sample graded consistently above that of the mean female high-school
graduate and below the mean female college graduate. In addition, there is a
remarkable downward trend (a drop from 0.60 to 0.46) in the mean female teacher
standard score.
Although this drop in the average relative skill of new
female teachers is of importance, it would also be helpful to determine how
admission into teaching profession altered differentially across the skill
distribution. This is observed using estimated five logit models where the
likelihood of going into teaching is assumed to be a function of an
individual’s test-score decile, age, and race.
With regards to the predicted probability that a white
female in each decile becomes a teacher (for each of the cohorts), it was
observed that for females in most deciles the likelihood of being distinguished
as a teacher declined by about half from 1964 to 1992 with a slight increase
between 1992 to 2000. Overall, women belonging in the top deciles are much less
probable to be teachers over time, while those belonging in the bottom are more
likely to become teachers.
An interesting consequence of the gender desegregation of
professions and the shift of skilled females into high-cognitive-ability
occupation is the possible substitution of highly-talented men into
teaching. Repeating the analysis on the
available but much smaller sample size of male teachers for the five cohorts
male graduates, the result are quite intriguing, if only indicative. Across the groups, the test-score ranking of
the mean male teacher increased from 1964 to 2000 by 6.6 percent. This
increment seems to be motivated by those at the top of the distribution:
although the possibility that any male graduate go into teaching dropped during
this period, the decrease in the probability is much less striking for those in
the top decile. These results however should be taken with caution, and are
worthy of further study
III. Conclusion
One of the observations, with using the centile rank in the
distribution of high-school graduates on a standardized test (of verbal and
mathematical skill), is a minor but noticeable decline in the relative skill of
the mean new female teacher. The extent of this drop is even larger when
ability is measured by standardized scores. Furthermore, across the period
(1964-2000), females close to the top of the test-score spread (assumed to be
those most likely to gain from labor-market desegregation) were less likely to
become teachers, in comparison to their peers (belonging in the middle of the
distribution).
These results suggest (if applied to a bigger population) a
given pupil in 2000, compared to a given pupil in 1964, could expect to find a
(female) teacher who is, approximately, of only somewhat lower academic
ability. On the other side, that pupil has a lower possibility of having a
teacher of the utmost academic skill than was a pupil in 1964.
Source:
Sean P. Corcoran, William
N. Evans, and Robert M. Schwab, “Changing Labor-Market Opportunities for Women
and the Quality of Teachers, 1957-2000”, American
Economic Review. Vol. 94, No.2, (May, 2004), pp. 230-235.
0 comments:
Post a Comment