Tsinghua 版 (精华区)
发信人: Landau (Dau), 信区: Tsinghua
标 题: Explanation of the Ranking
发信站: 紫 丁 香 (Thu Aug 13 20:29:51 1998), 转信
【 以下文字转载自 Flyingoverseas 讨论区 】
【 原文由 magic 所发表 】
BBS水木清华站∶精华区
发信人: myosotis (飞飞), 信区: AdvancedEdu
标 题: Explanation of the Ranking
发信站: BBS 水木清华站 (Tue Sep 23 10:17:55 1997)
[Image]
[Image]
[Image]
[Image]
The methodology
Primer | Criteria & Weights | Common Data Set
Rankings Index | Organizer | Add to Organizer
An explanation of the U.S. News rankings
Putting the numbers into context
BY J. J. THOMPSON AND ROBERT J. MORSE
A decade ago, U.S. News began ranking colleges. Every year since
then, in consultation with experts in higher education, we have
refined the methodology we use in this process. We now draw on
the results of our survey of higher education leaders about the
reputations of colleges and universities, and on 81 statistical
measures, including the number of applicants to the fall 1996
entering class, the average salary of full professors, and the
amount of money spent on student services.
Over time the ranking model has put less emphasis on input
measures, such as the average high school class rank of a
college's entering freshmen and the range of their standardized
test scores. The model now puts more emphasis on outcome
measures, such as the percentage of a college's entering class
that returns for a second year, and what we call "value added,"
which shows how well a school is educating students, regardless
of the level of their academic achievement before college. We
made this shift because, along with many in the higher education
community, we believe it is most important to measure the quality
of education provided by an institution, not the prior academic
achievements of those who enroll.
This year, U.S. News has not made any significant changes in its
methodology, but there are changes in the way the rankings are
presented. Most important, the final score of each ranked school
is rounded to the nearest whole number, which creates more ties
in the rankings. (In the past, the score was carried to one place
after the decimal point.) U.S. News made this change because
small statistical differences among institutions are not
significant in setting them apart. While we believe the rankings
have value as a source of information, we recognize that schools
whose scores vary by a few points are likely to be comparable in
academic quality. Also, a school's score in the reputational
survey is now shown on a scale from 4.0 (the highest) to 1.0 (the
lowest). In the past, each school's reputation rank was
displayed, such as 164th. We believe showing the actual score is
a more direct way of presenting the same information.
U.S. News arrived at its rankings in the following way:
We placed colleges and universities into one of four categories:
national universities, national liberal arts colleges, regional
universities, and regional liberal arts colleges. Our categories
are derived from those established by the Carnegie Foundation for
the Advancement of Teaching, which are accepted as the best
available classifications for higher education. The regional
schools were placed into one of four geographic regions, North,
South, Midwest, and West. We also print data for, but do not
rank, the service academies and specialty schools in performing
and fine arts, business, and engineering.
In early 1997, we sent 22-page questionnaires to some 1,400
four-year schools. When the surveys were returned, the data were
double-checked with the schools and cross-checked with data
collected by other sources. For example, we checked retention and
graduation rate data supplied by the schools against information
collected by the National Collegiate Athletic Association, and
checked financial data against information gathered by Moody's
Investors Service.
This year, 95 percent of the schools returned the questionnaires.
For those that did not, as well as for those who did not answer
all questions, U.S. News used comparable data from the U.S.
Department of Education, the Council for Aid to Education, the
NCAA, the American Association of University Professors, and
Wintergreen/Orchard House Inc. publishers, as well as data
collected by U.S. News in a prior year. Anytime data are drawn
from an alternative source, this fact is reported in a footnote.
If appropriate data were not available, U.S. News made
statistical estimates using available data points that correlate
highly with the missing pieces of data. For example, retention
and graduation rates are highly correlated. When it was not
appropriate to make such estimates, an estimate based on the
average for that data point among schools in the same category
(i.e., national universities) was used and an "N/A" is included
in the table.
In its methodology, U.S. News includes the following attributes:
Academic reputation. This counts for 25 percent of the final
score for schools in all categories. It is weighted this heavily
because we believe that a diploma from a college or university
well known for its academic excellence helps graduates get better
first jobs and gain acceptance to top graduate programs.
We determined academic reputation--the only subjective attribute
in our model--by surveying the president, provost, and dean of
admissions at each school. (Market Facts Inc., a Chicago-based
opinion research firm, did the actual collection of the
reputational data.) Officials evaluated schools in their own
category by placing an institution into one of four quartiles. If
an official didn't know enough about a school to evaluate it, he
or she could check "don't know" and that response was not
included in our calculations. Every time a school was placed in
the top quartile, it received 4 points, in the next quartile 3
points, and so on. We added all the points and divided them by
the number of responses for each school. The resulting average
scores, ranging from 4.0 to 1.0, are listed in the tables.
(Sixty-two percent of the respondents returned the reputational
survey.)
Retention. This attribute counts for 25 percent of the final
score in the regional categories and 20 percent in the national
categories. It indicates whether a school is retaining and
graduating those who enroll as freshmen. In the higher education
community, this is widely regarded as a measure of how satisfied
students are with a school and whether a school is providing
courses and other services needed for students to graduate within
six years.
There are two components to retention: graduation rates (80
percent) and freshman retention (20 percent). The graduation rate
is based on the proportion of a class that graduates in six years
or less; U.S. News uses an average of the graduating classes that
started as freshmen between 1987 and 1990. Freshman retention is
based on the average proportion of freshmen entering between 1992
and 1995 who returned the next fall.
Faculty resources. This is 20 percent of the final score and
represents the resources that schools devote to instruction and
faculty. Faculty resources is made up of five components: class
size (40 percent), faculty salaries (35 percent), faculty degrees
(15 percent), student-to-faculty ratio (5 percent), and the
proportion of full-time faculty (5 percent).
Class size is a measure of student-faculty interaction; there is
strong evidence that student-professor interaction is important
to learning. In measuring class size, we view classes under 20 as
a positive and classes of 50 and over as a negative.
We include faculty salaries and faculty degrees because they are
measures of the quality of the faculty. For salaries, U.S. News
asked schools to report the average salary-plus-benefits package
for full-time faculty during the 1995 and 1996 academic years.
These were averaged and then adjusted with indexes supplied by
Runzheimer International to diminish regional cost-of-living
differences. For faculty degrees, we calculated the proportion of
faculty holding the highest degree in their fields (usually a
Ph.D.).
The student-faculty ratio indicates how many faculty members are
available on campus for students. And the proportion of full-time
faculty is used because it is clear that the more full-time
faculty members at a school, the more likely it is that faculty
will be available to students and the more likely the school is
to have teachers who are deeply involved in the institution.
Student selectivity. This attribute counts for 15 percent of the
overall score. It consists of four components: test scores (40
percent), high school class standing (35 percent), acceptance
rate (15 percent), and yield (10 percent). The data are for the
fall 1996 entering class.
Both test scores and class rank are important because the
academic climate of a school is set by the abilities and
ambitions of its student body.
Some colleges do not require applicants to submit test scores and
others choose not to report scores for all of their
freshmen--perhaps in an attempt to make their scores look higher.
In such cases, U.S. News has used an estimated value. We put a
footnote by a school's submitted test scores whenever a college
has excluded any of its students' scores or when the average
score represents less than half of its freshman class.
Similarly, in high school class standing, if a school omitted
some students or has data for less than half its freshmen, we
substituted an estimated value in the model. In the national
categories, we use the proportion of freshmen who graduated in
the top 10 percent of their high school classes; in the regional
categories, the proportion in the top 25 percent.
Acceptance rate and yield are, respectively, measures of how
choosy and how successful schools are in admissions. (U.S. News
defines an applicant to a school as someone who has submitted all
information and fees and has been notified of the school's
admission decision.)
Financial resources. This attribute counts for 10 percent of the
overall score and is a measure of the total resources a school
devotes to education. It has two components: educational
expenditures (80 percent) and other expenditures (20 percent). We
consider how much each school spends per student for these
components based on an average of its fiscal 1995 and 1996
expenditures. For the educational part, a small total per student
may mean that a school has little to spend beyond tuition, while
a high figure means that tuition is probably supplemented by a
large endowment, enabling the college or university to offer a
rich educational program.
Value added. This is used only in the national university and
national liberal arts college categories and counts for 5 percent
of the overall score. It is a measure of the school's role in the
academic success of students as well as of how efficient the
school is in playing that role.
In computing value added, U.S. News uses a common statistical
formula to calculate a predicted graduation rate for each school
and then compares that rate with the school's actual graduation
rate. The predicted rate for a school's 1996 graduating class is
based on the average standardized test scores of the 1990
entering class and the average amount the school spent on
education while that class matriculated. If a school's actual
graduation rate is higher than its predicted rate, then the
school is considered to be enhancing the educational attainment
of its graduates and vice versa.
Alumni giving rate. This attribute counts for 5 percent of the
total and measures the average percentage of undergraduate alumni
who gave to their school during the 1995 and 1996 academic years.
The giving rate is a proxy for how satisfied graduates are with
their alma mater.
To get the final rankings of each school, U.S. News first
calculated a score for each attribute--reputation, retention,
faculty resources, selectivity, financial resources, value added
(for the nationals), and alumni giving. This was done by
assigning the school with the highest value for each component of
an attribute a score of 100 percent. Every other school's value
was then taken as a percentage of the top value. The resulting
scores were weighted and then totaled for the attribute scores.
Then the attribute scores were weighted and totaled for the final
score and sorted in descending order.
For the national universities, the method differs slightly
because the expenditures at institutions with large research
programs and medical schools are substantially higher than those
at the rest of the schools in the category. Therefore, we
weighted the attribute scores and ranked them in descending
order. The attribute ranks were then weighted and totaled.
As thorough as this approach is, we make no claim that it is
perfect. No system of rankings can be because it must be based on
judgments about the relative importance of a number of measures.
But we believe that our approach produces rankings that are
reliable and fair. While they are only one factor to evaluate in
considering colleges and universities, the rankings themselves
are the single best source of information because they allow
readers to compare the strengths and weaknesses of different
schools.
[Image]
[Image]
Comments? Questions? Contact us.
� 1997 U.S. News & World Report Inc. All rights reserved.
This site is engineered by AGTinteractive[Image]
--
--------# 飞飞 #--------[u
----------------------------------
| 神也是人来的 |
| 能掌握自己命运的人就是 - 神 |
----------------------------------
※ 来源:·BBS 水木清华站 bbs.net.tsinghua.edu.cn·[FROM: 166.111.78.114]
BBS水木清华站∶精华区
--
※ 来源:．紫 丁 香 bbs.hit.edu.cn．[FROM: 202.118.229.45]
--
※ 转载:．紫 丁 香 bbs.hit.edu.cn．[FROM: 202.118.226.8]
Powered by KBS BBS 2.0 (http://dev.kcn.cn)
页面执行时间：209.571毫秒