-14

dollar National Merit Scholarship Program who annually awards students with high PSAT scores.

In addition, the test Mr. Lowen had taken was made up of old SATS. Robert Moultrop, ETS spokesman, estimated that the scores of about 80,000 of the 316,000 persons who took tests using the "lousy item" in recent years should probably be raised by as much as 10 points. Their scores have never been corrected.

Mr. Lowen's discovery, to quote a New York Times headline, opened a "Pandora's Box." The Times and the testing companies received a "deluge of letters" suggesting a variety of new answers to the same questions. E. Bevin Williams, Director of Test Development at ETS in response to queries on the legitimacy of alternative solutions presented in the correspondence said with a shrug, "It seems as if every answer was correct except ours."

Just one week later on March 23, 1981, the fifth mistake, discovered by Michael Galligan, was announced by ETS. Galligan, a senior at Clarkstown South High School in New City, New York, discovered the error on the October SAT. Again the public was not informed until well after the error was discovered and rescored. In a letter to ETS, Mr. Galligan pointed out that a question about integers on the exam had two right answers. Despite admitting that Mr. Galligan was correct, Ernest Kimmel, Director of Test Development at ETS, declared, "We know the people who chose C (the other correct answer discovered by Galligan) were less able: they skipped over the obvious answer.

COPYRIGHT © 1980 BY ETS AND THE COLLEGE BOARD

A student pointed out that (C) should
have been accepted as a correct
answer, as well as the “official” (B).

-15

As a result of Mr. Galligan's discovery about 16,000 of some 84,000 "less able" test takers in New York State had their scores increased and several hundred thus qualified for state scholarships worth $1,000 over a four year period.

Mr. Lowen's discovery, coupled with Mr. Galigan's caused quite a disturbance in the testing industry. Just three days later, the College Board Trustees, faced with convincing evidence that opening up the testing process had and would improve test quality, not to mention the effect these discoveries would have on pending legislation, voted to partially extend the "Truth-inTesting" concept nationwide. Their new policy, however, is limited to selected test dates and unlike "Truth-in-Testing" legislation, provides no concrete guarentees that disclosure will continue in the future. It is interesting to note that College Board officials just three days earlier testified against Truth-in-Testing in both Massachusetts and Florida.

Many things concerning the sixth and seventh mistakes which occurred on the GRE in advanced engineering and advanced psychology are still unexplained. Both questions, according to ETS, were not released by the GRE Board because of the possibility of the questions being reworded and used in future administrations of these exams. Who discovered the errors is also open to debate. Edward Fiske, New York Times Reporter, on April 28, 1981 said that, "students successfully challanged questions on the GRE." ETS Public Relations person, Marilyn Ballis claimed that both mistakes were discovered as part of the ongoing review process after they were printed and administered. Assuming Mr. Fiske was correct, students who had discovered these mistakes would have gained no satisfaction, for the GRE Board voted not to change any incorrect scores unless differences of more than 10 points would result.

Ms. Ballis, in referene to the two errors remarked, "it was only a matter of four points" and was not statistically significant." Robert Altman, acting Vice-President for Higher Education Programs at ETS, in defense of the Boards decision stated, "We continually tell colleges that it is dangerous to make judgements about students on the basis of 10 point differences. If we sent out corrections of this magnitude, it would contradict everything we have been telling them." When asked about the GRE Board's decision, Thomas White, President of the LSAS, remarked, "I wouldn't want to apply the same logic to our test or to the SAT. There are instances, even if we do not consider them justified, in which 10 points can have an impact on a person's educational opportunities."

"

A recent FTC report entitled "Coaching for Standardized Admissions Tests" reprted that "42% of the private four year colleges and 37.9% of the public four year colleges responding (to their survey) indicated that they had minimum SAT score requirements below which applicants generally are not considered eligible for admission." In addition, although test makers "discourage" the use of cut-off scores, preliminary selection for National Merit Scholarships are based on "cut-off" scores on the PSAT. Finally, according to Bevin Williams, the testing industry isn't always aware of how test scores are used. He stated, "As long as there is any chance that not rescoring will affect any individual, then it is irresponsible not to do so.'

"

Of the 23,200 students who took the February 21 administration of the LSAT, almost 22,000 had to have their scores changed. The eighth error was

-16

discovered by Christopher Laucks of Brookline, Massachusetts when he realized
that their were two possible answers to a Geometry question on the exam, (see
diagram). The discovery resulted in 15,000 students' scores incresing, 1,600
scores remaining the same, and 6,600 scores dicreasing.

EXPLANATION

Figure 1 shows question 23 as it appeared on the LSAT. This question involves the comparison of two quantities, the length of XY and the integer 5. The credited response to the question was (C), that is, that the two quantities are equal. The intent of the question was to have the candidate observe that, if line segments labeled a and d were extended, they would meet in a right angle and form a right triangle with hypotenuse of length XY and sides of length a+b= 4 and c+d=3 (see Figure 2). The length XY would then be √4' + 3' = √25=5, and the answer would be (C) This reasoning is correct if, as many people assumed from the way the figure is drawn, the angle between the two triangles

where line segments labeled b and c meet is a right angle or if line
segments labeled a and b (or those labeled c and d) are known to
be parallel However, since this information is missing. there can
be other triangles that satisfy the given conditions, but for which
the length XY is not 5 The person who made the initial inquiry
pointed out one such case (see Figure 3). In this case line
segments labeled c and d are not parallel, and the length XY is
√3+2+√T2 + T2 = √13 + √2. which is greater than 5. Since
there are cases for which the length XY is 5 and other cases for
which the length is not 5, the answer to the question as it
appeared is (D), that is, the relationship cannot be determined
from the information given.

Three days after the mistake was reported, the LSAC, which represents 171 law schools, approved by the American Bar Association and is sponsor of the LSAT, announced drastic changes in the structure and scoring of the LSAT. According to the Associated Press, begining in the summer of 1982 students taking the LSAT would no longer be required to correct ungramatical sentences or solve geometry and other "quantitative" math questions. Also the scal used to report LSAT scores (200-800) would be replaced by a scale ranging from 10-50. Bruce Zimmer explained the council's actions are an effort to "discourage the use of small score differences to shape admissions decisions." It is important to note that generally Law Schools plac 60 to 70 per cent of their emphasis in admissions on LSAT scores. Only 30 to 40 per cent is placed on other criteria, like GPA and extra-curricular activities.

-17

DEMISTIFICATION

Truth-in-Testing legislation in New York is also having a positive impact on the availability of quality test preparation courses to low income and minority test takers. Patrick Shields of the East Harlem College and Career Counseling Program, a federally funded organization designed to provide additional assistance to college and career bound students, has been providing a three week coaching course for program participants. He characterized the materials available prior to the enaction of the law (published commercial materials and ETS sample booklets) as of questionable value. There was a very limited number of sample items of which many were of perhaps little value.

Due to the enaction of the law, Mr Shield maintains that having realistic SAT items available has enabled him to do a "better job." "Students," he described, "feel more comfortable about taking the tests now that they've had their hands on the real thing."

Jeannette Burgos-Servedio, Head Counselor of Talant Search, also feels that Truth-in-Testing is providing information to her which helps her to do a better job. In a letter to the Committee she states:

At the request of students in our program, we at Talent
Search Program of Bronx Community College sponsor an SAT coaching
course ever academic year. Presently, we charge the students
$10 for an average of eight two-hour sessions. We are now in
the process of writing the course into our project so that we
an offer the course at no cost and, therefore, make the course
available to an even greater number of our student population.

Throughout the 1980-81 academic year, we had four sessions;
and we got a opportunity to use actual SAT tests with the students
in the last session. Their responses, in comparison to those
students who only got an opportunity to use sample tests, were
more positive. They felt that having had an actual test before
them helped not only prepare them more adequately for the exam,
but they also felt that the level of familiarity they gained
by having the test would help relieve much of the anxiety they
would have gone into the exam with. This, in turn, would
enhance their chance of obtaining a better score.

As a counselor, I feel that, regardless of what others may
feel about coaching, coaching does help our student population.
That population being predominantly low income, minority students.
Coaching for these students gives them an opportunity to learn

test taking techniques they do not necessarily pick up in the high
schools. It also raises their level of confidence.

According to Luis DeGraffe, of the Puerto Rican Legal Education and Defense Fund which administers a coaching course to about 60 paw school candidates every year, Truth-in-Testing helps his students in two major

ways:

-18

1. students who take the exam more than once are enabled to see what part of the exam they are weakest in. They can then prepare accordingly.

2. it enables us to institute a simulation of a test situation. This can help to relieve much of the test anxiety which students generally experience.

The ultimate solution to the problem of disclosure is for the federal government to enact a nation-wide "Truth-in-Testing" law. This will eliminate the difficulties the testing industry faces having to comply with a variety of forms of state legislation and the costs incurred thereof. In addition, students now and in the future will be assurred that the testing process is conducted in a fair and open manner.