The Tech - Online EditionMIT's oldest and largest
newspaper & the first
newspaper published
on the web
Boston Weather: 63.0°F | Light Rain

Early Action Acceptance Rate Drops Slightly; Recentered SATs Don't Affect Admissions

By Christina Chu
Staff Reporter

The number of early action acceptances for the Class of 2000 remained about the same as last year, according to Associate Director of Admissions Elizabeth S. Johnson.

According to Johnson, a total of 511 students, or 29 percent of the applicant pool, were admitted early this year.

This year saw an increase of five percent in the number of early action applicants. The percentage of admitted students decreased from 33 percent, but that is a normal variation, Johnson said.

There was no significant change in the number of admitted underrepresented minorities. The percentage of female students admitted under early action was 39 percent, close to last year's 40 percent.

Admittees on new scoring scale

In 1995, students admitted under early action had an SAT verbal mean of 657. This year the SAT verbal mean is 722, Johnson said. All scores are out of a possible 800 points.

"This is not due to the fact that the group of admitted students are stronger, or that we have changed our criteria on admissions, but because the SATs have been recentered," Johnson said.

Since April of 1995, SAT scores have been rescaled so that nationwide verbal and math means fall around 500. "In the past, the math mean has been higher than the verbal mean, but now because of recentered scores, verbal scores are on average higher. As a result, we see an artificial inflation," Johnson said.

In the article "Recentering the SAT," published in the Faculty Newsletter, Director of Admissions Michael C. Behnke explained the reasoning behind recentering scores.

"The SAT is a norm-referenced test for use in making comparisons within a population of test takers at a particular time. Most such tests are renormed on a regular basis. The SAT, however, has also become a Œyardstick' ... to measure educational attainment over time," Behnke wrote.

The reference group used up until now was the original group of about 10,000 test takers in 1941.

"Since then the average verbal has gone down to the 420s and the average math has gone down to around 480. The changes have been attributed to changes in the make-up of the test-taking population [and] the increase in the volume of test takers," Behnke wrote.

Recentering affects admissions

The Admissions Office has studied converting the scores of this past year's applicants to the recentered scale and seeing how that might have affected decisions.

"It is important to note that the level of difficulty of the test will remain the same and that percentiles will be virtually unchanged," Behnke wrote.

"Because the scores had to be Œstretched' up to use the whole top of the scale and other scores were bunched near the bottom, the validity and reliability of the test were weakened."

But not all scores will be changed for the better. Apparently, the math scores from 660 to 710 go down after recentering. This will affect applicants that fall in this range and make them look weaker, Behnke wrote.

"If the 1995 verbal scores were recentered for students accepted under early action, then norm would fall at 717, instead of 657," Johnson said.

Changes in the math mean were not so dramatic. In 1995 the math mean was 756 - with recentering it would be 766. For the 1996 admittees, the mean is 764.