Roanoke Times
                 Copyright (c) 1995, Landmark Communications, Inc.

DATE: THURSDAY, April 11, 1991                   TAG: 9104120297
SECTION: EDITORIAL                    PAGE: A-10   EDITION: METRO 
SOURCE: 
DATELINE:                                 LENGTH: Medium

READING PART/ GUESSERS MAY DO OK ON SAT TEST

Wait a minute, you say. Where's the "above passage" I'm supposed to interpret? Lacking further clues, I might as well guess at the answer.

And you might do pretty well. If you're a high-school student taking the new Scholastic Aptitude Test to be used in 1994, you could get nearly half the questions in the reading-comprehension section right just by guessing.

That's the conclusion of Stuart Katz, a psychologist at the University of Georgia. He has his own way of checking the validity of SAT reading-comprehension exams, on which scores have been falling in recent years. He gives the multiple-choice questions to college students without the related reading passages. Then they wing it.

On the old SAT test, still in use, Katz's guessers got 38 percent right in reading comprehension. On the new test - reworked by the College Board to "measure higher-order analytical and evaluative skills" - the guessers got 43 percent correct. On a good reading test, says Katz, their scores would average 20 percent right.

These are not high school juniors or seniors but college students, which could make a difference; presumably they have bigger vocabularies. Lawrence Hecht, the College Board's senior scientist, is unimpressed by Katz's findings. "Reading comprehension," he says, "is a broad skill that involves deriving meaning from the written word - not just the passages, but the wording of the questions and answers." If some right answers can be inferred without the pertinent reading passage, he's not surprised.

It's a broad skill indeed that can get so many right answers from a limited range of materials - sort of like a cook baking a successful cake with an assortment of ingredients but no utensils. But then, the ability to make correct inferences might be one of those "higher-order analytical and evaluative skills" the board wants to measure. Call it educated guessing.

Our guess is that the 1994 test needs a bit more reworking.