Inhuman Test Corrector Has Perfect Score

Professor Henry S. Dyer '28, Director of the Office of Tcats, has a wondrous machine. Exame go into a little slot, the machine grumbles, sputters, coughs, and hiccups out the tests completely corrected. During the current examination period Dyer's IBM gadget is grading Economics 1, Psychology 1, Social Relations 1a, Biology 1, and sections of nearly all language finals.

The workings of the machine are fairly shuple. A stencil is placed over the test paper which covers all spaces except those where correct answers are supposed to be marked. Both the test paper and the stencil are then placed in a slot up against a panel of thick copper pins. When the machine is turned on, the pins pick up all pencil marks that show through the spaces in the stencil, because pencil marks made with special pencils, conduct electricity. The rest is simple. The machine just "counts" the number of electrical impulses and then stamps it on the exam paper.

All Papers Checked

Often the marks are not sufficiently "heavy, glossy, and black" for the machine and secretaries in the testing office have to go over the marks by hand. "We check each paper anyway to make sure that it was marked correctly by the student and we erase all the answers to a question if we find more than one has been made," one of Dyer's assistants remarked.

"We also have to make sure the machine is working properly," she went on, "so we rescore several papers from each batch by hand." Each set of tests is also run through the machine twice to make sure that there have been no errors in correcting. Thus far, Dyer claimed, the correcting apparatus has never made a mistake that wasn't caught by the human scorers in his University Hall headquarters.

Catches Mistakes

Mistaken as well as correct answers can be detected by Dyer's appliance. If a desperate student marked all answer blanks the machine, when set for counting errors, would turn in a score of 400 for every hundred questions.

At top speed the correcting machine can grade 800 papers an hour.

"Our chief purpose up here is not marking papers," Dyer pointed out. "What we're trying to do at present is to explore individual differences by means of statistical research. We predict the academic rank of each freshman after registration and then follow up his record. Our predictions are usually 78 percent accurate within a certain area."

Data for Dyer's research comes from five principle sources, including Kuder Preference Tests and College Board Aptitude Survey Tests, which freshmen will be asked to take on a voluntary basis shortly.