A READING COMPREHENSION TEST.

Comprehension accuracy, work time, and accuracy rate scores were determined for openbook tests of comprehension of hand translations and edited and unedited machine translations of representative areas of Russian scientific and technical documents in Physics, Earth Sciences, and Electrical Engineering. A total of 390 graduate students with at least a Masters degree background and 36 professional workers with a Masters degree and at least three years of field experience were used in the study. For general comprehensibility, hand translations were generally superior to unedited machine translations, but these differences were less consistent. Similar patterns were found in comprehension of categories of questions which differed in terms of abstractness, the kind of content to which questions referred, the presence of transliterated cognate words in the text, and scope. Performance of graduate students in Physics and Earth Sciences using machine translations who did not read Russian did not differ from the performance of graduate students who did read Russian or from the performance of professional workers. Similarly, Russian reader and Professional worker groups did not differ. Author