Here is the first comprehensive statement of a research method and its theory which were introduced briefly during a workshop at the 1953 AEJ convention. Included are findings from three pilot studies and two experiments in which “cloze procedure” results are compared with those of two readability formulas.

Cloze procedure" was introduced in 1953 , along with experimental evidence to support the conclusion that it is an effective and reli-' able method of quantitatively pretesting and contrasting the relative "readabilities"-for denned populations of readers-of various samples of English prose (4).The present article reports an experiment aimed at testing the validity of cloze indices of readability by determining the degree to which the cloze scores of individual subjects correspond to independent measures of spe-, cine knowledge and general aptitude.It was assumed at the outset that readability and comprehensibility are essentially synonymous terms. The statement that a particular prose passage is, for example, ''very readable" simply seems to mean that it is very easy for some "average" reader to understand.As will shortly be made more specific, a prose sample's cloze score depends on the individual cloze scores of all subjects with re-' gard to that sample. In turn, an individual's ' cloze performance appears to depend heavily on how well he understands the meaning of the materials administered-hence on the factors which affect comprehension, such as general language facility, specific knowledge and vocabulary relevant to the materials at hand, native ability to learn, attention, motivation, and so on. This experiment specifically tested the hypothesis that the cloze scores of individual subjects would correlate significantly with their performances on (a) carefully con-1 The research summarized here was supported by

A technique first presented as a way of testing the readability ofEnglish prose has been found to have many additional uses. I t also works in Korean and with spoken materials. Further, it indexes individual differences in comprehension, intelligence and knowledge; and it gauges the "entropy" of prose context. THE RESEARCH TECHNIQUE CALLED "cloze procedure" was introduced in 1953. In the summer of that year, the method and the results of an initial series of experiments were presented in a roundtable discussion at the AEJ convention at East Lansing, Mich.

Content categories were developed and used to explore what mass media are saying about mental health problems. Radio and TV samples were found to carry larger percentages of relevant material than newspaper and magazine samples, and certain kinds of attitudes seemed to dominate in all media.

No abstract