s, and Dissertation Abstracts, shows that there is a relatively standard set of categories (concept learning, problem solving, learning, critical thinking, etc.) for classifying learning studies. Each category might be thought to represent a closely bound set of ideas, but further problems emerge as one begins to read the studies. Take as an example the category learning. In a hypothetical study conducted by X and clearly labeled a discovery study in the title, second and third grade pupils were taught a series of on flowers. The pupils used programed directions written in the Initial Teaching Alphabet. The Mr. X Test of Discovery Learning was given as a preand posttest. The reader knows little about the test other than that it consists of twenty items constructed by X. After analysis of pre-post data, X reports that discovery learning was far superior to conventially taught lessons. In another hypothetical study, by Y, college freshmen in an introductory geology class were engaged in a series of lessons employing geological maps on field trips. An achievement test constructed by Y was given to classes using the discovery and to a control class using only the text and lectures. Analysis of data revealed significant differences in achievement in favor of the experimental discovery class.