BULLETIN NO. 34 BUREAU OF EDUCATIONAL RESEARCH ARE COLLEGE STUDENTS A SELECT GROUP? By CHARLES W. ODELL Assistant Director, Bureau of Educational Research PUBLISHED BY THE UNIVERSITY OF ILLINOIS, URBANA 1927 PREFACE Studies of the intelligence of high-school seniors and their intentions of attending college, notably the Indiana study by Book, have created the impression that practically no selection takes place between the high school and the college. This hypothesis has been strengthened by observation of college freshmen and by reports showing that large numbers of them fail to do satisfactory work. In interpreting Book's findings and the result of similar studies, most persons appear to have failed to realize that expressing the "intention of attending college" was not the same as "attending college," and that the group which enters college may differ in significant respects from the group which indicated the intention of attending college. In this Bulletin Dr. Odell presents information relative to a group of high-school graduates who actually entered college. In addition to the intrinsic value of the facts reported, the study demonstrates the need for more critical evaluation of data in conducting educational research. In many investigations, the researcher works with substitute data and frequently he changes the label attached to them without keeping in mind that he has done so. For example, Book and many of those who have used his data substituted "college entrance" for "expression of an intention of attending college" and failed to keep this fact in mind. Dr. Odell's study is significant in another respect. He calls attention to certain faults of his data and then shows that the nature of these faults is such that they do not seriously weaken his conclusions. The researcher in the field of education seldom works with perfect data; frequently they involve errors, both constant and variable; and in many cases they are lacking in validity. The researcher must know his data if he is to be scientific; he must ascertain their faults. But he should not stop here; he should inquire concerning the probable effect of these faults when the data are applied to his problem. Sometimes, as in the case of the present study, it may be shown that the existing faults do not seriously limit the conclusions. WALTER S. MONROE, Director. March 17, 1927. |