Cooperative Test Service Research Articles

John C. Flanagan and the development of psychological surveys: The U.S. aviation psychology program

An account is given of how John C. Flanagan became a major pioneer of psychological testing field research as a result of survey he conducted into factors affecting family size amongst U.S. Army Air Corps officers. Carried out on the initiative of a small New York nonprofit organization, this led to a massive government-sponsored psychological program covering U.S. Army pilots throughout World War II. KEY WORDS: John C. Flanagan, History of Psychological Surveys, Family Size, U.S. Army Air Corps, Dysgenics and War, Pioneer Fund. John C. Flanagan is a major name in the history of American psychology, having gained wide recognition for his pioneering role in the organization and supervision of extensive field surveys. The best known of these studies involved the psychological testing of U.S. Army pilots during World War II. This was a task he was asked to accomplish following his successful management of a 1938 U.S. Army Air Corps survey which sought to identify the factors responsible for decisions regarding family size made by Army Air Corps officers and their wives. Privately financed, but personally approved by General Hap Arnold and confirmed by the War Department, this study opened the door for Flanagan's massive World War II psychological surveys and subsequent research achievements. It was funded by a small New York foundation financed by Colonel Wickliffe Preston Draper, a textile heir who was himself a pilot and a decorated World War I combat hero. While representing the window of opportunity that served Flanagan so well in his career as a psychologist, the 1938 study is of intrinsic interest in itself. John C. Flanagan Born in 1906 in Armour, South Dakota, John C. Flanagan was the son of a Baptist minister who had graduated at the University of the son of a Baptist minister who had graduated at the University of Minnesota, and of a mother who had graduated from Cornell and had taught school before her marriage. At the age of eighteen, the young Flanagan was an enthusiastic football player who enrolled in the engineering school of the University of Washington, where he played as a quarterback on the college team. However, with only a few credits still required to earn a degree in electrical engineering, Flanagan transferred to the college of education when he found his studies interfering with his football practice. After graduating with a major in physics, he enrolled in graduate school at the University of Washington where he continued to take courses on a part-time basis while launching into a career as a High School science and mathematics teacher and football coach. Finding himself becoming interested in psychology, Flanagan enrolled in a summer program at Yale University in order to take courses in psychology under E. L. Thorndike, as well as Truman Lee Kelley and other notable scholars. Under their inspiration he decided - in the midst of the Great Depression - to leave his secure teaching position in Seattle to accept a fellowship in psychology at Harvard University where he earned a Ph.D. in mental measurement under the renowned Professor Kelley in 1934. Flanagan remained at Harvard to work for Walter F. Dearborn, who was collecting data in local schools for the twelve-year Harvard Growth Study. He also worked with Benjamin D. Wood, director of the Cooperative Test Service of the American Council on Education at Columbia University, where he was in charge of the annual achievement test program which collaborated with many leaders in the field of testing and measurement. Here he showed that teacher evaluations in fields such as mathematics and English were too often based on only very general assumptions, and that many teachers failed to evaluate student ability accurately. As a result, Flanagan's first major academic publication appeared as early as 1935: namely Factor Analysis in the Study of Personality, published by Stanford University Press. …

A Two-Stage Sequential Strategy in the Placement of Students in an Undergraduate Mathematics Curriculum

The purpose of the study was to demonstrate the effectiveness of applying decision theory in guidance for an undergraduate mathematics placement program. The strategy that was developed utilized the Scholastic Aptitude Test in Mathematics (SAT-M) and the Mathematics Test of the Cooperative Test Service (COOP-M) in the first and second stages, respectively. Application of the standard 2-stage sequential model resulted in a high utility factor by deriving a terminal decision for 50% of the 460 Ss at the first stage. A modified version of the decision model was recommended with a utility factor of 35% to allow for variance not accounted for by the predictor variables.

A Two-Stage Sequential Strategy in the Placement of Students in an Undergraduate Mathematics Curriculum

The purpose of the study was to demonstrate the effectiveness of applying decision theory in guidance for an undergraduate mathematics placement program. The strategy that was developed utilized the Scholastic Aptitude Test in Mathematics (SAT-M) and the Mathematics Test of the Cooperative Test Service (COOP-M) in the first and second stages, respectively. Application of the standard 2-stage sequential model resulted in a high utility factor by deriving a terminal decision for 50% of the 460 Ss at the first stage. A modified version of the decision model was recommended with a utility factor of 35% to allow for variance not accounted for by the predictor variables.

Research News and Communications

The merger of non-profit testing organ izations is now under consideration by the organizations concerned. This merger was originally proposed by the Commit tee on Testing of the Carnegie Founda tion. The affected organizations would include the College Entrance Examination Board, the Cooperative Test Service, The Graduate Record Office, and the Educa tional Records Bureau. Further steps will be announced in the near future.

An Analysis of the Results from the First Annual Edition of the National Teacher Examinations

This discussion is a report of the results of the first annual administration of the National Teacher Examinations. Discussions of the nature of the examination service and its development1 and of the administration of these examinations2 have been published else where. The National Committee on Teacher Examinations was appointed by the Amer ican Council on Education at the request of a group of superintendents in the spring of 1939. The Carnegie Foundation for the Ad vancement of Teaching provided funds for the development of the project, and the National Committee assigned to the Cooper ative Test Service of the American Council on Education the task of preparing the annual forms of the battery of examinations to be used. The service made available to superin tendents and boards of education by the National Committee on Teacher Examina tions has three principal objectives: first, the provision and administration of a battery of teacher examinations; second, the collation and distribution of information on the ap praisal of other factors to be considered in the selection of teachers ; third, the promotion of research in teacher education and guidance. It should be stated at the outset that the superintendents who inaugurated this project were well aware that no battery of tests could at the present time make possible a complete evaluation of all factors to be included in the selection of teachers. However, this group of superintendents felt that having objective and comparable information on certain aspects of training and professional preparation would be of great assistance to them with their in dividual problems of teacher selection. In common with all other persons who have been confronted with the task of teacher selection, these superintendents fully realized the diffi culty of interpreting the types of information which have previously been available to them. The selection of general areas to be tested was made chiefly through a series of confer ences with superintendents and a question naire sent to a larger group of superintend ents in the spring of 1939. Information about such traits as desirable attitudes and personal characteristics had to be eliminated from the proposed battery of examinations since satis factory techniques for measuring these traits objectively are not available. In general, the superintendents stated that they wished in formation about the extent to which candi dates could read with understanding, could express themselves clearly, and could handle numerical and non-verbal concepts intelli gently. They also wished to know to what extent the candidates had acquired a good liberal education and the extent of their awareness of contemporary affairs and cur rent social problems. They wished informa tion concerning the degree to which the candi date was aware of the social implications of education, his knowledge of the basic prin ciples of psychology as they apply to educa tion, and his familiarity with the techniques of guiding and assisting students. Lastly, they wished to know the extent to which the teacher was really a master of those fields in which he proposed to teach.

The Performance of the Mathematics Candidates in the 1940 National Teacher Examinations1

Probably every superintendent of schools has often wished for some measure of the the ability and knowledge of the applicants who offer themselves for positions on his staff which would be comprehensive, accurate, and comparable from applicant to applicant. In order to help meet this need, the American Council on Education in the spring of 1939 appointed a National Committee on Teacher Examinations and authorized it to supervise and delegate to the Cooperative Test Service the task of preparing a battery of objective tests for the examination of teacher candidates.

Age of Parent and Intelligence of Offspring

THIS article reports a study of the relation between age of parent and intelligence of offspring made through an examination of the records of the Georgia State Womans College. Differences between age of parent and of daughter were determined, and ratings were secured on several tests given at entrance into college. The Psychological Examination of the American Council on Education was used to test general ability. In the autumn of I937 the Cooperative Test Service tests in English, mathematics, social studies, and science were used. In the preceding years the tests used in these fields were prepared by the University System of Georgia. Tabulations were made according to migration of the family since the daughter's birth. Test ratings were tabulated according to whether the daughter was among the upper or the lower 50 per cent of the students taking a particular test. Data for non-migrating families are presented in Table I. These data show, with one exception, that the daughters ranking among the upper 50 per cent on each of the tests are more likely to have older parents than are the daughters in the lower 50 per cent. Conversely, the daughters in the lower 50 per cent on the tests are more likely to have young parents than are the daughters in the upper 50 per cent. Thus, on the American Council test 53.7 per cent of the daughters in the upper 50 per cent have fathers who are thirty or more years older than themselves, whereas only 42.2 per cent of those in the lower 50 per cent have fathers of comparable ages. Conversely, of the daughters in the lower half, 57.8 per cent have fathers who are not more than thirty years older than themselves, whereas of those in the upper half only 46.3 per cent have fathers this young. The same situation prevails when age of mother is considered in relation to age and rating of daughter. Thus, 50 617

Correlation of Achievement Scores and School Marks

Among the many investigations that have been made of the relation between achievement-test scores and school marks, a recent study reported the correlation between scores on a series of the tests published by the Cooperative Test Service and marks assigned on a percentage basis during the year 1934-35 in eight independent secondary schools for boys.' Cooperative Tests in the following subjects were included in the study: English, French, Latin, German, elementary algebra, intermediate algebra, plane geometry, solid geometry, trigonometry, biology, chemistry, general science, physics, American history, ancient history, modern European history, English history, world-history, contemporary affairs, and economics. The study showed that the median of 121 correlations was approximately .72 and that all except two of the correlations were statistically significant. The median correlations did not vary greatly from department to department. The median correlation in mathematics was the highest (.76), and the median in social science was the lowest (.68). The medians of the correlations in English, science, and foreign language were approximately equal, since the first two were .71 and the third was .72. Among the separate subjects, the median correlations varied from .90 in intermediate algebra to .58 in biology. These correlations between achievement-test scores and school

The Measurement Business

W HEN new developments in a field of educational policy and practice come to be discussed in the New Tork Times and in some of the popular magazines of the better group, it is a reliable indication that these developments are being considered seriously in numerous educational committees, associations, and journals, and are of interest and significance to the lay public. At the present time this is true of educational testing and educational measurements. For generations we have centered our attention on the mechanical arrangements designed to facilitate the pursuit of education without having developed anything approaching an adequate method of measuring validly and reliably the educational product. We have proceeded with a pious hope or blind faith that, if the educational machinery were designed as well as possible, the educational results would be the best attainable. Recent attempts to measure results more scientifically have been conducted with sufficient success to show that at many points our faith and hope have been misplaced-that results from a given educational set-up are not what we had hoped or taken for granted. Improved methods of testing educational outcomes which have recently been developed, have forced in many quarters a more precise definition of attainable educational objectives and a more critical examination of instructional methods-methods that have persisted for generations merely on presumption. It is my belief that the results of such enterprises of educational testing and measurements as those now being conducted by the Pennsylvania study, the Co-operative Test Service, the College Entrance

The Measurement Business

W HEN new developments in a field of educational policy and practice come to be discussed in the New Tork Times and in some of the popular magazines of the better group, it is a reliable indication that these developments are being considered seriously in numerous educational committees, associations, and journals, and are of interest and significance to the lay public. At the present time this is true of educational testing and educational measurements. For generations we have centered our attention on the mechanical arrangements designed to facilitate the pursuit of education without having developed anything approaching an adequate method of measuring validly and reliably the educational product. We have proceeded with a pious hope or blind faith that, if the educational machinery were designed as well as possible, the educational results would be the best attainable. Recent attempts to measure results more scientifically have been conducted with sufficient success to show that at many points our faith and hope have been misplaced-that results from a given educational set-up are not what we had hoped or taken for granted. Improved methods of testing educational outcomes which have recently been developed, have forced in many quarters a more precise definition of attainable educational objectives and a more critical examination of instructional methods-methods that have persisted for generations merely on presumption. It is my belief that the results of such enterprises of educational testing and measurements as those now being conducted by the Pennsylvania study, the Co-operative Test Service, the College Entrance

The Co-operative Test Service

The Co-operative Test Service

The Co-operative Test Service

The Co-operative Test Service

The Cooperative Test Service

The Cooperative Test Service