E. F. Lindquist (1901–1978)

One of the foremost applied statisticians and educational testing pioneers of the 1900s, Everet Franklin Lindquist received his Ph.D. from the University of Iowa in 1927 and was a member of the faculty there from 1927 until his retirement in 1969. During his long professional career, he made substantial contributions to the field of education in the areas of test development, test-scoring technology, measurement theory, and research methodology. His textbooks on design of experiments and statistical analysis had considerable influence on educational and psychological research.

Test Development

As director of Iowa Testing Programs at the University of Iowa, Lindquist was responsible for the development of the first editions of both the Iowa Tests of Basic Skills (ITBS) and the Iowa Tests of Educational Development (ITED). In the 1940s he served as an advisor to the United States Armed Forces Institute and to the American Council on Education (ACE). He played a major role in formulating policies with respect to granting academic credit for general educational growth during military service. These policies resulted in the development of the General Educational Development Test (GED). Lindquist oversaw the creation of the initial forms of the GED, which were modeled after the ITED. In the late 1950s Lindquist also was responsible for the design and construction of the first forms of the National Merit Scholarship Qualifying Test (NMSQT). In 1959 Lindquist and Ted McCarrel, registrar at the University of Iowa, cofounded the American College Testing Program (ACT) as an alternative to the College Entrance Examination Board. Lindquist was personally responsible for the design and development of the early editions of the ACT tests.

All of these tests continue to be published and used in the early twenty-first century. Although the tests have evolved over the years, it is a tribute to Lindquist's vision and creativity that he recognized the need for such tests and that the philosophy underlying them continues to be reflected in recent editions.

Test-Scoring Technology

In the late 1940s the Iowa tests developed under Lindquist's direction were being administered to large numbers of students in Iowa and throughout the nation. The demand to have the answer sheets of these tests scored at a central agency was growing. It became apparent to Lindquist that this demand would require a more efficient way of processing the answer sheets than the hand-scoring procedures then being used. Although crude scoring machines were available, they were not adequate for large-scale testing. Thus, in the early 1950s Lindquist undertook a major project to design a high-speed electronic scoring machine. He and his colleagues at the University of Iowa succeeded in building such a machine in the mid-1950s. The original version scored at the rate of 4,000 sheets per hour. They developed faster machines fairly quickly, and by 1970 the scoring machine could scan 40,000 sheets per hour. Without question, the availability of such scoring machines had a significant impact on testing practices in the United States.

Measurement Theory

Lindquist's major contributions to the field of measurement theory occurred primarily through two publications. The first of these was the influential volume Educational Measurement (1951), for which he served as general editor. Eighteen of the most distinguished measurement experts, including Lindquist himself, wrote chapters for this landmark publication. It served as the major reference for the educational measurement community until the second edition was published twenty years later.

Lindquist's second major contribution, and perhaps his most technical work in measurement, was published as the last chapter in his textbook Design and Analysis of Experiments in Psychology and Education (1953). In this chapter, "Estimation of Variance Components in Reliability Studies," Lindquist explicated techniques, which were not new in the statistical sense. But their relevance to educational and psychological measurement had not been appreciated. Lindquist saw these tools as a significant improvement over traditional methods of estimating measurement error variance and hit-or-miss methods of designing efficient, cost-effective measures. In essence, he provided the foundations for an area known as generalizability theory ten years before the first publications of Lee Cronbach et al. on this approach appeared in the literature of educational and psychological measurement.

Research Methodology

Lindquist's influence in the field of educational research occurred primarily through his textbooks. Early in his career, he concluded that a textbook constituted a more valuable vehicle than journal articles or conference presentations in reaching the educational research community. Thus many of his innovative ideas and original insights were published in his books. In Statistical Analysis in Educational Re-search (1940), he assumes the role of statistical translator for the educational research community. He was concerned about the slow adoption of the techniques of analysis of variance and covariance, including factorial designs, in the behavioral sciences. The implications of the work of Ronald A. Fisher had made little impression on educational and psychological researchers. Lindquist perceived this failure to be the result of a widespread inability of research workers to read the relevant literature and to translate terms such as blocks, plots, treatments, and yields into schools, classes or pupils, educational methods, and test scores. In this 1940 text, he undertook to facilitate this translation.

In a second influential book, published in 1953, Design and Analysis of Experiments in Psychology and Education, Lindquist again had the goal of translating statistical analyses developed in other fields into language understandable to researchers in psychology and education. However, he soon recognized that the adaptation of certain designs developed for use in industry and agriculture was by no means straightforward. The educational and psychological researcher's alternatives with respect to the choice of the unit of analysis often had no obvious analog in the original industrial and agricultural contexts. The phenomenon of repeated measurements on the experimental units took on a unique character in the behavioral sciences. Certain designs, such as Latin square designs, were seen to be potentially useful in the behavioral sciences. But their purposes and implementation differed significantly from the classical applications. The entire area of multidimensional, repeated-measurements designs had to be reworked. Analyses appropriate to factorial experiments that involve independent groups on some dimensions and repeated measurements on others had to be formulated almost without helpful precedent from other sciences. Thus his 1953 book represented far more than a statistical translation. It incorporated a number of original insights and formulations, particularly in the area of "mixed" designs involving both independent and dependent measurements. In sum, Lindquist's ability to bridge the statistical gap between education and other areas enabled him to exert a significant influence on educational research from the 1940s through the 1970s.