Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
⢠K k. b '1 R 11 T L J3 8 their work might save some future investigator the trouble of pursuing a promising line of re- search, to find that it is a cul de sac at the end of which the Union Jack has already been planted and waves in futility. I shall, therefore, take the liberty of referring to some of their material in the following comments. With a perspective given by five years away from the topic, one can observe some general characteristics of the research performed on night vision selection during the recent war. The most'conspicuous characteristic is the devotion of almost all who worked in the area to an equivocation. If we worked at levels of illumination below a certain value, we were working with "night vision," (or synonymously "night visual performance"), which was usually treated as a unitary and measurable property of the individual. Some people, we expected, should have "good night vision," and others "poor night vision." And our job was to find out "how much," or "how good" "night vision" various individuals had. By and large, our efforts were limited to a search for the best "test" of this "capacity." Some of us, at intervals, or should one say, at lucid intervals, would question this assumption of a unitary capacity, and point out the obvious fallacies involved. But such intervals were brief, the search for a test was resumed, and the experimental analysis of visual behavior which should have been made, never appeared. Back we went to the development of "a test of night vision," which would have been accepted as satisfactory, if only it gave statistically reliable results. If a reliable test was obtained, validation tests were planned, but there were questions as to the necessity of such a step. A second characteristic of the efforts was the variety of purposes with which different lab- oratories concerned themselves. Some sought the quickest and easiest method of eliminating the "night-blind". Others sought the most reliable test for grading the whole population, from "best" to "worst", without regard to cost, in personnel, time, or money, of the procedures of selection. Some tests were designed for use by field-hands in the woods on a reasonably dark night. Others required administration by the test designer himselfâno one else would, or could, for that matter, do. Each felt free to investigate someone else's test, which was designed for one purpose, with respect to its usefulness for his own quite different purposes, and, finding that it did not meet his own criteria, concluded that it was. worthless for any purpose whatsoever. Such criticism worked both ways; advocates of simplicity damned the complex tests with the same conviction that the ap- paratus-specialists displayed in describing the shortcomings of the simple devices. And in the cross-current of critical comparisons, it was not always clear that the tests were administered according to the procedures for which they were designed. With variety in purpose went variety in design,âa variety just great enough to produce a max- imum of confusion of results with a simultaneous minimum possibility of turning up a test suffi- ciently different from the rest to offer a fresh approach to the problem. Almost all the tests in- corporated an acuity figure, a Snellen T or a Landoldt ring. But the sizes employed varied. Sources of illumination and brightness levels varied. Viewing distances varied. Number of trials, and other details of procedure varied. The effort was not characterized by a multiplicity of approach, but of detail. Differences among tests there were, great enough to make compari- son difficult, but small enough to ensure that many visual functions went uninvestigated. Just as the purposes and methods of the various tests differed, so too did the types of analyses performed on the data. Split-half reliabilities and test-retest reliability studies were performed; *r', rho, tetrachoric "r's", eta's, epsilons, and Chi Squares were all computed in the effort to de- termine whether a particular test gave consistent results. Lack of homogeneity of reliability data, does not provide a satisfactory basis for the comparison of tests with one another. That it was often forced by differing test methods does not help any. It is, then, not surprising that the concrete results of the various research programs add up to so little for so much effort, and that the conclusions are conflicting. Some found learning highly important in the test situation, and urged repeated retests before disqualification of an in- dividual for one or another duty because of "Defective Night Vision" was recommended; others found tests sufficiently reliable to give useable results with one administration. One test was considered just as highly acceptable by some as others considered it worthless. Yet there is a sort of homogeneity in the results obtained with various tests. General conclusions which can stand scrutiny may be drawn. One generalization is, unfortunately, that tests of night vision are not very reliable. Indeed, one looks with extreme doubt on data showing reliability coefficients KK U T R ITTT-fi-eâ
