Empirical Studies in the Theory of Measurement

From the INTRODUCTION. ALMOST every student is familiar with the recent rapid growth of knowledge in the psychology and physiology of reading. It is a rather common remark, however, that this growth is not equally distributed between the two phases of the problem, progress on the strictly psychological side having scarcely kept pace with that on the side of physiology. From the very nature of the case one could hardly expect it to be otherwise, since so many of the psychological problems depend for their solution on questions of physiology, and since methods of investigating problems of a purely psychological character, in so far as they are not identical with those of physiology, are naturally less well developed because of the peculiar complexity of many of these problems. It is not surprising, therefore, that the criticism should be made that as yet many of our psychological theories of reading rest on a much too slender basis of fact and that the problems demand further studies of a more specialized character, employing methods calculated to yield a larger amount of objective data which may be rigidly dealt with in accordance with recognized statistical methods. However much students of the problem may differ in regard to the value of current psychological theories, probably all will agree that the time is opportune for such specialized studies. At any rate it was in this belief that the present study was begun. It was planned accordingly to limit the scope of the investigation to a single problem, and to rely as far as possible upon statistical methods for results. The particular problem chosen is the factors and processes involved in word perception as it takes place in adults in normal visual reading. As anticipated, however, this has turned out to be by no means a simple problem; in fact, it has proved to be much too comprehensive for a satisfactory treatment of its many phases within the brief scope of a study of this character. Attention has therefore been confined almost exclusively to those phases which seemed to lend themselves most readily to the statistical method and which thus gave promise of most immediate returns. This mode of procedure, to be sure, may tend to a certain incoherency and incompleteness of results but, in addition to the practical advantage mentioned, it should have a tendency to free the investigator from theoretical bias as far as that is possible. On the whole, the results obtained are in fairly close agreement with those of certain other studies. As might be expected, however, this is not true in all cases, and the writer has thus been led to make some independent generalizations, which, are offered merely as tentative, and may be of value only as hypotheses for further investigation.