aural mechanisms in their daily lives, still the auditory presentation of the material may have offered an unfair handicap to one of the groups. On this point we have no check. As arguments, however, for the merit of the instrument used by us to measure intelligence let us say, first, that among the 100 normal subjects examined, the university students (mostly sophomores) outstripped the high-school students (mostly juniors) by an average of 20 points, and, secondly, that the correlation between the scores made on our adaptation of the Army Alpha Test by the high-school students (normal group) who served in the maze experiment and the following criteria of intelligence: (a) score on the Miller Mental Ability Test, (b) the instructor's rating of intelligence, and (c) final grades in chemistry were, respectively, +.88 ± .04, + .80 ± .06, and + .44 ± .II. The test given included all or parts of the following six sub-tests of the Army Alpha (Form 9): the arithmetic, opposites, analogies, number-completion, information, and best-answers. All of the best-answers, opposites, and analogies tests were employed, as were also 16 of the arithmetic problems, 38 of the information, and 15 of the numbercompletion. Those tests given in part were not completed because their more difficult sections were so far beyond the capabilities of the blind subjects that a recital of them had the effect of completely discouraging the experimental group and of ruining its morale. Of the 100 university and high-school students who took the test and from among whom our subjects were selected, no one amassed more than 142 points. For the individuals whose performances are reported in this study 118 is the maximal score. The test was given rather informally to the blind, but still with adherence to most of the details of procedure usually observed. The problems were generally read aloud twice; and, if the group indicated a desire for a third reading, this also was given. Only when the blind subjects showed by their attitude that they had either completed an attempted solution or were unable to solve the problem was the next task presented. This method of administration makes of the test an achievement rather than a speed test. The test as conducted with the blind subjects furnished the standard of procedure for the normal group in regard to such details as instructions, time allotted for each individual problem, etc. The blind subjects used the Braille apparatus or the typewriter to record their answers, which in each case consisted of a response no more complex than a word, and usually merely a letter or number. The normal subjects wrote their answers with a pencil. How significant an effect this difference in routine may have had we do not know with certainty, though we question, since speed played so small a part in the performance, whether the variation is of any great moment. The blind group was handled by one experimenter; the normal, by the other. The authors attempted to prevent the personal equation from being introduced as a prominent factor into the results by carefully working over together the details of the administration of both of the tests. The maze used was a metal stylus maze designed by Dr. Koch, (see Fig. 1) and the general method of conducting the 5 This maze has two noteworthy characteristics: (1) the ease with which the distance traversed in it by the learner can be estimated; and (2) the multiplicity of different patterns which it affords. The distance covered can be estimated sufficiently accurately for group comparisons, at least, by merely counting the number of blocks passed by the learner en route to the goal and multiplying this count by the length of one block, as all the alley-ways are divided into segments of equal length. 6 maze experiment was the same as that described in her monograph, "The influence of mechanical guidance upon maze learning." The canopy which screened the maze from the view of the normal subjects was employed in the case of the blind subjects also, in order to keep the tactualkinæsthetic situation as nearly identical as possible for the two groups. Time was recorded with, a stop watch. A standard of mastery of three successive perfect trials was arbitrarily adopted. Tables III. and IV. show the maze performance of the blind (as measured by the three criteria, time, trials, and TABLE III AGES, INTELLIGENCE-TEST SCORES, AND MEASURES OF MAZE PERFORMANCE. possible by the employment of The dots in Fig. 1 indicate the A very great variety of maze patterns is made metal pegs at the junctions of the various run-ways. holes in the floor of the maze into any one or group of which the pegs may be inserted. Since the pegs fit snugly and are always set in flush with the walls of the run-ways the subject is unlikely to be disturbed by any roughness or break in a wall which the experimenter desires to be continuous. The positions of the pegs in the maze pattern employed in the study are indicated in the diagram by cross-hatching. Psychol. Monog., 1923, 32, no. 147. Pp. vi + 113. TABLE IV AGES, INTELLIGENCE-TEST SCORES, AND MEASURES OF MAZE PERFORMANCE. SUBJECTS WITH NORMAL VISION errors) to be inferior to that of the normal subjects, the two groups differing most in respect to their error scores. This relation was likewise observed in our earlier work. None of the group-differences, to be sure, are statistically significant, if a ratio of 3 or more between the difference of the means and the probable error of this difference is accepted as an index of reliability. We may question, however, the use of this index with a few cases. The consistency of the results and their close agreement with our earlier observations merit confidence in their representativeness of what would occur were many more subjects employed. One may press the question further and note too that, while the blind are almost of the same average age as those with normal vision, they have a very slightly lower intelligence average. The medians, on the other hand, show a reversal of these conditions. A later and more analytical presentation of the data (see Table VII.), moreover, suggests that these small differences in age and intelligence do not destroy the validity of the tendencies shown by the criteria of skill in maze learning. |