An experimental study in the psychology of reading

Front matter

4 o

^O^- rS

0^

Table Of Contents

PAGE CHAPTER

I. Introduction ^

II. Historical Sketch 4

Chapter I Introduction'

A recent investigation has shown that more than one-fourth of the time in the elementary schools of our cities is devoted to the teaching of reading (i)^. This proportion acquires material significance when viewed in the light of the fact that the teaching of reading has until recently been proverbially inefficient, especially in the case of the intermediate and grammar grades. The technique of teaching the mechanics of reading has of course undergone an almost complete revolution since 1870. The alphabet method in use for ages has been gradually discarded, while the phonic, phonetic, word, and sentence methods were being perfected and blended into a variety of superior combination methods (2). But even here there is still entirely too much clash of opinion and far too Httle scientific certainty. And to say the least, the most efficient method of teaching the mechanics of reading remains yet to be determined through scientific experimentation.

If we turn to the later stages in the process of teaching reading —the stages following upon the period when the mechanics of reading has been mastered — we find that improvement in practice

' Most of the data bearing upon this investigation were secured in the laboratory of the School of Education of the University of Chicago during the school year 1913-14. The writer gratefully acknowledges his indebtedness to Dr. F. N. Freeman for his introduction to the technique and methods of experimentation, for improvements in connection with the apparatus — particularly in so far as these bear upon the modifications which made possible measurement of eye-movement in the vertical plane — for much time in serving as a subject in connection with most preliminary and supplementary experimentation, and for much valuable advice and indispensable criticism in connection with the investigation and the manuscript. The writer is further greatly indebted to Dr. Charles H. Judd for suggesting the problem and for most valuable council and criticism. Grateful acknowledgments are also due to Principal F. W. Johnson, of the University High School, and to Principal H. O. Gillet, of the University Elementary School, for kindly co-operation in securing children, as well as to the many individuals who served as subjects in the course of the investigation.

2 Figures in parentheses refer to corresponding numbers in the bibliography given at the end of each chapter.

2 STUDY IN THE PSY GEOLOGY OF READING

has been much slower. In fact, training in this most important form of human behavior has until recently been guided almost exclusively by traditional and empirical principles. In consequence the results have been very unsatisfactory. Emphases have not infrequently been entirely misplaced, as in the case of silent and oral reading, for instance. In spite of the fact that silent reading is of far greater importance for adult life than oral reading, our schools have stressed the latter to the almost complete exclusion of the former. Similarly, adaptation to individual differences has been rather persistently ignored.

More recently, however, a marked change has been taking place. As a result, the reading process is rapidly being analyzed, and careful studies are being made of its several components, such as rate and comprehension. Individual differences are being determined, reading materials standardized, norms and standards of accomphshment established, the differences between silent and oral reading more definitely set forth, and new practices developed. This new and promising movement is due in part to the influence of the rapidly growing technical psychology of reading, the inception and development of which are traced in the next chapter, and in part it is part and parcel of the present scientific movement in education^ — the movement which is ruthlessly challenging all of our practices, in so far as these rest upon traditional and empirical rather than scientific principles.

Before the most advantageous and economical reconstruction of practice can be effected additional information is needed. To this end the reading process must be further analyzed and the boundaries of the psychology of reading must be materially extended. It is not sufficient, for instance, to know that reading rate is a variable. We must determine the factors which condition such variation; we must find out to what extent speed efl&ciency is dependent upon the number and to what extent upon the duration of fixation pauses ; how it is related to age and accomplishment, and how to efficiency in comprehension ; to what extent it is subject to improvement under judicious practice, and within what limits it may be developed in connection with continuous reading rather than scanning. It is essential, further, that the differences between

INTRODUCTION 3

silent and oral reading be carefully pointed out, and that the possibilities and limitations of each be more completely set forth.

Accordingly, this investigation was undertaken for the purpose of securing such information. An effort has been made to discover the dominant reading characteristics of a large number of individuals varying widely in age and accomplishment. Similarly, the differences between silent and oral reading — especially as these are revealed by variations in the number, duration, and location of fixation pauses and by differences in reading rate— have been singled out and described. Finally, a critical analysis of binocular behavior and adjustment affords information which is of particular interest to the technical psychology of reading.

1. Holmes, H. W. "Time Distribution by Subjects and Grades in Representative Cities," Fourteenth Yearbook of the National Society for the Stiidy of Education, Part I, pp. 21-27.

2. Huey, E. B. The Psychology and Pedagogy of Reading, chaps, xiii, xiv.

Chapter II

The scientific study of the reading process is of comparatively recent origin. The long neglect of this important field is, as Huey remarks, "a, curious instance of the failure of scientists to make first-hand observation except along certain lines that have become habitual." Once initiated, however, the movement has made considerable progress — so much so, in fact, that we have today an elaborate experimental technique and a wealth of scientific data at our command. The problem has been approached from many different angles, and has as a result given rise to a variety of experimental activities, each making its specific contribution toward the building up of a comprehensive psychology of reading. We are here most immediately concerned with three types of experimental investigation, namely: (A) those approaching the problem primarily through a study of eye-movements; (B) those dealing chiefly with the perceptual process; (C) those attacking the problem from the standpoint of individual differences in speed, comprehension, and reproduction.

A. EXPERIMENTAL INVESTIGATIONS OF EYE-MOVEMENTS

I. EVOLUTION OF EXPERIMENTAL TECHNIQUE FOR THE STUDY OF EYE-MOVEMENTS

The present complicated technique for the study of eye-movements has a rather interesting history. Although earlier attempts to attack the problem by means of mechanical devices are not wanting, yet up to 1898, when Huey succeeded in utihzing mechanical registration, the after-image method and the mirror method of direct observation gave the most satisfactory results. Volkmann (i) and Lamansky (2) used the after-image method during the third quarter of the last century. The speed of the eye's movement "was measured by counting the number of after-images produced during a given movement of a pencil of light flashed into the eye at regular intervals through the perforations in a rotating disk." Professor Javal, of the University of Paris, was among the first to use the mirror method of direct observation. As early as 1879 he called attention to the fact that the passage of the eye across the page from left to right is discontinuous, consisting of a series of movements and pauses (3). Somewhat later (1891) M. Landolt, also of the University of Paris, made extensive observations by means of the mirror method of direct observation (4). Still later (1897-98) Erdmann and Dodge carried on elaborate investigations at the University of Halle (5). Among the methods used was the mirror method of direct observation. A telescope was also used to advantage.

Meanwhile various mechanical devices were attempted. Lamare, working with Javal, succeeded in attaching a microphone to the upper eyelid which enabled him to count the sounds made by the movements of the eyes and thus to determine in a general way the number of pauses which the eye made in reading a given line (6). Somewhat later (1891) Dr. Ahrens, in studying eye-movements in connection with writing at the University of Rostock, attached an ivory cup to the cornea of the eye. By means of a pointer fastened to the cup he hoped to record on a smoked surface tracings of the movements of the eyes. The attempt, although not successful, was nevertheless suggestive (7). During the year 1897-98 Professor Delabarre, of the Harvard laboratory, resorted, at the suggestion of Dr. Lough, to the use of a plaster-of-Paris cup. This was attached to the cornea of the eye in somewhat similar fashion to that used with Dr. Ahrens' ivory cup. A light thread leading to a recording lever was fastened to the cup. To this lever was attached a thin elastic fiber in such a way that the horizontal movements of the eyes were recorded on the smoked surface of a kymograph cyHnder. However, no permanent records of eye-movements were made (8) . In the course of the same year Huey, at Clark University, devised, partly at the suggestion of Delabarre, a similar although more perfect apparatus (9). This marked a great step in advance. But, as Dodge points out, the apparatus was not entirely satisfactory. The eye necessarily worked under more or less unnatural conditions because of the

CHAPTER II HISTORICAL SKETCH

The scientific study of the reading process is of comparatively recent origin. The long neglect of this important field is, as Huey remarks, "a, curious instance of the failure of scientists to make first-hand observation except along certain lines that have become habitual." Once initiated, however, the movement has made considerable progress — so much so, in fact, that we have today an elaborate experimental technique and a wealth of scientific data at our command. The problem has been approached from many different angles, and has as a result given rise to a variety of experimental activities, each making its specific contribution toward the building up of a comprehensive psychology of reading. We are here most immediately concerned with three tj^es of experimental investigation, namely: (A) those approaching the problem primarily through a study of eye-movements; (B) those dealing chiefly with the perceptual process; (C) those attacking the problem from the standpoint of individual differences in speed, comprehension, and reproduction.

A. EXPERIMENTAL INVESTIGATIONS OF EYE-MOVEMENTS

I. EVOLUTION OF EXPERIMENTAL TECHNIQUE FOR THE STUDY OF EYE-MOVEMENTS

The present complicated technique for the study of eye-movements has a rather interesting history. Although earlier attempts to attack the problem by means of mechanical devices are not wanting, yet up to 1898, when Huey succeeded in utilizing mechanical registration, the after-image method and the mirror method of direct observation gave the most satisfactory results. Volkmann (i) and Lamansky (2) used the after-image method during the third quarter of the last century. The speed of the eye's movement "was measured by counting the number of after-images produced during a given movement of a pencil of light flashed into the eye at regular intervals through the perforations in a rotating disk." Professor Javal, of the University of Paris, was among the first to use the mirror method of direct observation. As early as 1879 he called attention to the fact that the passage of the eye across the page from left to right is discontinuous, consisting of a series of movements and pauses (3). Somewhat later (1891) M. Landolt, also of the University of Paris, made extensive observations by means of the mirror method of direct observation (4). Still later (1897-98) Erdmann and Dodge carried on elaborate investigations at the University of Halle (5) . Among the methods used was the mirror method of direct observation. A telescope was also used to advantage.

Meanwhile various mechanical devices were attempted. Lamare, working with Javal, succeeded in attaching a microphone to the upper eyelid which enabled him to count the sounds made by the movements of the eyes and thus to determine in a general way the number of pauses which the eye made in reading a given line (6). Somewhat later (1891) Dr. Ahrens, in studying eye-movements in connection with writing at the University of Rostock, attached an ivory cup to the cornea of the eye. By means of a pointer fastened to the cup he hoped to record on a smoked surface tracings of the movements of the eyes. The attempt, although not successful, was nevertheless suggestive (7). During the year 1897-98 Professor Delabarre, of the Harvard laboratory, resorted, at the suggestion of Dr. Lough, to the use of a plaster-of-Paris cup. This was attached to the cornea of the eye in somewhat similar fashion to that used with Dr. Ahrens' ivory cup. A Hght thread leading to a recording lever was fastened to the cup. To this lever was attached a thin elastic fiber in such a way that the horizontal movements of the eyes were recorded on the smoked surface of a kymograph cylinder. However, no permanent records of eye-movements were made (8) . In the course of the same year Huey, at Clark University, devised, partly at the suggestion of Delabarre, a similar although more perfect apparatus (9). This marked a great step in advance. But, as Dodge points out, the apparatus was not entirely satisfactory. The eye necessarily worked under more or less unnatural conditions because of the attachment. Moreover, the exactions of the technique were too severe, and the penalties of carelessness too great, to bring the apparatus into general use.

Meanwhile Dodge was at work at Wesleyan University devising a new apparatus. A satisfactory apparatus for recording eye-movements, he held, must necessarily meet the following requirements:

It must be capable of operating under normal conditions of binocular vision. It must be capable of registering both eyes simultaneously. The vmit of measurement must be i cr or less. The registering medium may have neither momentum nor inertia, whUe the eye must perform no extra work during registration and be subjected to no unusual conditions. The apparatus should be such as can be used to record the movements of a large number of eyes without serious inconvenience either during or after the experiments (lo).

It appeared to him that photography must ultimately offer the most satisfactory mode of approach. He felt convinced, too, that a pencil of Hght which had ''neither momentum nor inertia, which was absolutely safe and universally available," must constitute the final registering medium. But the development of a technique which would make this possible was no small task. The immediate outcome was an apparatus which enabled Dodge to photograph directly the horizontal movements of the eyes. This method was, however, open to serious objections. "The illumination tended to interfere with the normal eye-movement, the lines of demarkation between pigmented and unpigmented portions of the eye were not sharply defined, and the records could not be satisfactorily enlarged" (ii).

After considerable additional experimentation Dodge finally "hit upon the plan of utilizing the eccentric surface of the cornea as a reflector. Instead of photographing the eye directly," he now "photographed the movement of a bright vertical line as it was reflected from the surface of the cornea." The source of the corneal illumination was an arc lamp. The Hght was first passed through several thicknesses of pot-blue glass, which rendered it highly actinic and of low physiological intensity. Head-movement was eUminated as much as possible by fastening the head securely in a headrest, the latter consisting of an upper-jaw rest, a wooden nose-bridge, and a mastoid-rest. The pencil of light reflected from the surface of the cornea was focused upon a photograpJiic plate by means of a lens and camera. The plate, which was movable in the vertical plane, was secured within a carefully devised box so that its fall was absolutely constant and all lateral motion was excluded. The time of the eye's movement was recorded upon the falHng plate by means of a spring pendulum which vibrated directly behind and at the edge of a horizontal slit in the camera, alternately admitting and excluding the hght, and leaving a definite time record on the photographic plate (12).

Naturally the question of the reliability of the records of movements secured on the basis of corneal reflection presented itself. This question Dodge discusses at length, showing that limitations appear only in cases where reflection involves the use of the extreme and peripheral portions of the cornea. Records of reflection from the middle third of the cornea, to which clear vision is limited, are shown to correspond accurately to actual eye-movements (13).

Dearborn, who worked with this apparatus at Columbia University during the year 1904-5, introduced numerous minor modifications, especially by way of perfecting various parts of the technique. He was the first to use the apparatus extensively in experimenting with reading (14). Later on he set up the apparatus at the University of Wisconsin, and still later in the laboratory of the School of Education of the University of Chicago. While connected with the above-named institutions. Dearborn introduced further modifications. By the use of double lenses he was able to secure records of binocular reading. He also experimented successfully with the use of films, the latter being partly suggested by the kinetoscopic studies at Yale.

After Dearborn left the University of Chicago, Dr. F. N. Freeman became interested in the apparatus. He accordingly introduced various minor modifications, perfecting especially the use of the film device. It was at his suggestion, too, that in the course of the present investigation the film device v/as modified so as to make possible the registration of movements in the vertical plane.

While the Dodge photographic method was being used and perfected, another and in some respects superior method for the study of eye-movements was devised at Yale by Dr. Charles H. Judd, namely, the kinetoscopic photographic method. The Dodge apparatus had certain Hmitations; eye-movement, for instance, could be measured in only one plane, the horizontal. Dr. Judd overcame this difficulty by using a kinetoscopic camera. This made possible a rapid succession of discrete photographs — eight per second — the records showing the movements of the eye in all planes. Usually a double camera was used so as to get complementary records. The camera was driven by a mechanical device which assured absolutely uniform motion.

In order to get a fixed and identifiable spot on the eyeball — a spot which could be used as a point of reference in the successive photographs — a small particle of Chinese white was placed on the cornea. The photographs, when completed, were projected upon a screen by means of a lantern. By noting the successive positions of the white spot it was possible to measure and chart the varied movements of the eye.

Head-movement was eliminated as much as possible by securing the head in a firm headrest. Iron spectacles with highly polished steel beads fastened to the lower rim were worn by the subject; this made possible a comparison of eye- and head-movement. In order to eliminate error which might result from compensatory eye-movement, the points of reference or fixation were attached to the head, so that they invariably moved with the head. The possibihties of error were thus reduced to a minimum (15).

Since the Yale apparatus made possible the estimation and measurement of movement in the various planes, it was particularly well adapted for a minute study of fixations. McAllister's investigations accordingly dealt in a very thoroughgoing manner with this problem (16). The studies of illusions which were undertaken by Judd, Cameron, Steele, and Courten dealt indirectly with the same problem (17). Judd's final study dealing with convergence and divergence makes important contributions to various phases of the eye-movement problem (18).

2. RESULTS OF EXPERIMENTATION WITH EYE-MOVEMENTS

As previously indicated, up to the time that Huey succeeded in utilizing mechanical registration of eye-movements, the after-image method and the mirror method of direct observation gave the most satisfactory results. These results were, however, quite meager, since the methods were severely limited in several respects. As Dodge says:

The after-image method was limited to the span of memory, while any after-image sufficiently distinct to be serviceable would be a constant menace to clear vision, and the demands on the attention would more or less disturb

the processes it was serving to investigate Similarly, the invaluable

method of direct observation is limited by the memory span of the observer, and is irregularly interrupted by the pursuit movements of the observing eye (19).

Volkmann and Lamansky were particularly concerned with the estimation of the speed of eye-movement. Their estimates, however, were too low (20). Javal's discovery that eye-movement in reading is discontinuous, consisting of a series of alternate movements and pauses, was of course thoroughly significant. In fact, the chief interest of subsequent experimenters has centered about the nature, function, and relation of these movements and pauses.

a) Path of the eye's fixation point. — Erdmann and Dodge were the first to note that the path of the eye's fixation point was not coextensive with the line. The first and last fixations were found to be located within the line— the last more so than the first (21). Huey and Dearborn later confirmed this. The former found that the distance traversed by the eye's fix:ation point varied from 78 to 82 per cent of the length of the fine (22). The latter observed that these indentations varied more or less with individuals and with the nature of end-words (23).

Javal, upon observing on the basis of after-images the prominent part played by the upper part of letters, concluded that the fiLxation point moved across the fine evenly between the middle and the top of the small letters. Huey, with a more perfect technique, noticed that the path of the fixation point was more uneven than Javal had supposed, but concluded that the fixation lo STUDY IN THE PSYCHOLOGY OF READING

point did not wander far above or below the line that was being read (24). Dearborn, detecting on his photographic records indications of considerable movement in the vertical plane, concluded that the path designated by Javal was a "physical impossibility" (25). However, both admit the limitations of their technique. Neither apparatus was capable of registering accurately eye-movements in the vertical plane. The results of the present investigation, the technique of which made such measurement possible, will be referred to later.

b) Interfixation movements (26). — Interfixation movements, i.e., the movements of the eye from fixation to fixation, have been studied chiefly from the standpoint of extent and velocity. Javal concluded on the basis of his observations that the extent of these movements was rather uniform. Recent findings, however, have disproved this. Huey and Dodge found that the extent of these forward movements often varied in the ratio of i : 4 (2 7) . Dearborn found the same variation in extent, even though the formation of short-lived motor habits tended to render the number of pauses relatively uniform line after line (28).

The interest of investigators has, however, been centered largely upon the problem of determining the velocity of these movements. The chief question at issue was whether the velocity was such as to condition clear vision. Volkmann and Lamansky, as noted above, were interested in determining the angle velocity of eye-movement, but underestimated it (29). Javal, knowing that the passage of the eye from left to right is discontinuous, attempted by means of after-images to determine the velocity of interfixation movements. Although hampered by the inadequacy of his technique, he concluded that the movement was too rapid to condition clear vision (30). Erdmann and Dodge inferred on the basis of certain calculations that from 12/13 to 23/24 of the time of the eye's passage from left to right was spent in fixations. This velocity was considered too high to admit of significant stimulation (31). Further experimentation on the part of Dodge and Baxt showed that fusion of stimuli was ine\dtable during interfixation movements (32). Actual measurement of the velocity of these movements by Huey, Dodge and Cline, and Dearborn

HISTORICAL SKETCH ii

gave results which were in general agreement with previous findings izz)' There remained, however, one apparent difficulty. In spite of the fact that the conditions were such as to imply fusion, no gray bands were in evidence. This raised the question whether the conditions of fusion might not be different when the eyes, rather than the stimuli, were in motion. If so, perception and orientation might be quite possible during interfixation movements. Experimental evidence in support of this was not entirely lacking (34). However, Dodge presently secured experimental evidence which demonstrated clearly that fusion takes place during these movements (35). Holt likewise concluded on the strength of two experimental proofs that "voluntary movements of the eyes condition a momentary visual central anaesthesia" (36). About the same time Dearborn repeated the calculations of Erdmann and Dodge. He found them substantially correct, and he concluded, therefore, that "the duration of stimulation is well below the threshold necessary for producing a distinct visual impression" (37).

c) The fixation pause. — Significant stimulation is, as we have seen, hmited to the fixation pause. This naturally makes the pause the center of experimental interest. It has been studied thus far chiefly from the standpoint of number per line, duration, location, and nature. Javal inferred that there was uniformly one pause to every ten letters (about 5 pauses to a 90 mm. line), Landolt held that on an average 1.55 words were read per pause (about 6.5 pauses to a 90 mm. line). Dodge averaged 5 pauses for an 83 mm. line, and Erdmann 8 for a line of 122 mm. Two of Huey's subjects averaged 4.5 and 4.8 pauses, respectively, for lines 83 mm. in length. With 52 mm. newspaper lines two readers averaged, respectively, 3.4 and 3.8 pauses per fine. Dearborn's readers averaged from 3 to 7 . i pauses for somewhat longer newspaper fines, an average of from 1.9 to i . o words per pause. For lines not quite double this length from 7 . 5 to 9 . 4 pauses were made, averaging from 1.5 to i . 09 words per fixation. Dearborn is the first to stress strongly the fact that great individual variation exists.

The duration of pauses is quite as important a factor as their number. However, since the measurement of the former is a

12 STUDY IN TEE PSYCHOLOGY OF READING

more delicate task than the estimation of the latter, it is naturally more dependent upon an improved technique. As a result, the data regarding the duration of pauses are very recent. The average for several of Huey's ^readers appeared to be in the neighborhood of 185 (T per pause. However, his technique was not very well adapted for the measurement of duration. Dearborn was the first to measure duration extensively. His apparatus was such as to insure much greater accuracy. He found in reading the same newspaper passage that five subjects ranged from 160. 8 cr (average deviation 36.8) to 401.9 a (average deviation 163.7). Dearborn found, further, that these differences in duration time were very closely correlated with differences in the rapidity of reading.

Dodge has also devoted considerable attention to the investigation of the duration of pauses, especially as related to reaction time and to the total reading process (38). Although his reaction-time averages fall considerably below the averages for pauses, he feels, nevertheless, that some explanation is needed in the case of the large number of pauses the duration of which either approximates reaction time closely or falls below it. It would be absurd, he states, to attempt to crowd into such brief intervals "the complex apperceptive processes which condition the apprehension of the words as such, their contribution to the sum-total meaning consciousness, and the more or less definite mental and physical reactions to that meaning." ''We are forced," he continues, "to postulate a concurrent complication of the psychical processes of perception, extending through several fixations." He suspected, and later demonstrated experimentally, that the long initial pause, which Dearborn had found to be rather common, made possible a general initial survey of the line, and thus modified the duration of succeeding pauses. This, together with constant data from peripheral vision, he concludes, "tends to make the normal reading pause represent a comparatively late moment in the total process of perception."

The fixation pause has also been studied intensively from the standpoint of its location. The chief problem has been to determine with accuracy the location of pauses with respect to syllables, words, phrases, etc., and to arrive, if possible, at a law governing such location. Erdmann and Dodge's earlier experimentation led them to believe that fixations were centered upon words; in fact, usually upon the middle of words. Later and more accurate measurements by both Dodge and Huey seemed to indicate that the fixation point may be in any part of the word, or even between words. Dearborn, who made the most elaborate study of the location of fixation pauses, comes to similar conclusions. ''The exact point that is fixated," he states, "may be in any part of the words, nor does it occur more frequently in the first part of the sentence than in the last, and apparently pays little attention to many of the laws of apperception or the rules of the rhetorician." "The exact object of fixation," he continues, "is significant only as representing the point about which are grouped the 'block' of letters that are simultaneously perceived as one word or phrase complex. It more often falls in the first third than at the center of a given perception area." There were, however, in evidence certain factors which tended to determine the extent of the successive areas, and consequently, in a general way at least, the location of pauses. The perception area was found to be "large in the case of nouns and adjectives and verbs, and usually small in the case of the connective parts of speech, the conjunctions and prepositional phrases, the relative pronouns, the auxiliary verbs" (39). Certain other factors which will be mentioned later on were found to be influential at times in determining the more immediate location of the pauses.