151
— _ —— — ■■-■■—■ mmmimm mm mm mmm ■—
KTNESTHETTC PERCEPTION IN
BLIND ADULTS
S. P. Lindley, 1969
Published on demand by
wMmim
m hHHHHH
limmB ■-■'■'•■ .■■■' SB -'■■ ■
■■ ; H ■■■:■ , ,'■-;;.;
■HI
■■■ill
■•■■•■.;:'■'''■.■■''■■■■■■
','■■■■'.•'''-'.•-■■•■■■
ksSPIlF
■
fl!
''■;/■
8KHHH1iwjBH
Print
152
Lindley
Kinesthetic perception in blind
adults
This is an authorized facsimile and was produced by
microfilm-xerography in 1972 by University Microfilms,
A Xerox Company, Ann Arbor, Michigan, U.S.A.
70-25,980
LINDLEY, Sondra Patterson, 1939-
KINESTHETIC PERCEPTION IN BLIND ADULTS.
Case Western Reserve University, Ph.D., 1969
Psychology, clinical
University Microfilms, A XEROX Company , Ann Arbor. Michigan
© Copyright by
SONDRA PATTERSON LINDLEY
1971
KINESTHETIC PERCEPTION IN BLIND ADULTS
by
SONDRA PATTERSON LINDLEY.
Submitted in partial fulfillment of the requirements
for the Degree of Doctor of Philosophy
Department of Psychology
CASE WESTERN RESERVE UNIVERSITY
September 1969
Digitized by the Internet Archive
in 2011 with funding from
National Federation of the Blind (NFB)
http://www.archive.org/details/kinestheticperceOOIind
/
CASE WESTERN RESERVE UNIVERSITY
GRADUATE STUDIES
We hereby approve the thesis of
SONDRA PATTERSON LI1TUL5Y
candidate for the DOCTCR OF PHILOSOPHY
degree .
Signed
"3
Date S"(tj/u<\ .
KINESTHETIC PERCEPTION 321 BLIND ADULTS
by Sondra Patterson Lindley
An Abstract
The question of whether the standard against which body move-
ment and position are measured is visual was first raised by Head
(1920). In the present study, the relationship between vision and
perception of limb position and movement was approached by studying
kinesthetic perception in blind ac'iO.td who had had light perception
or less since early childhood. The blind subjects were compared to
an equal number of sighted adults whose kinesthetic perception was
measured while they were blindfolded. As the sighted group could use
visual imagery in the performance of the experimental task and the
blind group could not, inferences could be drawn about the role of
vision in the body schema, the mechanism by which body sensations
are perceived.
Subjects were required to reproduce waist-high arm movements
introduced by the examiner who moved the subject's arm along a
standard series of trajectories in a two-dimensional field. Subjects
were scored for accuracy both in reproducing the length of the move-
ment and in locating the target or end-point of each trajectory.
Three factors were analyzed in the study. The factor of
primary concern was vision and its effect on perception of movement
and position. Secondary factors examined were the effects on per-
ception of practice and of altering input by adding a constant weight.
ii
Results of the study indicated that the standard for measur-
ing movement and position is not necessarily visual, that learning
does not occur within one session without feedback, and that adding
a constant weight does not alter accuracy of perception. The first
conclusion was based on the findings of no difference between sighted
and blind groups in accuracy of reproducing position and of signifi-
cantly greater accuracy of blind subjects in reproducing movement
length. The conclusion that learning does not occur resulted from
the findings of no change from trial to trial on position scores.
The significant change from trial to trial on movement scores repre-
sented a decline in accuracy which was thought to result from lack
of feedback. The conclusion that altering input does not improve
accuracy of perception was drawn from the finding that adding a con-
stant weight did not change accuracy of reproducing either movement
length or position within either group.
Comparisons of sighted and blind subjects on accuracy of
locating limb position and of reproducing movement length indicated
that the standard against which position and movement are measured
1b not visual when the judgment does not involve cognition of space
outside the body. The superior performance of the blind on repro-
duction of movement length supports Gibson and Mowrer's (1938) theory
of ocular dominance in sighted adults and suggested that vision may
function as a regulatory mechanism in movements of sighted adults.
Equal accuracy of sighted and blind subjects in reproduction of
position suggested that sighted subjects used visual imagery to
iii
compensate for lack of sensitivity to kinesthetic cues possessed by
the blind. Comparison of the present findings with results of
earlier studies of maze learning in which congenitally blind per-
sons performed more poorly than sighted persons suggested that
vision is not a necessary element in forming a spatial model of the
body, but is necessary in constructing a model of extrapersonal
space.
iv
ACKNOWLEDGMENTS
Barry Lindley, Ph.D., gave frequent advice and assistance
throvighout the development and implementation of the project.
Leonard Pearson, Ph.D., provided encouragement during the
early phases of the development of the research. Arthur Rosner,
Ph.D., made suggestions regarding sampling and analysis of the data
and gave advice on the writing of the manuscript*
Dr. Lindley proposed the basic statistical analysis and re-
search design. Mrs. Elaine Keramidas, statistician, suggested the
specific statistical analysis that was used in the research and wrote
the programs for the analysis of the data. Douglas Schultz, Ph.D.,
gave advice on the analysis of the reliability and validity of the
data.
George Albee, Ph.D., and Norman Taslitz, Ph.D., read and com-
mented on the manuscript.
Mr. Cleo Dolan, director, and the staff of the Cleveland Soci-
ety for the Blind Sight Center rendered generous assistance in secur-
ing blind subjects for the study. Neil Shamberg, Ph.D., obtained
most of the control subjects. Appreciation is extended to individuals
who cooperated with the study by serving as research subjects.
Miss Esther Nolte, occupational therapist, gave advice on
the construction of the built-up pencil used in the research. Dr.
Lindley constructed the pencil.
▼
Dr. Lindley prepared the Illustrations. Mrs. G rattan Glesey
typed the manuscript.
I would also like to express my appreciation to the faculty
of the Department of Psychology for the opportunity of studying in
the department. My studies were supported by VRA Training Grant
71*5-2642.
TABLE OP CONTENTS
Page
ACKNOWLEDGMENTS v
LIST OF TABLES ix
LIST OF ILLUSTRATIONS xl
Chapter
I. THE RELATIONSHIP BETWEEN VISION AND BODY PERCEPTION . 1
Problem
Background
Bod/ Schema
Phantom Limb
Vision
Empirical Evidence
Rationale and Conceptual Hypotheses of the Present
Study
II. MEASUREMENT OF KINESTHETIC PERCEPTION 19
Background
Rationale of Method Used in the Study
Description of Method
Validity
Reliability
III. SAMPLE 32
Experimental Subjects
Control Subjects
IV. METHODS AND PROCEDURES 37
Administration
Scoring
Distribution of Scores
Research Design
Statistical Method
Estimation of Missing Data
vii
4
TABLE OF CONTENTS— Continued
Chapter Page
V. RESULTS 51
Position Scores
Movement Scores
Slghted-Blind Difference
Practice Effect
Influence of Weighting on Individual
Scores
VI. DISCUSSION 64
Influence of Visual Condition
Accuracy of Movement
Accuracy- of Position
Influence of Weighting
Influence of Practice
VII. SUMMARY 80
APPENDICES 83
I. 8k
II. 85
III. 91
LIST OF REFERENCES 100
vlii
LIST OF TABLES
Sable Page
1. Split-half reliabilities of position and movement
scores for sighted and blind subjects on weighted
and unweighted trials 31
2. Test-retest reliabilities of position and movement
•cores for sighted and blind subjects 31
3» Age and sex distribution of blind and sighted sub-
jects 35
k. Educational level of blind subjects 36
5» Design of typical session: first half 46
6* Design of typical session: second half 47
7* Schematic representation of research design (after
Winer, 1962) 48
8. Means and standard deviations of position scores for
sighted and blind subjects 51
9* Three-way analysis of variance: influence of visual
condition, weighting, and practice on position
■cores 52
10. Means and standard deviations of movement scores for
sighted and blind subjects 53
11. Three-way analysis of variance: influence of visual
condition, weighting, and practice on movement
scores 54
12* Influence of weighting on movement scores: change
In mean length of movement for sighted and blind
subjects with addition of a constant weight .... 59
13* Influence of weighting on movement scores: change
In mean error for sighted and blind subjects with
addition of a constant weight 59
LIST OF TABLES— Continued
Table Page
lit. Influence of weighting on movement scores: change
In mean length of movement with addition of a
constant weight for sighted and blind subjects
according to kind of error on unweighted trials . . 60
15. Influence of weighting on movement scores: change
in mean error with addition of a constant weight
for sighted and blind subjects according to kind
of error on unweighted trials 6l
16. Influence of weighting on position scores: change
In mean scores of highest and lowest scorers in
sighted and blind groups with addition of a con-
stant weight 62
17* Influence of weighting on position scores: direction
of score change among high and low scorers in
sighted and blind groups with addition of a con-
stant weight 63
id. F-ratios of practice effect and its interaction
terms for raw and transformed scores 78
19. Ranges of standard deviations on trials one to forty
for sighted and blind subjects on weighted and un-
weighted trials for raw data and transformations . 90
LIST OF ILLUSTRATIONS
Figure Page
1. Twenty random movements used In the study ...... 2k
2. Remaining twenty random movements used in the study . 25
3. Measurement of the actual magnitude of the error . . kO
km Measurement of the response trajectory , . kl
5. Influence of visual condition: mean movement scores
of sighted and blind subjects on trials one to
forty on weighted and unweighted trials 56
6. Influence of practice: mean movement scores of
sighted and blind subjects combined on trials one
to forty with weighted and unweighted trials com-
bined 57
7» Means of weighted and unweighted movement scores on
trials one to forty for sighted and blind groups
combined 77
8. Model of transformation of position score by
squaring 87
9. Model of transformation of movement score by
squaring 88
10. Distribution of mean position scores of sighted sub-
jects on weighted trials (N ■ 20) 92
11. Distribution of mean position scores of sighted sub-
jects on unweighted trials (N » 20) 93
12. Distribution of mean position scores of blind sub-
jects on weighted trials (N » 20) 9k
13« Distribution of mean position scores of blind sub-
jects on unweighted trials (N = 20) 95
Ik. Distribution of mean movement scores of sighted
subjects on weighted trials (N - 20 ) 96
xi
LIST OF ILLUSTRATIONS— Continued
Figure Page
15* Distribution of mean movement scores of sighted
subjects on unweighted trials (N = 20) 97
16. Distribution of mean movement scores of blind
subjects on weighted trials (N ■ 20) 98
17 • Distribution of mean movement scores of blind
subjects on unweighted trials (N ■ 20) ...... . 99
xli
CHAPTER I
THE RELATIONSHIP BETWEEN VISION
AND BODY PERCEPTION
Concern with the problem of how people perceive their bodies
has a long history in psychiatry and neurology. Some writers have
focused on the effects of personality on organization of body aware-
ness (Fenichel, 19^5; Freud, 1957; Lewis, 1958; Linn, 1955; Machover,
1949; Szasz, 1957); others have emphasized the mechanisms by which
body sensations are perceived (Head, 1920; Head and Holmes, 19U);
and still others have focused on the interaction between personality
and the mechanisms of perception ( By chows ki, 19^3; Federn, 1952;
Fisher and Cleveland, 1958; Kolb, 1959a; Kolb, 1959b; Scheerer,
195^; Schllder, 1950; Smythies, 1953; and Witkin, 1965).
Inferences about the mechanisms by which body sensations
are perceived have been based on an Impressive array of clinical re-
ports of disturbances in body awareness (Bors, 1951; Critchley,
1950; Critchley, 1953; Gerstmann, 19^; Head, 1920; Head, 1963;
Head and Holmes, 1911; Hebb, i960; Henderson and Smyth, 19^; Kolb,
1959b; Nielsen, 1938; Schllder, 1950; Slmmel, 1956a; Simmel, 1956b;
Ullman et ad., i960). An early and currently prominent theory
(Gerhard, 1968; Teuber, i960) of body perception was proposed by
Head (1920) who believed that body sensations are not experienced
directly, but are mediated by an organizing mechanism called the
"body schema."
Although Head's theory has remained the basis for explaining
clinical phenomena of body perception, certain issues in his concep-
tualization have not been resolved. Primary among these issues is
the question of how the body schema develops. Various writers who
have considered the specific problem of how the senses contribute to
the development of the body schema have given special attention to
the role of vision in the formation of the body schema (Critchley,
1950; Gerstmann, 19^2; Hebb, i960; Kolb, 1959b; Schilder, 1950;
Traub and Orbach, 1964). A major unresolved Question, first raised
by Head (1920), is whether the standard against which body orienta-
tion is measured is visual or not.
Few empirical investigations have been designed to answer
the question of whether the standard against which body orientation
is measured is visual. Little research has taken advantage of the
opportunity for studying the relationship between vision and body
perception that is to be found in congenitally blind adults. Re-
search on the use of somesthetic cues by blind persons includes
studies of maze learning in congenitally blind persons (Berg and
Worchel, 1956; Duncan, 1934; Knotts and Miles, 1929; and Koch and
Ufkess, 1926) and a few studies of body perception in congenitally
blind adults (Renshaw, Wherry, and Newlin, 1930; Slinger and Horsley,
1906). Whereas maze learning studies suggested that the standard
for measuring body orientation is visual, the perceptual studies
Indicated that it Is not.
Problem
The present study of kinesthetic perception in early blind
adults investigated the question of whether the standard against
which body orientation is measured Is visual. The problem has both
theoretical and practical importance. The results of the study pro-
wide a basis for making inferences about the role of vision in the
body schema, the mechanism by which the body is perceived. The find-
ings are applicable to rehabilitation practices with the blind where
greater understanding of the causes of poor mobility and orientation
in blind persons is needed.
Background
The present study of kinesthetic perception in early blind
adults contributes to the clarification of Head's (1920 ) concept of
the body schema. Studies of phantom limb have provided clear evi-
dence for the body schema as Head conceptualized it. However, pre-
cise details of the schema were never worked out by Head.
An important question that has remained unanswered is whether
vision contributes to the body schema. Opposing points of view have
been taken as to whether the standard of reference in measuring body
orientation is visual. This is an intriguing problem particularly
In regard to recognition of posture because the model of the body is
primarily spatial. The assumption that vision is the sense through
which space is perceived leads to the conclusion that the standard
of reference against which perception of limb movement and position
la measured must be visual. Opponents of this view contend that
postural orientation prt.ed^s visual perception of space and that
vision is not a necessary component of body orientation.
Studies of congenital! y blind persons whose postural model
of the body has developed without visual standards of reference have
provided conflicting evidence on the relationship between vision and
body perception. Poorer performance of congenitally blind on com-
plex tasks requiring use of body sensations indicated that the
standard of reference for recognizing tactile and kinesthetic cues
is normally visual. A group of studies of less complex tasks of
body perception showing equal or superior performance of congenitally
blind compared to slghts-l persons indicated that the standard is not
visual.
Body Schema
From extensive studies of clinical cases, Head (1920) de-
veloped the theory that direct perception of body sensation is im-
possible. Proprioceptive stimuli are recognized only after they
have been brought into connection with an unconscious model of the
body called the body schema. The schema is a standard against
which all changes in body perception are measured. It is a dynamic
entity which is built up and changed by the addition of every fresh
group of sensations. These sensations Include the ones underlying
postural recognition and appreciation of passive movement; tactile
differentiation other than contact and texture, for example, weight;
spatial discrimination in the form of recognition of two-point simul-
taneous stimulation and recognition of size and shape; localization
of spot stimulated on the skin; and recognition of thermal stimula-
tion (Head, 1920). Subsequent writers (Kolb, 1959b; Smythles, 1953)
have reduced this scheme of body sensations to tactile, kinesthetic,
and visual sensations.
The body schema is an unconscious phenomenon. Awareness of
change is all that is accessible to consciousness once a sensation
has been brought Into relation with the schema.
Recognition of posture and passive movement implies combination
of every fresh group of sensations with postural schemata out-
side the central field of attention. The change in conscious-
ness which corresponds to this combination is immediate recogni-
tion of an altered position (Head, 1920, p. 604).
Although postural recognition is not usually the center of attention
of the individual, one is capable of becoming conscious of body posi-
tion at any time through the process described.
Smythles (1953) clarified the distinction between the per-
ceived body and the body schema. The perceived body is the "somatic
sensory field . . . present In direct experience." The body schema,
which coordinates posture and voluntary and automatic movements,
exists below the threshold of consciousness. It is not witnessed
or experienced, but its presence is Inferred from introspective re-
ports and behavioral observations especially of persons with dis-
turbances in body perception.
Commentators on Head's theory criticized the rudimentary
nature of his conceptualization. Oldfield and Zangvill (1942)
pointed out that Head never defined hov the schema is built up, an
issue to which the present research is related. The authors also
pointed out that Head did not explain hov the schema produces a re-
sponse and at the same time creates awareness of change. In addi-
tion, Head failed to define the physiological processes underlying
the body schema. Although Oldfield and Zangvill may be Justified in
criticizing Head for his lack of theoretical attention to the prob-
lems of consciousness and physiological processes, these issues are
not ones which can presently be submitted to empirical study. The
authors also pointed out that Head's combination of functional and
material elements in a single concept of the body schema constituted
a weakness of his theory. This criticism overlooked Head's real
effort to develop a concept which would account for integration of
past experience vith present perception.
In spite of its weak points, Head's concept of the body
schema has remained useful for understanding phenomena of body per-
ception. Evidence from clinical reports, especially from studies of
phantom limb, indicate that his conceptualization of the process of
body perception is sound.
Phantom Limb
Observations of phantom limb, one of the most thoroughly
studied phenomena of body image, provide the best evidence for in-
ferring the existence of the body schema. Most vriters agree that
phantom limb is a clear manifestation of the body schema (Head and
Holmes, 1911; Gerhard, 1968; Schilder, 1950; Simmel, 1956a; and
Teuber, i960). In the case of phantom limb, an amputated limb is
felt to be intact. Phantom limb studies avoid a problem which
plagues studies of the agnosias where body perception also fails to
correspond to body structure—differentiating between the effects
on body perception of neurological lesion and psychological state
(Critchley, 1953; Gerstmann, 19^2; Nielsen, 1938; Ullman, et al.,
I960).
In studying 300 cases of amputation, Henderson and Smyth
(19^8) found that phantom limb was an almost universal occurrence
following such a procedure. They described three kinds of phantom
limb: a mild tingling; a stronger tingling produced by stimulation
of the stump neuroma; and painful phantom.
Writers in the field have raised the question of whether
the phantom limb represents a central or peripheral phenomenon
(Bychowskl, 1943; Henderson and Smyth, 19^8; Schilder, 1950). Re-
search showing that phantom limb represents a manifestation of a
central process supports Head's (1920) speculation that maintenance
of the body schema is a cortical function. The finding that there
is greater stump sensitivity where there is greater cortical repre-
sentation led Henderson and Smyth (19^8) to conclude that mainten-
ance of the phantom limb is a cortical function. Haber's (1955)
observation that sensitivity of the stump is greater than sensitivity
of the homologous part of the opposite arm or of the homologous part
of the arms of controls in spite of equal density of receptors led
him to a similar conclusion.
Observations of the vicissitudes of the phantom limb con-
firmed Head1 8 (1920 ) speculation that the body schema Is built up
and altered through the addition of new body sensations. Normally
the phantom limb gradually recedes until perception of the body
corresponds to its structure (Henderson and Smyth, 1946). Bors
(1951) observed that patients with paraplegic phantom or with ampu-
tation of paralyzed parts did not experience this telescoping
phenomenon. Bors' report that amputees with partial cord lesions
did experience telescoping of the phantom led to the conclusion that
sensory input is necessary for shaping the body image to correspond
to structural change. Slmmel's (1956a; 1956b) studies led to similar
conclusions. She found no incidence of phantom limb in leprosy
patients who lost digits by absorption while digital amputees re-
ported such phenomena (1956b). She speculated that leprosy patients
have an opportunity to alter body Image during the gradual loss of
the digit while those with sudden loss continue to experience the
digit as present until new experience can change the body image to
correspond to actual structure. Slmmel (1956a) also reported that
phantoms did not occur in children with congenital absence of limbs
or where amputation occurred before five years of age.
Although agreeing that phantom limb represents a manifesta-
tion of the body schema, Bychowski (19U3) also felt that emotional
components contribute to the maintenance of the phantom limb. The
"hallucination" has a compensatory meaning. With gradual loss of
limb, there is no phantom because the Individual has emotionally ac-
cepted or adjusted to his loss. The problem of differentiating be-
tween emotion and body schema as the mechanism responsible for main-
taining the phantom limb cannot be resolved easily. Interpretation
of the preceding studies as evidence of the body schema must be quali-
fied vith Bychovski's interpretation that emotion may be a factor
responsible for maintaining the phantom.
While Bychovski's Is not the generally accepted view of phan-
tom limb, writers In the field agree that emotional processes play a
role In maintaining the painful phantom (Henderson and Smyth, 1948;
Simmel, 1956a). Bailey and Moersch (l94l) pointed out that since
neurotomy did not obliterate the painful phantom, it is probably pro-
duced through an emotional process. Although discussing but a few
cases, Kolb (195*0 presented some evidence for the motivational basis
of the painful phantom which he treated successfully with psycho-
therapy.
Vision
Phantom limb phenomena provide good evidence for the body
schema as Head (1920) described it. The body schema mediates body
perceptions by relating them to a model of the body which has been
built up from previous sensations. Precisely how this process occurs
and how the various sensory modalities contribute to the body schema
was not clearly defined by Head, a fact which has been the focus of
criticism by some writers (Oldfield and Zangwill, 1942).
10
Although he failed to delineate the development of the body
•chema In detail, Head (1920) gave special consideration to the role
of vision in the development of the body schema. He questioned
whether the standard to which reference is made when a fresh posi-
tion is recognized is visual and concluded that it is not. He
reached this conclusion after observing a patient who retained the
ability to visualize a limb but who lost the power of recognizing
posture and passive movement. Although the patient could localize
a spot touched on his arm, he referred to the position of the ini-
tially visualized arm rather than to the position to which his arm
had been moved by the examiner after the patient had closed his eyes.
Since the visual Image remained intact while the power of recogniz-
ing posture was lost, Head (1920 ) concluded that the standard against
which change is measured is not visual, but is constituted of pos-
tural images.
Schilder (1950 ) reported similar phenomena in his patients,
but contrary to Head, felt the observation led to the conclusion that
vision plays an Important part in the body schema. Such observations
Indicated to Schilder that it is necessary for perceptions to be
brought into connection with an optic image of the body in order for
them to be meaningful. He believed that tactual localization is im-
possible without the optic factor, but qualified his point of view
with the statement that kinesthetic experiences may take the place
of optic factors. Subsequent writers have restated Schilder' 8 point
of view by indicating that vision contributes equally (Critchley,
11
1950; Gerstmann, 19^2) or secondarily (Hebb, i960; Kolb, 1959b) to
kinesthetic and tactual perception in the formation of the body
schema.
The problem of the contribution of vision to the body schema
is an especially intriguing one because vision is the sense by which
space is perceived, and space is one aspect of the body schema. Head
(1963) and other writers (Brown and Gotein, 19^3; Kolb, 195^; Mac-
Donald, i960; Scheerer, 195*0 have pointed out that the body schema
is a plastic model through which body parts are related to one
another in an integrated fashion and their spatial relationships and
positions are cognized. By excluding vision as a necessary component
of the body schema, Head (1920 ) implied that kinesthetic and tactual
sensations are equally efficacious in forming a spatial model of the
body. Senden (i960) proposed the opposite view that space as a con-
cept does not exist without vision and questioned whether a congenit-
ally blind person can cognize his body as a whole "figured object in
space." Studies of the way blind persons describe the body through
verbal and artistic media have confirmed the speculation that the
organization of the surface model of the body may lack morphological
rectitude or may vary greatly from descriptions given by sighted per-
sons (Bennett, i960; Lindley, unpublished data; V. Lowenfeld, 1939)-
Two opposing points of view have arisen on the relationship
between visual and somesthetic perception. One view proposes that
postural cues are genetically prior, but visual cues eventually be-
come dominant In body orientation (Gibson and Mowrer, 1936). Gibson
12
and Itowrer's theory predicts that sighted adults deprived of vision
will use body cues less veil than congenital 1y blind persons because
their standard of reference is visual. Coinciding with this view is
the idea that the lack of vision In blind persons leads to tactile
and postural hypersensitivity (Critchley, 1950).
The contrary point of view which has developed from empirical
observations proposes that the ability to translate kinesthetic and
tactual sensations into visual Imagery increases the ability of a
sighted person to use body cues (Duncan, 193^; Koch and Ufkess,
1926; Worchel, 1951) • This point of view predicts that sighted adults
with vision occluded will utilize body cues better then congenitally
blind persons who are unable to translate these cues into visual
images.
Gibson (1952) later retracted his theory that postural orien-
tation precedes visual spatial perception. He concluded that the
question of whether posture or vision is the primary framework for
orientation is insoluble and that the two components interact. In
stating the issue in an insoluble form, Gibson overlooked the contri-
bution to the problem of the relationship between body orientation
and vision that could be made by studies of congenitally blind per-
sons. The type of evidence cited by Gibson and Mowrer (1938) about
the relationship between vision and postural orientation involved
only sighted subjects.
13
Empirical Evidence
Studies of sighted persons with altered visual input nave
yielded conflicting evidence on the relationship between vision and
body orientation. Stratton's (1896) report of his experience with
Inverted retinal images suggested that the standard against which
movement and orientation are measured is visual and that when visual
and kinesthetic images conflict, the visual prevails. In a study of
adaptation to distorting lenses, Harris (1963) found that with altered
visual input, the visual image prevailed over the proprioceptive.
Other work (Wooster, 1923) with adaptation to distorting
lenses indicated that the standard against which body position is
measured is kinesthetic. With competing visual and kinesthetic cues,
subjects made smaller initial errors than in tasks where tactual or
auditory input were the competing cues. Wooster concluded that adjust-
ing to changed visual condition with the use of kinesthetic cues was
effective even without practice as required for the other modalities.
Another approach to studying the relationship between vision
and 8omesthetic perception in body orientation has been the comparison
of sighted and blind persons on various tasks requiring use of body
cues. Observations that blind persons lack facility in the use of
their bodies (Brieland, 1950; Buell, 1950; Fulcher, 19^2; B. Lowenfeld,
1963; Horris, Spaulding, and Brodle, 1957; !• M. Siegel, 1966) sug-
gested that in the absence of visual Images tactile and kinesthetic
images are an inadequate standard of reference for measuring change In
body position and movement. Studies comparing congenitally blind, late
blind, and sighted with occluded vision on maze learning (Berg and
Ik
Worchel, 1956; Duncan, 193^5 Knotts and Miles, 1929; Koch and Ufkess,
1926) and on tactual form recognition and tactual space relations
(Drever, 1955; Worchel, 1951) provide an unusually high degree of
agreement that the ability to translate body cues into visual images
gives the late blind and the sighted a decided advantage over the
congenitally blind on these tasks.
Contrary to maze learning studies, research on body percep-
tion indicates that performance among the congenital ly blind is equal
or superior to that of the sighted. Blind subjects have performed
the same as sighted on perception of the upright (Bitterman and
Worchel, 1953 )> perception of tactile stimulation (Bender, Green,
and Fink, 195*0, perception of kinesthetic stimuli (Jastrow, 1886),
and perception of tactile -kinesthetic stimuli (Bartley, Clifford,
and Calvin, 1955 )• The blind samples were superior to the sighted
subjects on other tasks of tactile (Renshaw, Wherry, and Newlin,
1930), kinesthetic (Jastrov, 1886; Slinger and Horsley, 1906), and
tactile -kinesthetic (Hunter, 195*0 perception. The blind performed
more poorly than sighted controls on only one task of kinesthetic
perception (Jastrow, 1886).
Discrepancies in the results of studies of perception may
be attributed to differences in samples, in perceptual modality
studied, and in treatment of the data. One study (Hunter, 195*0 vhich
showed the blind to be significantly better than the sighted on per-
ception of straightne8s (described as a tactile-kinesthetlc task),
combined data from congenitally blind and late blind and thereby
15
failed to Isolate the influence of visual imagery on the performance
of the task. Studies reporting no difference between congenitally
blind and sighted on tactile -kinesthetic size perception (Bartley,
Clifford, and Calvin, 1955)* on perception of double simultaneous
tactile stimulation (Bender, Green, and Fink, 195*0 » and on percep-
tion of the upright (Bitterman and Worchel, 1953) used children ae
subjects and did not take account of the possible Influence of develop-
mental factors on perception. In the latter study, Bitterman and
Worchel (1953) reported that congenitally blind were superior to the
sighted if their data were analyzed with a different method of scor-
ing which makes their results even more difficult to interpret.
Studies of congenitally blind adults on tactual localization
(Renshaw, Wherry, and Newlin, 1930) and on kinesthetic perception
(Slinger and Horsley, 1906) showed the blind groups to be superior to
the sighted groups. However, the authors of these two studies did
not present statistical analyses of their data. Jastrow (1886),
who found no differences between sighted and blind on one task of
kinesthetic perception, found the blind superior on another task,
and Inferior on yet another task. Jastrow did not present either
the characteristics of his blind sample or statistical analysis of
bis data. Since the task on which the blind performed more poorly
than the sighted was a complex one involving the learning of a
standard series of lengths and the production of movement from verbal
instructions, the observed difference may have been due to the nature
of the task rather than to group differences.
16
Although the findings of studies of body perception in the
blind seem to be at odds with one another, they do suggest that the
blind are equal or superior to the sighted in use of body perception.
The results of the perceptual studies are contrary to findings of
poorer performance of the blind In the use of body cues in maze learn-
ing. Whereas perceptual studies suggest that the standard of refer-
ence for body orientation is not visual, maze learning studies sug-
gest that it is,
Rationale and Conceptual Hypotheses
of the Present Study
The present study of kinesthetic perception in early blind
adults Investigated the question of whether the standard against
which body orientation is measured is visual or not. The study was
designed to correct problems in method, sampling, and treatment of
data found In previous studies of blind persons. Kinesthetic per-
ception, the sense through which body movement and position are per-
ceived, was assumed to provide data from which inferences could most
clearly be drawn regarding the relationship between vision and body
orientation in the body schema. A simple task of duplication of arm
movements was used so that results could be clearly attributed to the
functions of body perception without confounding of other factors
such as Intelligence, memory, and experience which might play a role
In maze learning and In tasks involving more than one sense. Observa-
tion of early blind adults and sighted adults provided comparison of
17
body perception in a group of individuals who had no visual standard
of reference for body orientation with a group of individuals who
could develop such a standard of reference. Using only early blind
subjects avoided problems in interpreting results introduced when
mixed groups of blind persons are used. Study of adults rather than
children provided observations from persons in whom body perception
is stable. The study added statistical analysis of the data which
was lacking in the two previous studies of body perception in con-
genitally blind adults (Renshaw, Wherry, and Newlin, 1930; Slinger
and Horsley, 1906).
The present study complemented previous studies of kines-
thetic perception in the blind, by adding refinement in sampling
lacking in J as trow' s (1666) study and by providing statistical
analysis of results not included in the two preceding studies of
kinesthetic perception in the blind (Jastrow, 1866; Slinger and
Horsley, 1906). The method used in the present study was simpler
than two of the methods used by Jastrow and was designed to provide
information on both perception of position and perception of move-
ment. Slinger and Horsley's study provided data only for perception
of position and Jastrow* s study gave data only for perception of
movement .
In addition to examining the basic question of the role of
vision in the perception of movement and position, the present study
examined changes in kinesthetic perception resulting from practice
and from altered input. Altering input by adding weight or resis-
tance to enhance perception has a history in studies of kinesthetic
18
perception (Bahrick, Fitts, and Schneider, 1955; Bahrick, Bennett,
and Fltts, 1955; Cohen, 1958b) and in rehabilitation practices with
the blind. Weighting was included in the study primarily to evaluate
the clinical practice of using weight to improve kinesthetic percep-
tion in the blind. Since Renshaw, Wherry, and Newlin (1930) found a
significant improvement with practice on a task of tactual localiza-
tion, the data were analyzed for such an effect. Results of the
analysis of the effect of practice are relevant to rehabilitation
procedures as well.
The study Included three conceptual hypotheses.
1. Kinesthetic perception is more accurate in con genital ly
blind adults than in sighted adults.
2. Sighted adults increase their accuracy with practice
more than blind adults do.
3. Helther group increases accuracy of kinesthetic per-
ception with addition of a constant weight.
CHAPTER II
MEASUREMENT OF KINESTHETIC PERCEPTION
The assumption that duplication of arm position, the method
used in the present study, was an adequate measure of kinesthetic
perception was based on previous research in the field. Perception
of body movement and position has frequently been used as a measure
of kinesthetic perception. Duplication of arm position appeared to
be aa acceptable task which provided maximum Information about
kinesthetic perception. Ayres' (1966) test of kinesthetic percep-
tion in which subjects duplicate arm movements on a table top was
modified to meet the requirements of the present research. The
method used appeared to have acceptable construct validity, and
reliabilities estimated from the data were high.
Background
Studies of kinesthetic perception come from three areas:
neurological examination for the presence of central nervous system
pathology; psychological examination of individual differences, and
psychophysical studies of perception. In all three fields of re-
search, perception of limb position has been used to measure kines-
thesia.
Neurologists use appreciation of passive movement and posi-
tion sense as measures of kinesthetic perception. Position sense
19
20
is assessed through duplication of position and through the finger-
to-finger test in which the individual is required to align the index
finger of one hand with the index finger of the other hand (Ruch,
1965).
Researchers studying psychological and psychophysical aspects
of kinesthetic perception have most often employed duplication of arm
position as a measure of kinesthesia (Brown, Knauft, and Rosenbaum,
19W; Harris, 1963; Ronco, 1963; Smith and Smith, 1962; Wyke, I965).
A significant amount of work in kinesthetic perception has been done
by persons interested in developing tests for predicting performance
In physical education. Arm position sense has been used by these
authors singly or in combination with other tests (Phillips and
Summers, 1954; Roloff, 1953; Scott, 1955; Wettstone, 1938; Witte,
1962). Position sense of other parts of the body, usually the leg,
has also been used for assessing kinesthetic perception (Scott,
1955; Young, 1945)* Other dimensions of kinesthetic perception
which have been included in test batteries are balance (Scott, 1955;
Bass, 1939; Roloff, 1953; Young, 19^5), force (Henry, 1953; Scott,
1955; Young, 19^5), and weight (Roloff, 1953; Scott, 1955; Young,
19^5 )• Young (19^5) discarded weight perception because she found
It did not discriminate sufficiently among subjects.
Rationale of Method Used in the Study
The method chosen for the study required subjects to dupli-
cate waist-height arm movements in a two-dimensional plane. To
21
increase the acceptability of the task to the subjects, the ana,
rather than other parts of the body, was used. Duplication of posi-
tion avoided the problems introduced when alternative methods are
used vith blind subjects. Movements in a two-dimensional plane
provided more information than could be obtained from linear move-
ments ..
A pilot study of body concept in blind persons (Lindley,
unpublished data) showed that acceptability of the task to the sub-
jects is essential in obtaining their cooperation. A method requir-
ing seated subjects to make arm movements on a table top appeared to
be an acceptable task. Subjects did not have to exert themselves
to any great extent in making movements in the space around the body
normally used in daily activity.
Duplication of position was chosen from the three methods
that have been used to measure kinesthetic perception through arm
movements: target pointing, duplication 'of position, and assuming a
position from verbal direction. Target pointing, which requires
locating position Initially through sight, was eliminated because
its use is inappropriate with blind subjects. Requiring subjects to
locate a position from verbal direction is also Inappropriate with
blind subjects whose concepts of Instructions such as "twice as long
as" or "ninety degree angle," may be quite different from the con-
cepts of sighted people. Having subjects duplicate a position avoided
the problems Introduced by the other two methods.
Of the two prevalent techniques for duplicating arm position,
22
use of a slider (Brown, Knauft, and Rosenbaum, l$kQ; Ronco, I963)
and placement of subject's arm by the examiner (Ayres, 1966; Cohen,
1936a), the latter allows movements In a two-dimensional plane. In
contrast to linear movements, movements In a two-dimensional plane
provide information about perception both of position and of move-
ment. Because of the added information obtained, placement of sub-
ject's arm by examiner in a two-dimensional plane was used in the
present study.
Description of Method
The method used in the present study is a modification of
Ayres' (1966) test of kinesthetic perception for children. Seated
at a table, the blindfolded subject has his arm moved along each of
eight paths by the examiner. The subject attempts to repeat each
movement and is scored for his accuracy in locating the target.
Ayres' test was modified by increasing the number of movements made,
by having subjects hold a pencil instead of pointing with the Index
finger, by observing movements with the subject's holding a weighted
as well as an unweighted pencil, by having only sighted subjects wear
a blindfold, and by scoring subjects for accuracy of reproducing both
movement length and position.
Ayres' method was modified by using a larger number of
paths that were randomly drawn. A greater number of paths than
Ayres uses was required by the research purposes of the study. The
paths were randomly drawn to meet the requirements of the research
23
design.
Movements vere made in an area kO.J cm by 50. 8 cm (sixteen by-
twenty inches) after results of a pilot study (Lindley, unpublished
data) Indicated that this is the maximum area within which arm move-
ments can be comfortably made. The area was. divided into a matrix
80 by 100 and a random pattern of forty paths was drawn using pairs
of random numbers (Attneave and Arnoult, 1956). Figures 1 and 2
show in reduced scale the forty random paths used in the study. One
cm on the figure represents 2.9 cm on the original field. The
trajectories are shown in two parts to make their presentation in
reduced scale clearer.
Subjects held a pencil mounted In a plastic bar while making
movements. The pencil recorded the point where the subject landed
when he repeated the movement. The bar made it easier for the subject
to hold the pencil and waa filled with lead shot for the weighted
movements.
lb test the hypothesis that altering input changes kines-
thetic perception, movements were observed with the subject's holding
a weighted pencil and an unweighted pencil. The weighted pencil was
constructed by filling the plastic tube in which the pencil was
mounted with thirteen ounces of lead'shot. The weight was selected
Consecutive pairs of random numbers from a teble of random
numbers from one to one hundred defined the point of origin and the
target point of each path. The two points were Joined by a line for
ease of administration. The path is referred to as the trajectory.
2k
B
I
25
K
uj
a.
x
UJ
26
on the basis of a pilot study (Lindley, unpublished data) which
Indicated that the maximum weight that can be lifted for forty
trials without fatigue is about a pound.
Subjects were scored for their accuracy in reproducing both
movement length and position. Clinical impressions that blind peo-
ple are constricted in body movement suggested that it would be use-
ful to measure the movement parameter of kinesthetic perception.
Validity
While an estimate of the empirical validity of the present
data would be desirable, the lack of independent data in the study
prohibited this approach to the validity problem. An alternative
approach consisting of logical review of earlier studies of kines-
thetic perception (Bahrick, 1957; Bahrick, Bennett, and Fitts, 1955;
Boring, 1942; Browne, Lee, and Ring, 1954; Cohen, 1956; Cohen,
1958a; Cohen, 1958b; Gardner, 1964; Morgan, 1965; Provins, 195fl;
Wyke, 1965) showed that duplication of position and movement is a
valid method for measuring kinesthesls. Research relevant to the
validity problem includes general studies of joint receptors and of
changes in perception with altered kinesthetic input as well as
studies specific to the present method which show that some possible
sources of constant error do not contribute to observed performance.
Sensitivity to position and movement of the body, along
with sensitivity to weight, resistence, and effort, constitute
kinesthetic perception (Boring, 1942). Anatomical studies showing
the existence in the Joints of nerve endings sensitive to movement
27
(Gardner, 1964; Morgan, 1965 ) provide a physical basis for assuming
kinesthesia represents a separate entity of experience. At the be-
havioral level, Cohen (1958b) confirmed that Joint receptors make a
greater contribution to perception of arm position than either tac-
tile receptors or muscle spindles and tendon organs.
Studies shoving that alteration in kinesthetic input leads
to difference8 in behavior lend validity to the concept of kinesthetic
perception. Input can be altered through direct interference with
the nervous system and through change of the force applied when a
movement is made (Bahrick, 1957) • Studies of faradization and
anesthetization of joints have shown that sensitivity to passive
movement is reduced when the nerve endings are interfered with by
these techniques (Boring, 19^42; Provins, 1958).
The subjective experience of movement and poaition may be
conveyed by verbal expression or by duplication of movement and posi-
tion. If the latter method, which was used in the present study, ia
a valid measure of kinesthesia, changes in performance should occur
with altered kinesthetic input. Bahrick, Bennett, and Fitta (1955)
found a change in accuracy of duplication of arm poaition with
altered kinesthetic input when the force on the control stick was
changed relative to the size of the movement to be made.
While the preceding atudiea provide a basis for accepting
the validity of the concept of kinesthetic perception and of duplica-
tion of position as a method for measuring it, other research haa
examined sources of constant error that might occur in the method
28
used in the present study. Requiring subjects to match active move-
ments to passive movements may provide a source of constant error.
The present method involved two steps, a passive movement when exam-
iner moved the subject's arm from start to target and an active move-
ment when the subject repeated the movement. Browne, Lee, and Ring
(195^) proposed that active and passive movements are not equivalent
since perception of movement and position varies under the two condi-
tions. If the proposition were true, the present method would have
introduced error due to the differences in perception of passive
input and active output. Provins* (195Q) research showed that per-
ception of movement is equivalent under active and passive conditions.
He studied sensitivity to movement In the finger Joint when the finger
was relaxed and when the muscles were tensed In both the flexed and
extended positions and found no significant differences in perception
of movement under the two conditions.
Poor motor ability may constitute another source of constant
error (Cohen, 1956). Considering this problem in a study of kines-
thesia in the arm, Cohen had subjects point to the target with eyes
open as a control. Since he found that all subjects could locate
the target without error with their eyes open, he concluded that in-
accuracies in target location that occur when the subject is blind-
folded are due to inaccuracies in kinesthetic perception and not to
lack of general motor control.
In the present study, selection of subjects was designed to
minimize invalid results by excluding from the sample persons with
29
irrelevant characteristics which might correlate with performance of
the task. Persons with psychiatric disorder, intellectual deficit,
and nervous system damage were excluded from the study and the age
of subjects was limited to a range within which body perception is
•table.
Although the preceding studies showed that the present method
had an acceptable degree of construct validity, uncontrolled sources
of error remained. Fatigue (Bahrick, 1937) end movement of the rest
of the body (Cohen, 1958*) were possible sources of error. Wyke's
(1965) research indicated that the latter variable may be an Important
source of error in observed responses. An additional source of con-
stant error may have been the compounding of the error of the initial
perception with the error of perception In the duplicated movement
(Cohen, 1956a).
lb the extent that the preceding variables contributed random
error, their effect was adequately controlled as demonstrated by high
split-half reliabilities. To the extent that they represented con-
stant error, their effects reduced validity of the results.
Reliability
Because there were no estimates of reliability for the present
method as applied to the populations studied, this index was calcu-
lated from the present data. Split-half and test-retest reliabilities,
estimated with Pearson product -moment correlation coefficients, were
satisfactorily high.
30
Since the Instrument was assumed to be a homogeneous test,
split-half reliability, vhlch gives an estimate of Internal consis-
tency, was an appropriate measure of the reliability of the data ob-
tained. Split-half reliability was estimated by correlating the
sums of the scores of odd-numbered trials for subjects with the sums
of their scores on the even-numbered trials. The tvo halves of the
test were assumed to be equivalent because the items were randomly
ordered and variability in scores due to item differences should have
correlated zero with the sums of the scores. Observed correlation
between halves of the test could be assumed due to stability of the
trait within the person being tested.
Four split-half reliabilities were estimated for position
scores and four for movement scores: sighted subjects on weighted
trials; sighted subjects, unweighted trials; blind subjects, weighted
trials; and blind subjects, unweighted trials. Weighted and un-
weighted, trials were analyzed as separate tests and the Spearman-
Brown formula was applied to each split-half reliability to correct
for underestimation due to shortening of the test. Table 1 shows
the split-half reliabilities are satisfactorily high, ranging from
,728 to .9«9.
Test-retest reliabilities were estimated by correlating
scores from the weighted trials with scores from the unweighted
trials for the sighted group and for the blind group* The two sets
of trials were assumed to be equivalent because the order of pre-
sentation of items was Identical In the two sets for any given
31
TABLE 1
SPLIT-HALF RELIABILITIES OF POSITION AND MOVEMENT SCORES
FOR SIGHTED AND BLIND oUBJECTS ON WEIGHTED
AND UNWEIGHTED TRIALS
Split-Half Reliabilities
Group Position Scores Movement Scores
Sighted
Weighted
Trials
.572
.728*
•731
.845*
Unweighted
Trials
.721
.838*
.679
.809*
Blind
Weighted
Trials
.861
.925*
.867
.929*
Unweighted
Trials
.692
.818*
.842
.914*
Corrected with Spearman-Brown fonaala.
subject. Lover correlations of .519 to .877 shown in Table 2 indicate
that the assumption of equivalence may- not be tenable. The issues
raised by this finding are dealt with in Chapter V.
TABLE 2
TEST-HETEST RELIABILITIES OF POSITION AND MOVEMENT
SCORES FOR SIGHTED AND BLIND SUBJECTS
Test-Retest Reliabilities
Group Position Scores Movement Scores
Sighted .520 .710
Blind .877 .519
CHAPTER in
SAMPLE
Two groups of twenty adults were subjects for the present
study. One group was composed of early blind persons and one of
sighted persons. Although the blind subjects were randomly selected
according to rigorous criteria, the sighted group constituted an
Incidental sample selected because the individuals were the most
available.
Experimental Subjects
Blind subjects were selected according to two major criteria.
The visual history had to be such that the subject could not use
visual imagery in the performance of the experimental task. Persons
with irrelevant characteristics which might have correlated with per-
formance were excluded from the sample.
Only blind persons who did not have object or form perception
were included in the experimental group. Blind subjects possessed
light perception or less, up to but not including 2/200. Since studies
of visual Imagery in the blind showed that conscious (Schlaegel, 1953)
and unconscious (jastrow, 1901) visual memory is not retained if blind-
ness occurs before age five, the experimental group included only sub-
jects blind before this age.
Only persons from sixteen to forty-five years of age were
32
33
Included In the sample. The lover limit was set on the -basis of studies
(Fink and Bender, 1953; Renshaw, Wherry, and Newlin, 1931; Wapner and
Warner, I965) which indicated that body perception is a developmental
phenomenon that stabilizes in early adulthood. The upper limit was
chosen to avoid contamination with processes of aging.
Persons with known psychiatric disorders or with central ner-
vous system dysfunction were excluded from the sample because studies
have shown that such conditions can affect body perception (Bennett,
1956; Critchley, 1953; Head and Holmes, 1911; Gerstmann, 19^2; Nielsen,
1938). To rule out variations in intellectual capacity that would
bear on the experiment, blind persons with verbal IQ of seventy-nine
or less or who were known to be mentally retarded were excluded from
the sample.
To meet the requirements of the research design, the experi-
mental group contained equal numbers of men and women.
Subjects for the experimental group were selected from the
files of the Cleveland Society for the Blind Sight Center according to
five criteria:
1. Age sixteen to forty-five;
2. only light perception or less since age five or younger;
3* absence of known central nervous system dysfunction;
k. absence of known psychiatric disorder;
5. average or above intelligence.
Although the files of the Sight Center contained several thousand
entries, only a very small number of this group — fourteen men and
5k
fourteen vomen--fit the above criteria. Twelve men who fit the cri-
teria were selected randomly and asked to participate in the study.
Of this number, one refused and one was unavailable. Of the eleven
women selected randomly from the list, one refused to participate.
Subjects were contacted initially by letter, then were tele-
phoned to obtain their cooperation. Appendix I contains the contents
of the letter sent to the prospective subjects. Although the subjects
were asked to come to the Sight Center for the experimental session,
nine were unable to do so and were tested in their homes.
Control Subjects
Subjects for the control group were selected on the baiia of
two criteria. The control group included only persons without visual
defects or with visual defects which were not handicapping. The sub-
jects in the control group were matched by sex and age to the subjects
in the experimental group as shown in Table 3.
Unlike the experimental group, the control group was an inci-
dental sample selected because the individuals were the most available.
It was assumed that since the control subjects were apparently ade-
quately functioning individuals, they would meet the criteria of ave-
rage intelligence and absence of central nervous system dysfunction
and psychiatric disorder. Since all of the control subjects possessed
vision which was the most significant property under investigation, it
was assumed that the results from the sample could be generalized to
the sighted population.
35
TABLE 3
AGE AND SEX DISTRIBUTION OF BLIND
AND SIGHTED SUBJECTS
Males
Females
Age
Sighted
Blind
Sighted
Blind
16-21
6
6
4
k
22-27
-
-
1
1
28-33
1
1
3
k
3^-39
3
2
1
■-
1*0-45
-
1
1
1
Mean Age
25-9
25.9
26.3
27.9
The control group contained a disproportionately large number
of college students. Six subjects were obtained through the Psychol-
ogy Department at Cuyahoga Community College, eleven through the
Psychology Department at John Carroll University, and three through a
friend. On the basis of a study by Laidlav and Hamilton (1937) which
shoved no differences in kinesthetic perception between two groups of
adults with differing levels of education, it was assumed that educa-
tion did not correlate with performance. In addition, the educational
level of the blind group was surprisingly high as shown In Table k.
36
TABI£ k
EDUCATIONAL LEVEL OF BLIND SUBJECTS
Educational Level Men Women Total
Failed to Graduate from
High School 1 1
High School Student 2 3 5
High School Graduate - 3 3
High School Graduate with
Some College 2-2
College Student k 1 5
College Graduate 1 1
Graduate Degree 2 13
CHAPTER IV
METHODS AND PROCEDURES
Procedures for administering the present test of kinesthetic
perception were based on Instructions in Ayres' (1966) manual. Two
■cores were calculated for each response, one to measure accuracy of
position and one to measure accuracy of movement. Inspection of the
scores shoved that their distributions were normal (Appendix III).
Since it vas expected that the observed scores would depend on a num-
ber of factors, only a few of which were relevant to the study, the
research was designed to take account of the factors eliminated from
the statistical analysis. The position and movement scores were sub-
mitted to analysis of variance.
Administration
A standard procedure based on the instructions for adminis-
tering Ayres' test (1966) was used for all subjects. The subject was
seated at a table and duplicated arm movements on the table surface
introduced by the experimenter.
A separate sheet of paper with the trajectories drawn on it
was used for each subject. The paper was aligned with the table edge
with the midpoint falling at the middle of the subject's body. Sighted
subjects were blindfolded before the paper was set out.
The subject was given a pencil to hold in his hand for
37
38
recording the point of his response. The pencil was inserted at right
angles to a plastic bar. The bar was grasped horizontally with the
palm down and the pencil between the index and middle fingers.
The following instructions were given to the subject:
I am going to move your arms one at a time from one spot on the
table to another. I want you to repeat the movement. We will
•tart with your right (left) hand.
After positioning the subject's hand, the experimenter said:
This is the starting place. I will now move your arm to another
place which we will call the target. I will leave your arm at
that place a few seconds so you can feel where it is. Remember
where it is so you can come bade to that place.
After moving the subject's hand to the target point, the experimenter
released the subject's hand and allowed three seconds of silence for
the subject to concentrate on feeling where his arm was. The experi-
menter then said:
low I am going to move your arm back to the starting place.
After returning the subject's arm to the starting point, the experi-
menter said:
low I want you to move your arm to the place where I left it for
a few seconds.
If the subject did not lift his pencil from the paper, he was instructed
to do so. Rill Instructions were repeated as often as necessary for
the particular subject. For each movement, the experimenter continued
to Identify verbally the starting place and the target point by saying:
Here is the starting place. Here is the target place.
Scoring
Previous studies have examined kinesthetic perception mainly
39
In the linear dimension. Ihe present study provided data from a two-
dimensional plane allowing examination of two parameters of kinesthe-
sia, perception of position and of movement. Of the two scores calcu-
lated, measurement of movement length represented a new way of looking
at the responses while assessment of position accuracy has been fre-
quently used in past studies.
Two scores were calculated for each item. As shown in Figure 3,
accuracy of position was computed by measuring the distance between
the subject's response point and the target point (actual magnitude
of the error or AME). The second score calculated was the length of
the response movement relative to the length of the target trajectory
(R/Traj). Measurement of the response trajectory is shown in Figure k.
o
These two scores were subsequently submitted to analysis of variance.
The actual magnitude of the error has been the most widely
used measure of accuracy of kinesthetic perception (Ayres, 1966;
Bahrick, Bennett, and FittB, 1955; Bahrick, Fitts, and Schneider,
1955; Cohen, 1958a; Cohen, 1956b; Msrton, 1961; Phillips, 1941;
Phillips and Summers, 195^; Roloff, 1953; Scott, 1955; Wettstone,
1938; Wiebe, 1954; Witte, 1962; tyke, 1965; Young, 19^5). Tne length
of the response trajectory relative to the length of the target
2
To increase the probability that the assumptions underlying
the analysis of variance would not be violated, the scores were exam-
ined to determine whether the raw or transformed data would give the
best fit to the model. Several transformations were suggested by the
physical model of the method and by previous research. Since it was
found that the original scores gave the greatest reduction of vari-
ance, the transformations were discarded. Appendix II contains a de-
tailed description of the transformations.
fco
TARGET
AME
RESPONSE
POINT
START
Fig. 3* —Measurement of the actual magnitude of the error.
Ul
START
Fig. h. — Measurement of the response trajectory.
U2
trajectory represents ..a new way of examining the individual respon-
ses. A ratio score was used because the length of response trajectory
is not Independent of the length of the target trajectory. The possi-
bility that the relative length of response movement would be an Im-
portant parameter to investigate was suggested by the clinical impres-
sion that congenitally blind persons are constricted in the use of the
space around the body.
Distribution of Scores
Since early qualitative observations by Wooster (1923) indi-
cated that subjects fall into two groups according to their ability to
utilize kinesthetic cues effectively, the mean scores of individuals
were examined to determine whether the distributions were normal. An
attempt was made to detect deviations from normality by constructing
a histogram for each group of subjects on the weighted and unweighted
trials and determining by inspection whether the distributions were
bimodal or unimodal. As shown in Appendix 111, the distributions did
not show any marked deviations from normality.
Research Design
It was expected that many factors would contribute to the
variance of the observed scores. The research design was balanced
to insure that the mean differences of the effects eliminated from
the analysis of variance would be zero.
It was expected that any given score would depend on the
following factors:
*3
1) visual condition of the subject, sighted or blind
2) experimental condition, weighted or unweighted
3) ordinal position of the trial
k) sex of the subject
5) hand being used, preferred or nonpref erred
6) characteristics of the target trajectory: length, quadrant of
origin, and direction.
Ordinal position of the trial was included to analyze practice effects
which occurred during the experimental session.
Of the above effects, the ones which were relevant to the
present research were: visual condition, experimental condition, and
trial number. Although the other effects— sex, hand, and trajectory-
needed to be balanced Into the design, it was not necessary to esti-
mate them for the purposes of the study. With the elimination of
these factors, the reduced model with interaction terms added was:
m(iqr)
where,
JIn words, the model Indicates that any observation is the sum
of the grand mean, (fA) plus the variance contributed by group member-
ship (Qt), condition (&), trial number (Y), and the interaction of
these effects (a/3, ftY , CXY , and CL/3Y ). IT describes the varia-
bility due to the particular subject.
H
i « 1, 2 sighted, blind
q ■ 1, 2 weighted, unweighted
r ■ 1, 2 . • . kO trial number
as ■ 1, 2 . . .20 subject number
Ihe term € Is the residual error which Is assumed to be uncorrelated.
Since the research design needed to be balanced to Justify
the assumption that the mean differsoces of the eliminated effects
(sex, hand, trajectory) were zero:
1) There were equal numbers of men and women In both sighted and
blind groups.
2) Preferred and honpreferred hands were used in equal numbers
of trials by all subjects. There was a standard order for the use of
the hands with change of hand occurring after every ten trials for
all subjects. The hand which was used first was determined randomly
for each subject with half the subjects in each group starting with
the preferred hand and half with the nonpref erred hand.
3) A standard series of trajectories was used for all subjects.
The trajectories for all subjects had the same distributions of
length, direction, and quadrant of origin. The order of presenta-
tion of trajectories was randomized for each subject to prevent se-
quence effects from being confounded with the main effects being
estimated (Winder, 1962).
The experimental condition which occurred first in the ses-
sion for any given subject was determined randomly with one half of
the subjects in each group starting with the weighted condition and
*5
one half with the unweighted condition. There were four possibilities
for the Initial trial:
1) unweighted, preferred hand;
2) unweighted, nonpreferred hand;
3) weighted; preferred hand;
k) weighted, nonpreferred hand.
In each group of twenty subjects, the four possibilities occurred in
the initial position randomly an equal number of times.
Eighty trials were administered to each subject, forty
weighted and forty unweighted. This number was selected after the
results of a pilot study (Lindley , unpublished data) suggested that
eighty responses are the maximum that can be made in an hour without
fatigue to the subject. Since the total number of trajectories used
was forty, each occurred twice in the series of eighty trials for any
given subject. The order of presentation of the forty trajectories
was randomized for each subject with the same random order of pre-
sentation occurring under both weighted and unweighted treatments for
any given subject.
A typical session is described in Tables 5 and 6.
Statistical Method
The data were analyzed by analysis of variance. The analysis
translated conceptual hypotheses of the study Into a number of statis-
tical hypotheses.
The analysis of variance used was one suggested by Winer
k6
5
S3 1
M
§
v4
e*
■P
h
3
O
is
8
u
u
e H
■5(5
+ ♦ + ♦♦♦♦♦ + + ■«'* + 4- + + + + + +
*^ *^ *w *o *^ *w *o *w ^3 *0
seseesssss
vvvvvvvvvv
B B B B£ £££ B B
cvj trwost t— c—co t-vo ct\
HWHwn cocvj cvj cvj
cvj co-* i/wo t— co o\o
CVJCVJCVJCVJCVJCVJCVJCVJco
'OtJ'tiTJia'd'iSTjTi'd
E
b
I
u
E
BB
I
8
u
u
u
81
(1
0)
V
V
0) D
V
0)
0)
U
r-i
'■M
«M
4h
<H f*
ft
u
ft
O
S
0)
»-l
B
b b b b
11
u
Pi
ft
a,
Pi ft
ft p.
Pi
ft
ft
a
o
c
o
a
a
C
o
c
o
gg
a
o
a
o
c
o
ksssssssssssssssss;
oo co h <?\--t -* c\J o\ o trv
CM CVJ or, OJonnH
H cvj co-* u-wo t— co CT\0
CO CO CO CO CO CO CO CO CO-^
+ ♦ + + ♦ + + + + + ++++++++++
'O *^3 t? *o *U ^3 '^ ^3 ^3 *o
BBBBB BBBB B
GEbliiiiiEIChh
<hV«*hViViViVi%hVi*h
BBBB£S£]jJiJ
(S££(£,(XpM(iifta<£
^ ^ ^ *^
BBBB
l< t. t, V,
ft ft ft ft
Sc c c
o o o
s z z: s
Sh (-, V, (h (h J-;
V ft) 1) 1) V 1)
ft ft ft 4-1 4-1 ^H
B B B B K
a & p< & a.
a c c a a
o o o
u
s sz; ss S5 S3 85
OWOr|WCONrlO\CO
CUcOHHcOCviH i-l
i-l CM CO^t ITNVO t-CO 0\O
a
co-* ltwo r— co on o
r-tHHHHr-lrHCVJ
•a.
V7
i i i i i i
*0 *0 *0 Tj *o *^ *u *^ *o 'o
EEEEEEEEEE
8 8 8 8 8 8 8 8 8 8
•«
T) T3 T3 T3 Ti T3 T5
73 tJ
E
888888
tj
ee
j* u
u
u, u u
u
4)
0) V
V
0) 0) u
«J
V V
<m
«M <*-(
u
M ^1 tl
<u
=H <W
e
88
p
u
8 £ 2
M
88
&
ft ft
ft
ft ft ft
ft
ft p<
c
d c
a
c a c
c
a a
o
o o
o
o o o
o
o o
ssssssBSsssssssas
■S o
2
ifNVO -* t- t-CO t—^O ON
CVJHCJCO COOJCVJCU
roooH^J'jw o\o i/\
cu c\i m cu mm H
5
h cu on-* trwo t— co ono
•H <M CO-* t/NVO t—CO ONQ
[— C- C— C— C— c- 1- 1- t~co
*W *w *^J ^3 *^ *0 *tj t^ rCj *^
8888888888
(IdHUtlVilltlliV
8888888888
£ftftft£aia,;i,<i,,i,
i ■ i
*w TJ *0 *^ *0 *^J *0 *w *0 "O
8S888S888S
8888888888
ftaftftftftftftftft
Saacecceca
ooooooooo
S|
£
3
cvj i/nvo ro ir\ o <n\o .*
ro rn _* h roH
OCUOr-IHCOt-.HCT\CO
wmHHmwH i-i
H CU O-* IAVD f-CQ ONO
-*-*-*-*-*-5^*2F-* ir\
H C\J CO-* 1/NV0 C— CO 0\0
UM*M/Mr\uM/\ir\ir\i/\vo
ltd
(1962) for the special case of a three-factor experiment with re-
peated measures on two of the factors. This analysis is especially
appropriate for experiments designed to study learning rates as a
function of experimental condition. Table 7 is a schematic repre-
sentation of the present design as a p x q x r factorial experiment
with repeated observations on the last two factors.
TABI£ 7
SCHEMATIC REPRESENTATION OF RESEARCH DESIGN
(AFTER WINER, 1962)
Treatment (q)
Weighted Unweighted
Bi Y
Group
(P)
.Trial
(r)
Cl C2 V ' ,C40
Cl C2 V ' 'ChO
Sighted
sl?lsl- • -sl
•
•
•
^O^O^O* * ,S20
si si sr • -si
S20S20S20* " *S20
Blind
*2
S S S . . .3
°1 1 1* ' * 1
S2 S2 V ' *S2
s s s . . .s
1 1 1 1
S S S . . .3^
2 2 2 2
S20S20S20* * *S20
S20S20S20* * *S20
In the experiment, there are p ■ 2 groups of n ■ 20 subjects with all
k
forty subjects observed under all eighty combinations of qr.
For details of computational procedures, see Winer, 19&2,
PP. 319-337.
*9
The analysis of variance translated the conceptual hypoth-
eses stated In the first chapter into seven statistical hypotheses.
These hypotheses vere tested Independently for the two scores, AME
and R/TraJ.
1)
<& .0
2)
<# -o
3)
*# -°
*)
<r* . 9
5)
(TftV-O
6)
Gjjff- 0
7)
<£?y«0
In words, these hypotheses tested the following relationships:
1) There is no difference between the mean scores of the sighted
and blind groups.
2) There is no difference between the mean scores under the dif-
ferent treatment conditions, weighted and unweighted.
3) There is no interaction between treatment condition and group
members hip.
k) No significant contribution is made to the total variance by
practice effect.
5) There is no interaction between practice effect and group
membership.
50
6) There Is no Interaction between practice effect and treat-
Beat.
7) There is no interaction among practice effect, group mem-
bership, and treatment.
Alpha was set at .05.
Estimation of Missing Data
It was necessary to estimate data for two of the subjects
for whom one of the eighty trials administered vas accidentally
omitted. The data were estimated for statistical purposes to meet
the requirement of the analysis of variance for equal numbers of ob-
servations in all cells. The subject with missing data was assigned
the mean of his group on the corresponding trial and treatment that
had been omitted in his experimental session. This mean was used as
the best estimate for taking account of hypothesized differences be-
tween groups, conditions, and trials. Data were estimated for
sighted subject one on unweighted trial six and for blind subject
nine on weighted trial fourteen.
C
CHAPTER V
HESULTS
Analysis of variance shoved no significant differences on
the position scores. On the movement scores, the blind were signi-
ficantly better than the sighted and there vas a significant change
In mean score for both groups combined from trial one to trial forty.
Since errors Increased rather than decreased with succeeding trials,
the observed practice effect could not be called learning. Analysis
of the effect of weighting on individual scores shoved the effect
among individuals vas random.
Position Scores
Table 8 gives the means and standard deviations of the posi-
tion scores for the sighted and blind subjects. Table 9 shows that
TABI£ 8
MEANS AND STANDARD DEVIATIONS OF POSITION SCORES FOR
SIGHTED AND BLIND SUBJECTS
Mean Score Standard Deviation
Group In cm in cm
Sighted 2.962 1.930
Blind 3.IU9 2.318
51
52
TABLE 9
THREE-WAY ANALYSIS OF VARIANCE: INFLUENCE OF VISUAL
CONDITION, WEIGHTING, AND PRACTICE
ON POSITION SCORES
Source of Variation
df
Mean
Square
Group Membership:
Sighted, Blind
27.863
0.579
Error
. 38
48.118
Experimental Condition:
Weighted, Unweighted
1
1.532
0.226
Group-Condition
Interaction
1
6.242
1.215
Error
38
6.785
Practice Effect
39
4.266
0.763
Group-Practice Effect
Interaction
39
6.605
1.181
Error
1482
5.590
Condition-Practice Effect
Interaction
39
2.473
1.031
Group-Condition -Practice
Interaction
39
2.783
1.160
Error
1482
2.398
none of the F-ratios calculated for the position scores was signifi-
cant at p < .05. Since no differences were found between main
effects and within interaction terms, none of the hypotheses stated
In the preceding chapter can be rejected.
53
Movement Scores
Table 10 gives the means and standard deviations of the
movement scores for sighted and blind subjects.
TAELE 10
MEANS AND STANDARD DEVIATIONS OF MOVEMENT SCORES
FOR SIGHTED AND BLIND SUBJECTS
Group Mean Score Standard Deviation
Sighted 1.069 0.175
Blind 0.962 0.185
As shown in Table 11, analysis of variance of the movement
scores gave significant F-ratios for two of the main effects. The
•ighted and blind groups were significantly different from each other
on this score. There was a significant practice effect from trial
one to trial forty on the movement scores. Since the F-ratios for
differences between weighted and unweighted conditions and for inter-
action terms were not significant, only statistical hypotheses one
and four can be rejected.
Sighted-Blind Difference
The vary large F-ratio of 18.138 for differences between
groups indicated that the alpha-error involved in rejecting hypoth-
esis one was very small. Although the difference in the mean scores
of I.069 and .982 of the sighted and blind groups respectively was
5h
TABLE 11
THREE-WAY ANALYSIS OF VARIANCE: INFLUENCE OF VISUAL
CONDITION, WEIGHTING, AND PRACTICE
ON MOVEMENT SCORES
Source of Variation
df
Mean
Square
F
Group Membership:
Sighted, Blind
1
6.124
l8.138b
Error
38
0.338
Experimental Condition:
Weighted, Unweighted
1
0.081
0.964
Group-Condition
Interaction
1
0.009
0.113
Error
38
0.084
Practice Effect
39
O.06Q
1.511*
Group-Practice Effect
Interaction
39
0.035
0.882
Error
1482
0.040
Condition-Practice Effect
Interaction
39
0.018
1.139
Group-Condition -Practice
Interaction
39
0.017
1.048
Error
11*82
0.016
p < .05
.01
not great in absolute terms, the large F-ratio was due to the small
standard deviations within groups. The difference between the two
55
groups Is quite clear in a graph of the mean movement scores by
trial as shown in Figure 5«
The blind subjects were significantly better than the sighted
subjects in reproducing movement length. The group means show that
the relative length of movement of the blind subjects was smaller
than that of the sighted group and more closely approximated the per-
fect score of one. The blind subjects characteristically made
smaller, more accurate movements than the sighted subjects did.
Practice Effect
Because there was a tendency for errors to increase rather
than decrease with succeeding trials as shown in Figure 6, signifi-
cant findings on practice effect cannot represent learning as had
been originally hypothesized. The linear regression of the means of
trials one to forty showed a positive slope of .0059 per trial with
a standard error of .0011. The t-value (5.36) is significant at
p < .001. The direction of change in the positive slope of the
line is away from the perfect score of one.
Influence of Weighting on
Individual Scores
Although the effect of weighting was not significant between
groups, the data were examined to determine whether any pattern in
the effect of weighting on performance of individuals could be de-
tected. The possibility that such a pattern might be found was sug-
gested by the test-retest reliability coefficients which were lower
56
O
O
O
I
R
o
to
9
O
VD
UJ
2 £
o 0 o o
3 5. o °\
O
00
0
•a
9
■P
■a
2
o •
o u
•P Tl
a u
4) <tf
S*P
B-a
s?
8S
■H _
•P t3
Jg +»
o
o o
1 1
a
P
s
1
57
O
UJ
1
o .
* 5!
- o
— r
o
3V00S 1N3W3A0IN NV3IN
58
than the split-half reliability coefficients. High split-half re-
liabilities for the weighted trials indicated that the effect of
weighting on performance within individuals was consistent, but
lower correlations between the weighted and unweighted trials sug-
gested that weighting affected performance among individuals in a
less consistent fashion. It was hypothesized that the less consis-
tent effect of weighting among individuals might have been due to
Identifiable differences among subjects. If a pattern of differen-
tial effect of weighting could be found, it would be useful in re-
habilitation in identifying persons who would benefit from the use
of weighting to improve kinesthetic perception.
Movement scores were examined to determine whether there was
any pattern to the way weighting affected scores within the sighted
and blind groups. Both position and movement scores were examined
to determine whether the effect of weighting differed according to
the type of error the individual made on the unweighted trials. Both
analyses showed that weighting affected performance among individuals
randomly.
Tables 12 and 13 show that change in mean length of movement
and average size of error on the movement scores was random within
the two groups. The chi-square one sample test is not significant
for either group at p < .05. The effect of weighting on the move-
Bent scores was also random regardless of the kind of error a subject
made on the unweighted trials— overshoot, undershoot, or approach
target lengths. Qualitative examination of Tables 1^ and 15 showed
59
TABUS 12
INFLUENCE OF WEIGHTING ON MOVEMENT SCORES: CHANGE IN MEAN LENGTH
OF MOVEMENT FOR SIGHTED AND BLIND SUBJECTS
WITH ADDITION OF A CONSTANT WEIGHT
Direction of Change
Group Increase* Decrease* Ho Change*
Sighted 8 5 7
Blind 6 8 6_
Total Ik 13 13
a >.025 mean per trial.
TABLE 13
INFLUENCE OF WEIGHTING ON MOVEMENT SCORES: CHANGE IN MEAN ERROR
FOR SIGHTED AND BLIND SUBJECTS WITH
ADDITION OF A CONSTANT WEIGHT
Direction of Change
Group Increase* Decrease* No Change*
Sighted 7 5 8
Blind __J ; k 9
Total Ik 9 17
* >.025 mean per trial.
that the effects of adding a constant weight were random even when
Individual characteristics in initial level of performance were con-
60
•idered. Among the blind subjects who overshot, Increase in size
of error from unweighted to weighted trials appeared to prevail, but
an examination of the individual scores showed the effect was an in-
consistent one because two of the subjects increased their errors by
lengthening their average movements and two by shortening their
average movements.
TABLE Ik
INFLUENCE OF WEIGHTING ON MOVEMENT SCORES: CHANGE IN MEAN LENGTH
OP MOVEMENT WITH ADDITION OF A CONSTANT WEIGHT FOR
SIGHTED AND BLIND SUBJECTS ACCORDING TO KIND
OF ERROR ON UNWEIGHTED TRIALS
Kind of Error
on Unweighted
Trials
Group
Increase
Direction of Change
Decrease* No
Change
Sighted
6
5
k
Overshoot
Blind
2
k
1
Total
8
9
5
Sighted
1
-
1
On Target
Blind
-
2
-
Total
1
2
1
Sighted
1
-
2
Undershoot
Blind
k
2
5
Total
5
2
7
>.025 mean per trial
'Because of small expected cell frequencies, it was not
Bible to run a chl-square test (S. Siegel, 1956).
pos-
61
TABLE 13
INFLUENCE OF WEIGHTING ON MOVEMENT SCORES: CHANGE IN MEAN ERROR
WITH ADDITION OF A CONSTANT WEIGHT FOR SIGHTED AND BLIND
SUBJECTS ACCORDING TO KIND OF ERROR
ON UNWEIGHTED TRIALS
Kind of Error
on Unweighted
Trials
Group
Increasea
Direction of Change
Decreasea No Changea
Sighted
6
5 5
Overshoot
Blind
1*
1 2
Total
10
6 7
Sighted
1
-
On Target
Blind
2
-
Total
3
-
Sighted
-
3
Undershoot
Blind
1
3 7
Total
1
3 10
a >.025 mean per trial.
Correlations between weighted and unweighted position scores
were high for the blind group and lower for the sighted group. As
shown in Table 16, adding a constant weight constricted differences
among the sighted group and attenuated differences among the blind
subjects by decreasing the errors of the best subjects and increasing
the errors of the poorest subjects. Results in the blind group
62
TABLE 16
INFLUENCE OF WEIGHTING ON POSITION SCORES: CHANGE IN MEAN SCORES
OF HIGHEST AND LOWEST SCORERS IN SIGHTED AND BLIND GROUPS
WITH ADDITION OF A CONSTANT WEIGHT
Mean Error on
Unweighted
Rank
Group
Trials
Direction
Amount
Highest
Sighted
4.52
Decrease
1.5
Scorers
4.025
Decrease
0.628
3.525
Increase
0.55
3.45
Decrease
0.575
Blind
5.7
Increase
1.275
4.2
Increase
0.555
4.075
Increase
0.555
3.65
Decrease
0.555
Lowest
Sighted
2.45
Increase
0.025
Scorers
2.1
Decrease
O.025
2.05
Increase
0.175
2.025
Increase
~ 0.55
Blind
2.35
Decrease
0.1
2.325
Increase
0.425
2.275
Increase
0.80
2.125
Decrease
0.75
mean per trial.
63
suggested that adding a constant weight had a differential effect
on a subject's performance according to whether his errors were
large or small relative to his group. However, Table 17 shows that
increases and decreases in error with addition of a constant weight
were distributed randomly among high and low scorers.
TABLE 17
INFLUENCE OF WEIGHTING ON POSITION SCORES: DIRECTION
OF SCORE CHANGE AMONG HIGH AND LOW SCORERS
IN SIGHTED AND BLIND GROUPS WITH
ADDITION OF A CONSTANT WEIGHT
Direction of High Scorers Low Scorers
Change in
Average Error Sighted Blind Total Sighted Blind Total
Increase 5 5 10 7 6 13
Decrease 5 5 10 3 k 7
CHAPTER VI
DISCUSSION
While some findings of the study were expected, others had
not been predicted by the conceptual hypotheses. The finding that
neither group improved on either measure of kinesthetic perception
with addition of a constant weight vas an expected result. Failure
of either group to show learning with practice vas unexpected and
the finding that accuracy of movement scores declined with succeed-
ing trials was surprising. The latter finding most likely resulted
from lack of feedback. Superiority of the blind subjects on accur-
acy of reproducing movement length was a predicted result. Although
the finding of no difference between groups on accuracy of duplica-
ting position had not been predicted, the results of group compari-
sons on both measures of kinesthetic perception support the conclu-
sion that the standard of reference in the body schema for measur-
ing movement and position is not visual when the judgment does not
Involve cognition of space outside the body.
Influence of Visual Condition
The predicted superiority of the blind subjects on kines-
thetic perception was observed only in accuracy of reproducing move-
ment length. The blind and sighted subjects were equally accurate in
duplicating position. Both findings indicated that the standard for
6>»
6
65
measuring position and movement is not visual when the Judgment does
not Involve cognition of space outside the body.
The unexpected finding that the two groups compared differ-
ently on the two measures suggested that movement and position consti-
tute Independent elements of kinesthetic perception. The superiority
of the blind subjects in reproducing movement length suggested that
the blind are more accurate in using kinesthetic perception as a
regulatory mechanism. The complementary hypothesis is that vision
assumes the regulatory function In sighted adults. Failure of the
blind subjects to show superiority in duplication of position was
Interpreted as due to the ability of the sighted subjects to use
visual Imagery in performance of the task. Comparisons of the re-
sults of the present study with results of maze learning studies sug-
gested that vision plays a role in perception of extrapersonal space,
but Is not a necessary component of the spatial model of the body.
Accuracy of Movement
As predicted, the blind were significantly more accurate than
the sighted in reproducing length of arm movement. The difference in
performance between the sighted and blind groups on movement length
Indicated that the blind utilize kinesthetic information more accur-
ately to reproduce movement. A possible interpretation of this find-
ing is that the blind make better use of kinesthetic information as
a regulatory mechanism. If the dominant regulatory mechanism in
sighted adults is visual, they would perform less well when deprived
66
of vision and made dependent on kinesthetic sensation to regulate
movement. This finding supports Gibson and Mowrer's (1938) theory
that postural cues are genetically prior, but visual cues become
dominant in sighted adults.
The finding in the present study that the movements of blind
subjects were more constricted than those of the sighted confirmed
the clinical impression that congenitally blind persons are constricted
In use of space around the body. It also agrees with Jastrov's (1SS6)
findings of constriction In movement by blind subjects. The extent
of constriction reported by Jastrow was not observed in the present
study. His finding that the "motion-inch" of the blind is about one-
half inch may have been due to his method.
Accuracy of Position
Contrary to expectation, there was no difference between
sighted and blind subjects in accuracy of reproducing arm position.
The average error observed in the study was greater than the expected
error in kinesthetic perception and was thought to be due to the diffi-
cult nature of the task of recognizing a stimulus. Failure to find
a difference between groups may have been due to the ability of the
sighted subjects to use visual imagery to compensate for lack of
sensitivity to kinesthetic cues possessed by the blind subjects.
The average errors for the sighted and blind groups of 2.96
Jastrow had subjects learn distances by feeling pegs set on
the inch marks of a ruler. He then had them produce arm movements
of various lengths from verbal direction.
67
and 3*15 cm respectively were consistent with average errors reported
in other studies (Ayres, 1966; Cohen, 1958a; Cohen, 1956b). The con-
sistency across studies on the average size of error in reproducing
arm movements raised the question of whether the observed magnitude
of error is the expected error in perception of kinesthetic sensation.
Errors in perception of just-noticeable-differences of up and
down Joint movement in a group of subjects under age forty were re-
ported by Laidlaw and Hamilton (1937) • Although the up and down
movements are not strictly equivalent to those which occurred in the
present study, they are the best data available to indicate an order
of magnitude of the expected error in perceiving change of limb posi-
tion. Errors in perception of up and down movement in the shoulder,
elbow, and wrist range from .100 to .556 cm. This order of magni-
tude is much smaller than the error found in the present and previous
studies of reproduction of arm position. Even considering that the
error involved in a movement of the whole arm was compounded by the
error at the different joints, the total still is not great enough
to account for the observed error in these studies.
The nature of the task used in the present study might ac-
count for the discrepancy between observed and expected error.
Gerhard (1968) suggested that recognition of a stimulus, as opposed
'The findings for shoulder, elbow, and wrist were transformed
from degrees as reported by Laidlaw and Hamilton with the formula for
calculating the length of a chord subtending an angle (= 2rsin 0/2)
where r ■ the length of the appendage involved in the movement and
0- size of angle. Appendage length was estimated by taking measure-
ments from one man and one woman to obtain a range of magnitude of ex-
pected error.
68
to discrimination of a stimulus, constitutes a more difficult task,
I.e., the expected error is larger. While the method used in the
studies under discussion Is not strictly equivalent to that by which
Gerhard defined the task of recognition, his model is useful In inter-
preting the results of the research. Gerhard proposed that when one
stimulus can be compared to another stimulus as in the discrimination
task, the amount of information transmitted relative to the number of
response choices is much greater than that transmitted when the stimu-
lus is compared to a subjective standard as in recognition. The
amount of Information that can be used by a person reaches a plateau
quickly even though the number of response choices may increase as
occurs when one moves from discrimination to recognition of stimuli.
Thus, the accuracy of response reaches a plateau beyond vhlch the
Individual cannot use the greater amount of Information available in
the recognition task. In other words, a person is unable to recog-
nize finer categories of input.
The findings of Slinger and Horsley (1906) support Gerhard's
idea that recognition and discrimination are different perceptual
tasks. Comparing a group of twenty-one adolescents and young adults
blind before age four with a group of twenty-five sighted individuals
on a discrimination task of kinesthetic perception, the authors found
that the blind were more accurate than the sighted and that the
average error of the blind subjects was smaller than that observed
in the present study. They reported an average error of 1.3 cm in
8
In the method used by Slinger and Horsley, the examiner
69
eight blind subjects and 2.3 cm in the sighted group for locating arm
position in the vertical plane at waist level at finger tip, wrist,
9
and elbow extensions. For locating arm position at the same level
in the horizontal position they reported an average error of 2.k cm
10
among thirteen blind subjects and 3. 3 cm among the sighted group.
Die average error for both series for the blind group was 2.1 cm
and for the sighted group, 2.8 cm.
Die question remaining to be answered is why there should be
differences between sighted and blind subjects on recognition and
discrimination of body position. This question is part of the larger
problem of reconciling the different findings of studies comparing
sighted and early blind subjects. The findings show that blind sub-
jects are more accurate on reproduction of movement and on discrimina-
tion of position (Slinger and Horsley, 1906). However, blind and
sighted subjects are equally accurate on recognition of position. In
contrast, sighted subjects are more accurate than blind subjects in
placed the subject's arm on the target on a glass plate. The subject
localized the position of his arm by pointing to the other side of
the plate with his other arm.
a
These figures were calculated from the original figures re-
ported by Slinger and Horsley of 20.7/2 cm for the blind group and
27.8/2 cm for the sighted group. They identified the figures as the
sums of the average errors of the groups. It was assumed that the
sums included six averages, extension (3) x hand (2) since this cal-
culation gave the average error for the sighted group for this posi-
tion that the authors reported elsewhere in the article.
10
The figures were calculated from the original figures re-
ported by Slinger and Horsley of 28.8/2 cm for the blind subjects and
39.6/2 cm for the sighted subjects according to the method described
in the previous footnote.
70
learning mazes from body cues (Duncan, 193**; Koch and Ufkesa, 1926;
Worchel, 1951).
The different findings Indicate that the role of vision In
perception of body sensations depends on the nature of the task.
One possible interpretation suggests that vision is not a standard
of reference for recognizing isolated stimuli in simple perceptual
tasks, but it is a necessary element in tasks requiring complex in-
tegration of cues in maze learning tasks. However, this interpre-
tation does not satisfactorily account for differences in blind
subjects between recognition and discrimination of body position.
A more satisfactory interpretation suggests that when the
task requires cognition merely of body cues without regard to space
outside the body, vision is not a necessary component. However,
when cognition of space outside the body is required in the task,
vision plays an Important role in interpreting body cues.
The latter Interpretation reconciles the findings of the
various studies mentioned. It is reasonable to assume that repro-
duction of movement involves a regulatory function of klnesthesis
and that discrimination of limb position involves sensitivity to
kinesthetic cues without regard to space outside the body. In the
case where cognition of space outside the body is not involved, the
blind perform better than the sighted because the blind, who are de-
pendent on body cues for orientation, are more sensitive to pro-
prioceptive stimuli.
In recognizing limb position, the blind lose their
71
superiority. Assuming that tfc^jy retain the sensitivity to kines-
thetic cues demonstrated In discrimination of position and reproduc-
tion of movement, their loss of superiority compared to sighted sub-
jects can he accounted for by the ability of the sighted subjects to
use visual Imagery to compensate for their lack of sensitivity to
body cues. It is possible for the sighted subjects to compensate in
this manner because recognition of position involves relating the
limb to space outside the body. In the maze learning task, vhlch
definitely requires cognition of space outside the body, the sighted
gain superiority over the blind by their capacity to use visual
Imagery to interpret body cues.
The preceding interpretation suggests the hypothesis that
vision is not a standard of reference In the spatial model of the
body, but it is the standard of reference for perception of extra-
personal space. When the task requires primarily recognition of
extrapersonal space as In maze learning, the blind are at a disad-
vantage, but vhen performance of the task depends primarily on the
spatial model of the body, they are not.
The Interpretation that the difficulty for the congenitally
blind person lies in orientation to extrapersonal space rather than
to his own body has Important implications for rehabilitation prac-
tices. Findings from studies of maze learning and body perception
in the blind suggest that the observed lack of facility in the use
of the body by the congenitally blind person is due to problems in
relating himself to extrapersonal 6pace. Rehabilitation should
72
emphasize spatial orientation rather than body awareness. An inter-
esting area for further research is investigation of the hypothesized
differences in perception of personal and extrapersonal space.
The findings of the present study that the blind subjects
were better in reproduction of movement than the sighted subjects
and that the blind were as accurate as the sighted in recognizing
limb position indicates that vision is not a necessary component of
the body schema. The findings lead to the conclusion that kinesthetic
and tactual clues are as efficacious as vision in forming a spatial
aodel of the body. Ihe observation that the blind are better than
the sighted in reproduction of movement supports Gibson and Mowrer's
(1938) proposal that vision becomes the standard of orientation
among sighted adults.
Influence of Weighting
As predicted, adding a constant weight did not alter subjects'
accuracy in duplicating movement and position. The finding, which
agreed with previous studies, suggested the conclusion that adding a
constant weight is not a useful rehabilitation practice. Examination
of the effect of weighting on Individual performance showed that add-
ing a constant weight, which did not alter group performance, also did
not alter performance of individuals in any predictable way.
The finding that adding a constant weight did not alter
accuracy of kinesthetic perception in a predictable way is congruent
with similar studies of sighted persons. Bahrick, Fitts, and Schneider
73
(1955) found that adding a constant spring loading to a control stick
did not alter subjects' accuracy In reproducing circular and triangular
movements. Cohen (1956b) reported that adding a constant velght sig-
nificantly Increased the size of error made in reproducing arm posi-
tion. His method differed from that used in the present study in
that the target point was first localized without the weight. The
size weight he used (one kilogram) was probably large enough to In-
crease error through mechanical fatigue.
The results of studies which show that adding a constant
weight does not improve accuracy of kinesthetic perception indicate
that the practice of using constant weights In rehabilitation may
not be useful. A study by Bahrick, Bennett, and Fltts (1955) sug-
gest* that a more sophisticated approach is needed. In a study of
accuracy of positioning a horizontal arm control by sighted persons,
the authors found that alteration of kinesthetic input improved ac-
curacy of response if the input was changed relative to the size of
movement to be made. An area for further research would be to deter-
mine whether similar results would be found with blind persons.
Influence of Practice
Contrary to expectation, learning did not occur during the
experimental session. Practice effect was not significant for the
position scores. Although it was significant for the movement scores,
change was in the direction of increased error and cannot be charac-
terized as learning.
lh
The expectation that learning would occur was based on the
findings of Renshaw, Wherry, and Newlin (1930 ) who found the subjects'
ability to localize tactile sensations improved from session to session.
Improvement in Renshaw, Wherry, and Newlin 's study may have been due
to the fact that the task itself provides knowledge of results which
were lacking in the present study. Thorndike (19^0) pointed out that
knowledge of results rather than mere repetition is necessary for
learning to occur. Several studies (Cole, 1929; Thorndike, 1927;
Trowbridge and Cason, 1932) support the validity of this notion.
Cole (1929) compared two methods of locating tactile stimuli on the
skin and found that when a subject received knowledge of the results
of his response, he improved in accuracy more than when he did not
receive such knowledge. In the method where subjects received no
knowledge of results, a glass plate was inserted over the skin before
the subject responded so he had no basis for comparing his response
to the original stimulus. In the knowledge of results method, sub-
jects responded by touching the skin which presumably gave them a
basis for comparing accuracy of response to the original stimulus.
In the present study, subjects received no knowledge of re-
sults. This could account for the observation that they did not im-
prove in accuracy during the testing session. An area for further
study would be to determine whether knowledge of results improves
accuracy of kinesthetic perception and what kind of knowledge of
results is most useful in improving accuracy. Slinger and Horsley
(1906) found that simply telling a person whether he was right or
75
wrong was not helpful in Improving accuracy of kinesthetic perception
as corrective movements were only rarely made in the right direction.
In a study of accuracy in reproducing a three-inch line, Trowbridge
and Cason (1932) found that telling the person the size and direction
of his error led to much greater improvement in accuracy than simply
telling him whether he was right or wrong. This finding suggests that
a similar type of knowledge of results would be required to improve
accuracy of kinesthetic perception.
The quite unexpected decline in performance on movement
scores must be accounted for. There are several possible explana-
tions for the findings that errors increase with succeeding trials.
The most plausible explanation is that increased error results from
an actual loss of accuracy of perception when a long series of move-
ments is made without feedback or knowledge of results. Alternative
explanations attribute increased error to fatigue and to variability
in test items.
Contradictory observations have been made of the effects of
fatigue on kinesthetic perception. One study (Slinger and Horsley,
1906) ascribed subjects* loss of accuracy in perceiving kinesthetic
sensation within a session to fatigue. Another author (Cohen, 1958a)
felt that fatigue did not affect subjects* ability to localize kines-
thetic sensations. Both observations are qualitative and details of
the changes in performance during sessions were not reported in
either study. In the present study, subjects' spontaneous comments
suggested that fatigue in the form of loss of ability to concentrate
76
might account for decrement in performance, but a similar phenomenon
was not observed on the position scores. Since it does not seem
likely that failure to concentrate would affect position and move-
ment accuracy differentially, fatigue does not represent a good ex-
planation of the observed change in means across trials.
Another factor which might account for the observed change
is muscular or "mechanical" fatigue (Bahrick, 1937)* If mechanical
fatigue were the cause of the observed decline in performance, errors
should have been more pronounced in the weighted trials which re-
quired greater muscular effort. Although the weighted and unweighted
trials did not show a significant interaction with practice effect, a
graph of the means on trials one to forty for both conditions was
examined to see if any difference between the two conditions could
be detected. Since the elopes of the means of the two conditions
appeared to be parallel as shown in Figure 7, mechanical fatigue did
not appear to be an adequate explanation of the observed decline in
performance on the movement score .
Decline in performance could be an artifact of the instrument
used. Decline could have resulted from variability in the items
rather than from variability in the subjects. This suggestion is
based on the observation that decline in performance occurred only
in the ratio scores. As shown in Table 18, practice effect and its
interactions were significant only on the scores in which the
"^Significant interaction between weighting and practice
effect on AME/Traj and (AME/Traj)2 scores had raised the possibility
that a similar trend might be detected on the movement scores.
77
4-0
: a
uj
K
3:
o
g
a:
o
s
1
5:
-O
— r
O
o
o
3UO0S 1N3A0IN NV3W
o
t
I
9
as
?i
■is
£9
78
trajectory length appeared. Since there was no change from trial
o
one to trial forty on the AMB and (AME) scores, the mean absolute
value of the error was constant from trial to trial. When trajectory
length waa included in the AME/TraJ and the (AME/Traj) scores, the
practice effect and/or its interactions showed significant decline
In performance. Given a constant absolute error, the increased
■cores could be caused by a decrease in the size of the trajectory
length. If the error remained constant, but the trajectory length
became shorter, the error relative to the trajectory length would
increase.
IABI£ 18
F-RATIOS OF PRACTICE EFFECT AND ITS INTERACTION
TERMS FOR RAW AND TRANSFORMED SCORES
Score
Practice
Effect
Practice x
Weighting
F-Ratios
Practice x
Group
Practice x
Group x
Weighting
AMB
•763
1.031
1.181
I.I60
(AME)2
.849
.803
1.218
1.318
AME/TraJ
1.432ft
1.409*
1.046
I.O69
(AME/Traj)2
1.376
1.526*
1.159
1.004
R/Traj
1.5lla
1.139
.882
1.048
(R/Traj)2
1.494a
1.223
.900
1.141
p < .05
79
A decrease In the average trajectory length could also account
for the Increasing error observed in the movement scores. One study
(Brown, Knauft, and Rosenbaum, 19W) showed that subjects reproducing
movements on a slider overshot the smaller trajectories and undershot
the larger ones. If a constant error of this sort vere affecting
performance in the present study, the decline in accuracy could be
due to a decrease in the average length of the target trajectories
from trials one to forty. However, the randomization of order of pre-
sentation of trajectories makes unlikely any significant correlation
of trajectory length with trial number.
Decline in the movement scores may represent a true loss of
ability to interpret kinesthetic perceptions when a long series of
movements is made without feedback or knowledge of results. The
fact that such a decline did not occur on the position scores sug-
gests that this explanation is not wholly adequate. Use difference
between the two scores may underscore the independence of the two
measures. The difference between scores may also be due to a lack
of precision in the position scores. If the size of the AMB is de-
pendent on the length of the trajectory as shown by previous research
(Brown, Khauft, and Rosenbaum, 19**6)> a true increase in error size
could be masked by the randomization of trajectory length. The find-
ing that the AME/TraJ and (AME/Traj) scores did show such an in-
crease in error supports this explanation. To evaluate these explana-
tions more fully it would be necessary to perform another study com-
paring changes within a session with and without knowledge of results.
CHAPTER VII
SUMMARY
The question of whether the standard against vhich body move-
ment and position are measured is visual or not was first raised by
Head (1920). For the present study, the relationship between vision
and perception of limb position and movement was approached by study-
ing kinesthetic perception in blind adults who had had light percep-
tion or less since early childhood. The blind subjects were compared
to an equal number of sighted adults whose kinesthetic perception
was measured while they were blindfolded. Since the sighted group
could use visual imagery in the performance of the experimental task
and the blind group could not, inferences could be drawn about the
role of vision in the body schema, the mechanism by which body sen-
sations are perceived.
Subjects were required to reproduce waist-height arm move-
ments introduced by the examiner, who moved the subject's arm along
a standard series of trajectories in a two-dimensional field. Sub-
jects were scored for accuracy both in reproducing the length of the
movement and in locating the target or end-point of each trajectory.
Three factors were analyzed in the study. The factor of
primary concern was vision and its effect on perception of movement
and position. Secondary factors examined were the effects on per-
80
81
ception of practice and of altering input by adding a constant weight.
Results of the study indicated that tne standard for measur-
ing movement and position is not necessarily visual, that learning
does not occur vrciiln one session without feedback, and that adding
a constant weight does not alter accuracy of perception. The first
conclusion was based on the findings of no difference between
sighted and blind groups in accuracy of reproducing position and of
significantly greater accuracy of blind subjects in reproducing move-
nent length. The conclusion that learning does not occur resulted
from the findings of no change from trial to trial on position
scores. The significant change from trial to trial on movement
scores represented a decline in accuracy which was thought to result
from lack of feedback. The conclusion that altering input does not
improve accuracy of perception was drawn from the finding that add-
ing a constant, weight did not change accuracy of reproducing move-
ment length or position within either group.
Comparisons of sighted and blind subjects on accuracy of
locating limb position and of reproducing movement length indicated
that the standard against which position and movement are measured
is not visual when the Judgment does not involve cognition of space
outside the body. The superior performance of the blind on repro-
duction of movement length supported Gibson and Mowrer's (1938)
theory of ocular dominance in sighted adults and suggested that
vision may function as a regulatory mechanism in movements of
sighted adults. Equal accuracy of sighted and blind subjects in
82
reproduction of position suggested that sighted subjects used visual
imagery to compensate for lack of sensitivity to kinesthetic cues
possessed by the blind. Comparison of the present findings with
results of studies of maze learning in which congenitally blind per-
sons performed more poorly than sighted persons suggested that vision
is not a necessary element in forming a spatial model of the body,
but Is necessary in constructing a model of extrapersonal space.
APPENDICES
AFPEN DEC I
The following letter was sent to prospective blind subjects
to obtain their cooperation with the project.
Dear
I am writing to ask for your assistance on a research project I
am doing in cooperation with the Cleveland Society for the Blind
Sight Center. The purpose of the study is to help us understand
how people without vision coordinate body movements. In order to
examine this problem, I will be studying simple arm movements.
Your contribution to the project is Important because problems
of mobility are significant in rehabilitation of blind individuals.
I hope that the results of the study will increase our understand-
ing of methods used in rehabilitation. I intend to use the results
of the study in my doctoral dissertation at Case Western Reserve
University so your participation will also contribute to my pro-
fessional training.
I will especially appreciate your individual assistance in the
study because I can get the best understanding of the problem
through participation of persons who have been blind since early
childhood. The results of your particular contribution will be
kept confidential.
The study will take about one hour of your time. It will be
most helpful if you can come to the Sight Center at East 101 and
Chester. I can pay you for your transportation expenses. If
you cannot come to the Sight Center, I will be glad to make
other arrangements with you.
I will call you in a few days to discuss your interest in
assisting in the project.
6k
APPENDIX II
To Increase the probability that the assumptions underlying
the analysis of variance vould not be violated, the scores were
examined to determine whether the raw or transformed data vould
give the best fit to the model. Several transformations were sug-
gested by the physical model of the method and by previous research.
It was found that the original scores gave the greatest reduction
of variance.
The following scores and their transformations were examined
to determine which would give the greatest reduction of variance:
1) AME (actual magnitude of the error)
2) AME2
3) AME/TraJ (actual magnitude of the error relative to the
trajectory length)
k) (AME/Traj)2
5) R/Traj (response trajectory length relative to target
trajectory length)
6) (R/Traj)2.
Although the usual transformation is the square root rather
than the square of the original observation (Winer, 19^2 )> the
physical model of the present method suggested that the square of
the score would be the more appropriate transformation. In a two-
dimensional space, the area within which a response can fall
85
86
Increases by the square of the magnitude of the distance. In the
case of the AME, the area encompassed by an error of any given
magnitude Is actually a circle as shown in Figure 8. The area en-
compassed within an error of any given magnitude equals 1T r where
Y is equal to (AME) . Since 7T is a constant, it was dropped from
the transformation. In the case of R/Traj, the area encompassed
by a response is 1/2 7T r as shown in Figure 9. Since all subjects
were observed to move in the direction of the target rather than away
from it, the area encompassed by any given response trajectory was a
semicircle. Since l/2 TT represents a constant, it was droppe''. from
the transformation forsula.
A ratio of the actual magnitude of the error to the length
of the target trajectory was also calculated. This transformation
was based on previous experimental work on kinesthetic perception
in the arm which indicates that the magnitude of the error is not
independent of the length of the movement made in positioning the
arm (Brown, Khauft, and Rosenbaum, I9I+8; Ronco, 1963).
A computationally simple method of selecting the best trans-
formation is the use of the range statistic (Winer, 1962). The goal
of the transformation is to make the variance more uniform. The
range statistic, which tends to be proportional to the variance,
gives an indication of which transformation gives the greatest re-
duction of variance.
The range of the standard deviations on each of trials one
to forty was calculated for sighted subjects on weighted trials;
87
2 cm error
I cm error
area of A+ IT (l cm)2
2) drea of B , 7T (2 cm)'
Fig. 8.— Model of transformation of position score by-
squaring.
88
4 cm response trajectory
area 1/2 IT (4 cm)2
5 cm response trajectory
area 1/2 IT (5 cm)2
squaring.
Fig. 9. — Model of transformation of movement score by
89
sighted, unweighted trials; blind, weighted trials; and blind, un-
weighted trials by subtracting the smallest standard deviation from
the largest in each category. The resulting figure was a range of
column standard deviations which gave an estimate of variability
across subjects as well as across trials. The uniformity of the
four ranges of standard deviations within a given transformation
was estimated by calculating the ratio of the smallest to the larg-
est. This ratio made the various measures comparable in spite of
absolute differences in size and is similar to the coefficient of
variation, V( ■ C/M), which indicates the variability among measures
relative to their average size (Scott and Wertheimer, 1962). Since
the most uniform ranges of standard deviation were found in the
original scores as shown in Table 19, data from these two scores
were submitted to further analysis.
ON
90
CO
r4
■p
■a
■H
%
a
S
ON VO ON
vo ir\ cvl
*
o
OJ
c—
a
E* O
v
-p
■a
-p
d
VO
S
00
-p
-a
00 H O
C— t«- -o
t» H OJ
a
CO
on
OJ
S. > 5
APPENDIX III
The following histograms give the distributions of mean
scores for Individuals by group, experimental condition, and kind
of score. Chapter IV contains a more detailed discussion of these
figures.
91
92
o g
Is
2i
i i — i — i — i — — i i > — -»
1.0 2.0 3.0 4.0 5.0 6.0
MEAN POSITION SCORE cm
Fig. 10. — Distribution of mean position Bcores of sighted
Subjects on weighted tria?s (N f 20).
93
6
o jo
§ to
lO 2.0 3.0 4.0 5.0
MEAN POSITION SCORE , cm
6.0
Fig. 11.— Distribution of nean position scores of sighted
subjects on unweiehted trials (N = 20).
9*
o jo
2-\
1.0 2.0 3.0 4.0 5.0
MEAN POSITION SCORE > cm
6.0
Fig. 12 — Distribution of mean position scores of blind
subjects on weighted trials (ll = 20).
95
o jj
g I/)
2
i i i i — i — i —
1.0 2.0 3.0 4.0 5.0
MEAN POSITION SCORE M cm
6.0
Fig. 13.— Distribution of mean position scores of blind
subjects on unweighted trials (N = 20).
96
t 1 1 r
~l
.8 .9 1.0 I.I 1.2
MEAN MOVEMENT SCORE
Fig. lU. — Distribution of mean movement scores of sighted
subjects on weighted trials (N = 20).
97
04
a 3
.8 .9 10 /./ /.a
M5AW MOVEMENT SCORE
Fig. 15* — Distribution of mean movement scores of sighted
subjects on unweighted trials (N = 20).
98
.8 .9 1.0 U 1.2
MEAN MOVEMENT SCORE
Fig. 16. — Distribution of mean movement scores of blind
subjects on weighted trials (N = 20).
O K
ft*
|w2
99
.6
.9
1.0
I.I
1.2
MEAN MOVEMENT SCORE
Fig. 17. — Distribution of mean movement scores of blind
subjects on unweighted trials (N « 20).
REFERENCES
Attneave, F., and M. D. Araoult. 1956. The quantitative study of
shape and pattern perception. Psychol. Bull. 53 '- 452-471.
Ayres, A. Jean. I966. Southern California kinesthesia and tactile
perception tests: manual. Western Psych., Beverly Hills,
Calif. 48 p.
Bahrick, H. P. 1957* An analysis of stimulus variables influencing
the proprioceptive control of movements. Psychol. Rev. 64:
324-328.
Bahrick, H. P., W. F. Bennett, and P. M. Fitts. 1955. Accuracy
of positioning responses as a function of spring loading in
a control. J. Exp. Psychol. 49:437-444.
Bahrick, H. P., P. M. Fitts, and R. Schneider. 1955. Reproduction
of simple movements as a function of factors influencing
proprioceptive feedback. J. Exp. Psychol. 49:445-454.
Bailey, A. A., and F. P. Moersch. 1941. Phantom Limb. Can. Med.
Ass. J. 45 :n-k2.
Bartley, H. S., L. T. Clifford, and A. D. Calvin. 1955 . Effect of
visual imagery of tactual and kinesthetic 6pace perception.
Fere. Mot. Skills. 5:177-184.
Bass, Ruth I. 1939* An analysis of the components of tests of
semicircular canal function and of static and dynamic
balance. Res. Quart. 10:33-52.
Bender, M. B., M. A. Green, and M. Fink. 1954. Patterns of percep-
tual organization with simultaneous stimuli. Arch. Neurol.
Psychiat. 72:233-264.
Bennett, D. H. 1956. Perception of the upright in relation to body
image. J. Msnt. Sci. 102:467-506.
Bennett, D. H. i960. The body concept. J. Went. Sci. 106:57-75«
Berg, J., and P. Worchel. 1956. Sensory contributions to human
maze learning: a comparison of matched blind, deaf, and
normals. J. Gen. Psychol. 54:61-93.
100
101
Bitterman, M. E., and P. Worchel. 1953* The phenomenal vertical
and horizontal in blind and sighted subjects. Araer. J.
Psychol. 66:598-602.
Boring, E. G. 1942. Sensation and perception in the history of
experimental psychology. Appleton, New York. 644 p.
Bora, E. 1951* Phantom limbs of patients with spinal cord injury.
Arch. Neurol. Psychiat. 66:610-631.
Brieland, D. M. 1950. A comparative study of the speech of blind
and sighted children. Speech Monogr. 17:99-103.
Brown, E. T., and P. L. Gotein. 1943. The significance of body
image for personality assay. J. Nerv. Ment. Dis. 97:
401-4o6.
Brown, J. S., E. B. Khauft, and G. Rosenbaum. 1948. The accuracy
of positioning reactions as a function of their direction
and extent. Amer. J. Psychol. 6l:l67-l82.
Browne, K. , J. Lee, and P. A. Ring. 1954. The sensation of pas-
sive movement at the metatarso-phalangeal Joint of the great
toe in man. J. Physiol. 126:448-456.
Buell, C. 1950. Motor performance of visually handicapped children-
New Outlook Blind. 44:256-253.
Bychowski, G. 1943. Disorders in the body-image in the clinical
pictures of psychoses. J. Nerv. Want. Dis. 97:310-334.
Cohen, L. A. 1956. Analysis of position sense in humans. Fed.
Proc. 15:38.
Cohen, L. A. 1958a. Analysis of position sense in human shoulder.
J. Neurophysiol. 21:550-562.
Cohen, L. A. 1958b. Contributions of tactile, musculo-tendinous
and Joint mechanisms to position sense in human shoulder.
J. Neurophysiol. 21:563-568.
Cole, L. E. 1929. The localization of tactual space: a study of
average and constant errors under different types of local-
ization. Genet. Psychol. Monogr. 5:339-450.
Critchley, M. 1950. The body image in neurology. Lancet. Jan.-
June:335-340.
Critchley, M. 1953. The parietal lobes. Arnold, London. 480 p.
102
Drever, J. 1955* Early learning and the perception of space.
Amer. J. Psychol. 68:605-6l4.
Duncan, Bertha K. 1934. A comparative study of finger-maze learn-
ing by blind and sighted subjects. J. Genet. Psychol, 44:
69-94.
Federn, P. 1952. Ego psychology and the psychoses. E. Weiss (ed. )
Basic Books, New York. 375 P«
Fenichel, 0. 1945 • The new psychoanalytic theory of neuroses.
Norton, New York. 703 p.
Fink, M. , and M. B. Bender. 1953* Perception of simultaneous tac-
tile stimulation in normal children. Neurology. 3:27-3^.
Fisher, S., and S. E. Cleveland. 1953- Body image and personality.
Van Nostrand, Princeton, N. J. 420 p.
Freud, S. 1957* The ego and the id, p. 210 to 235* In J. Rickman
(ed. ) A general selection from the works of Sigmund Freud.
(Transl. from German) Doubleday, Anchor, Garden City, N. Y.
Fulcher, J. S. 1942. "Voluntary" facial expression in blind and
seeing children. Arch. Psychol. 36: (272): 1-49.
Gardner, G. 19^3 • Fundamentals of neurology. Saunders, Phila-
delphia. 3^9 p.
Gerhard, W. 1968. The study of sensation in physiology, p. 1643
to I67I. In Mountcastle, V. B. (ed.) Medical physiology.
Vol. II. Mosby, St. Louis.
Gerstmann, J. 1942. Problem of imperception of disease and of
impaired body territories with organic lesions. Relation
to body schema and its disorders. Arch. Neurol. Psychiat.
U8:690-913.
Gibson, J. J. 1952. The relation between visual and postural de-
terminants of the phenomenal vertical. Psychol. Rev. 59 '•
370-375.
Gibson, J. J., and 0. H. iMowrer. 1938. Determinants of the per-
ceived vertical and horizontal. Psychol. Rev. 1+5:300-324.
Haber, W. B. 1955. Effects of loss of limb on sensory functions.
J. Psychol. 40:115-123.
Harris, C. S. I963. Adaptation to displaced vision: visual,
motor, or proprioceptive change? Science. I40:8l2-8l3.
103
Head, H. 1920. Studies in neurology. Vol. II. Oxford, London.
862 p.
Head, H. 19^3. Aphasia and kindred disorders of speech. Vol. I.
Hafner, New York. 1+30 p.
Head, H., and G. Holmes. 1911. Sensory disturbances from cerebral
lesions. Brain. 34: 102-254.
Hebb, D. 0. i960. The American Revolution. Amer. Psychol. 15:
735-7^5.
Henderson, W. R., and G. E. Smyth. 1948. Phantom limbs. J.
Neurol. Neurosurg. Psychiat. 11:68-112.
Henry, F. M. 1953 • Dynamic kinesthetic perception and adjustment.
Res. Quart. 24:176-167.
Hunter, I. M. L. 1954. Tactile -kinesthetic perception of straight-
ness in blind and sighted humans. Quart. J. Exp. Psychol.
6:149-154.
Jastrow, J. 1666. The perception of space by disparate senses.
Mind. 11:539-554.
Jastrow, J. 1901. Fact and fable in psychology. Houghton, New
York. 375 p.
Knotts, Josephine R. and W. R. Miles. 1929. The maze-learning
ability of blind compared with sighted children. J. Genet.
Psychol. 36:21-50.
Koch, Helen L., and Jennette Ufkess. 1926. A comparative study of
stylus maze learning by blind and seeing subjects. J. Exp.
Psychol. 9:118-131.
Kolb, L. C. 1954. The painful phantom. Thomas, Springfield, 111.
50 p.
Kolb, L. C. 1959a« The body image in the schizophrenic reaction,
p. 87 to 97. In A. Auerback (ed.) Schizophrenia: an
integrated approach. Ronald, New York.
Kolb, L. C. 1959b. Disturbances of the body-image, p. 749 to 769.
In S. Asieti (ed.) American handbook of psychiatry. Vol.
I. Basic Books, New York.
Laidlaw, R. W. and Mary Alice Hamilton. 1937. A study of thresh-
holds in apperception of passive movement among normal con-
trol subjects . Bull. Neurol. Inst. 6:268-273.
104
Lewis, Helen B. 1958. Over-differentiation and under-indlvidua-
tion of the self. Psy choanal. Rev. 45:3-23.
Linn, L. 1955* Some developmental aspects of the body image. Int.
J. Psychoanal. 36:36-42.
Lowenfeld, B. I963. The visually handicapped. Rev. Educ. Res.
33:38-41.
Lovenfeld, V. 1939. The nature of creative activity. Harcourt,
New York. 272 p.
MacDonald, Joanne Catherine, i960. An investigation of body scheme
in adults with CVA. Amer. J. Occ. Ther. 14:75-79.
Machover, Karen. 1949 • Personality projection in the drawing of
the human figure. Thomas, Springfield, 111. l8l p.
Merton, P. A. I96I. The accuracy of directing the eyes and the
hand in the dark. J. Physiol. 156:555-577.
Morgan, C. T. 1965« Physiological psychology. McGraw, New York.
687 P.
Nielsen, J. M. 1938. Distrubances of the body schema: their
physiologic mechanism. Bull. L. A. Neurol. Soc. 3:127-135.
Norris, Miriam, Patricia J. Spaulding, and Fem H. Brodie. 1957.
Blindness in children. U. Chicago, Chicago. 173 P«
Oldfield, R. S., and 0. L. Zangwill. 1942. I Head's concept of
the schema and its application in contemporary British
psychology. Brit. J. Psychol. 33:58-64.
Phillips, B. E. 194l. The relationship between certain phases of
kinesthesis and performance during the early stage of acquir-
ing two perceptuo-motor skills. Res. Quart. 12:571-586.
Phillips, Marjorie, and D. Summers. 1954. Relationship of kines-
thetic perception to motor learning. Res. Quart. 25:45b-
469.
Provins, K. A. 1958. The effect of peripheral nerve block on the
appreciation and execution of finger movements. J. Physiol.
143:55-67.
Renshaw, S., R. J. Wherry, and J. C. Newlin. 1930. Cutaneous
localization in congenitally blind versus seeing children
and adults. J. Genet. Psychol. 38:239-247.
105
Renshaw, S., R. J. Wherry, and J. C. Nevlin. 1931. Studies on
cutaneous localization: III The age of onset of ocular
dominance. J. Genet. Psychol. 39:493-496.
Roloff, Louise L. 1953* Kinesthesia in relation to the learning
of selected motor skills. Res. Quart. 24:210-217.
Ronco, P. G. 1963. An experimental quantification of kinesthetic
sensation: extent of arm movement. J. Psychol. 55:227-238.
Ruch, ?. C. 1965. Somatic sensation, p. 302 to 317. In T. C.
Ruch and H. D. Patton (ed.) Physiology and biophysics. 19th
ed. Saunders, Philadelphia.
Scheerer, M. 1954. Cognitive theory, p. 91 to 142. In G. Lindzey
(ed.) Handbook of social psychology. Addison-Wesley,
Cambridge, Mass.
Schllder, P. 1950. The image and appearance of the human body:
studies in the constructive energy of the human psyche.
(Transl. from German) Int. Univs., New York. 353 p.
Schlaegel, T. F. 1$53« Tbe dominant method of imagery in blind as
compared to sighted subjects. J. Genet. Psychol. 63:265-277.
Scott, M. Gladys. 1955 • Measurement of kinesthesis. Res. Quart.
26:324-341.
Scott, W. A. and M. Wertheimer. 1962. Introduction to psychological
research. Wiley, New York. 445 P'
Senden, M. von. i960. Space and sight. (Transl. from German.)
Free Press, Glencoe. 348 p.
Siegel, I. M. 1966. Posture in the blind. Amer. Found. Blind Res.
Ser. 5(l5):l-39-
Siegel, S. 1956. Non parametric statistics for the behavioral sci-
ences. McGrav, New York. 312 p.
Simmel, Marianne L. 1956a. On phantom limbs. Arch. Neurol.
Psychiat. 75:637-647.
Simmel, Marianne. 1956b. Phantoms in patients with leprosy and in
elderly digital amputees. Amer. J. Psychol. 69:529-545.
Slinger, R. T. and V. Horsley. 1906. Upon the orientation of points
in space by the muscular, arthrodial, and tactile senses of
the upper limbs in normal individuals and in blind persons.
Brain. 29:1-28.
106
Smith, I. U., and W. M. Smith. 1962. Perception and motion: an
analysis of space structured behavior. Saunders, Phila-
delphia. 3^1 p.
Smythie6, J. R. 1953* The experience and description of the human
body. Brain. 76:132-1^5.
Stratton, G. I896. Some preliminary experiments on vision with in-
version of the retinal image. Psychol. Rev. 3:6ll-6l7.
Szasz, T. S. 1957. Pain and Pleasure: a study of bodily feelings.
Basic Books, Hev York. 301 p.
Teuber, H. i960. Perception, p. 1595 to 1668. In Field, J. (ed.)
Handbook of physiology. Vol. III. Amer. Fnysiol. Soc,
Washington.
Thomdike, E. L. 1927. The law of effect. Amer. J. Psychol. 39:
212-222.
Thorndlke, E. L. 19^0. Human nature and the social order.
MacMillan, Hew York. 1019 p.
Traub, A. C. and J. Orbach. 190*. Psychophysical studies of body-
Image. Arch. Gen. Psychiat. 11:53-66.
Trowbridge, Margery Hayden and H. Cason. 1932. An experimental
study of Thorndike's theory of learning. J. Gen. Psychol.
7:21*5-258.
UHman, M., E. M. Ashenhurst, L. J. Hurwitz, and A. Green, i960.
Motivational and structural factors in the denial of hemi-
plegia. Arch. Neurol. 3:306-316.
Wapner, S., and H. Werner. I965. An experimental approach to body
perception from the organismic -developmental point of view,
p. 9 to 25. In S. Wapner and H. Werner (ed.) The body per-
cept. Random, New York.
Wettstone, E. 193d. Tests for predicting potential ability in
gymnastics and tumbling. Res. Quart. 9:115-125-
Wiebe, V. R. 195i*. A study of tests of kinesthesis. Res. Quart.
25:222-230.
Winer, B. J. 1962. Statistical principles in experimental design.
McGraw, New York. 672 p.
107
Wltkin, H. A. I965. Development of the body concept and psychol-
ogical differentiation, p. 20 to Vj\ In S. Werner and H.
Werner (ed. ) The body percept. Random, New York.
Witte, Fae. 1962. Relationship of kinesthetic perception to a
selected motor skill for elementary school children. Res.
Quart. 33:^76-484.
Wooster, Margaret. 1923* The development of a new spatial coordin-
ation. Psychol. Monogr. 32(l46) :l-96.
Worchel, P. 1951. Space perception and orientation in the blind.
Psychol. Monogr. 65 (332): 1-28.
Wyke, Maria. 1965- Comparative analysis of proprioception in left
and right arms. Quart. J. Exp. Psychol. 17:14-9-157.
Young, Olive G. 1945. A study of kinesthesis in relation to
selected movements. Res. Quart. 16:277-267.
SmYSl
ETC
■H
■
$$8$
■HH
■:'■■.■■.■-.,:■.■
■;''•■■ v ■'■■ '
«;'•••■■. 'v . whs
■ * • ' ' ■ I ' :- ■•.■■ ■,''"■■ ■ ■-
'■'''' ;' '"'■■■ '■ ■
■■'■''•''"'.-.',•'■
V;' 'V'/V-, : ': ' $
' ' ' '■ I I ■'''■''■.',-.■ Jw.
■ * v ' QJfB Wat'
?.!SkAf wife' ^tfSSSB i7v^3L%m