University of Southern California Dissertations and Theses

An Application Of Adaptation Level Theory To The Response Bias Of Falsification

(USC Thesis Other)

An Application Of Adaptation Level Theory To The Response Bias Of Falsification

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content This dissertation has been 64— 3103
m icrofilm ed exactly as received
PALMER, Theodore B ernard, 1931—
AN APPLICATION OF ADAPTATION LEVEL
THEORY TO THE RESPONSE BIAS OF FALSI
FICATION.
University of Southern California, Ph.D., 1963
Psychology, general
University Microfilms, Inc., Ann Arbor, Michigan
C o p y r ig h t b y
THEODORE BERNARD PALMER
1964
AN APPLICATION OF ADAPTATION LEVEL
THEORY TO THE RESPONSE BIAS
OF FALSIFICATION
by-
Theodore Bernard Palm er
A D issertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In P artial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Psychology)
August 1963
UNIVERSITY O F S O U T H E R N C A L IFO R N IA
G RADUATE SC H O O L
U N IV ER SITY PARK
LOS A N G E L E S 7. C A L IF O R N IA
This dissertation, written by
.................
under the direction of h...S.3.Dtssertation C om
mittee, and approved by all its members, has
been presented to and accepted by the Graduate
School, in partial fulfillment of requirements
for the degree of
D O C T O R OF P H IL O S O P H Y
.............
« ^ Dean
D ate A u S U S .t... 3 1 #....1 9 .6 3 .................
DISSERTATION COMMITTEE
«-» / * C h a irm a n '
* >
A C K N O W LE D G E M E N TS
I would like to express my gratitude to the many students
who served as the subjects for this study.
Also, I would like to thank the following individuals for their
sound advice, particularly during the planning phases of the study:
D rs. W. B. Michael, J. P. Guilford, W. H. W erkm eister,
A. Jacobs, and L. E. Longstreth.
When all is said and done, of course, the unsung heroes
can be none other than the people who invent, and who rep air,
machines such as the IBM 101 C ounter-Sorter, and its many
unusual relatives.
Last, and not least, I thank my very patient wife, M ildred.
CO N TEN TS
P ag e
ACKNOWLEDGEMENTS.......................................................................... ii
LIST OF TABLES.. . .................................................................................. v
LIST OF ILLUSTRATIONS..................................................................... x
Chapter
I. BACKGROUND OF THE STUDY..................................... 1
II. DEVELOPMENT OF HYPOTHESES, AND
OPERATIONS............................................................................. 19
in . THE Q UESTIO NNAIRE, SA M PL E , AND TESTING 34
IV. PRELIMINARY ANALYSES................................................. 46
V. MAIN FINDINGS........................................................................... 60
VI. SUPPLEMENTARY ANALYSES........................................ 84
VII. DISCUSSION AND SUMMARY............................................ 104
BIBLIOGRAPHY............................................................................................ 127
APPENDICES
Appendix A : Comments regarding the Matching
P rocess............... 133
Appendix B : Procedures relating to Indices Substituted
for College Entrance Examination Board
Scores........................................................................... 135
Appendix C : Copy of Note Requesting Participation
of Subjects Not Present on Day the Study
had been Conducted.............................................. 137
P ag e
Appendix D: Introductory Remarks to Subjects.............. 138
Appendix E : GZTS Items which Differentiated Subjects
as to Amount of Shift........................................ 140
Appendix F : Booklet Number in GZTS, and Type of
Scale on GZTS, of T est-item s Appearing
in Experimental Booklet. ............... 143
Appendix G: Test Form I: Introduction, Ratings,
and Rankings........................... 146
Appendix H : Test Form II: Instructions fof: Different
Experimental G roups.......................................... 154
iv
LIST OF TABLES
T a b le P a g e
1. Means, Sigmas, and Sex Ratios of Matching
Variables, by Group. ........................................................... 40
2. Intercorrelations of Matching V ariables.............................. 42
3. Indices of Reliability of T rait-scales, for
Individual Groups............................................................................... 48
4. Indices of Reliability for Groups Requested to
Falsify, Taken Together, and for Groups Not
Requested to Falsify, Taken Together, by Individual
T rait-S cales......................................................................................... 49
5. Reliability of Differences between Means of Independ
ent M easures : Within- and Between-Trait Compar
isons........................................................................................................... 51
6. Reliability of Differences between Means of Ranked
Importance of T raits....................................................................... 55
7. Reliability of Differences between Means of GZTS
T rait-S cales......................................................................................... 57
8. Self-Ratings of Actual Positions on T rait........................... 59
9. Relationship between Individual’s Self-Rating of
Actual Position on T rait and Amount of Shift on
Same T rait .......................................................... ^2
v
T a b le P a g e
10. Relationship between Individuals Intensity of
Adaptation to T rait and Amount of Shift on same
T rait.......................................................................................................... 65
11. Relationship between Importance of T rait to Individ
ual Subjects and Amount of Shift on same T rait............. 66
12. Reliability of Differences between Coefficients of
Correlation between Position on, Value of, and
Intensity of Adaptation to P airs of T raits and the
Amount of Shift on respective Criterion T raits, in
connection with Average, Personal Importance of
T raits.......................................... 68
13. Sets of Groups Compared on Experimentally Mani
pulated Variables, for Testing Hypotheses 6, 7, and 8. 72
14. Means, Sigmas, Medians, and Sem i-interquartile
Ranges of Number of C orrect Responses to T rait-
scales, by Individual Groups................................................... 73
15. Critical ratios for Mann-Whitney test of Reliability
of Differences in Amount of Shift on T raits, between
Contrasting Groups.......................................................................... 74
16. Critical ratios for Mann-Whitney test of Reliability
of Differences in Amount of Shift on T raits, between
P airs of Individual Groups.......................................................... 75
vi
T ab le P ag e
17. (A) : Relationship between Self-ratings of Desired
Position on Trait and Amount of Shift on same Trait. 79
(B): Relationship between Individual*s Judgment of
Adults1 Actual Position on Traits and the Individual's
Amount of Shift on same T rait.................................................. 80
(C) : Relationship between Individual's Judgment of
Adults’ Desired Position on Trait and the Individu
al's Amount of Shift on same T rait........................................ 81
18. Results of Critical ratio Tests of Reliability of
Differences between Coefficients of Correlation
between Position on, Value of, and Intensity of
Adaptation to Pairs of T raits and the Amount of
Shift on respective Criterion Traits, in Connection
with Average Importance of Traits to Adults................... 83
19. Relationship between the Amount of Difference
between Types of Position on Trait, and the Amount
of Shift on same T rait.................................................................... 86
20. Relationship between the Direction of Difference
between Types of Position on Trait, and the Amount
of Shift on same T ra it. .......................................................... 87
21. Relationship between the Amount of Difference
between Position on Trait and Tolerance Limits for
Trait, and the Amount of Shift on the same T ra it.. . . “ 89
vii
T a b le P a g e
22. Relationship between Matching Variables and Amount
of Shift on T raits.................................................................................. 91
2 3 . Relationship between Matching Variables and Im por
tance of T rait, Position on T rait, and Intensity of
Adaptation to T rait............................................................................. 93
24. Reliability of Differences between M ales1 and
Females* Judgments of Importance of T raits to Self
as Compared with Importance to Adults............................. 96
25. Relationship between CEEB and Judgments of Males
and Fem ales, as to Relative Importance of T raits.
Relationship between AGE and Judgments, of Males
and Fem ales, as to Relative Importance of T raits. . . 97
26. Relationship between Intensity of Adaptation to T rait
and the Judged Importance of and Position on the
same T rait.................... 99
27. Relationship between Importance of T rait and Judged
Position on same T rait................................................................... 100
28. Relationship between Judgments of Adults' Actual,
Adults' Desired, Own Actual, and Own Desired
Positions on given T raits............................................................. 102
29. Intercorrelations of the C riterion M easures.................... 109
v iii
T a b le
30. Standard Deviations of Importance of Trait, Position
on T rait, and Intensity of Adaptation to Trait, for
Males Compared with Females, and for Groups
Requested to Falsify Compared with Groups Not
Requested to Falsify........................................................................
P a g e
110
ix
LIST OF ILLUSTRATIONS
C h a r t Page
1. Model for Experimental Manipulation of Groups of
Subjects, in Two Types of Falsification Tasks............ 32
2. Contents of T rait-scales. Type and Quantity of
Item in Each Task............................................................................... 36
x
C H A P T E R I
BACKGROUND OF THE STUDY
The quality of interpretation, of em pirical findings, often
rise s or falls in conformance with the proportion of known to un
known factors in a body of data. Unwanted factors may, sim ilarly,
influence the generality of findings. As the rules and demands of
the social sciences grow more rigorous, there is increasing re c
ognition and concern with factors--som e unknown, most unwanted - -
which lim it the specificity of interpretation, and the generality,
of otherwise well-designed studies.
High on the list of these factors are the "response biases. "
In the context of any questionnaire, a biased response is one which
has been "altered in such a way that it indicates something other
than [what] we intended it to m easure. " (20) These responses are
generally thought of as the product of mental sets which may p e rsist
throughout a questionnaire, often in spite of the content of particular
test item s. Among these would be the set to disagree, the set to
gamble, and the set or tendency to acquiesce. With increasing
frequency, investigators refer to the presence of biased responses
as they qualify the conclusions of their experiments--including
experiments which have made use of the better validated and more
readily interpretable testing instrum ents.
In response to this situation, attempts have been made in
recent years to identify more precisely the'varied forms of response
bias. Research has, in fact, proliferated in this area since the
m id-fifties. Its central concern has been that of demonstrating
the existence, the generality, and the stability of the phenomena.
During this period, however, no general, yet rigorous
theoretical structure was emerging which might begin to subsume
1
and predict such phenomena. Together with the availability of
little normative data from which to draw regarding the extent to
which specific conditions were associated with differing response
b iases--th is situation led investigators to rely alm ost totally upon
inference, or else upon ad hoc explanation, as they attempted to
(a) attach new names to response patterns which appeared to be
biased, and as they (b) speculated on the nature of the underlying
factors which might account for the relationships found between
these presumed biases and the variables involved in their own
research . A single theory--that of Loevinger*s--has appeared
only recently.
Two levels of need thus exist. The general need is that of
system atizing, to some extent, the numerous independent research
endeavors within the area of bias. The related and more specific
need is that of clarifying the interpretation of. particular types of
bias. The strongly worded closing statement in Loevinger's 1959
review of theories and techniques of assessm ent, points to the
same needs:
Proliferation of tests of high-sounding psychological constructs
in disregard of response bias is conspicuous waste of research.
(37)
The aim of the present study is to determine whether a p a r
ticular theory--one which has been generating research in several
areas of psychology--holds prom ise of meeting these needs.
The widely recognized response bias of falsification will be
singled out for system atic investigation. Together with that of
"gambling, " this is the oldest of all biases; it had been recognized,
in the context of testing, long before Cronbachts classic statement
of the general problem of bias, in 1946. (7) Its influence is
reflected in recent statem ents such as;
The opportunity for willful biasing of responses and the fact
that examinees take advantage of the opportunity has been the
m ost serious criticism of personality inventories. The problem
is not a simple one by any means. (21)
Beginning as early as Steinmetz1 study of falsification on the
Strong Vocational Interest Blank, in 1932, the ability of individuals
to alter their scores on self-report inventories has repeatedly been
documented. In 1946, E llis' review of twenty-five investigations
concluded with an affirm ative answer in all but three instances to
the question of whether falsification had taken place, in a variety
of experimental situations involving self-report instrum ents. (14)
However, the occurrence of falsification, once demonstrated,
was generally treated as self-evident--as requiring little further
explanation--both in these twenty-two studies and in later investi
gations. In consequence, specific findings which were part of these
studies--such as differences within a given sample as to relative
^No hard and fast definition of falsification was adhered to
in these studies. The meaning, generally, was tacitly assumed to
be identical with the common-sense concept of m isrepresentation.
A handful of investigators further stipulated that this could occur
either consciously or unconsciously. But the m ajority felt that
the w isest course was to think of falsification as "prim arily the
effect of a conscious attempt on the p art of the subjects to influence
their scores. " (32) This view, stated by Kelly, Miles, and
Term an, took into account the fact th at--in the typical investigation--
subjects (a) would be rath er definitely requested to falsify, or
change, their scores, and (b) would be given either a general or
a specific idea about the manner in which they were to falsify.
amount of falsification--seldom received special comment, or sep
arate follow-up in later studies. Concern was not with pursuing
the theoretical leads which had developed in earlier investigations.
It centered, instead, upon the direct threat to the efficacy of specific
self-report inventories. As a result, the typical investigation was
prim arily an attempt to determine whether, or under what conditions,
various traditional as well as more current personality inventories
2
would be susceptible to falsification.
The following are representative of the type of term s and con
cepts which were most often used to account for the occurrence of
falsification. "Defensive attitude" (58), "test-sophistication" (l),
"desire to cover up defects" (45), "oblivescence of the disagree
able" (32), "seeing oneself through rose-colored glasses" (34),
"rationalization" (42), "repudiation" (48), "underlying tendency
to conform" (55), and so forth. Whereas some of these were
essentially restatem ents or descriptions of the process of falsifica
tion, others referred to postulated antecedents to that process;
and in still others, the distinction between description and explana
tion remained uncertain.
The conceptual base for research in this area changed but
little from the thirties through the late fifties; and, during this latter
period, the differences among conceptualizations did not outweigh
the sim ilarities. In retrospect, the following appears to stand out
in connection with the earlier period: In the main, the overall
approach to the problem lacked rigor. With few exceptions, individ
ual concepts such as the above would bear the burden of explanation.
2
Many questionnaires have been looked at, in this context.
Included, among others, are the Humm-Wadsworth Temperament
T est, B ernreuter Personality Inventory, Thurstone*s Neurotic
Inventory, Allport*s Ascendance-Submission Reaction Study,
Minnesota Personality Scale, California Psychological Inventory,
Jurgensen Classification Inventory, Minnesota Teacher Attitude
Inventory, Strong Vocational Interest Blank, Kuder Preference
Record, Guilford-Zimmerman Temperament Survey, Minnesota
Multiphasic Personality Inventory, Edwards* Personal Preference
Schedule, and the California F test.
5
Hypotheses, of a testable nature, were not derived from underlying
and organizing principles. There appeared only one article whose
main concern was with the implications of the logically possible--
and intuitively recognized--differing functions of falsification. (51)
All in all, relatively little interest was apparent in the literature,
in te rm s of ongoing dialogue.
The m ajority of explanations which focused upon a single mo
tive or circum stance, and which implicitly included assertions as
to the functional relationship between antecedents and consequents,
may be classified as follows:
1, Falsification aimed at gratification
a. Self-deception for purpose of ego enhancement:
b. Deception of others for purpose of obtaining or maintaining
ego enhancement from others;
c. Deception of others for purpose of obtaining or maintaining
m aterial gain.
2. Falsification aimed at relief
a. Distortion (directed at self) of attributes perceived as
deficiencies;
b. Misinforming of others, to avoid censure.
(Either a minimization of unwanted characteristics or an exaggera
tion of desired characteristics, or both, may be involved in any
category found in this classification. )
Investigators typically would note that one or both of the follow
ing had been found in their studies: (a) all subjects did not falsify,
and (b) all subjects did not falsify to the same extent. These conclu
sions were supported by studies conducted under conditions of real-
life motivation ( 2,10, 18,29 ) as well as by those--considerably
m ore numerous--conducted in the classroom or laboratory, where
subjects would be requested to answer questions in the way they
imagined certain, specified individuals would answer them, and
3
where differing rew ards would generally prevail.
-
Included among the more recent of these latter studies would
be those of Cross ( 8 ), Dicken (9), Herzberg (30), Kimber (33),
6
Despite there being little disagreem ent with the above findings,
or perhaps "because" of it, the fact that there were individual differ
ences among those subjects who would falsify, did not appear to
arouse strong interest among the more theoretically inclined inves
tigators. This state of things was nicely reflected in a comment,
m ade--w ith reference to a research finding--alm ost twenty years
after the Steinmetz study:
. . .there is probably a continuum of the degree or extent to
which each individual will attempt to or (and) succeed in faking.
. . . To what extent does ability to fake successfully vary with
degree of possession of the tra it? Is it positively correlated?
Here is an area of research which needs a very definite answer
which can readily be obtained. (18)
These questions were an outcome of the fact that little had
been done to pursue--w hether system atically or otherw ise--specific
findings, and logical distinctions, such as the following--(even to
pursue them in term s of the relatively general, single-motive
explanations summarized above):
Kelly, M iles, and Term an, in 1935, found that on the Stanford
T est of Mental M asculinity-Femininity, male subjects shifted their
scores significantly farther than female subjects in the direction of
femininity, w hereas female subjects shifted farther in the direction
of the opposite pole. (32) Kimber, using the California Test of
Personality, found, in 1945, that males and females differed in the
degree to which they portrayed, upon instruction, a "happy and well
adjusted [college] student"; it was concluded that the latter carried
out the request better than the form er. (33)
Rosenzweig had suggested that part of the within-sample,
test-sco re variation might spring from differing aims on the part
Longstaff and Jurgensen (38)* Mais (41), Noll (46), Quiggle (48)»
Rabinowitz (49), and Sundberg (56). In both the form er and the
latter types of study, the subjects would be given directions which
ranged from broadly stated requests to the effect that the subject
make a ’good overall im pression, * through to explicit instructions
as to the type and degree of trait-to-be-conveyed. The following,
abridged instructions would be illustrative of the middle part of this
continuum:
I want you to help me to find out how wanting or trying to be a
of subjects: Whereas some individuals could be responding in term s
of "actual traits" ("what a man thinks he is" or "what a man is"),
others might be responding in term s of "ideal traits" ("what a man
thinks he would like to be" or "what a man would like to be"). (51)
H eidbreder!s earlier study lent credence to this distinction.
She had shown that self-ratings of one*s actual degree of possession
of a tra it "may be regarded as constituting a very different reaction
from that involved in ratings which avowedly indicate preferences. "
(25) Heidbreder*s report contained another significant facet.
Implicit in her discussion of the study was the earliest suggestion
of a specifiable functional relationship between the occurrence of
falsification and certain clearly operationalized variables--in this
case, actual and preferred degrees of possession of traits. In
pointing out that the subjects had preferred the trait of extraversion
to that of introversion, she added that--according to self-ratings--
they (a) were more likely to believe they possessed those tra its
which they also preferred to possess, and that, at the same tim e,
(b) the subjects who considered themselves the most introversive
(and the ones who felt they had the most "inferior attitude") were
those m ost likely to feel there was the greatest discrepancy between
their actual and preferred positions on these traits. Studies which
might have pursued such leads were not forthcoming, however.
Yet, whether phrased more generally or phrased more in term s
of specified variables, neither the gratification hypotheses nor the
relief hypotheses appeared capable of representing the findings of
these investigators. Their findings, for one thing, appeared closely
linked with subject-characteristics which cut across the factors
involved in the above types of explanation. In addition, studies such
certain way or to belong to a certain type, such as a masculine
or a feminine type, can influence the score. . . . make yourself
test as masculine as you possibly can. Mark each item in the
manner you imagine a hard-boiled, extremely masculine he-m an
would m ark it. (32)
as Goldstein’s (17), which dealt with malingering, and Kimber’s
(33), which related to subjects’ differential "insight" into the
falsification-potential of given item s--these made it increasingly
clear that it would be well to at least supplement the explanatory
concept of differential desire to falsify, with a cognitive factor, viz. ,
differential ability to falsify.
Progress in the direction of greater conceptual breadth, as
w ell as rigor, was evident in Adams’ exhaustive review, in I960,
of the numerous "response tendencies. " ( 1 ) In reviewing the more
recent work^ done on "the problem of fakability of self-report
inventories, " she pointedly spoke in term s of functional relation
ships between antecedents and consequents, even though the nature
of the relationships as well as the dimensions of the antecedent
5
conditions were spoken of in broad term s. Adams distinguished
five "factors affecting the probability and effectiveness of faking. "
These were (a) motivation to fake, (b) test-w iseness, (c) tra n sp a r
ency of the inventory, (d) examinee’s awareness of correction scores,
and (e) attitude of the examinee. (Factor 'e ’--which will assume
greater significance in later ch ap ters--referred to findings that
pointed up how subjects who, for example, were already fairly high
on ’tolerance, ’ would, for that very reason, find it difficult to "fake
good" [tolerance] . )
This list--w hich included conative as well as cognitive dimen
sions--brought new conceptual problem s to the fore, and pointed up
the older, unresolved ones. For one, it raised the possibility that
relation ships of theoretical importance might exist between the five
factors, and, more generally, between the conative and cognitive
dimensions. Also, there was the question of which conditions might
in turn account for any one--or several--of these five factors.
For instance, in regard to either the within- or the betw een-factor-
_
Eighteen studies, fifteen of which took place between 1954 and
1959.
5
For example, "The m ore highly motivated examinees are,
the m ore likely they are to modify their responses to test item s. "
relationships, it would be plausible to raise the question of whether
any or all of these factors could be thought of as expressions, in
turn, of other, underlying processes. It would be conceivable, for
example, that item s could be more "transparent” (*c*) to individuals
who have--in connection with greater motivation to falsify (*aI) as
well as greater opportunity to do so (*e*) --developed a greater
sensitivity to questions which can be interpreted as bearing upon
attitudes and activities of special significance to them.
Another major source of the increasing number of studies
relating to falsification, in the later fifties, was the interest gener
ated by Edwards* work on the social desirability of traits. (12)
The basic finding had been that a strong, positive relationship existed
"between judged desirability of a tra it and the probability that the
tra it will be endorsed. " (11) The basic implication was that the
perceived social desirability of an item , or of a tra it, would also be
related, differentially, to the likelihood that the item or tra it would
be falsified. A study by Lunneberg, of fourth, fifth, and sixth grade
youngsters, in 1961, gave specific support to this implication. (40)
Beginning with the earliest investigations of falsification, a number
of studies--using older subjects--have come to substantially the
same conclusion. (24, 33, 38, 48, 52, 59)
From Edwards* point of reference (12,p. 53ff. ), falsification
was conceptualized as follows:
Without concern for the distinction between conscious and un
conscious distortion, it may be noted that a subject*s responses
may be falsified in such a way that he obtains either a higher or
a lower score on a particular variable than he would if his
responses were completely accurate. . . , If a subject fakes
good, then we may expect him to obtain a higher score on a
scale than he otherwise would, if the variable measured by the
scale is itself a socially desirable one, that is, if the keyed
responses are socially desirable responses.
Faking good on personality inventories, without special instruc
tions to do so, I would consider equivalent to the tendency to give
socially desirable responses in self-description.
10
If a su b jects responses under standard instructions are already
prim arily influenced by considerations of social desirability,
special instructions to attend to social desirability should not
result in any marked shifts in his responses. (Italics supplied. )
If no change occurs in scores obtained under standard instruc
tions and under instructions to fake good, this [also] does not
necessarily mean that faking was not taking place under standard
instructions. It could equally well be interpreted as meaning
that the tendency to give socially desirable responses was influ
ential to such an extent under standard instructions that special
instructions to fake good could not in any way increase the tend
ency. Failure to find shifts in scores under the two conditions
of adm inistration, however, would surely seem to be a more
unusual resu lt than to observe that the shifts did occur. No
m atter what the result, however, we can only conclude that
studies of faking good have little or no bearing upon the problem
of whether or not scores obtained under standard instructions
have been distorted by the tendency to give socially desirable
responses.
As a resu lt of Edwards1 discussion, and despite its attempt at
distinguishing social desirability and falsification, investigations of
a type that had previously been conceptualized and phrased in term s
of falsification, often would now appear in a context of social d esir
ability. This was in large m easure the result of the latter concepts
being sufficiently broad to cut across those aspects of both the
gratification and the relief hypotheses which had relevance to individ
uals other than oneself. In the context of self-report inventories,
it had particular relevance to Gratification Hypothesis lb --
"deception of others for purpose of obtaining or maintaining ego
enhancement from others. " The influence of social desirability
could also be considered a m ajor source of Adams1 factor *a*--
"motivation to fake--which appeared to be one of the main factors
in her classification.
Throughout the period of increasing interest in social d esir
ability, it was recognized that the personal side of the 1 equation1
should not be neglected. Whereas Rosen1 s study had, for example,
pointed to a low, though significant correlation between judgments of
personal and of social desirability (50), ^ H eidbrederIs earlier
----------- r ------------------------------------------------------------------------------— ---------------------------------—
An average correlation of + . 30 was obtained, using MMPI
item s with 200 subjects. Summarizing, Rosen stated that the
11
investigation had illustrated the substantial difference which existed
between the actual and the desired positions which individuals p e r
ceive for them selves. (25) All in all, it was apparent th at--
although a strong relationship had been demonstrated between the
mean social desirability of numerous trait-attributes, and the prob
ability that those same attributes would be considered descriptive of ;
the self--the following caution was still in order:
. . .these results pertain to averages; individuals may not show
as much correlation between descriptions and judgments of
desirability of traits. (21)
Loevinger was one of the students of response bias most
concerned with pursuing the personal side of the equation. She also
took into account the research done on social desirability. Recently,
she set forth a general theory of personality development which cen
te rs around the changes, with age, in the ability of individuals to
form an accurate picture of themselves. (36) The nature of these
changes are seen to be largely a function of the manner in which
individuals respond to the standards of society. Particular traits
of personality are seen as accompanying, and expressing, each of
the following three stages of development, in a more or less fixed
sequence. It is in term s of these traits that the theory has relevance
to the problem of falsification, and to that of other forms of bias.
1. The early--childhood--phase of ego development is characterized
by little ability to conceptualize oneself accurately. In addition,
external standards and controls are not yet internalized. These
conditions manifest themselves in the form of impulsivity and a
pattern of activity marked by resistance to authority, and relativ
ely little concern with the external consequences of one’s behavior.
2. During this middle phase, Loevinger sees the individual as iden
tifying with authority and as trying to react in term s of socially
desired stereotypes. Here, accordingly, is where the greatest
"intercorrelations between the three forms [self-appraisal, personal
desirability, and social desirabilityj were low enough to show that
the three tasks were genuinely different. The concept of desirability
may be subdivided into personal desirability and perceived social
desirability." (50) He added, with respect to these judgments,
that "the greatest homogeneity appeared on personal desirability
and the least on self-appraisal for both men and women. Women
were more homogeneous than m en." (50 )
12
amount of falsification, as well as acquiescence, is said to occur.
3, The final stage is characterized by a well articulated, fairly
realistic self-im age. In general, adherence to social stereo
types, and other standards, now takes place on a more discrim
inating basis than it had during the previous phase. Idiosyncratic
tendencies in the self have at this point been better identified and
have become more acceptable to the individual; this results in
less need for facade, and other forms of defensiveness.
The degree to which an individual will attempt to cover up--
by means of a false front, or "facade" (1)--personality character
istics with which he is dissatisfied--this, according to Loevinger*s
theory, may be expected to rise with an increase in level of educa
tion, this being a factor to which fuller self under standing, and also
greater age, should relate. One study, for example, found a de-
crease in response stereotypy through the college years. (6)
At the same tim e, this finding does conform to the more basic
g
relationship postulated between facade (as well as acquiescence,
and deviant response set) and self understanding. A sim ilar re la
tionship was found in a study in which--on an Adjective Check L ist--
there appeared a decreasing tendency, during the college years, to
describe oneself in favorable, or defensive term s. At the same
tim e, there was no decrease in the degree of self-acceptance.
This latter finding was in general accord with predictions
7
Response stereotypy was considered to be an index of
defensiveness. A reversed F scale was used, with four groups of
subjects. Level of education ranged from one year of college
through graduate training.
o
Recent studies suggest that the relationships between social
desirability and acquiescence are not simple and, probably, are not
linear. Strieker, in 1962, was unable to find the predicted, signifi
cant positive correlations between m easures of acquiescence and of
social desirability, using a sample of ninety-three college students.
(55). Using a sample of 104 undergraduates, McGee, in 1961, r e
jected his prediction that "individuals with high acquiescence tenden
cies are socially oriented individuals who will (a) shift their position
in response to social pressure, and (b) withold negative attitudes in
order to gain social acceptance, more than individuals with low
acquiescence tendencies." (43) He suggested that a distinction be
drawn between "social acquiescence" (conformity, suggestibility,
persuasibility, and so forth) and "response acquiescence" (a tend
ency to agree with psychometric test item s, irrespective of their
content).
13
relating to the third stage in Loevinger*s theory. (53) On the other
hand, in a more recent, related investigation, using 110 male
9
patients and involving a wider range of education and I. Q. , no
significant relationships were found between an index of social d e sir
a b ility ^ and the subjects1 level of education, or their I.Q. (15)
The prediction that the degree of facade will be inversely
related to I. Q. has usually not been borne out--within, at least, the
relatively restricted ranges included in the m ajority of studies;
these studies generally involve college students. (2,32, 33, 54)
Despite this restriction, it is not improbable that the postulated
relationship will turn out not to be monotonic, even when a broader
age range is included.
Despite the complexities brought to light in these studies,
LiOevinger's theory, in its broad outlines, appears both plausible
and prom ising. The present difficulties might to some extent be
met if the nature of the relationships between the motivational
aspects of the theory (the response to external standards and lim its)
and the cognitive aspects (the stages in ego development) were more
fully worked out with reference to specific situations, aim s, or
objects (such as traits).
Individuals who, for example, typically try to conform to the
expectations of others, or try to conform with respect to "crucial"
m atters - -these individuals may under specifiable conditions be m oti
vated to express considerable "deviance" in connection with particu
lar issues and activities. ^ Such expressions of deviance, in spe
cific situations, could occur irrespective of the individuals* over-
9 ; ] ! ! ;
Fifty neuropsychiatric, thirty duodenal ulcer, and thirty
mixed medical, non-neuropsychiatric patients.
10
Derived from a structured Q -sort,
■^Whether these individuals would appear to be conformists
or deviants, on te sts of personality, would then depend in large
m easure on the nature and levels of specificity of the activities
and attitudes referred to on the test.
14
all level of self understanding; the same would apply with respect
12
to expressions of conformity, on specific issues. On the cognitive
side, a plausible hypothesis would be that--w hatever the stage of
ego development, or the level of intelligence--individuals will have
m ore accurate self-conceptions with respect to certain of their
13
, tra its than with respect to others.
It may be possible to approach the problem of delineating
specific relationships between conative and cognitive dimensions,
in term s of variables which could be considered--and which might
prove to be--relevant and representative indices or attributes of
each of these dimensions. Thus, for example, an individuals (a)
position on a particular tra it might be one meaningful type of index
12
Studies which support the thesis that situational factors can
modify the response patterns otherwise expected on the basis of
individuals1 past experience, include those by Blake e ta l. ( 3 ),
Blake et al. ( 4 ), Helson et al. (27), and Helson et al. (28).
13
A study by Goldfarb et al. (16) supports this hypothesis.
The Otis Self-Administering Test of Mental Ability, six Aptitude
tests selected from the Employee Aptitude Survey, the Kuder P re fe r
ence Record, the Guilford-Zimmerman Temperament Survey, and
others, were adm inistered to thirty-nine college students applying
for vocational counseling. "The results of this study suggest that
self-aw areness, as defined in term s of the ability to estim ate one's
perform ance on objective tests, does not represent a unitary phen
omenon. Rather, it appears to be unique to the area in which it is
m easured. Subjects who are accurate in estimating their perform i
ance on aptitude tests are not necessarily accurate in their estim a
tion of scores on tests of interest or tem peram ent. .. . The results
suggest further that intellectual ability does not appear to play an
im portant role in the ability to estim ate one’s score on tests of
tem peram ent or interest nor does the ability to estim ate one’s score
on aptitude or interest tests seem to be related to differences in
tem peram ent," (16 ) Subjects had been divided into Low-, Middle-,
-and High-Awareness groups on the basis of their ability to estim ate
th eir perform ance on the objective tests. Those in the Low group
did not differ significantly from those in either the Middle or High
groups in th eir ability to accurately estimate their actual GZTS
position on the traits of R estraint and Thoughtfulness.
15
of his current stage of cognitive development, and adaptation; the
(b) intensity with which he adheres to his position on the tra it may
be seen as a m easure of the motivational components; and(c) the
importance which he consciously attaches to the tra it could be seen
as bridging both the motivational and the cognitive/developmental
dimensions. Variables such as these might provide a basis for
making differential predictions about the behavior, in concrete
situations, of individuals who--at a more general level of ab strac
tion, or in term s of longer-range patterns of activity--m ay still be
referred to in such term s as "subm issive,1 1 or "authoritarian, "
Further, each such variable would lend itself to quantification.
The theory around which the present study has been organized,
makes it possible to address these several considerations. It will
now be described.
Helson*s Adaptation Level Theory
The theory which Harry Helson has developed over the past
twenty-five years was not designed with the problem of response
bias in mind. Instead, it has come to represent both a general
method of and framework for doing research, and a body of em piri
cal findings from which inferences may be drawn which can relate
to response biases, including falsification.
Helson sums up the general approach of the adaptation level
theory:
The relative importance of inner and outer determinants of
behavior is treated as a m atter of fact to be determined experi
mentally by evaluating the weighting coefficients for focal,
contextual, and residual stimuli in specific cases. (26)
The theory can perhaps best be described in Helsonls own
words (26, p. 567ff. ) :
The first problem that confronts the system atist who attempts
to deal with adjustment in a quantitative way is to find an opera
tional criterion or measure of it. From the results of numerous
experiments it was apparent that the level of performance is
reflected in responses to situations indicating neutral, indifferent,
or balanced states of the organism . These responses, it was
found, could be closely approximated as a weighted geo-
16
14
m etric mean of all stim uli affecting the organism within
certain tim e lim its. It is convenient to distinguish three broad
classes of stimuli operative in all behavior: (1) the stimuli
responded to and in the immediate focus of attention; (2) all
other stim uli immediately present and forming a background
or context for the focal stim uli and often affecting them p ro
foundly; and (3) all determ inants of behavior having their
locus within the organism , such as effects of past experience
and constitutional and organic factors which interact with present
stimulation and are treated as residuals since they are not
ordinarily under experim ental control. These three classes
of stim uli pool to form a single level to which all responses
are ref e rr able. This level we call the adaptation level because
it represents the adaptation or adjustment of the organism to
given conditions of stimulation.
The adaptation level is approximated as a weighted log mean of
all stim uli affecting the organism.
All dimensions of present and residual stimuli are related to
the adaptation level. In some cases, only frequency, intensity,
area, o rd er, and spacing of stimulation need to be taken into
account in determining AL; in others, properties such as diffi
culty, beauty, prestige, significance, quality, affective value,
and so on, must be included.
Fixed stim uli do not have constant effects on the organism.
P roperties of stim u li.. . depend upon the relations of stimuli
to the prevailing AL.
The weighted log mean form ula defines the value of AL and makes
possible determination of the relative contributions of focal,
background, and residual stimuli to resultant ALs. It does not
yield an exact prediction of responses to particular stim u li.. . .
To obtain quantitative predictions of responses to specific
stim uli, it is necessary to derive equations embodying the
relations between stim uli, AL, and responses.
Additional comment is necessary in regard to the three
general sources of stimulus variation, referred to above. These
may be classified and defined as follows :
15
F o c a l: These are "stim uli responded to and in the immediate
focus of attention.1 1 (26) An example would be a series of weights
which the subjects are to place in order, from "lightest" to "heav-
14
Subsequently changed to a logarithmic mean.
^ A lso referred to in the literature as "experimental, " or
"se rie s" stim uli.
17
iest, " in accordance with a set of instructions. In some experi
ments, subjects are left to assum e--but in others they are told--
that such stimuli, including instructions, are to be their prim ary
or exclusive concern. Typically, the experimenter expects that
subjects will tacitly assume that--unless otherwise indicated--
their judgments are to be based exclusively or essentially upon
properties and perceived implications of whichever are the objects
and instructions to which th eir attention has been thusly directed,
rather than being based upon other stimulus conditions in their
immediate or past environment. Individual differences come into
play in connection with this tacit assumption, a result being that
the following classes of stim uli are also able to exert varying
degrees of influence:
Background;^ As apart from the focal stimuli, these are "all
other stim uli immediately present and forming a background or
context for the focal stimuli. . . . " ( 26 ) A variety of stimulus
conditions are included here, ranging from such things as (a)
neutrally-colored screens against which colored disks (focal
stimuli) are placed and then judged as to either absolute or relative
intensity, to (b) statements describing social ideals and norm s on
political issues about which the subject's opinion is requested, in
the form of responses to w ritten statements and instructions (focal
stimuli). Thus, background stimuli are usually used to create
standards of comparison which are representative of everyday
experience and everyday levels of stimulation, and, more basically,
to help the subject discrim inate the focal stimuli and continue to
physically perceive them in a reliable manner or conceptually
relate them to some circum scribed universe of discourse. These
bases of comparison need not be explicitly stated as such by the
17
experim enter, nor recognized as such by the subject.
. j ■ - ■ ■ , — H " "
Also referred to in the literature as "contextual, " "anchor, "
"standard," and "comparison" stimuli.
17
It is apparent that the term "immediately present, " above,
refers to mental processes of varied degrees of cognitive complex-
18
Residual: These refer to "all determinants of behavior having
their locus within the organism .. . . " (26 ) T raits of personality
exemplify this class of stimuli. In general, the emphasis is on
the effects of the past--including those of constitutional and of
organic factors--upon the individual1 s present levels of adaptation
to background and focal stimulus conditions.
ity, ranging from a more simple -appearing perception of the physi
cal properties of objects, to a consideration of the intentions and
beliefs of other individuals.
C H A P T E R II
DEVELOPMENT OF HYPOTHESES, AND OPERATIONS
Hypotheses relating to two types of falsification will be p re
sented in this chapter, and the variables in term s of which they are
to be evaluated will be introduced.
An individual who has falsified on a personality questionnaire
has knowingly responded to its item s differently than he would
respond had he carried out an intention to convey an accurate picture
of him self. Two common types of consciously directed, inaccurate
self-portrayal may be distinguished. In one type, an individuals
prim ary concern may be to avoid conveying a particular picture of
him self; this may lead him to attempt to convey a picture which
contrasts with the one he does not wish to convey.
The second type does not center around the task of conveying
a contrasting kind of picture. Instead--as in laboratory experiments
--the individual may see his task as that of answering questions in
the way he believes certain kinds of individuals would answer them,
or in a way which he him self might respond if he were a given kind
of individual.
In the first type, the individual may, for example, attempt to
appear not Restrained, by instead appearing un-Re strained. With
the second type, his concern may just be that of appearing un-
R estrained, or less Restrained than he subjectively perceives him
self to be; he need not be trying to shift away from a contrasting
position. What rem ains common to both situations, however, is
the subjective awareness of responding to the questions in a way
which differs from , and which may bring about a shift from , an
accurate self-portrayal. ^
"''The term "shift" will be used synonymously with the term
"falsify"; the form er connotes more closely the resulting, measured
effect on scale scores. Kelly et al. , in 1935, were the first to use
this term in the context of falsification. (32)
19
20
Neither of these situations need be explicitly understood to
involve the socially or personally disapproved act of falsification;
however, either one may be understood in this way. In either of
the situations, individuals may, in addition, be told that--in answ er
ing the questions - -they (a) are to be carrying out their own person
al decisions, or that they (b) are to be carrying out the decisions
of some other individual.
Predictions relating to either type of falsification may be
developed on the basis,of the following considerations. Operation
ally, tra its descriptive of personality are bipolar in the sense of
having a high end and a low end, to represent greater and lesse r
amounts of the tra it. As a result, the main course of action avail
able to an individual wishing to answer a set of questions yet not
wishing to appear to possess high (or low) degrees of a tr a it- -
"very much (little) Restrained, " for example--would be that of trying
2
to shift toward the opposite end of the trait.
The response of shifting may, however, vary as to degree.
F o r any sample of subjects, individual differences may be expected
in the degree of shift away from , and toward, particular points on
any tra it. In the first type of falsification, described above, such a
"point" or region would be the one with which the individual wishes
to contrast him self; in the second type, it would refer to the person
al im pression of either his own or some other individual's actual
location on the tra it.
Given the above bipolarity and the resulting limited course
of action open to a subject who wishes to convey an inaccurate self-
portrayal, the following appears plausible: The degree as well as
the direction of shift which rem ain available to an individual may in
p a rt be limited by and thereby a function of the subjective im pression
If he did not wish to appear to possess m oderate degrees of
the tra it, he could choose to shift toward either one of the two poles.
21
which, at the outset, the individual has concerning his actual loca
tion on the trait, together with his beliefs about what would be a
desirable level of the tra it to appear to possess.
As an impetus to change--both from and toward moderate as
well as extreme points on traits--th e se actual or desired positions
which individuals perceive, and to which they may hold, would
appear to function as would a set of p rio r committments and inten
tions, regarding personal or social issues. Both would appear to
rep resen t a backdrop against which alternative courses of action
may or may not be discrim inated, or appear significant, urgent,
and feasible. Also, they could constitute both the jumping-off point
and the limiting conditions to be overcome, prior to change. During
the course of any change, they may provide reference points in rela
tion to which the degree of change is evaluated, as to sufficiency.
In addition to the possible influence of actual and desired
positions upon the degree and direction of shift, these considerations
would lead one to expect that the likelihood of occurrence of any
m easurable shift could be influenced by such additional factors as
the intensity of an individual*s adaptation--regardless of the particu
la r locus of his adaptation--and by the importance which specific
tra its generally possess for the individual.
If regarded as concomitants or expressions of either an individ
ual* s hypothetical average or of his 'real* adaptation level, subjective
perceptions of actual and desired status on a tra it can be justified
as providing operational indices of such a level; their utility then
becomes an em pirical question. The dimension of intensity could,
for example, furnish one way of characterizing the adaptation level.
F or the purpose of experimental manipulation, these perceptions
may, in addition, be postulated as possible sources of the adaptation
level, as may the individual1 s perceptions about the importance of
a given tra it.
Our main thesis is that response variation in degree of falsi
fication can be predicted in part by knowing the status of individuals
with respect to a number of m easurable characteristics of, and
22
postulated sources of their level of adaptation to given traits.
Perceptions about the social desirability of given traits may, for
instance, perhaps provide individuals with reference points which
influence the direction of their shifting, as well as influence the
2
degree of moderation with which they will portray them selves.
The general hypothesis of the present study is that:
Hypothesis 1
For a given sample of subjects, variation in degree of inaccurate
self-portrayal on personality questionnaire item s is a function of
specifiable focal, background, and residual stimulus conditions.
Further hypotheses will accompany the presentation of the
variables chosen to represent these stimulus conditions. These
variables will now be introduced.
Indices were developed to represent the following aspects of
individuals’ adaptation on traits: Position on trait, Importance of
tra it, Intensity of adaptation to tra it. For Position and Importance,
an attem pt was made to distinguish between the adaptation of the
subjects them selves and those which the subjects believed were
characteristic of adults in general. The distinction was designed to
bring out possible differences associated with residual as against
norm ative or background factors. A further distinction was made
between actual and desired Position in order to bring out (a) re la
tionships which falsification might have with possible differences in
the degree to which individuals regard their adaptations as being
either stabilized or completed to their satisfaction, or (b) relation-
^An ’extrem e’ self-portrayal may be defined in general as one
which deviates greatly from the central point on any trait. This defi
nition is based on the finding that, on the average, individuals believe
them selves to actually be situated about halfway between either ex
trem e of such traits as R estraint, Ascendance, and Masculinity-
Femininity. (47) For this reason, m ost of the subjectively p e r
ceived shift that occurs on a bipolar tra it should take place from the
central regions, out toward either of the poles, rather than vice
v ersa. In this context, the most extrem e self-portrayals may also
be considered the most inaccurate, m ost falsified ones, once again
from the point of view of subjective aw areness.
23
4
ships in connection with the direction of any such differences.
This distinction cut across both residual and background conditions.
These differentiations were also designed to increase the specificity
with which significant relationships might be interpreted.
5
The following concepts were the result:
1. Social adaptation level (position; background)
2. Socially desired optimum level (position; background)
3. Personal adaptation level (position; residual)
4. Personally desired optimum level (position; residual)
5. Social desirability of tra it (importance; background)
6* Personal desirability of tra it (importance; residual)
7. Range of adaptation (intensity of adaptation; residual)
These were defined by, and m easured in term s of, ratings and
rankings made by all subjects in response to essentially the following
7
questions, respectively:
1. Where do you think most adults actually stand on (each) tra it?
2. Where do you think most adults would like to stand on (each)
tra it ?
3. Where would you estim ate yourself to actually stand on (each)
tra it ?
4. Where would you like to stand on each trait?
5. Which of the (three) tra its do you think most adults consider
the m ost im portant, the next most important, and the least
important ?
-
For example, possessing ’’ too much, " as against "not enough"
of a trait.
5
See Appendix G for the complete presentation of the ratings1
and rankings1 task. Included within the parentheses which follow
each of the above and below concepts, is the classification of the
experimental index, and, next, its classification in term s of stimulus
conditions dealt with in adaptation level research.
^The operational equivalents of these seven concepts are later
referred to as the "independent m easures. " They are used in te s t
ing the first five hypotheses.
24
6. Which of the (three) traits is most im portant to you personally,
which is next most important, and which least important ?
g
7. Indicate the TWO points (on each trait) below and above which
you would probably feel quite uncomfortable or anxious--that is,
if you actually possessed those degrees of (the) tra it. . . . You
might think of these two points as being maximum and minimum
’tolerance lim its. ’ ^
The hypotheses relating to background and residual factors
will be based upon the following general considerations, both of
which are commonly reflected in the choice of variables and in the
type of hypothesizing which appear in adaptation level research.
The first is that change, in adaptation, may be conceptualized
as occurring within the context and lim its of whichever are the
points of reference about which subjects have some degree of aw are
ness and to which they attach some degree of significance. Apart
from characteristics of physically manipulable objects, the major
type of dimension along which reference points are located in adapta
tion level studies, is that of normative dispositions--w hether
perceived by the subjects as being those of their own or those of
other individuals.
In virtually any social or personal context, specific norm s
represent concrete instances of more general regulative principles
or established ideals; as such, they may be said to imply a set, or
sets, of preferences and p rio rities, which individuals may under-
-----------7------------------------------------------------------------------------------------------------------------------------
See Appendix G .
g
On a ten point scale.
□
The num erical difference between these two points was consid
ered inversely related to intensity of adaptation. Thus, a sm aller
’range of adaptation’ became, by definition, a more intense adapta
tion, mainly on grounds that it should indicate, albeit roughly, less
willingness or readiness to vary in one’s position on a tra it, for
whatever reason.
25
stand and accept in differing ways and differing degrees. With
respect to such standards, it is the direction in which an individ
u a ls preferences lie --h is favoring, for example, more of, or less
of, a tra it--th a t is used as a key reference point in predicting d irec
tion of change in adaptation, and in predicting preferences among
possible alternate adaptations; and, the greater the relative or
absolute amount of preference, the larger will be the predicted
change. These considerations should, m oreover, apply whether
an individual were shifting from or toward moderate, as well as
extrem e, positions on a tra it.
In the present study, the "socially desired optimum level" and
the "personally desired optimum level" are the reference points
which m ost closely represent the above situation with respect to
preferences as to direction, while "social desirability" and "personal
desirability" of a trait, relative to other tra its, provide a basis for
predicting relative amount of change.
Secondly, interaction is always occurring within as well as
between groups of residual and background stim uli. A cumulative
influence upon the adaptation level will result if two or more such
stimuli a re , simultaneously, operating in the same direction. Such
influence should increase the likelihood of occurrence of change in
adaptation level, as well as the degree. Whether or not such stimulus
conditions operate in the same direction, such a situation may be
used to advantage by adding to research hypotheses which predict the
type of influence of either of these factors, an assertion about the
state of the remaining factor at the time of such influence. Thus, it
might in some experiments be possible, and theoretically appropriate,
^ ^Implicit is the assumption that the conditions to which an
individual is adapted are--o n the average--m ore rather than less
likely to also be preferred by that individual. The results of a p re
test of the present study support this assumption. (47) Also im plicit
is the assumption that the adaptation level and the preferred conditions
are both less likely to be relinquished and more likely to be sought--
other things being equal.
26
to predict an increase (or decrease) in the strength of relationship
with respect, for instance, to the personally desired optimum level,
in the further event that other conditions also obtained for the
personal adaptation level, or with respect to the social desirability
of a trait.
Some final points, by way of explanation. In the hypotheses
which follow, the "amount of shift" will be m easured in term s of
the total number of responses which subjects make in a direction
specified by a given set of instructions. In light of the definition of
falsification on personality tests--inaccurate self-portrayal,
perceived as such--shift is conceptualized as occurring relative to
the point or region which the individual believes to be his actual
position, on given tra its. This subjective point would be approxi
mately the same for sam ples drawn from a population of subjects,
regardless of the nature of whatever test instructions these different
samples might (in any given study) be responding to, and regardless
of whether the testing situation had been structured for the subjects,
by the experim enter, as being one which involves falsification.
This subjective point, then, would operationally be the adaptation
level which subjects bring to the testing situation, and with which
the instructions and the other stimulus conditions would interact.
Relative to such a point, the larg er the number of correctly keyed
responses, the greater is the amount of falsification.
Hypothesis 2 is based on the prem ise that individuals' concep
tions of given positions on a tra it--fo r example, "fairly Restrained, "
or "very A scendant"--are influenced by perceptions which they have
about their own positions on the tra it, and that, as a result, different
individuals, who perceive them selves to be located at differing
positions on a trait, can interpret differentially the meaning of
externally specified, defined points on a trait.
Thus, for example, an individual who considers himself to be
a ’7' on. a 10 point scale of R estraint, ^ could be expected to p e r
ceive a request to convey the position '8* in a m anner which differs
in degree from that of an individual who considers him self to be a
’S’ on the trait. In order to portray himself as being more R estrain
ed than he considers himself to actually be, the form er individual
may feel it necessary to answer ’tru e’ to questions which the latter
might only answer ’tru e ’ to in the case of a request to portray him
self as being a 1 9 * or a *10'. The form er individual would thus
obtain a higher "shift" score.
Hypothesis 2
a* Differential shift as a function of individuals position on tra it
Significant differences will occur in the amount of shift on traits,
on the part of individuals who perceive them selves as actually
possessing higher as against lower amounts of the respective
tra its.
b. Amount of shift as a function of individual’s position on tra it
Amount of shift occurring in the direction of either extreme of
a bipolar trait, will be proportionately greater for individuals
who perceive themselves as actually possessing amounts of the
tra its which them selves are located nearer to, rath er than
farther from , the respective pole. Thus, (1) the subjects who
rate them selves as being relatively higher on a tra it, will be the
ones to shift farthest in the direction of the high end of that same
tra it; the same applies with respect to (2) the low end of a
tra it, in the case of individuals who rate them selves relatively
low on that tra it. ^
■^With ’10' being extrem ely Restrained.
12
The same prediction is made with respect to one’s own desired
position on given tra its, and with respect to the actual and desired,
estim ated positions of most adults. In order to make the presenta
tion of hypotheses and findings less cumbersome, these latter p re
dictions will be considered separately.
13
Either the social and personal, or the actual and desired
levels, or both, may at tim es appear in connection with the same
hypothesis, apart from the separate presentation mentioned in n. 12.
This is in light of the p re-test, which indicated that these four a s
pects of the adaptation level--w hile differing significantly in degree--
are with a single exception located on the same side (relative to the
midpoint) of the three bipolar tra its which are to be used in the
present study. Thus, in the present hypotheses these dimensions
28
Hypothesis 3
a. Differential shift as a function of individuals intensity of
adaptation to tra it
Significant differences will occur in the amount of shift on
tra its, on the part of individuals who manifest a relatively
higher as against a relatively lower intensity of adaptation
to the respective tra its .
b. Amount of shift as a function of individuals intensity of
adaptation to tra it
The amount of shift on given traits will be proportionately
less for individuals who manifest greater strength of adapta
tion to those tra its.
Hypothesis 4
a. Differential shift as a function of importance of tra it to individual
Significant differences will occur in the amount of shift on tra its,
on the p art of individuals for whom the respective traits have
higher as against lower degrees of personal importance. ^
b. Amount of shift as a function of importance of tra it to individual
The amount of shift on given traits will be proportionately
greater for individuals who judge those traits to have a greater
relative degree of personal importance.
Hypothesis 5
Amount of shift as a function of average, relative importance
of tra it to a sample of individuals
The relationships hypothesized in ,2b1, 13b‘, and *4bf, above,
will be significantly greater when occurring in the context of
tra its which the sample of individuals judge to have a relatively
greater, average degree of personal importance,
receive equivalent status, unless otherwise specified. More than
one such dimension will not appear as p art of the same hypothesis
when there is reason to expect that only one of the dimensions will
differentiate individuals on the criterion m easures.
14
The same predictions apply in the case of the tra its1 im por
tance to adults.
15 1 £ >
Cf. n. 14. Cf. n. 14.
29
Three remaining stimulus conditions were investigated by
1 7
varying the task and the instructions:
18
1. Presence versus absence of definite request to falsify; (focal)
2. Source of request to complete a task in specified ways:
(a) Hypothetical individual ; (background) ; (b) Self ; (residual)
3. Direction of shift ; ^ (focal)
These three conditions are representative of those which may
occur in the case of the two common types of falsification: (a) the
situation may or may not be externally structured as being one which
involves falsification, (b) the subjects may be told that they are to
carry out either their own, personal decisions or those of some
other individual, (c) the subjects may be asked to portray them
selves in a way which contrasts with some other, defined portrayal,
whether of them selves or of other individuals, or they may be asked
to imagine given kinds of individuals (themselves or others), and
then to respond as these individuals would respond. ^
Hypothesis 6 is based upon the following considerations. In
term s of most social and personal norm s, the act of (a) falsifying
would be considered less representative of optimally desired behav
ior than would the alternative action, viz. , that of portraying one
self (b) accurately. Thus, although the two approaches--’a1 and
•b1--appear equally capable of allowing subjects to carry out any
' 'p 7
See Appendix H for the complete presentation of tasks and
instructions given to different groups of subjects.
18
Class of stimulus condition.
19
From greater to lesse r amounts of a trait, or vice versa.
20
These three conditions have to occur in specific combinations.
In the present study, for example, the subjects who were asked to
portray a contrast, were those for whom the task was structured as
as involving falsification. For those who were not asked to portray
a contrast, the situation was not structured as one involving falsifi
cation. These particular combinations appeared to be more typical
of situations which subjects would become involved in than were the
rev erse combinations. Condition 'b* cut across both of the above
combinations of conditions.
30
particular request, relatively m ore resistance, or less support,
might be expected to accrue to the request to falsify. In light of
each subject’s ideas about his true location on a given trait, however,
conscious shift would be involved, in the typical case, in carrying
out either type of request, in the event that the request required
subjects to shift away from relatively moderate regions of a trait.
Yet, although both tasks involve falsification in a specific sense of
the term , only one of these tasks is structured as involving, in a
direct sense, accurate self-portrayal; and relatively less resistance,
or more support, should result from this additional factor, viz. ,
the latter request to portray oneself accurately.
Hypothesis 6
a. Differential shift as a function of presence or absence of
request to falsify
Subjects who have been requested to falsely portray them selves
as possessing specified levels of a trait, will portray them selves
as being at a level of the tra it which differs significantly from
that of subjects who have been requested to convey the same level
of the tra it, but who have not had the task structured as one which
involves falsification.
b. Amount of shift as a function of presence or absence of
request to falsify
Subjects who have been requested to portray them selves as
possessing relatively high or low levels of a tra it, without
falsifying, will have a tendency to shift farther in the direction
of these requested levels than will subjects who have been request
ed to falsely portray them selves as possessing the same levels
of the trait; the latter group will have a tendency to convey more
moderate, relatively less extrem e levels of the tra it.
The next hypothesis is based on the fact that, to date, the
influence of background factors outweighs that of residual factors
in the clear m ajority of adaptation level studies in which subjects
21
were to express a personal opinion.
21
Regardless of whether the situation called for extensive or
intensive face-to-face contact between subjects and research person
nel.
31
Hypothesis 7
a. Differential shift as a function of source of request
Significant differences will occur in the amount of shift on tra its,
on the part of subjects to whom requests which relate to shifting
are structured as having a common, external origin, rather than
as representing personal preferences of the individual subject.
b. Amount of shift as a function of source of request
The amount of shift will be proportionately greater in the case
of requests which are structured as having a common, external
origin, rather than as representing personal preferences of the
individual subject.
Hypothesis 8
a. Differential shift as a function of direction of shift in relation
to desirability of trait
Significant differences will occur in the amount of shift on tra its ,
between individuals who are shifting in the direction of the pole
in which is located the Personally and Socially desired optimum
levels, and individuals who are shifting in an opposite direction. ^
b. Amount of shift as a function of direction of shift in relation
to desirability of trait
The amount of shift on the part of individuals will be proportion
ately greater when the shift occurs in the direction of the pole,
of any tra it, in which is located the Personally and Socially
desired optimum levels, rather than in an opposite direction.
Whereas (1) correlational procedures which were unaccom
panied by direct experimental manipulation were the means of eval
uating the hypotheses relating to (a) Position on trait, (b) Im por
tance of trait, and (c) Intensity of adaptation to trait, (2) experi
mental manipulation of instructions and of task, as represented in
Chart 1, was employed, together with correlational procedures, as
22
Shifting (a) in a specified direction, relative to the starting-
level, and shifting (b) to a specified degree--that is, toward a stated
point on the particular trait-scale.
32
Chart 1 . — Model f o r E x p er im e n ta l M a n ip u la tio n o f Groups
o f S u b j e c t s , i n Two Types o f F a l s i f i c a t i o n T asks
S t a r t i n g - l e v e l
( p o s i t i o n
on
t r a i t )
H igh
I n s t r u c t io n s
F a ls ify *
Do n o t f a l s i f y
Source
P e r so n a l
E x te r n a l
P e r so n a l
E x te r n a l
D e s ig
n a t io n
o f
Group
- K
- L
- S
- T
Low
a
F a l s i f y
Do n o t f a l s i f y
P e r so n a l
E x te r n a l
P e r so n a l
E x te r n a l
M
R
W
Z
In a s p e c i f i e d d i r e c t i o n , r e l a t i v e to th e s t a r t i n g -
l e v e l , and to a s p e c i f i e d d e g r e e , i . e . , toward a
s t a t e d l e v e l .
33
the means of studying (d) Presence versus absence of request to
falsify, (e) Source of request to falsify, and (f) Direction of shift.
The two starting-levels, two kinds of instructions, and two
types of request--w hich were involved in 'd 1 , ’e1 , and ^ ’--called for
a total of eight different groups of subjects. As seen in Chart 1,
groups K and L . were instructed to shift from high to low points, on
the bipolar traits, while M and R were to shift from low to high;
groups S, T, W, and Z were not instructed to shift from their
specified starting levels. In attempting to carry out either of these
instructions, all eight groups would be shifting from th eir average,
subjectively perceived actual position on the given tra its.
The reference points, "high" and "low," were represented
23
to subjects in term s of specified scale points.
^A nd also by means of the qualitative term, "fairly, " as for
example, "fairly Restrained, " or "fairly ‘low1 on Ascendance. "
See Appendix H .
C H A P T E R III
THE QUESTIONNAIRE, SAMPLE, AND TESTING
The criterion m easures employed in evaluating all hypotheses
were drawn from the Guilford-Zimmerman Temperament Survey
(GZTS), which is a representative personality inventory, often used
in psychological research. Of the ten scales in this inventory,
three were selected for use in the present study. They are
(l) "R estraint versus Impulsiveness, " (2) "Ascendance versus
Submissiveness, " and (3) "Masculinity versus Femininity.-"
These were selected in view of the following advantages, and
in view of certain requirem ents of the present design and hypotheses:
(a) Considerable factorial purity; reliability; predictive validity.
(21,39)
(b) A relatively wide distribution of subjects on each of the scales.
(23,47) A p re-te st indicated that typical sem i-interquartile
ranges were present for the scales of R estraint, Ascendance,
and Masc. -Fern. In addition, published norms indicate a
relatively norm al distribution of subjects on each of these
scales.
(c) An average, unfalsified scale score which is about equally
distant from either pole of the given traits. Both the published
norm s and the p re-test indicate that this is the case. The
latter, in addition, indicates that subjects perceive themselves
to actually be located approximately halfway between the
extrem es on these three scales, as m easured by the subjects1
self-ratings and also by th eir responses to the three GZTS
scales.
■^General Activity (G), R estraint (R), Ascendance (A), Sociabil
ity (S), Emotional Stability (E), Objectivity (O), Friendliness (F),
Thoughtfulness (T), Personal Relations (P), and Masculinity-
Femininity (M),
2
H ereafter abbreviated as ,Masc. -Fern. 1
34
35
(d) The traits should differ from one another as to average social
and personal desirability, thus allowing for a relative ranking
of traits in these respects. The p re-test indicated that sub
stantial differences were present in mean desirability scores,
for these tra its.
(e) Subjects should not have considerable difficulty in carrying out
their own decision to follow the request to either falsify or not
falsify--either in full or in partial accordance with the in stru c
tions. The p re-te st indicated that subjects were able to carry
out these req u ests--at least, to either shift or not shift--when
3
given a proper cue as to the nature of the tra its.
(f) It would be desirable to obtain subjects* responses to all three
scales, separately, during a single session. The p re-test
indicated that this was possible, with a proper arrangem ent
of test item s.
Involved in the questionnaire were four main aspects. These
w ere presented to all subjects in the following sequence: Short
introduction, definition of term s and task; instructions as to ’how
ratings work* ; fifteen ratings and two rankings; 210 questions from
the GZTS. (See Appendix F for coded listing of GZTS questions. )
For all subjects, these questions formed three separate tasks,
each pertaining to only one of the three tra its, and each 70 questions
in length. The tasks were presented to all subjects in the same
order. Essentially, the task was for the subject to answer the 70
questions so as to portray himself, with respect to a trait, in the
manner described in the instructions pertaining to the given trait.
This procedure was then repeated with the next group of 70 item s,
which focused upon a different trait; and sim ilarly with the final
70 item s. (See Chart 2.)
- — — ■ , . -
The cue consisted of short, descriptive phrases which p ro
vided all subjects with a shared definition of term s. See Appendix
G .
36
C hart 2 . — C o n ten ts o f T r a it - S c a l e s : Type and Q u a n tity o f
Item i n Each Task
R e s t r a in t Task
(ite m -g r o u p I)
A scendance Task
(item -g ro u p I I)
a s B n m f t n n e n M
Q u e s tio n s 1 - 7 0
(70 ite m s)
Q u e stio n s 71 - 140
(70 ite m s)
M asc.-Fem . Task
^ ^ ^ e ^ g r o u ^ ^ I I ^
Q u e stio n s 141 - 210
(70 ite m s)
R e s t r a in t
(20 ite m s)
A scend ance
(2 0 ite m s)
M asc.-Fem .
(2 0 ite m s)
R e s t r a in t
(20 ite m s)
A scendance
(20 ite m s)
M asc.-Fem .
(20 ite m s)
R e s t r a in t
(20 ite m s)
A scendance
(20 ite m s)
M asc.-Fem .
(2 0 ite m s)
10 ite m s from
o t h e r s c a l e s
10 ite m s from
o th e r s c a l e s
( d i f f e r e n t ite m s)
10 ite m s from
o th e r s c a l e s
(a g a in d i f f e r e n t )
37
Twenty key item s were interspersed among each group of 70.
The criterion m easure, for each of the three tra its, was drawn from
only these form er item s, and consisted simply of the unweighted
number of correct answers, in term s of the standard scoring keys
of the GZTS.4
The identical twenty item s relating to each of the three tra its --
60 item s in all--reappeared in each of the three groups, but every
item within each of these groups appeared in random order. (35)
From scales other than R estraint, Ascendance, and Masc. - Fern. ,
ten item s were selected^ which, on a rational basis, appeared to
differ considerably from the rem aining items in term s of the traits
involved. These item s were included so that (a) subjects would
not find the three tasks unduly repititious, and (b) so that subjects
would have fewer cues as to which items might be crucial to c a rry
ing out these tasks.
For each of the three tra its, twenty of the standard number of
thirty item s were selected from the GZTS scales. One reason for
this selection was the tim e factor: subjects would otherwise have
had 300 rather than 210 questions to respond to --in addition to the
ratings and rankings--within the available period of approximately
The instructions appearing on the answer sheets varied slight
ly from ihose which appear on the standard G Z .'i'fc > torm , The form er
did not include the standard statement which indicates that the sub
ject may give the answer, " ?, " if unable to decide between "Yes"
and "No. " The instructions were: "If the statem ent you read is
m ore true than false, or if you agree with it, make a mark above the
' Yes1 . If the statement you read is more false than true, or if you
disagree with it, make a m ark above the 'No'.
^Ten different item s for each of the three groups of questions--
30 item s in all. For these three groups, 1, 2, and 3, respectively,
the number of such item s contributed by each scale is given in the
parentheses following the name of the scale: General Activity (0, 1,
l); Sociability.(1, 1, l); Emotional Stability (1, 3, 1); Objectivity
(3, 1, 2); Friendliness (1, 1, 2); Thoughtfulness (0, 1, 2); Personal
Relations (4, 2, 1). There were no significant differences in the deg
ree to which these several scales were represented in the three
groups of questions. (Kruskal-W allis one-way analysis of variance:
H = 0, 500; 2 d. f . ; p > . 70 < . 80. )
38
fifty minutes. The second consideration also supplied the main
basis for reducing to twenty the number of item s. This involved an
attem pt to exclude questions which, in term s of content, appeared
relatively sim ilar to any of the other questions pertaining to the same
tra it. The presence of sim ilar pairs of questions could have provid
ed subjects with an unwanted type of cue as to which item s might be
crucial to carrying out the tasks, the result of which could have been
a disproportionate contribution to the criterion m easures on the part
of such item s, for reasons not central to the study. The remaining,
secondary basis for reducing the number of item s also applied to the
selection of item s from scales other than R estraint, Ascendance,
and Masc. - Fern. This was an attempt to exclude questions which
had some likelihood of upsetting many subjects to the point of altering
the direction of their motivation, or engendering a persistent defen
siveness which might distract them from the task at hand.
The Subjects
213 subjects took part in the study. There were 127 males
and 86 fem ales. All subjects were college undergraduates. With
ten exceptions, they were enrolled at the University of Southern
California, in Los Angeles. ^
Approximately 180 of the subjects were either freshm en or
sophomores, and 197 were currently enrolled in different sections
of an introductory psychology course. Students enrolled in such a
course customarily expect to be asked to take part in one or two
"experim ents'1 during the school sem ester; given this situation,
students as well as class time w ere readily available. The rem ain
ing subjects were enrolled either in an elem entary statistics course
^The ten exceptions were students at the Immaculate Heart
College, in Los Angeles; they w ere enrolled in an elem entary
statistics course.
39
or in an introductory, survey-of-history course; tlxey were, according
to their instructors, willing volunteers. Neither restrictions nor
other types of conditions were placed upon any subjects, or groups
V
of subjects, as to eligibility to participate in the study.
The eight groups into which subjects were placed, were matched
on the variables of College Entrance Examination Board scores
(CEEB), ® Age, and Sex. B artlett's te st for homogeneity of v a r
iances indicated that these eight "subgroups" could statistically be
considered a single group with respect to CEEB ( B ® 0.094 ;
7 d. f. ), Age ( B = 0.946 ; 7 d.f. ), and ratio of m ales to fem ales
( B = 1. 963 ; 7 d.f. ) .9 The manner in which this relatively close
matching of groups was achieved is described in Appendix A .
In regard to these matching variables, the basic descriptive
statistics for the eight groups appear in Table 1 ; Table 1 provides
summary information about the four groups which were requested
to falsify, taken together, and the four groups which were not
7
In all, there were nine different classes of subjects. B art
le tt's test, with correction included, revealed that these classes
could be regarded, statistically, as a single group with respect to
variance on two of the three matching variables, viz. , CEEB ( B =
1 .7 7 0 ; 8 d .f. ), and Age ( B - 12. 563 ; 8 d. f. ). With 8 d. f. , B
m ust equal 15. 507 to perm it rejection of the hypothesis of equal v a r
iance, at the . 05 level of confidence. The variances were not homo
geneous with respect to the sex ratios among the nine classes of
subjects. (One class contained nine women and one man. )
g
In all cases, these were the unweighted total scores - -Vocab
ulary plus Quantitative. Related m easures were substituted in the
case of 12 subjects for whom CEEB scores were non-existent or
otherwise unavailable. The form er m easures were either the high
school grade point average, the college grade point average, the
Vocabulary score on the English Classification Test, or, in one
instance, the School and College Ability Tests (SCAT). See Appendix
B, in regard to these related m easures.
9
With 7 d. f, , a B of 14. 067 would be required to reject the
hypothesis of equal variance at the . 05 level of confidence.
T able 1 . — Means, Sigm as, and Sex R a tio s o f M atching V a r ia b le s , by Group
D e sig
n a tio n
o f
Group
CEEB Age
Sex
Males
N
Fem ales
N N Mean Sigma Mean Sigma
K 31 1 0 2 2 .2 136.6 2 3 8 .2 6 .7 2 19 12
L 32 1068.7 1 6 6 .4 2 3 6 .6 7 .6 0 19 13
M 31 1045.7 1 02 .6 2 3 2 .9 4 .6 3 18 13
R 32 1025.8 1 0 8 .3 2 37.8 5 .3 4 19 13
S 22 1034.7 1 1 5 .5 2 4 0 .1 7 .8 5 13 9
T 22 1008 .9 137 .7 2 3 5 .5 5 .2 8 13 9
W 22 10 2 3 .2 1 1 3 .9 2 4 3 .1 1 0 .6 4 13 9
Z 21 104 1.6 1 3 2 .0 238.7 6 .6 4 13 8
m
126 1040.7 1 3 3 .1 2 3 6 .4 19.86 75 51°
s * v
w * zb
87 1026 .9 1 25 ,7 2 3 9 .3 2 5 .0 2 52 35d
T o ta l 213 1 0 3 5 .1 1 3 0 .3 2 3 7 .6 2 2 .1 6 127 86e
a b
Groups i n s t r u c t e d to f a l s i f y Groups n o t in s t r u c t e d to f a l s i f y
c M a le/fem a le r a t io : 1 .4 7 : 1 d M ale/fem ale r a t io : 1 .4 9 :1
e M a le/fem a le r a t i o : 1 .4 8 :1
41
requested to falsify, taken together. The average age of individuals
in the sample was approximately 19. 8 years; the mean CEEB
total score was approximately 1035, which is close to the average
of entering students at the University of Southern California. (3l)
The ratio of males to fem ales was roughly 3 to 2.
The chi-square test of the norm al-distribution hypothesis
indicated that all eight groups, taken as a single group { N = 213 ),
were norm ally distributed as to CEEB scores ( X^ = 1 7 . 17 ; 12 d .f;
p < .20 > . 10 ). ^ According to the chi-square test, however, the
total sample was not norm ally distributed as to age, as computed in
six-m onth intervals ( = 60.90 ; 11 d. f. ; p < . 01 ). Here, the
distribution was leptokurtic but nevertheless relatively sym m etrical,
and unimodal. For the purpose of later applying standard statistical
tests to the variable of age, the distribution was normalized by con
verting the number of months into T -scores. The same conversion
was also applied to the CEEB scores, as a m atter of general sta tis
tical procedure for improving scales of m easurement. (19) '
A s a group, males were higher on the CEEB, and they were
also older. The mean CEEB scores for the 127 males and 86 females
were 1045.29 and 991. 58, respectively. This difference is signifi
cant beyond the . 01 level of confidence ( z = 3. 16 ), as is the
difference between mean ages of the males and females ( 234. 17
and 228. 31 months, respectively; z = > 4. 89 ). The intercorrelations
among the matching variables are shown in Table 2 . There was no
relationship between the age of individuals in the sample and their
CEEB score.
The p re-te st revealed that both age and sex were related
significantly to several of the ratings and rankings given the traits
^Computed from May 1, 1962 for 189 subjects who took part
in the study between April 16th and May 15th. Computed from June
1st in the case of subjects who took part after May 15th.
^C om puted at intervals of 50 points. Includes equivalent CEEB
scores of the 12 subjects for whom related m easures were substi
tuted.
42
T ab le 2 . — I n t e r c o r r e l a t i o n s o f M atching V a r ia b le s (N»213)
M atching
V a r ia b le CEEB Age Sex
CEEB • • .ooa + .2 1 b
Age • • • • + .3 ? b
Sex • • • • .. • «
a P earso n product-m om ent c o e f f i c i e n t
b P o i n t - b i s e r i a l c o e f f i c i e n t
43
under study. These results, together with the substantial sex
differences in mean age and CEEB of the present sample, made it
especially im portant to achieve a rather close matching of the eight
groups of subjects.
Circumstances Of The Testing
The study was conducted between April 18th and May 24th,
12
1962. Although in the final analysis participation in the study
was voluntary--both from the point of view of the experimenter and,
probably, from that of most subjects--certain differing expectations,
which probably could not have been circumvented, did exist between
some of the classes of students. Instructors of three of the nine
classes of subjects ( N = 96 ) informed the experimenter that their
students understood, from the beginning of the sem ester, that
participation in any experiment would result in a sm all bonus of
points to be counted toward the final course grade. Secondly, during
the individual testing of certain subjects--described below --it
becam e clear that students in one of these three classes believed
th at participation in at least one experiment was a course req u ire
m ent. These are both common practices in many kinds of classes,
but they were not present for the rem ainder of the sample ( N = 117).
Although the specific effects which these differing expectations
m ay have had upon the criterion m easures can be neither gauged nor
corrected for, two things should be borne in mind. F irst, as was
13
the case with all subjects in the remaining classes, the subjects
in these three classes had no p rio r idea that any study was to be held;
they came to th eir regular classroom expecting to hear a regular
lecture. Approximately 95 per cent of the students who entered
the classroom rem ained, after learning that an experiment was to
be held. This was in spite of the fact that an estim ated one-half of
—
Specific dates of testing were selected by the instructors.
13
Except for the Immaculate Heart students.
44
the subjects in two of these three classes ( N = 60 ) had already-
taken part in two classroom experiments during the sem ester.
From this, it may be considered unlikely that the three classes
represented a sub-sample with characteristics which differed
substantially from those of the remaining classes; it will be recalled
that, at least, there were no non-chance differences in variance
among the nine classes, with respect to age and CEEB,
Secondly, the subjects in question had been distributed among
the eight experimental groups in the same manner as had all other
subjects. Thus, any differences with respect to the criterion scores
of those 9 6 subjects were likely to be fairly evenly distributed among
the eight groups used in the study.
Approximately 8-12 students, from five of the classes, left
the classroom as soon as they realized that something other than
the regular lecture was about to take place. No attempt was made
to prevent them, or others, from leaving. One subject discontinued
the questionnaire after about fifteen minutes, with the comment that
he could not see any relation between the stated task and the way
one answered the questions. His responses could not be used.
Fourteen subjects either came late or could not finish within
the regular class period and class intermission. These subjects
completed their questionnaire in an office of the psychology
department, generally on the same day.
As the class attendance - -part way through the study--was
nearly 30 per cent below that which had been anticipated, an attempt
was made to contact those "absent" students who might turn up at
14 1 ^
either of the subsequent two lectures. A note which requested
the subjects1 participation was passed out, or read, to the students
by the instructors, at these lectures. A large number of subjects
responded-as-hoped to this request--33 in all, from six classes.
. . 15
A rbitrarily determined. Reproduced in Appendix C.
45
Subsequently, these subjects were tested individually, in offices of
the psychology departm ent, under conditions which were otherwise
1 6
as sim ilar as possible to those in the classroom .
In all instances ( N = 213 ), the introduction and the question
naire were adm inistered by the experim enter alone. The experi
m enter explicitly stated at the outset that the way in which an indi
vidual responded on the questionnaire had no bearing on his grades
in the course. The experim enter attempted to set a tone which was
m oderately relaxed, yet relatively serious. Confidentiality was
17
assured. The spoken introduction appears in Appendix D .
These subjects had also been distributed throughout the eight
groups. All subjects who had been tested individually ( N * 47 )
were included as a single group in the earlier-reported test of homo
geneity of variances for the nine classes of subjects. While subjects
were drawn from actually only eight classroom s, the individually-
tested subjects were regarded as a "ninth class. "
17
Subjects quickly received their pre-assigned questionnaire
by means of an alphabetized classlist--distributed at the beginning
of the class--upon which the letter designation of the appropriate
questionnaire appeared, next to the individual's name. Subjects
were not asked to put their names on the questionnaires or on the
answer sheets. The name of the subject who turned in any given
questionnaire and answer sheet was determined by the experimen
te r 's checking--against his own lists--th e (1 ) birthdate and (2 ) sex
which the subjects had indicated on their booklets; also, (3), the
form of the questionnaire provided a third check.
C H A P T E R IV
PRELIMINARY ANALYSES
P retest
A p re-te st of the questionnaire was made in January of 1962,
using a sample of 67 undergraduates, all of whom were enrolled in
an introductory psychology course at the University of Southern
California. In addition to results already referred to, it was found
that (1) the order of appearance of the tra its--e ith e r R estraint
first or Ascendance first--w as unrelated to the means of all ratings
and rankings; (2 ) the order of appearance of the individuals being
rated or ranked--either Oneself or Adults appearing first--w as
unrelated to the means of all ratings and rankings; (3) instructions
which (a) directed one group of subjects to first respond to the
GZTS questions truthfully and then respond to identical questions
in a specified untrue manner, and (b) directed other subjects to
only falsify, in the same specified manner, were not associated with
differences in mean GZTS scores on the R estraint and Masc. -Fern.
scales, but were associated with significant differences on the
2
Ascendance scale ( p < . 01 ).
■^Order was not varied with respect to Masc. -Fern, as well,
in view of the size of the sample.
2
As compared with the test used in the present study, the
ratings-and-rankings1 section in the p re-test was identical as to
task and content. The identical traits appeared, and the GZTS
questions were identical in all respects. The setting for the fa lsi
fication instructions which preceeded these questions, differed from
that of the present form in that it involved "job seeking" on the part
of the subjects, rather than participation in a "research" task.
Some of the p re-test form s also differed as to the order of appear
ance of traits and of individuals to be rated, as indicated above.
Considerably more detail was included in the trait-definitions given
to the subjects. Because this degree of detail may have helped make
the falsification of scales too simple a task, it was decided that the
definitions to be used in the final form should contain a sm aller
number of term s, and should be stated at a more general level. In
the p re -te st, subjects were also instructed to falsify to a slightly
greater degree than in the present study.
46
Reliability Of The Criterion M easures
All estim ates of reliability were computed with the basic
3
Kuder-Richard son formula number 20. These internal consist
ency estim ates relate to the twenty keyed item s comprising the
criterion m easures involved in each of the three tra it-sc a le s--
R estraint, Ascendance, and Masc. -Fern.
In Table 3, the reliabilities are presented, by trait, for each
of the eight groups tested, together with overall reliabilities. In
Table 4, the reliabilities of each of the three traits are presented
for the four groups which were requested to falsify, taken together,
and sim ilarly for the four remaining groups.
The reliabilities for individual groups, on all three tra it-
scales, are quite sim ilar. They appear to be sufficiently high for
differentiating groups of individuals, although in three cases they
fall below ,70, No significant differences exist between the re lia
bility indices of the three trait-scales. The reliabilities of the
* Falsify* - as against the 'non- Falsify* - groups, for all traits
taken together, are very sim ilar. The overall reliability index--
which includes all eight groups and all three tra its --is . 931. This
level of reliability is usually considered satisfactory for most
testing purposes, given scales which are relatively valid.
R estraint item s were, on the average, less often responded
to correctly--that is, in the manner requested by the instructions --
than were either the Masc. -Fern, item s or the Ascendance items
( p < . 0 1 , in both cases )--with the mean percentage-of-m isses for
all scale-Item s being approximately 22.4, 13.8, and 9.4, resp ec
tively. R estraint item s were, in this sense, the most difficult of
the three. Masc. -Fern, items were on the average significantly
more difficult ( p < . 01 ) than w ere Ascendance item s. On the
average, item s on all three scales were responded to in the manner
requested by the instructions approximately 85 per cent of the tim e.
^The formula is
- in
Q
Table 3 . — I n d ic e s o f R e l i a b i l i t y o f T r a i t - s c a l e s , fo r I n d iv id u a l Groups
D e s ig n a tio n
o f
Group N
R e s tr a in t
r t t
Ascendance^
r t t
M asc.-Fern.b
T
t t
K 31 .846 .864 .919
L 32 .829 . 846 .799
M 31 .702 .885 .748
R 32 .726 .730 .816
S 22 .853 .808 .605
T 22 .724 .853 .860
W 22 .812 .734 .675
Z 21 .883 .507 .86 3
E ig h t Groups
Combined 213 .797 .83 0 .826
O v e r a ll r t t (8 groups and 3 s c a l e s ) = .93 1
a A l l e s tim a t e s o f r e l i a b i l i t y were computed from item s t a t i s t i c s , u s in g
th e R uder-R ichardson form ula No. 20
Twenty item s in th e s c a le
c Groups K,L,M ,R,S,T,W , and Z
T able 4 . — I n d ic e s o f R e l i a b i l i t y , f o r Groups R equ ested to F a l s i f y ,
Taken T o g eth er, and fo r Groups Not R equested to F a l s i f y ,
Taken T o g e th e r, by I n d iv id u a l T r a i t - s c a l e s a
T r a it - s c a l e
R e s tr a in t .775 .3 1 2
Ascendance .846 .817
M3sc.-Feta, .8 3 4 .8 1 4
r t t ( f o r th r e e .845 .8 3 4
t r a i t - s c a l e s )
a K uder-R ichardson form ula No. 20 was u sed t o compute a l l
r e l i a b i l i t i e s
Groups r e q u e ste d to f a l s i f y (N"126)
° Groups n o t r e q u e ste d t o f a l s i f y (N«87)
K + L + M + R
r t t
S + T + W + Z
r t t
50
Differentiations Among Ratings And Rankings
In the case of twenty-five out of thirty basic com parisons--
as seen in Table 5 --an affirmative answer may be given to the
question of whether the present sample of subjects reliably differ
entiated between and within different background and residual
factors, as represented by such concepts as "importance of tra it
to adults" versus "importance of tra it to self, " and so on. The
actual and the desired positions on traits appear relatively well
differentiated in term s of mean scores, as do the ratings which
refer to self as compared with those relating to adults.
4
The critical ratios, together with the correlations which are
presented, indicate that the variables chosen for the present study
were perceived by the subjects in term s of separable, albeit over
lapping contexts, and that they in this sense possess some m easure
of specific cognitive meaning, or meanings, for this sample of
subjects. The relationship of these separable subjective percep
tions to the amount of falsification, will be dealt with later.
Relative Importance Of T raits
A test of Hypothesis 5 required the presence of reliable differ
ences as to ranked importance of the three tra its, so that the most
im portant and the least important traits could be distinguished.
Table 6 indicates that the required differences were present.
The tra it of Masc. -Fern, was judged to be, on the average, the
m ost important of the three traits personally; R estraint was consid
ered the next most important, personally. Ascendance was consid
ered the least important.
All high/low groupings reported in the present study represent
either median splits, or splits which were as close to the median as
possible. In over 90% of these latter splits, the divisions, p and pl,
fell within the lim its of . 4 to . 6 ; and all splits which involved the
criterion m easures fell within p and p1 ranges of .46 to . 54. All
cosine-pi estim ates of the tetrachoric r were corrected for non
median dichotomizations, where necessary. { 5 )
T a b le 5 . — R e l i a b i l i t y o f D i f f e r e n c e s b e tw e e n Means o f
In d e p e n d e n t M easu res: W ith in - and B e tw e e n -T r a it
C om p arison s (N“ 213)
T r a i t -
s c a l e and ■ V a r ia b le
Z - r a t i o h
b etw een
g P a ir s o f
Mean Sigma 12 V a r ia b le s
Im p ort, to A d u lts
R e s t r a in t v s
Im p ort, t o S e l f
2.16® .8 1
A .2 9 2 .7 5
1.99® .7 8
Im port, to A d u lts
A scen d an ce v s
Im p ort, t o S e l f
1 .9 3 e .8 1
.29 3 .8 6
2.18® .8 2
Im p ort, to A d u lts
M asc.-F em . v s
Im p ort, to S e l f
1.91® .8 1
.5 9 1 .6 0 °
1.83® .8 1
A d u lt s ' a c t u a l p o s.
R e s t r a in t v s
A d u l t s ' d e s ir e d pos.
5 .4 2 1 .4 6
.1 4 6 .9 1
6 .5 2 2 .0 8
A d u lt s ' a c t u a l p o s.
R e s t r a in t v s
Own a c t u a l p o s.
5 .4 2 1 .4 6
.2 0 2 .7 1
5 .8 3 1 .9 9
A d u lts ' a c t u a l p o s.
R e s t r a in t v s
Own d e s ir e d p o s.
5 .4 2 1 .4 6
- . 0 4 5 .8 4
6 .4 9 2 .3 0
A d u lts' d e s ir e d p o s .
R e s t r a in t v s
Own a c t u a l p o s .
6 .5 2 2 .0 8
.2 8 4 .1 4
5 .8 3 1 .9 9
A d u l t s ' d e s i r e d p o s.
R e s t r a in t v s
Own d e s ir e d p o s .
6 .5 2 2 .0 8 ,
.7 8 0 .2 9
6 .4 9 2 .3 0
Own a c t u a l p o s.
R e s t r a in t v s
Own d e s ir e d p o s.
5 .8 3 1 .9 9
.4 8 4 .4 8
6 .4 9 2 .3 0
A d u lts' a c t u a l p o s.
A scen d an ce v s
A d u l t s ' d e s i r e d p o s.
4 .6 6 1 .4 8
.0 7 1 5 .6 2
7 .0 0 1 .7 6
T a b le 5 (c o n tin u e d )
52
T r a i t -
Eicale and V a r ia b le
Z - r a t i o h
b etw een
r g P a ir s o f
Mean Sigma 12 V a r ia b le s
A d u l t s ' a c t u a l p o s .
A scen d an ce v s
Own a c t u a l p o s.
4 .6 6 1 .4 8
- . 1 2 6 .3 5
5 .7 9 2 .0 1
A d u l t s ' a c t u a l p o s .
A scen d an ce v s
Own d e s i r e d p o s .
4 .6 6 1 .4 8
- . 0 2 1 5 .5 2
7 .0 6 1 .7 6
A d u l t s ' d e s i r e d p o s .
A scen d an ce v s
Own a c t u a l p o s .
7 .0 0 1 .7 6
.38 8 .5 4
5 .7 9 2 .0 1
A d u lts' d e s ir e d p o s .
A scen d an ce v s
Own d e s i r e d p o s .
7 .0 0 1 .7 6 .
.67 0 .6 5
7 .0 6 1 .7 6
Own a c t u a l p o s .
A scen d an ce v s
Own d e s i r e d p o s .
5 .7 9 2 .0 1
.7 1 1 3 .0 7
7 .0 6 1 .7 6
A d u l t s ' a c t u a l p o s .
M asc.-Fem . v s
A d u l t s ' d e s i r e d p o s.
5 .8 4 1 .6 9
.5 4 1 1 .7 1
7 .2 6 2 .0 0
A d u lts' a c t u a l p o s .
M asc.-F em . v s
Own a c t u a l p o s .
5 .8 4 1 .6 9
.6 2 3 .9 5
6 .2 7 1 .9 6
A d u l t s ' a c t u a l p o s .
M asc.-Fem . v s
Own d e s i r e d p o s .
5 .8 4 1 .6 9
.5 0 9 .4 7
7 .0 2 1 .9 6
A d u l t s ' d e s i r e d p o s .
M asc.-Fem . v s
Own a c t u a l p o s.
7 .2 6 2 .0 0
.4 5 7 .0 4
6 .2 7 1 .9 6
A d u l t s ' d e s i r e d p o s .
M asc.-Fem . v s
Own d e s ir e d p o s .
7 .2 6 2 .0 0
.7 9 2 .6 9
7 .0 2 1 .9 6
Own a c t u a l p o s.
M asc.-F em . v s
Own d e s ir e d p o s .
6 .2 7 1 .9 6
.6 3 6 .6 1
7 .0 2 1 .9 6
T a b le 5 (c o n tin u e d )
53
T r a i t -
s c a l e and V a r ia b le Mean Sigma
r s
12
Z - r a t i o h
b etw een
P a ir s o f
V a r ia b le s !
R e s t r a i n t
v s I n t e n s i t y o f adap.
A scen d a n ce
4 .8 9 f
4 . 7 7 f
1 .6 5
1 .6 8
.7 5 1 .3 6 °
R e s t r a i n t
v s I n t e n s i t y o f adap .
M asc.-F em .
4 .8 9 *
4 . 1 0 f
1 .6 5
1 .7 2
.5 4 6 . 98a
A scen d an ce
v s I n t e n s i t y o f adap.
M asc.-F em .
4 . 77*
4 .1 0 f
1 .6 8
1 .7 2
.5 0 5 .6 5 a
R e s t r a in t
v s Min. t o l . l i m i t
A scen d a n ce
3 .2 7
3 .5 6
1 .4 8
1 .6 8
.4 3 2 .5 7 b
R e s t r a in t
v s Min. t o l . l i m i t
M asc.-F em .
3 .2 7
4 .7 6
1 .4 8
2 .1 0
.4 6 1 1 .5 0 a
A scen d an ce
v s Min. t o l . l i m i t
M asc.-F em .
3 .5 6
4 .7 6
1 .6 8
2 .1 0
.26 7. 72a
R e s t r a in t
v s Max. t o l . l i m i t
A scen d an ce
8 .1 7
8 .3 5
1 .3 3
1 .4 3
.3 3 1 .6 5 °
R e s t r a in t
v s Max. t o l . l i m i t
M asc.-F em .
8 .1 7
8 .9 0
1 .3 3
1 .3 8
.2 4 6 .4 7 a
A scen d a n ce
v s Max. t o l . l i m i t
M asc.-F em .
8 .3 5
8 .9 0
1 .4 3
1 .3 8
.4 2 5 . 36a
a p < .0 1 b p < .05 c p < .2 0 » .1 0
d
p > .5 0
e •1■ - h ig h e s t rank; ’3' = lo w e s t rank
^ I n v e r s e l y r e l a t e d to i n t e n s i t y o f a d a p ta tio n
T a b le 5 (c o n c lu d e d )
54
® C o s in e - p i e s t . o f r fc. W ith 211 d . f . , r ’ s o f .1 4 and
.19 a r e r e q u ir e d f o r a c ce p ta n c e o f t h e h y p o th e s is o f
n o n -ch a n ce c o r r e l a t i o n a t th e .0 5 and .0 1 l e v e l s o f
c o n f id e n c e , r e s p e c t i v e l y .
S tan d ard e r r o r s o f th e means were c o r r e c t e d f o r th e
e f f e c t s o f m a tc h in g . The m u ltip le c o r r e l a t i o n b etw een
the t h r e e m a tch in g v a r i a b l e s and ea ch o f th e in d ep en d e n t
m easures was u se d i n t h i s c o r r e c t io n .
T a b le 6 . — R e l i a b i l i t y o f D if f e r e n c e s b e tw e en Means o f
R anked Im p o rta n ce o f T r a it s (N »213)
Judgment and T r a it
Z - r a t i o
Mean b etw een
Im portance P a ir s g f
o f T r a i t 0 Sigma T r a i t s 0
R e s t r a in t
Im port, to A d u lts v s
A scend ance
2 .1 6 .8 1
3 .0 2
1 .9 3 .8 1
R e s t r a in t
Im p ort, to A d u lts v s
M asc.-Fem .
2 .1 6 .8 1
3 .2 4
1 .9 1 .8 1
A scend ance
Im port, to A d u lts v s
M asc.-Fern.
1 .9 3 .8 1
0 .2 4
1 .9 1 .8 1
R e s t r a in t
Im p ort, to S e l f v s
A scen d an ce
1 .9 9 .7 8 ,
2 .5 9
2 .1 8 .8 2
R e s t r a in t
Im port, to S e l f v s
M asc.-Fem .
1 .9 9 .7 8 .
2 .0 5
1 .8 3 .8 1
A scen d an ce
Im port, to S e l f v s
M asc.-Fem .
2 .1 8 .8 2
4 .5 8
1 .8 3 .8 1
c '1 ' = m ost im p o rta n t; '3 ' = l e a s t im p ortan t
^ C o r r e la tio n term n o t in c lu d e d in e s t im a t e o f sta n d a r d
e r r o r , due to n e c e s s a r y in te r d e p e n d e n c e o f th e r a n k in g s
56
The tra its of Masc. -Fem . and Ascendance were of equal
average importance to m ost adults ( z = 0. 24 ), whereas R estraint
was considered the least im portant of the three traits as far as most
adults were concerned, as judged by the sample of subjects.
Differentiations Among Criterion M easures
The manner in which the subjects portrayed the requested
trait-positions--for instance, "rather R estrained," or "8 on a 10
point scale"--varied significantly among each of the three GZTS
scales, as measured in term s of the unweighted mean total number
of correct, shift responses. Table 7 indicates that the most
extreme portrayals occurred on the trait of Ascendance ( 18. 12
points out of 2 0 possible ), and the least extreme on the tra it of
R estraint, although identical positions had been requested for each. ^
Such differences provide a basis for rank-ordering these tra its as
to the amount of shift--the mean number of correct responses--
associated with each. A reliable rank-ordering was essential for
testing the assertions relating to greater relative amounts of shifting
on particular traits as compared with other tra its.
Subjects’ Actual Position On The Traits
The hypotheses of the present study were based on the prem ise
that shift, for each trait, took place from the same, average p re
test adaptation level in the case of all eight groups, irrespective of
the test instructions which these groups received. B artlett’s te st
5 '
Out of a maximum possible number of twenty such responses.
These were responses which were in the direction--on the bipolar
tra its - -specified by the instructions. A score of twenty would
represent the most extreme possible portrayal, whether in connec
tion with the lowest point (e. g. , "not at all Restrained") or the
highest point (e. g ., "very much Restrained") on these tra its.
^"Extrem ity of portrayal" - -that is, a greater as against a
le sse r amount of shift--w as always determined as follows: Of all
the subjects who--in accordance with instructions- - shifted in the
direction of a high point on a tra it, the 50% (approx. ) receiving the
larg est number of shift responses were combined with the 50%
(approx. ) who had likewise received the largest number of shift
T ab le 7 . — R e l i a b i l i t y o f D if f e r e n c e s b etw een Means o f
GZTS T r a i t - s c a l e s (N -213)
T r a i t - s c a l e s
ffffT 'I I l l i 411 I,., . . . T C g
Mean Sigma r 12°
Z - r a tio
betw een
T r a i t - s c a l e s
R e s tr a in t
v s
A scendance
15.27
1 8.12
3 .7 1
2 .8 5
.4 7 b 1 2 .0 0 a
R e s tr a in t
v s
M asc.-Fem.
15.27
17.18
3 .7 1
3 .2 8
.2 7 b 6 . 57a
A scendance
v s
M asc.-Fem.
18.1 2
1 7 .18
2 .8 5
3 .2 8
. 31b 3 . 79a
S i g n i f i c a n t beyond th e .0 1 l e v e l . A lt e r n a t e s o l u t i o n s ,
u s in g th e Mann-Whitney U t e s t , y ie ld e d s im ila r z - r a t i o s —
9 .7 2 , 6 .4 6 , and 3 .7 0 , r e s p e c t i v e l y .
b A l l c o e f f i c i e n t s s i g n i f i c a n t beyond the .0 1 l e v e l
c C o s in e -p i e s tim a te s o f r t
58
for homogeneity of variances perm its us to accept the hypothesis of
no difference in perceived actual position-on-trait between the eight
•7
experimental groups, taken together, for each of the three traits.
The means and standard deviations of these self-ratings are
O
presented in Table 8 . The means are sufficiently close to the
midpoints of the traits to justify the prem ise that, on the average,
the more extreme self-portrayals also involved the greater degree
of falsification and that, as a result, the total number of responses
made in the direction specified by a set of instructions would be a
A
meaningful index of the amount of shift on a trait.
responses in accordance with instructions referring to a low point
on the same trait. These subjects, taken together, constituted the
high-shift group. All remaining subjects constituted the low-shift
group. The number of subjects comprising the high- and low-
shift groups for R estraint, Ascendance, and Masc. -Fem . were
112/101, 115/98, and 106/107, respectively.
equalled 0. 332, 3.680, and 2.891 for the traits of
R estraint, Ascendance, and M asc.-F em ., respectively. ( 7 d.f.
in each case. )
8
On the tra it of Masc. -Fem . , a rating of '0 r ( "very much
Feminine" ), for fem ales, was considered equivalent to a rating of
*10' ( "very much Masculine" ), for m ales. Females who, for
example, chose a rating of ' 2 ’, had responded in term s of the same
number of scale-differentiations--relative to the maximum and
minimum possible number of scale differentiations--as did males
who chose a rating of ‘S1 . As a result, Masc. -Fem . ratings of 0,
1 , 2 , and so on, on the part of fem ales, could with some justifica
tion be changed to scores of 1 0 , 9 > and so on, for statistical
purposes.
9
Given the request that subjects convey a position which was
located closer to either extreme of the tra it than it was to the m id
point.
T ab le 8 . — S e l f - r a t i n g s o f A c tu a l P o s i t i o n
on T r a it (N -213)
T r a it Mean3 Sigma
R e s t r a in t 5 .8 3 1 .9 9
A scen d a n ce 5 .7 9 2 .0 1
M asc.-F em . 6 .2 7 1 .9 6
3 A l l s c a l e s range from '0 ' (low )
th ro u gh '1 0 1 (h ig h )
C H A PT E R V
MAIN FINDINGS
The general hypothesis of the present study was that variation
in the degree of inaccurate self-portrayal is a function of specifiable
i focal, background, and residual stimulus conditions. T his--
Hypothesis 1 - -was evaluated in term s of the seven remaining hypoth
eses, which dealt with specific stimulus conditions.
All of these stimulus conditions--residual and background
conditions, in particular--w ere found to be significantly related to
variations in degree of inaccurate self-portrayal, and were related
in the manner predicted, thus lending support to the general hypoth
esis, It will be noted, however, that each of the stimulus conditions
had been represented by more than one variable and had appeared
in connection with more than one hypothesis--thus making it possible
for all three stimulus conditions to receive support in term s of one
or another of these variables and hypotheses, or in term s of them
all. It was also possible for any given hypothesis to receive support
in term s of one or more of the traits, or in term s of them all. In
view of the fact that two of the specific hypotheses did not receive
support, but that four of the remaining five hypotheses did receive
support in connection with at least two of the three traits, it would
appear fair to conclude that Hypothesis 1 was given partial yet
relatively general support in term s of the presently explored v a ria
bles.
The bases for the above generalization will now be presented,
in connection with Hypotheses 2 through 8 . To simplify the p resen
tation, we will first present the findings which relate to subjects1
judgments about their own actual position on each trait, the personal
importance of each tra it, and so on. Further evidence^ w ill then
be given in the section entitled "findings relating to judgments about
adults" ; this later section will also contain findings relating to sub-
^Bearing on Hypotheses 2, 4, and 5.
60
61
jects' judgments about their own desired positions on each trait.
Findings Relating To One's Own Perceptions Of A T rait
Hypothesis 2a. - -Differential shift as a function of individual's
position on tra it: Table 9 indicates that for the total sample of
; 213 subjects, this hypothesis was supported in the case of R estraint
( P < . 01 ) and Ascendance ( p < . 05 ), but not in the case of
Masc. -Fem . The subjects who perceived them selves to be re la
tively high (upper one-half) on the two form er traits differed
significantly, as to degree of shift on those tra its, from the subjects
who perceived them selves as being relatively low (lower one-half)
on the same tra its . M oreover, perceived position-of-self on the
tra it of R estraint was related to the total amount of shift on an
unweighted combination of all three tra its , and tended to be thusly
related in the case of Ascendance.
Hypothesis 2b. - -Amount of shift as a function of individual's
position on tra it: 1. The prediction that the subjects who rate
them selves as being relatively higher on a tra it will be the ones
who shift farthest in the direction of the high end of that same trait,
is seen, in Table 9 , to receive support in the case of R estraint
( p < . 05 ). For the tra it of Ascendance, however, the reverse of
this prediction tends to be supported ( p « .10 > .05 ), In the case
of Masc. -Fem . , the prediction is not supported.
2. With respect to the traits of Ascendance and Masc. -Fem . ,
Table 9 indicates no support for the prediction that the subjects who
rate them selves as being relatively lower on a tra it will be the ones
who shift farthest in the direction of the low end of that trait. The
reverse of this same prediction receives support in the case of
R estraint ( p < .01 ); here, subjects who rated them selves as being
relatively higher on R estraint were those also more likely to shift
farthest in the given direction.
62
T a b le 9 . — R e la t io n s h ip b e tw e en I n d i v i d u a l ' s S e l f - r a t i n g o f
A c t u a l P o s i t i o n on T r a it and Amount o f S h i f t on
Same T r a it d
T r a i t - s c a l e s
( c r i t e r i o n
m easu res)
S e l f - R a t in g s
o f
A c t u a l P o s i t i o n
R e s t r a in t A scend ance Masc . -Fem.
T o t a l
Sam ple R e s t r a in t . 23a » • • •
o f
E ig h t A scend ance • • - , 1 4 b • •
Groups
(N=213) M asc.-Fem . « • • « .0 6
1 RAM ® . 28a - , l l b .0 3
R e s t r a in t 5 . 20b • • • *
Groups
M+R+ A scen d an ce5 » « 16° • •
S+T
(N»107) M asc.-F em .5 • • • • .0 3
R e s t r a in t . 35a • ♦ • •
Groups
K+L+ A scen d an ce • * - . 1 2 • •
W+Z
(N“ 106) M asc.-Fem . * • • • .1 3
a p < .0 1 b p < .0 5 c p < .1 0 > .0 5
J
C o s in e - p i e s t i m a t e s o f r fc
e An u n w eigh ted t o t a l o f th e i n d i v i d u a l ' s s c o r e s on
R e s t r a i n t , A sce n d a n c e, and M asc.-Fem . S ee T ab le 29
f o r i n t e r c o r r e l a t i o n s among t h e s e m ea su res, in c lu d in g
I RAM.
^ Convey 'lo w ' p o in t on t h i s t r a i t . ('L ow 1: ' 2 ' , on a
T a b le 9 (c o n tin u e d )
63
10 p o i n t s c a l e . )
® Convey 'h ig h ' p o in t on t h i s t r a i t . ( ’H ig h 1:
10 p o in t s c a l e . )
Hypothesis 3a. - -Differential shift as a function of individual's
intensity of adaptation: Table 10 indicates that for the total sample
of 213 subjects, either support or a tendency for support was given
to this hypothesis in the case of all three traits--R estra in t
( p < . 05 ), Ascendance ( p < . 05 ), and M asc. -Fern. ( p < . 10 >
. 05 ). In addition, the intensity of adaptation of subjects to the
tra its of R estraint and Ascendance was related to the total amount
of shift on the unweighted combination of all three traits ( p < . 0 1 ,
in both cases), and tended to be thusly related in the case of Ascen
dance ( p < ,10 > , 05 ).
Hypothesis 3b. - -Amount of shift as a function of individual's
intensity of adaptation: With respect to the tra it of Ascendance,
Table 10 indicates support for the prediction that the amount of
shift on given traits will be proportionately less for the individuals
who m anifest greater strength of adaptation to those traits ( p < . 05).
Amount of shift associated with intensity of adaptation to the tra its of
R estraint and Masc. -Fern, tends to also support the prediction
2
(p < .10 > .05 in both cases).
Hypothesis 4a. - -Differential shift as a function of importance
of tra it to individual: Table 11 indicates that for the total sample
of 213 subjects, support was given to this hypothesis in the case of
Ascendance ( p < . 05 ), and tended to be given in the case of
R estraint (p < .10 > .05). Furtherm ore, the value of the tra it of
Masc. -Fern, to individual subjects was related to the total amount
of shift on all three tra its combined, and tended to be related in the
same manner in the case of R estraint.
2 “ |
The inter cor relations among the three tra its, as to intensity
of adaptation, are all in a positive direction and are well beyond the
.01 level of significance. The cosine-pi estim ates are .75, .54,
and . 50 for R estraint vs. Ascendance, for R estraint vs. Masc. -
Fern. , and for Ascendance vs. Masc. -Fem, , respectively (N 1 213).
65
T a b le 1 0 . — R e la t io n s h ip b e tw e e n I n d i v i d u a l ' s I n t e n s i t y o f
A d a p ta tio n t o T r a it and Amount o f S h i f t on Same
T r a it® (N »213)
T r a i t - s c a l e s
( c r i t e r i o n
m easu res)
* *
I n t e n s i t y o f A d a p ta tio n
R e s t r a in t A scend ance M asc.-Fem .
R e s t r a in t - . 1 3 ° • • • •
A scendance • • 15b
M asc. -Fern. • • - . 1 3 °
I BAM - . 22a - . 12c - . 19a
a p < .01 b p < .05 c p < .1 0 > .0 5
j
C o s in e - p i e s t im a t e s o f r t
66
T a b le 1 1 . — R e la t io n s h ip b etw een Im p ortan ce o f T r a it to
I n d iv id u a l S u b j e c ts and Amount o f S h i f t on
Same T r a it c (N“ 2 1 3 )
T r a i t - s c a l e s
( c r i t e r i o n
m easures)
Im portance o f T r a it
R e s t r a in t A scendance M asc.-Fem . ;
R e s t r a in t . 12b
A scendance . 16a
M asc.-Fem . . . . . .0 1
I RAM . 12b .07 - . 17a
a p < . 05 b p < . 10 > . 05
c C o s in e -p i e s t im a t e s o f r t
Hypothesis 4b. - -Amount of shift as a function of importance
of tra it to individual: With respect to the tra it of Ascendance,
Table 11 indicates support for the prediction that the amount of
shift on given traits will be proportionately greater for the indivi
duals who judge those traits to have a greater relative degree of
personal importance ( p < . 05 ). Amount of shift associated with
the personal importance of the tra it of Restraint tends to also sup
port the prediction ( p < .10 > .05). No support is given in the
case of Masc. -Fern.
Hypothesis 5a. - -Amount of shift as a function of average, relative
importance of tra it to a sample of individuals: This hypothesis
a sse rts that the strength of the relationship which exists between
a tra it which is considered more valuable personally and the amount
of shift which occurs on that same trait, will be greater than will
be the relationship which exists between the traits which are less
valued personally and the amount of shift which occurs on these
latter tra its . For the sample of 213 subjects, the present tra its
were ordered in the following manner as to average personal
importance: Masc. -Fern, (first), R estraint (second), Ascendance
(third).
1. In regard to the relationship between Position on a tra it
and amount of shift in the direction of the high end of the tra it
( Hypothesis 2b-l ), Hypothesis 5 tends to be supported when com
paring the most important with the least important trait: Table 12
indicates that the relationship between position on tra it and the
amount of shift tends to be stronger, in the predicted direction, for
the tra it of Masc. - Fern, than for the tra it of Ascendance ( p < . 10 >
. 05 ). The comparison of R estraint with Ascendance also supports
the above prediction ( p < . 05 ). No support is given when com
paring the most important trait, Masc. -Fern. , with R estraint, ■
which is the second most important trait.
3 ~ ~ “—
In the direction hypothesized for the particular stimulus
conditions. Stated differently, the predicted direction of the above
relationships is significantly more likely to occur in connection with
the tra its which have greater personal value to the sample of
T able 1 2 .— R e l i a b i l i t y o f D iff e r e n c e s betw een C o e f f ic i e n t s o f C o r r e la tio n betw een
P o s i t io n on , Value o f , and I n t e n s i t y o f A d a p ta tio n to P a ir s o f T r a it s and
th e Amount o f S h i f t on R e s p e c tiv e C r ite r io n T r a i t s , i n C on nection w ith
A v era g e, P e r so n a l Im portance o f T r a its ^
T r a i t - s c a l e s
Mean D iffe r e n c e
betw een C o r r e la tio n
C o e f f ic i e n t s o f
T r a i t - s c a l e s w ith
C r ite r io n Measures
Standard E rror
o f D iffe r e n c e
betw een
C o r r e la tio n
C o e f f ic i e n t s
(< r dr) d
Z - r a tio between
P a ir s o f
T r a i t - s c a l e s
P o sitio n ®
(N -107)
M asc.-Fem . v s A sc.
R e s tr . v s A sc.
M asc.-Fem . v s R e s tr .
.19
.36
.17
.1 1 0
.147
.129
1.73®
2 .4 5 b
1 .3 2
U
P o s i t io n
(N-106)
M asc.-Fem . v s A sc.
R e s tr . v s A sc.
M asc.-Fem . v s R e s tr .
.25
.47
.2 2
.1 1 1
.142
.122
2 . 25b
3 . 31a
1.80®
I n t e n s i t y o f
A d a p ta tio n
(N-213)
M asc.-Fem . v s A sc.
R e s tr . v s A sc.
M asc.-Fem . v s R e str .
.0 2
.0 2
.0 0
.069
.049
.067
0 .2 9
0 .4 1
0 .0 0
Im portance
o f T r a it
(N-213)
M asc.-Fem . v s A sc.
R e s tr . v s A sc.
M asc.-Fem . v s R e s tr .
.15
.0 4
.1 1
. 096e
. 095e
.0 9 6 e
1.5 6
0 .4 2
1 .1 5
a p < .0 1 b p < .05
e C o r r e la tio n term n o t in c lu d e d in
c p < .1 0 > .05
e s tim a te o f stan dard
d r ~ x------------------------—
\ 2 r r < r < r crv
V 11 34 iH 00 ;
e r r o r , due to n e c e ss a r y i n t e r -
T a b le 12 (c o n tin u e d )
dependence o f th e r a n k in g s.
f
C o s in e -p i e s t im a t e s o f r t
® P o s i t io n on t r a i t v s Convey
L
P o s i t io n on t r a i t v s Convey
1 lo w 1 p o in t ( 12 1)
'h ig h ' p o in t ( ' 8 ' )
o
x O
The above statem ents relate to ’Own Actual* position on tra it.
In addition to this variable, Hypothesis 5a tends to be supported
(in the context of 2b-1) when comparing R estraint with Ascendance
in connection with 'Own D esired* position on tra it ( p < .10 > . 05 );
the opposite of the prediction is supported when comparing Masc. -
Fern, with Restraint ( p < . 05 ). In the same context, support
occurs when comparing R estraint with Ascendance, and when com
paring Masc. - Fern, with R estraint, in connection with ’Adults* D e
sired* position on tra it ( p < .01 and p < .05, respectively).
2. Regarding the relationship between Position and amount of
shift in the direction of the low end of a tra it ( Hypothesis 2b-2 ),
the opposite of Hypothesis 5a receives support in the case of M asc. -
Fern, versus Ascendance ( p < . 05 ), and in the case of R estraint
versus Ascendance ( p < . 01 ). However, Table 12 indicates a
tendency for support of the prediction when comparing Masc. - Fern,
with R estraint { p < . 10 > . 05 ). ^
In connection with ’Own D esired* position on trait, Hypothesis
5a is supported, in the context of 2b-2, when comparing Masc. -Fern,
with R estraint ( p < . 05 ), and tends to be supported when com
paring M asc. -Fern, with Ascendance ( p < . 10 > . 05 ). The
opposite of the prediction is supported, however, when making the
latter comparison in connection with ’Adults* Actual* position on
tra it ( p < . 01 ), and also when comparing R estraint with Ascen
dance in this same context ( p < . 01 ).
3. In regard to the relationship between Intensity of Adapta
tion and amount of shift ( Hypothesis 3b ), Hypothesis 5a receives
individuals. This may occur whether or not any of the relationships
between the particular stimulus conditions and the respective c rite
rion m easures have, in th eir own right, reached levels of statistical
significance.
4
As compared with the tra it of Masc. - Fern. , the trait of
R estraint is significantly m ore likely to be positively related to the
amount of shift--in accordance with Hypothesis 2b-2.
71
no support from any of the three possible combinations of tra its,
as seen in Table 12 .
4. Regarding the relationship between the Importance of a tra it
for individual subjects and the amount of shift on that same tra it
(Hypothesis 4b), Hypothesis 5 receives no support from any of the
; three possible combinations of p airs of tra its . Table 12 indicates
: that statistical reliability was p resen t in the case of none of the
differences between the coefficients of correlation of each m em ber
of the above p airs, and its respective criterion m easure.
Hypotheses 6, 7, and 8
The hypotheses which follow w ere studied by means of
experim ental manipulation of instructions and of task, for eight
groups of subjects. They relate to (a) the presence versus absence
of request to falsify, (b) the source of request, and (c) the d irec
tion of shift. The p airs of groups which were compared with one
another--both individually and in com bination--in order to te st these
5
hypotheses, are shown in Table 13 .
The mean scores received by the eight groups on the criterion
m easures are presented in Table 14 , together with other descriptive
information. The Mann-Whitney U te s t was selected for testing the
significance of differences between groups, as to number of correct
responses on individual criterion m easures, and on all criterion
m easures combined. The results of this test are presented in
Tables 15 and 16 .
In general, only a moderate amount of support was given to
Hypotheses 6a, 7a, and 8a, in which were predicted the presence
of differential shift among given groups. Almost no support was
5
A lis t of standard deviations for the judgments about Im por
tance of and Position on tra its, and fo r the Intensity of Adaptation to
tra its, appears in Table 30 , in relation to the four groups which
were requested to falsify as compared with the four groups which
were not requested to falsify. With few exceptions, these m easures
of variability are comparable in degree.
72
T a b le 1 3 . — S e t s o f Groups Compared on E x p e r im e n t a lly M anipu
l a t e d V a r i a b l e s , f o r T e s t in g H y p o th e se s 6 , 7 ,
and 8 .
' F a l s i f y
Groups 1 Task
'Do n o t
F a l s i f y
G r o u p s' Task
K 8 - 2 , Pa W 2 - 2 , pb ;
Hypo L 8 - 2 , I z 2 - 2 , I
t h e s i s 6
M 2 - 8 , P s 8 - 8 , P
R 2 —8 , I T 8 - 8 , I
'P e r s o n a l
S o u rce
G ro u p s' Task
'Im p e r so n a l
S o u r c e
G roups 1 T ask
K 8 - 2 , P L 8—2 , 1
Hypo M 2 8 , P R 2 - 8 , I
t h e s i s 7
S 8 —8 , P T 00
I
00
H
W
2 - 2 , P Z 2 - 2 , 1
' S h i f t -
Away-From-
D e s i r e d -
P o le G ro u p s' Task
'S h ift-T o w a r d
D e s i r e d -
P o le G ro u p s' T ask
K 8 - 2 , P M 2 - 8 , P
Hypo L 8 - 2 , I R 2 - 8 , 1
t h e s i s 8
W 2 - 2 , P S 8 - 8 , P
Z 2 - 2 , I T 8 - 8 , I
a F or e x a m p le , Group K ' s h i f t s ' from '8 ' (o n a 10 p o i n t
s c a l e ) tow ard ' 2 ' , f o r P e r s o n a l r e a s o n s .
b F or e x a m p le , Group W — w it h w h ich Group K i s b e in g com
p a r e d — ' s h i f t s ' from '2 ' tow ard ' 2 ' , a l s o f o r P e r s o n a l
r e a s o n s . The l e t t e r ' I 1, i n o t h e r c e l l s , d e s i g n a t e s
'I m p e r s o n a l' ( E x t e r n a l) r e a s o n s .
Table 1 4 . — Means, Sigm as, M edians, and S e m i- in t e r q u a r t ile Ranges o f Number o f C orrect
R esponses to T r a i t - s c a l e s , by I n d iv id u a l Groups8 (N*213)
Group D e sig n a tio n
T r a it - s c a l e s N -
K
31
L
32
M
31
R
32
S
22
T
22
W
22
Z
21
Mean 15.1 0 16.06 1 4 .7 1 15.25 1 5 .3 2 1 5 .7 3 14.50 15.4 8
R e s t r a in t
Sigma 4 .0 8 3 .6 2 3 .2 1 3 .2 2 3.9 7 3 .0 8 3 .8 8 4 .4 5
Median 16.5 8 1 7 .6 3 1 5 .6 3 1 5 .8 3 1 6 .5 0 1 6 .0 0 15.50 17.2 5
Q
2 .9 3 1 .9 5 2 .7 3 2 .2 0 2 .8 5 3 .0 4 3 .3 8 4 .5 9
Mean 17 .6 5 1 8 .0 6 1 7 .9 4 18.7 8 18.27 16.55 18.68 19 .0 0
Ascendance
Sigma 3 .3 4 2 .9 6 3 .2 5 1 .8 3 2 .5 6 3 .7 3 1 .9 6 1 .3 1
Median 19.0 6 1 9 .1 4 1 9 .1 4 19.2 8 19.0 0 1 6 .5 0 1 9 .5 0 1 9 .6 3
Q
1 .4 9 0 .9 2 1.45 1 .2 9 1 .3 6 3 .2 0 0 .9 5 1 .2 6
Mean 17.39 1 7.69 16.97 17.63 17.6 8 1 5 .1 4 17.6 4 17.2 9
M asc.-Fem.
Sigma 4 .1 4 2 .8 3 2 .8 0 2 .8 8 2.08 4 .3 2 2 .2 9 3 .4 4
Median 1 9 .0 4 18.67 1 8 .0 0 1 8 .5 0 18.17 15.25 1 8 .10 18.86
Q
1 .8 6 1 .4 7 2 .1 9 1 .8 3 1 .2 3 2 .9 4 1.1 6 2 .0 0
a
Maximum p o s s ib le sc o r e * 20
T ab le 1 5 .— C r i t i c a l R a tio s fo r M ann-W hitney T e s t o f R e l i a b i l i t y o f D if f e r e n c e s in Amount
o f S h i f t on T r a it s , b etw een C o n tr a s tin g Groups (N=213)
Combined Groupings____________ C r i t i c a l R a tio s______________
(1) (2) R e s t r a in t Ascendance M asc.-Fem . I R A M 6
H ypoth-
e s x s 6
K. L. M, R vs
( ’ f a l s i f y ')
(N = 126)
W. Z, S , T
('d o n o t
f a l s i f y ')
(N = 87)
0 .2 2 x
0 .3 4 1 .6 4 $
1 ‘ 20k
H ypoth
e s i s 7
K, M, S , W vs
( 'p e r s o n a l
s o u r c e ')
(N « 106)
L, R, T, Z
( 'e x t e r n a l
s o u r c e ')
(N * = 107)
1 .4 7 c 0 .° 2 x 0 .0 6
X
0 .6 2
X
Hypoth
e s i s 8
K, L, W, Z vs
( ' s h i f t away from
d e s ir e d p o l e ')
(N ~ 106)
M, R, S , T
( ' s h i f t toward
d e s ir e d p o l e 1)
(N ■ 107)
1 .2 7 d
X
° .9 3 x
2 * 0 ° x
1 .6 4 $
An 'x ' f o llo w in g any c r i t i c a l r a t i o i n d ic a t e s th a t th e f i r s t o f th e two combined group
in g s to be l i s t e d in any row d id a g r e a t e r o v e r - a l l amount o f s h i f t i n g , i r r e s p e c t i v e o f
w hether th e d if f e r e n c e betw een grou p in gs was s t a t i s t i c a l l y s i g n i f i c a n t .
a p < .0 5 b p < .1 1 > .1 0 C p < .2 0 > .10 d p < .25 > .20
e An unw eighted com b ination o f th e s c o r e s fo r R e s t r a in t , A scend ance, and M asc.-Fem.
T ab le 1 6 .— C r i t i c a l R a tio s f o r M ann-W hitney T e s t o f R e l i a b i l i t y o f D if f e r e n c e s in Amount
o f S h i f t on T r a i t s , b etw een P a ir s o f I n d iv id u a l Groups
Groups
T o ta l
C r i t i c a l R a tio s
(1)
(2)
N R e s tr a in t A scendance M asc.-Fem. IRAMG
K v s w 53
° ‘ 75x
0 .9 2 1 .4 8 ° 0 .9 5
Hypoth L v s z 53 0 .0 9
x
1 .0 3
° ' 29X
0 .5 9 x
e s i s 6
M v s S 53 1.07 0 .2 0 0 .5 8 0 .8 9
R vs T 54 0 .4 3
2 *43x
2.38® 1.46®
K v s L 63 1 .0 7 0 .2 9 0 .4 4 0 .6 2
Hypoth M v s R 63 0 .6 8 0 .6 8 1 .2 1
0 .8 5
e s i s 7
S v s T 44 0 .0 8 1 . 5 1 1 1 .8 6 * 1 .2 6
X
w v s Z 43 1 .0 3 0 .3 3
° - 52x
0 .4 8
K vs M 62 0.94* 0 .3 3 1 .6 7 * 1 . 34c
X
Hypoth
e s i s 8
L v s R 64 1 .3 0 ° 0 .3 5 0 .1 6 0 .9 2
W v s S 44
X
0 .9 2 0 . 72x
X
0 .0 4
X
0 .5 5
Z vs T 43 0 .2 9 2 . 37a
X
0 . ! 9 x 1 .2 5
X
An ' x 1 fo llo w in g any c r i t i c a l r a t i o in d ic a t e s th a t th e f i r s t o f th e two groups l i s t e d in
any row d id a g r e a te r amount o f s h i f t i n g , i r r e s p e c t i v e o f w hether th e d if f e r e n c e betw een
groups was s t a t i s t i c a l l y s i g n i f i c a n t . Thus, in the top row, group K d id more s h i f t i n g
than group W.
a p < .0 5 b p < .1 0 C p < .2 0 > .1 0 £
e An unw eighted com b in ation o f th e s c o r e s fo r R e s t r a in t , A scend ance, and M asc.-Fem.
76
given to Hypotheses 6b, 7b, and 8b, which consisted of predictions
as to which groups would manifest the greater degree of shifting;
instead, some support was given to the opposite of these predictions.
Hypothesis 6a. --D ifferential shift as a function of presence versus
absence of request to falsify; In comparing groups K, L, M, and
R, taken together, with groups W, Z, S, and T, taken together,
no support was given, on the traits of Restraint and Ascendance,
to the prediction that "subjects who have been requested to falsely
portray them selves as possessing specified levels of a trait, will
p ortray them selves as being at a level which differs significantly
from that of subjects who have been requested to convey the same
level of the tra it, but who have not had the task structured as one
which involves falsification." ( p.30)» Table 15 indicates that a
tendency to support this prediction was very nearly obtained in the
case of Masc. -Fern. ( p < . 11 > . 10 ); but no support was received
in the case of all three traits combined.
Table 16 indicates that, at the level of individual groups, only
the comparison of groups R and T provided support for Hypothesis
6a ( p < .05 for Ascendance, and p < . 05 for M asc.-Fem . ).
The remaining three pairs of comparison groups provided no support.
Hypothesis 6b. - -Amount of shift as a function of presence versus
absence of request to falsify: Table 15 indicates that a tendency
to support the opposite of the prediction was very nearly obtained
in the case of Masc. -Fern. ( p < . 11 > . 10 ) ; The subjects who
had been requested to portray them selves as possessing relatively
high or low levels of this trait, without falsifying ( groups W, Z,
S, and T ) , were slightly less likely to shift as far in the requested
direction as were the subjects who had been requested to falsely
portray them selves as possessing the same levels of the trait
( groups K, L>, M, and R ).
At the level of individual groups, Table 16 indicates support
for the opposite of the prediction, when comparing groups R and T
on the traits of Ascendance and Masc. -Fem . ( p < . 05, with both
tra its).
Hypothesis 7a. - -Differential shift as a function of source of request:
Table 15 indicates no support for Hypothesis 7a, when groups K, M,
S, and W, taken together, are compared with groups L, R, T, and
Z, taken together. At the level of individual groups, only the com
parison of S and T resulted in any tendency to support the hypothesis
this occurred with the tra it of Masc. -Fem . ( p < .10 > . 05 ).
?
Hypothesis 7b. - -Amount of shift as a function of source of request:
Table 15 indicates that the comparison of groups K, M, S, and W,
taken together, with groups L, R, T, and Z, taken together, gave
no support to the prediction that the amount of shift would be propor
tionately greater in the case of requests which were structured as
having a common, external origin, rather than as representing
personal preferences of the individual subject.
When comparing individual groups, the opposite of the present
prediction tended to be supported in the case of the two groups, and
the single trait, referred to in the discussion of Hypothesis 7a.
Hypothesis 8a. - -Differential shift as a function of direction of shift
in relation to desirability of trait: ( As we have seen, the P erson
ally and the Socially Desired Optimum levels, of all three tra its,
w ere located in the direction of the high end of the tr a its .) Table 15
indicates support for Hypothesis 8a in the case of Restraint ( p < . 05)
and a near tendency in the direction of support in the case of all
th ree tra its combined ( p < .11 > .10), when comparing groups
K, Li, W, and Z, taken together, with groups M, R, S > and T, taken
together. When comparing individual groups, support is given in
the case of only two of the twelve basic combinations of groups.
Four groups for each of three traits.
78
Hypothesis 8b, - -Amount of shift as a function of direction of shift
in relation to desirability of trait- This hypothesis is not supported.
Although some reliable differences do occur between individual
groups in regard to direction of shift--Hypothesis 8a--the direction
of all such differences is contrary to Hypothesis 8b, as seen in
Tables 15 and 16 , Thus, the amount of shift which occurred in the
direction of the Personally and Socially Desired Optimum points
was proportionately less than the amount of shift which occurred
in the opposite direction.
7
Findings Relating To Judgments About Adults
The findings which relate to judgments about adults, and to
judgments about one’s own desired position on given tra its, in
general provide less support for Hypotheses 2 and 4 --but more for
Hypothesis 5--than do those relating to subjects’ judgments about
their own actual position on each trait, and about the personal
importance which they attach to these traits.
For the sample of 213 subjects, Table 17 indicates some
support for Hypothesis 2a--differential shift as a function of position
on trait--w ith respect to:
1. Own desired position on Masc. -Fem. (p < .10 » .05)
2. Adults’ actual position on Ascendance ( p < . 01 )
3. Adults’ desired position on Restraint ( p < . 10 > . 05 )
In the remaining six possible combinations of ’Position by
T ra it,1 the hypothesis is not supported.
The prediction that the subjects who rate them selves as being
relatively higher on a tra it will also be likelier to shift farthest in
the direction of the high end of the same trait--H ypothesis 2b-l--
received no support. ( Table 17 .) The opposite of this prediction
received support with respect to:
1. Own desired position on R estraint ( p < . 05 )
2. Adults* desired position on R estraint ( p < . 01 )
7
Also to judgments of personally desired position on tra its.
T a b le 1 7 ( A ) .— R e la tio n s h ip b etw een S e lf - R a t in g s o f D e s ir e d P o s i t io n on T r a it and Amount
o f S h i f t on Same T r a it d
T r a it - S c a le s
( c r i t e r i o n m easures)
S e lf - R a t in g s o f D e sir e d P o s it io n
R e s t r a in t A scendance M asc.-Fem .
T o ta l R e s t r a in t .06 . . • •
Sample Ascendance . . .05
• • n
o f M asc.-Fem. • • « * .1 2 c
E ig h t I RAM .0 4 .0 2 - .0 8
Groups
(N -213)
Groups R e s tr a in t 20b . . • *
M + R + A scendancee .05 « *
S + T M asc.-Fem.
• • • • .1 2
(N»107)
Groups
f
R e s t r a in t £ . 19b • •
K + L + A scendance£ .08 • •
W + Z M asc.-Fem . • • • • - .0 9
(N -106)
3 p < .0 1 b p < .0 5 c p < .1 0 > .05 ^ C o s in e -p i e s tim a te s o f r.
t
e Convey 1 low' p o in t ( 12 ' , on a 10 p o in t s c a le ) on t h i s t r a i t
^ Convey 'h igh 1 p o in t ( ' 8 ' , on a 10 p o in t s c a l e ) on t h i s t r a i t
Table 1 7 ( B ) .— R e la tio n s h ip betw een I n d i v i d u a l' s Judgment o f A d u lts' A c tu a l P o s i t io n on
T r a its and th e I n d iv id u a l's Amount o f S h i f t on Same T r a it^
T r a it - S c a le s R a tin g s o f A d u lts' A c tu a l P o s i t io n ______
( c r i t e r i o n m easures) R e s tr a in t Ascendance M asc.-Fem.
T o ta l R e s tr a in t .07 • • * •
Sample Ascendance • • - . 18a * •
o f M asc.-Fem. • • • • .01
E ig h t I R A M .09 - .0 9 - .1 0
Groups
(N=»213)
Groups
R e s t r a in t e - .0 7 • » • •
M + R + A scendancee « « .00 • •
S + T M asc.-Fem. « • « • .01
(N*“107)
Groups
f
R e s t r a in t £ . 21b
• • a
• •
K + L + Ascendance^ » • - .3 4 • •
W + Z M asc.-Fem. • • « • .0 3
(N=106)
a p < .0 1 ^ p < .05 c p < .1 0 > .05 ^ C o s in e -p i e s tim a te s o f r t
e Convey 'low ' p o in t ( ' 2 ' , on a 10 p o in t s c a le ) on t h i s t r a i t
£
Convey 'h ig h ' p o in t ( ' 8 ' , on a 10 p o in t s c a l e ) on t h i s t r a i t
T ab le 1 7 ( C ) .— R e la tio n s h ip b etw een I n d iv id u a l’ s Judgm ent o f A d u lts ' D e s ir e d P o s i t io n on
T r a it and th e I n d iv id u a l's Amount o f S h i f t on Same T r a it^
T r a it - S c a le s
( c r i t e r i o n m easures)
R atin gs
R e s tr a in t
o f A d u lts 1 D e sir e d
Ascendance
P o s it io n
M asc.-Fem.
T o ta l R e s tr a in t - . 12c • * • •
Sample Ascendance « * •
o
O O
* •
o f M asc.-Fem. ¥ ¥ • • .05
E ig h t IRAM - .0 5 •
o
O O
- .0 8
Groups
(N-213)
Groups
0
R e s tr a in t - , 4 2 a * * • •
M + R + A scendancee • • .06 • •
S + T M asc.-Fem. • • • *
i
•
o
•O
(N-107)
Groups
f
R e s tr a in t £ . 21b • • • •
K + L + A scendance£ • • .1 4 ■ •
W + Z M asc.-Fem. • • • • .1 2
(N-106)
a p < .0 1 k p < .05 c p < .1 0 > .0 5 ^ C o sin e -p i e s t im a te s o f r t
p
Convey 'low ' p o in t ( ' 2 ' , on a 10 p o in t s c a le ) on t h i s t r a i t
^ Convey 'h ig h ' p o in t ( ' 8 ' , on a 10 p o in t s c a le ) on t h i s t r a i t
82
The prediction that the subjects who rate them selves as being
relatively lower on a tra it will be the ones who shift farthest in the
direction of the low end of that trait--H ypothesis 2b-2--received
support with respect to:
1. Adults* actual position on Ascendance ( p < . 01 )
However, the opposite of this prediction received support
with respect to:
2. Own desired position on Restraint ( p < . 05 )
3 * Adults* actual position on R estraint { p * . 05 )
4. Adults’ desired position on R estraint ( p < . 05 ).
The hypothesis of differential shift as a function of importance
of tr a it- - Hypothesis 4 a , applied to adults--tends to be supported
with respect to the tra it of Masc. - Fem . ( p < . 10 > . 05 ). How
ever, this difference is in a direction opposite to that predicted by
Hypothesis 4b (when applied to adults) , viz. , that the amount of
shift will be greater fo r individuals who judge given traits to have a
g reater relative degree of importance to adults.
Whereas Hypothesis 5a received a fair amount of support and
a moderate amount of opposition in connection with the personal
importance of tra its, it received considerable support and almost
no opposition in relation to the judged importance of traits to adults.
For the sample of 213 subjects, these traits were ordered in
the following manner, as to average importance to adults: Masc. -
Fem . and Ascendance w ere, statistically, tied for first, while
R estraint was judged to be the least im portant of the three. Table
18 sum m arizes the contexts in which Hypothesis 5a was either
supported, not supported, o r directly opposed (reverse of prediction
supported).
8 3
Table 1 8 .— R e s u lts o f C r it ic a l r a t io t e s t s o f R e li a b il it y
o f D iffe r e n c e s betw een C o e f fic ie n t s o f C orrela
tion® between P o s it io n on , Value o f , and
I n te n s ity o f A d ap tation to P a irs o f T r a its and
th e Amount o f S h if t on R esp ec tiv e C r ite r io n
T r a it s , in C onnection w ith Average Im portance
o f T r a its to A d u lts^
Own Own A d u lts 1 A d u lts '
A ctu a l D esir ed A ctu a l D esired
P o s itio n P o s itio n P o s itio n P o s itio n
on on on on
T r a it-S c a le s T r a it T r a it T r a it T ra it
Hypoth. 2 b -1
Masc.-Fem. v s R estr. • • sb . . s a
Ascend, vs R e str . 0b s c . . sb
Hypoth. 2b-2
Masc.-Fem. v s R estr. s c s b
Ascend, vs R e str . s a
_a
. . S . .
Hypoth. 3b
Masc.-Fem. v s R estr. • •
Ascend, vs R e str . • •
Not a p p lic a b le
Hypoth. 4b
Masc.-Fem. v s R estr. • *
Ascend, v s R e str . • •
S: h y p o th e s is supported
0: o p p o s ite o f h y p o th e sis supported
Blank c e l l : h y p o th e sis n o t
supported
a p < .01 b p < .05 c p < .10 > .05
^ A ll c r i t i c a l r a t io s were computed by means o f th e
form ula: _ /“ ZZ I
^ ^ + ‘ V , - " V
£ 2 v 1 1 J * » l i
C o sin e-p i e stim a te o f r t
C H A P T E R VI
SUPPLEMENTARY ANALYSES
Findings which will aid in the interpretation of those given
in the previous chapter, will now be presented.
Relationships Between Independent Variables
And C riterion Measures
Importance of tra it for self as compared with importance of tra it
for adults. - -Subjects who ranked the tra it of R estraint as being of
greater relative importance to them selves than to adults, did a
g reater amount of shifting on the tra it of R estraint than did the sub
jects who did not consider the tra it to be more important to them
selves than to adults ( p < . 01 ). 1 The same was found for the
tra it of A scendance--greater importance to self than to adults,
g reater shifting ( p < . 05 ) --but was not found for Masc. -Fem .
It was seen in the previous chapter that less shifting was
associated with greater intensity of adaptation, for all three tra its.
The present findings would then suggest that the subjects who have a
greater intensity of adaptation to R estraint and to Ascendance may
be those who consider these tra its more important to adults than to
them selves. This relationship did, in fact, occur for the tra it of
R estraint ( p < . 01 ), and did not occur for Masc. - Fem. Con
tra ry to expectations, however, no significant relationship occurred
in the case of Ascendance.
Amount of difference between types of position on trait, for self
and for adults. - -By subtracting the rated score given to Adults1
desired position on tra it from that of Own actual position on tra it,
as one example, an unweighted difference score was obtained for
this and, sim ilarly, for other combinations of self/adults by actual/
desired positions on given tra its. Significant differences in amount
^Cosine-pi estim ate of tetrachoric r. ^Cf. n. 1.
Cf. n. 1. q a
85
of shift were associated with three of the resulting six possible
4
combinations of pairs of positions, as seen in Table 19 .
With one exception, it was found that the greater is the abso
lute difference between such positions, the greater is the amount of
shift for these three com binations--irrespective of the direction of
these differences. The subjects who believed th ere were sm aller
differences between own-and-adults*, actual-and-desired positions,
in general did less shifting in the case of these combinations.
It may be noted that each of these particular combinations of
positions differentiated subjects on at least two of the three tra its,
as to amount of shift, whereas none of the remaining three combina
tions differentiated subjects on any of the traits. It would appear
that the form er combinations, or relationships, are particularly
influential within the present context of subjects and tasks, taken
together.
Direction of difference between types of position on trait, for self
and for adults. - -The ratings which an individual gave to the separate
m embers (*X * and * Y * , for the present ) of any of the six combina
tions of positions just discussed, could produce one of three possible
directional relationships between those m em bers: X > Y , X = Y ,
or X < Y.
Table 20 contains the significant relationships between amount
of shift and these directional relationships, for the five position-
combinations which differentiated within the sample of subjects. ^
The three significant combinations which appeared in connection
with "amount of difference, " discussed above, all reappeared.
With each of these combinations, greater shift was likely to occur
4
Only the absolute difference in scale scores is considered in
the present analysis; the direction of such differences is considered
in a subsequent analysis.
5
Only the direction of difference, and not the degree, is con
sidered in the present analysis.
T ab le 1 9 .— R e la tio n s h ip b etw een th e Amount o£ D if f e r e n c e b etw een Types o f P o s i t io n on
T r a it , and th e Amount o f S h i f t on Same T r a it d (N“ 213)
Types o f P o s itio n on T ra it R estra in t Ascendance Masc.-Fem.
Own A ctual minus Own D esired • • . 14b . 14b
A d u lts' A ctu al minus A d u lts' D esired 13° . 24a . 13°
A d u lts' A ctu al minus Own D esired • • . 23a . 27a
a p < .01 b p < .05 c p < .10 > .05
^ C o sin e-p i estim a te s o f r t « R ela tio n sh ip s which do n ot reach the .10 le v e l o f
s ig n ific a n c e are n ot in clu d ed .
□ 0
o
Table 2 0 .—R ela tio n sh ip between the D ir e c tio n o f D iffe r en ce between Types o f P o sitio n
on T r a it, and the Amount o f S h if t on Same T raith (N ^ IS )
D ir e c tio n a l R ela tio n sh ip between
Types o f P o s itio n on T ra it R estra in t Ascendance Masc.-Fem.
Own A c tu a l> Own D esired • • • • - . 21a
A d u lts1 A c tu a l> A d u lts1 D esired - . l l + c • • 16b
A d u lts' A ctual* Own A ctual 20a • • • •
A dults A c tu a l’ O wn D esired • * 23a # *
A d u lts' D e sir e d ’ O wn A ctual ~. 13° • • . 18a
3 p < .01 b p < .05 C p < .10 > .05
b C o sin e-p i estim a te s o f r t . R ela tio n sh ip s which do not reach the .1 0 le v e l o f
s ig n ific a n c e are not in clu d ed .
88
when desired positions had been judged higher than actual positions;
and, conversely, less shift occurred with these same combinations
when the individual’s actual position on any of the three tra its had
been judged to be greater than his desired position.
Additionally, in the case of four of the seven findings presented
in Table 20 , the relationship between the p airs of positions on
given tra its and the intensity of adaptation to the respective tra its
attained or approached statistical significance in the direction
expected on the basis of the direction of relationship between the
form er positions and the amount of shift. ^ No support was given
7
for the remaining relationships.
Position on trait, and tolerance lim its for tra it. - -Certain reg u lari
ties were suggested by an analysis of the absolute difference in the
num ber of scale points between the eight possible combinations of
position on tra it and (a) the maximum and (b) the minimum
"tolerance lim its" which subjects perceived themselves as p o ssess
ing with each trait.
Table 21 indicates that, with a single exception, the significant
differences within the sample as to the amount of shift on each of the
three tra its , occurred in connection with each of three different kinds
of trait-positions, respectively: Less shift occurred on Masc. -Fern,
when either tolerance lim it was closer to the point at which subjects
personally want to be located, on that trait. On Restraint, the sub
jects who did less shifting were those who (a) believe that their
minimum tolerance lim it is located n earer to the point which they
The following p airs of positions were related to intensity of
adaptation, in the expected direction:
Adults’ act. pos. vs. Adults’ des. pos. on R estr. (p « . 05);
Adults’ act. pos. vs. Own act. pos. on R estr. (p < .01);
Adults’ act. pos. vs. Adults’ des. pos. on Masc, -Fern, (p < .05);
Adults' des. pos. vs. Own act. pos. on M asc.-Fern, (p < . 10 > .05).
7
Not included in Table 20 , for purpose of simplification.
g
Without consideration of the direction of the difference.
89
T able 2 1 .— R e la tio n sh ip betw een th e Amount o f D iffe r e n c e
betw een P o s itio n on T r a it and T oleran ce L im its
fo r T r a it, and th e Amount o f S h if t on th e same
T r a itc (N=213)
Types o f P o s it io n
on T r a it,
and
T oleran ce L im its
C r ite r io n M easures
R e s tr a in t A scendance M asc.-Fem.
Own A ctu a l
minus
Minimum T o le r .
. 22a
Own A ctu a l
minus
Maximum T o le r .
19a
Own D esired
minus
Minimum T o le r .
. 29a . 16b
Own D esired
minus
Maximum T o le r .
. 15b
A d u lts' A ctu a l
minus
Minimum T o le r .
. . - . 27a . .
A d u lts' A ctu a l
minus
Maximum T o le r .
. 24a
a p < .0 1 b p < .05
c C o sin e -p i e s tim a te s o f r t » R e la tio n sh ip s w hich do not
reach th e .1 0 le v e l o f s ig n if ic a n c e are n o t in c lu d e d .
90
estim ate to be tbeir own actual position on this tra it, or who
(b) believe that th eir maximum tolerance lim it is located farther
from this same, actual position. The reverse of these relationships
(those applicable to the tra it of Restraint) apply in the case of
A scendance--this tim e, however, with reference to adults* actual
g
location on that tra it.
Total amount of difference between all possible positions on tra it. - -
Some subjects believed there was little difference, in scale scores,
as to 'own actual, * 'own desired, ' 'adults' actual, * and 'adults'
desired* positions on given traits; other subjects believed that con
siderable differences were present. Using a simple, unweighted
index of these perceived differences, ^ for all six possible com
binations of self/adults by actual/desired positions, it was found
that subjects who believed there was greater difference, or
"discrepancy," between these four positions on the tra it of R estraint,
tended to also shift less on this trait ( p < . 10 > . 05 ), whereas the
subjects who had tended to shift less (p < .10 > .05) with respect
to Ascendance, were the ones likelier to, on the contrary, perceive
fewer such discrepancies for this latter tra it. No significant
differences were found in relation to Masc. -Fern.
Relationships Between Matching Variables
And C riterion M easures
CEEB, age, and sex in relation to amount of shift. - - Table 22
indicates that, on the traits of R estraint and Masc. -F ern., fem ales
shifted significantly more than did m ales. The females also shifted
more on a combination of all three traits, taken together. There
- - -
No significant differences occurred in connection with
adults' desired position on tra it.
■^The index was: (own actual minus own desired) + (own actual
minus adults* actual) + (own actual minus adults' desired) + (own
desired minus adults* actual) + ■ (own desired minus adults' desired) +
(adults' actual minus adults' desired). Algebraic signs were d is
regarded, to exclude from consideration the direction of differences.
T ab le 2 2 .— R e la tio n s h ip b etw een M atching V a r ia b le s and Amount o f S h i f t on
T r a its ^ (N -213)
T r a it - S c a le s M atching V a r ia b le s
M u ltip le R
betw een
CEEB + AGE + SEX
and
C r ite r io n Measure ( c r i t e r i o n m easures) CEEB AGE SEXS
R e s t r a in t . 13c - .0 4 - . 18a
h
. 25aa
Ascendance .06 .0 4
- . 0 5 h .0 0
Masc.-Fern. .07 - .1 2 °
_ ig a
‘ i y h
. 27aa
I RAM f . 14b - . 14b - 17b
* h
. 31aa
33 A - l 3 A * 1
p < .0 1 p < .0 1 b p < .05
c
P <
.10
^ C osin e-p i e stim a te s o f r t , u n le ss otherw ise s p e c ifie d .
p
M u ltip le R, co rrected fo r sample s iz e (b ia s ). 209 d.f.
f
T o ta l amount o f s h i f t on a l l t r a i t s combined.
® N egative c o r r e la tio n s s ig n if y more s h if t by fem ales.
P oin t b is e r ia l r .
92
were no differences as to amount of shift on Ascendance, considered
individually.
The two form er findings are strengthened by those in Table 23 , ;
which indicate that fem ales1 intensity of adaptation to R estraint and
to Masc. -Fern, is significantly less than that of m ales. E arlier
findings indicated that a greater intensity of adaptation to these
traits is associated with a lesser degree of shift.
Table 23 also shows that fem ales considered R estraint and
Masc. -Fern, to be of greater value than did m ales. The finding
with respect to Restraint is in line with earlier findings to the
effect that the subjects who considered R estraint to be of greater
value, ^ were also likelier to shift more on this trait. ( p. 67).
The findings with respect to Masc. -Fern, receive no support from
this sam e, earlier source.
Subjects with higher CEEB scores tended to shift more on
R estraint, and shifted more on all traits combined, despite the
12
earlier-rep o rted relationship between Board scores and sex.
In addition, a significant multiple correlation exists between
the matching variables and the criterion m easures, for two of the
three tra its . This dem onstrates the importance of having obtained
the present close matching between experimental groups, as to the
above variables.
Relationships Between Matching Variables
And Independent Measures
Table 23 indicates that one or more genuine multiple c o rrela
tions also exist between the above matching variables and the judged
importance of and position on each tra it, as well as the intensity of
adaptation to one of the tra its.
11 ' ~ [
Subjects were divided into high and low groups, as to judged
importance of Restraint.
12
The present sample of males scored higher on the CEEB.
T ab le 2 3 .— R e la t io n s h ip b etw een M atching V a r ia b le s and Im p ortan ce o f T r a it , P o s it io n
on T r a it , and I n t e n s i t y o f A d a p ta tio n to T r a it (N=213)
T r a it and
J u d g m en t I n d e p . Mea s u r e
M atching V a r ia b le s
CEEB AGE SEX
x
M u ltip le R
betw een c
CEEB AGE SEX
and Independent
Measures**
R e s t r a in t
A scendance
Masc. -Fern.
R e s t r a in t
Ascendance
Masc.-Fern.
R e s t r a in t
A scendance
M asc.-Fem.
R e s tr a in t
A scendance
M asc.-Fem.
Im portance
to
A d u lts
I m p o r ta n c e
to
S e l f
A d u lts 1
A c tu a l
P o s i t i o n
A d u lts 1
D e sire d
P o s it io n
0 )
§
•
1
- . 1 5 b
e
- 17b
' i
. 28aa
•09e
e
.18® .16
.0 0
e
.03
e
.13^ .06
- . 0 5 .
- 10.
- 17b
* J '/ i
. 21bb
,1 9 f
e • 14e
.25® . 2 1 * *
- 14b
* e
- . 0 4 . -.19® . 26aa
.27®
s - • 0 7 s
.05
m
. 26aa
- .0 7
s
13c
s
.0 1
m
.09
.06
s
- .0 4
s
.08
m
.00
- 23a
s - 07S
.0 5
m
. 21bb
23a
•01S ■16»
m
. 21bb
- .0 6
s
* 13g
.24®
m
. 22bb
v O
1 0
T ab le 23 (C o n tin u ed )
T r a i t an d
J u d g m en t Xndep. Measure
M atching V a r ia b le s
CEEB AGE SEX
x
M u ltip le R
betw eenc
CEEB AGE SEX
and Independent
M easures^
R e s t r a in t
Ascendance
M asc.-Fem.
Own
A c tu a l
P o s i t io n
- .0 6
s
15b
s
.1 1
s
.1 2 °
s
• 14b
s
.06
s
.09
m
* 1 8 m
m
. 19a
m
.06
. 21bb
.16
R e s t r a in t
Ascendance
M asc.-Fem,
Own
D e sir e d
P o s i t io n
" 31s
. 14b
s
.0 2
s
• 0 5 S
. 19a
s
.0 0
s
.0 4
m
. 28a
m
. 20a
m
. 31aa
. 29aa
.18
R e s tr a in t
A scendance
M asc.-Fem.
I n t e n s i t y
o f
A d ap tation
- 0 3 S
- .0 8
s
• n s
.07
s
.1 1
s
.07
s
,1 5 b
m
, 27a
m
. 15b
m
.1 2
. 27aa
.1 1
aa p < .0 1
bb A[- a
p < .0 5 p < .0 1 b p < .05 C p < .10
M u ltip le R c o r r e c te d fo r sample s i z e ( b i a s ) . 209 d. f .
t r i s e r i a l r (19) 1 Phi c o e f f i c i e n t
m
P o in t b i s e r i a l r
x
I f s ig n i s ' + ' , m ales sc o r e h ig h e r or rank t r a i t as b e in g more
im p ortan t than fem a les; v i c e v e r sa i f s ig n i s
C o s in e -p i e s t i
mate o f r t
v O
95
Strikingly, differences occurred in all possible instances,
with respect to each matching variable, in the case of subjects1
Own Actual and Own Desired positions on Ascendance. Reliable
differences also occurred between males and females in connection
with the judged importance of all three traits, both for adults and
for oneself. ^
Tables 24 and 25 reveal substantial differences as to the judged
importance of tra its for adults as compared to self, particularly in
regard to sex of judge. Whereas the males tended to believe more
often than did fem ales that Masc. -Fem . was relatively more im por-
14
tant to adults of one's own sex than it was to oneself ( p < .10 »
. 05 ), the females considered the reverse to be the case ( p < . 01 ).
Males who scored higher on the CEEB tended to consider
R estraint of less importance to adults than did those who scored
low er. This difference was not found in the case of fem ales,
however.
On the tra it of Ascendance, the males who scored higher on
the CEEB also considered Ascendance more important, both to
them selves and to adults, than did those who scored lower. In this
case, the judgments of females were more clearly in the same
direction as those of the m ales, but still did not attain statistical
significance.
Age differences appeared in connection with the judgments of
m ales--but not of fem ales--as to the importance, to adults, of
R estraint and Ascendance. Older males believed that adults con
sidered Ascendance relatively more important, and that adults con-
• 15
sidered R estraint relatively less important.
13
A list of standard deviations for males and for fem ales, on
each of the above independent m easures, appears in Table 30. In
m ost instances, substantial differences do not exist between these
m easures of variability; striking exceptions are present, however.
14
As compared with the remaining two traits.
15
Judgments about the relative importance of the three traits,
whether to adults or to oneself, were negatively intercorrelated as a
necessary result of the manner in which these ratings were obtained.
T able 2 4 .— R e l i a b i l i t y o f D if f e r e n c e s betw een M ales' and F em ales' Judgments o f
Im portance o f T r a it s to S e l f a s Compared w ith Im portance to A d u lts (N=213)
Im portance to A d u lts
v s
Im portance to S e l f ,
o f T r a it
S u b je c ts
Judging
Im portance
o f T r a it d
Z - r a tio
Mean betw een
Im portance ^ P a ir s o f
o f T r a it e Sigma r -^ Judgments
Imp. to A d u lts
R e s tr . v s
Imp. to S e l f
Male
2 .2 7 .80
.4 0 2 .9 9
2 .0 3 .8 2 x
Imp. to A d u lts
R e s tr . vs
Imp. to S e l f
Female
2 .0 0 .79
.4 0 0 .8 9
1 .9 2 .7 2
Imp. to A d u lts
A scend. v s
Imp. to S e l f
Male
1.8 9 .77
• 50 1 - 77xx
2 .0 2 .8 2 xx
Imp. to A d u lts
A scend. v s
Imp. t o S e l f
Female
1.9 9 .87
• 37 4 - 53xx
2 .4 4 .76 xx
M a sc.- ImP- t0 A d u lts
vs
*era* Imp. to S e l f
Male
1 .8 4 .82
.6 3 1 .7 6
1 .9 5 .8 0 xx
M asc.- Imp- A d u lts
Fem* Imp. to S e l f
Female
2 .0 1 .79
.4 3 3 .9 6
1 .6 5 .77 x
x: ju d ged more im portant to S e l f than to A d u lts.
xx-: judged more im portant to Ad u lt s than to S e l f .
a p < .0 1 p < .0 5 c p < .1 0 > .0 5 127 m ales and 86 fem ales
e ' l Ia=most im p ortan t; ' 3 ' = l e a s t im portant p h i c o e f f i c i e n t s
T able 2 5 .— R e la tio n s h ip betw een CEEB and Judgments o f M ales and F em ales, as to
R e la tiv e Im portance o f T r a it s . R e la tio n s h ip betw een AGE and Judgm ents,
o f M ales and F em a les, as to R e la t iv e Im portance o f T r a its ^ ( ^ 2 1 3 )
CEEB AGE
T r a it and Judgment M ale6 Fem ale6 Male6 Fem ale6
R e s t r a in t
Im portance
to
A d u lts
- .1 5 ° .0 2 - . 1 5 ° - .1 2
Ascendance
Im portance
to
A d u lts
. 20b .1 2 . 20b .0 5
M asc.-Fem.
Im portance
to
A d u lts
- .0 7 - .0 8 - .1 1 .1 0
R e s t r a in t
Im portance
to
S e l f
- .1 4 .07 - .0 7 - .1 2
Ascendance
Im portance
to
S e l f
. 26a .1 4 - .0 5 .12
M asc.-Fem .
Im portance
to
S e l f
- .0 1 - .1 3 .0 3 .0 0
a p < .0 1 k p < .05 c p < .1 0 > .0 5 ^ P o in t b i s e r i a l r
e 127 m ales and 86 fem ales
vD
- J
98
Relationships Among Independent M easures
Intensity of adaptation to tra it, and the judged importance of
and position on same trait. - -Table 26 indicates th at the intensity of
adaptation to Masc. -Fem. is negatively related to the judged im por
tance of this tra it, both for adults ( p < .10 > . 05 ) and for self
( p < .05 ), and is, in the case of R estraint, positively related to
the judged importance for adults { p < . 05 ).
Subjects with a greater intensity of adaptation to Masc. -Fem .
were also more likely to see both adults and them selves as actually
possessing greater degrees of this trait--"m o re masculinity, " if
m ale, and "more femininity, " if female.
No significant relationships were found between intensity and
position for the remaining tra its , nor were any found between
intensity and importance, in the case of Ascendance.
Importance of tra it in relation to position on same tra it. - -Table 27
reveals significant relationships in the m ajority of cases, and indi
cates that each of these are in a positive direction. Subjects who
felt that adults either were or wished to be high on Restraint or on
Ascendance, ^ were also likelier to believe that adults would con
sider those tra its significantly more important than would those who
indicated lower points on the tra its, for adults. The same applied
in the case of subjects* Own Actual and Own Desired positions on
all three tra its, in relation to the tra it’s importance to Self.
For the three traits generally, the frequency of significant
relationships increased, respectively, in the following contexts:
Adults* Actual, Adults' D esired, Own Actual, Own Desired position
on tra it. In the last-m entioned instance, statistical significance was
attained or approached with each of the six combinations of variables
to be interrelated.
Subjects were dichotomized into upper and lower (approx. )
50 per cent on R estraint and Ascendance.
99
T able 2 6 . — R e la t io n s h ip b etw een I n t e n s i t y o f A d a p ta tio n to
T r a it and th e Judged Im portance o f and P o s i t i o n
on th e Same T r a it d ( n=213)
Type o f Judgment
P o s i t i o n on T r a it
o r R a tin g R e s t r a in t A scendance M asc.-Fem .
Im portance t o A d u lts . 15b - . 0 5 -.11+*
Im portance t o S e l f .0 9 .0 2 - . 15b
A d u l t s 1 A c tu a l P o s i t i o n - . 0 1 - . 1 1 . 25a
A d u l t s 1 D e s ir e d P o s i t i o n .0 8 .0 0 .0 3
Own A c tu a l P o s i t i o n .09
o
1 — 1
•
. 27a
Own D e s ir e d P o s i t i o n .1 0 .08 . 06
a p < .0 1
b p < .0 5
c p < .1 0 > .0 5
b C o s in e - p i e s t i m a t e s o f r fc
Table 2 7 .— R e la tio n s h ip betw een Im portance o f T r a it and Judged P o s i t io n on
Same T r a itd (N=213)
P o s it io n on T r a it
Judgment T r a it
A d u lts 1
A ctu a l
P o s it io n
A d u lt s 1
D e sir e d
P o s it io n
Own
A c tu a l
P o s it io n
Own
D e sir e d
P o s it io n
R e s tr a in t . 32a .1 0 . 12c . 15b
Im portance
to Ascendance .0 4 . 33a - .0 5 ,1 6 b
A d u lts
M asc.-Fem. .10 - .0 6 . 14b .1 3 °
R e s tr a in t .0 2 . 16b . 27a . 32a
Im portance
a
to Ascendance - .1 0 . 22 .40 . 37
S e l f
M asc.-Fem. . 20a . 19a . 18a . 36a
a p < .0 1 b p < .05 c p < .1 0 > .05
^ C o s in e -p i e s t im a t e s o f r t
101
A c tu a l and d e s ir e d p o s itio n s on t r a i t , f o r s e lf an d a d u lts . - -W ith few
exceptions, significant and positive relationships were found between
: the four judged positions on any given trait. Despite this, an inspec
tion of Table 28 suggests that sufficient variance remained unaccoun
ted for--betw een almost all pairs of positions--to statistically jus-
■ tify the presence of the different types of position. Inspection fu r
ther suggests that the strongest relationships are to be found between
judgments of adults* desired position and one's own desired position
on the present tra its.
Item Analysis
The biserial r, corrected for item -total overlap (22), was
used to determine whether specific questions were likely to be
responded to differently by subjects who shifted to greater or to
lesse r degrees. The findings are given in Appendix E.
Results indicate that the frequency of occurrence of differenti
ating item s, parallels, on each scale, the relative difficulty of the
respective scale, as m easured by the mean percentage-of-m isses
17
for all item s of each scale. (p. 47)- In all, 31 item s differenti
ated high-from low -shifters on the tra it of R estraint. This included
18
19 of the 20 R estraint item s, and 12 item s from other scales.
Only three item s--all from the Ascendance scale--differentiated
subjects as to amount of shift on Ascendance. Eight Masc. -Fem .
item s, and two item s from other scales, differentiated as to amount
19
of shift on Masc. - Fem.
1 7
The complete list of differentiating item s, together with
correlation coefficients, appears in Appendix E .
18
Unkeyed item s, mainly from the Ascendance scale. Such
item s, from "other" scales, usually differentiated less sharply than
item s from criterion scales.
19
Each of the three sections of the questionnaire contained 70
item s. Twenty of these were the key items used in determining the
degree of falsification on the traits of R estraint, Ascendance, and
Masc. -Fem . The present sample of subjects did not, however,
consider the remaining 50 item s, in each of these sections, to be
neutral with respect to any of the three falsification-tasks. The
102
T ab le 2 8 . — R e la tio n s h ip b etw een Judgm ents o f A d u lts' A c t u a l,
A d u lts ' D e s ir e d , Own A c t u a l, and Own D e s ir e d
P o s i t i o n s on G iven T r a its ^ (N=213)
Type o f Judgment
T r a it and Judgment
A d u l t s '
D e s ir e d
P o s i t i o n
Own
A c tu a 1
P o s i t i o n
Own
D e s ir e d
P o s i t i o n
R e s t r a in t
A d u lts 1
. 14b . 20a - . 0 4
A scendance A c tu a l
P o s i t i o n
.07 - . 12c - . 0 2
M asc.-Fem . . 54a . 62a . 50a
R e s t r a in t
A d u l t s '
• • . 28a . 78a
A scendance D e s ir e d
P o s i t i o n
• • . 38a . 67a
M asc.-Fem . • • . 45a . 79a
R e s t r a in t
Own
• * • • .4 8 a
A scendance A c tu a l
P o s i t i o n
• • • « . 71a
M asc.-Fem . • « • # . 63a
a p < .0 1 b p < .0 5 C p < .1 0 > .05
C o s in e - p i e s t im a t e s o f r t
103
total sample of subjects definitely saw in these item s implications
for whichever happened to be the tra it they were concerned with at
the tim e. Thus, for example, had the 213 subjects believed that a
given item was not relevant to the tra it of Ascendance, and had they,
as a result, answered either Yes or No to that item purely on the
basis of chance, roughly 107 Yes responses, and about the same
number of No responses, could then be anticipated. However, it
was common for the item s in all three sections to receive nearly
tw o-thirds of one or the other kind of response. Tables of the
cumulative binomial probability distribution (57) indicated that,
with p = .5 , the number of unkeyed item s receiving an amount of
Yes or No responses which could be attributed to chance alone
( p < , 05 ) was 7, 8 , and 13, in the case of these 50 Ascendance,
M a sc .-F e m ., and R estraint item s, respectively.
C H A P T E R VII
DISCUSSION AND SUMMARY
The three classes of stimulus conditions which are distinguish-
: ed by adaptation level theory, were represented as follows in the
! present investigation:
Residual
1. Own actual position on tra it (personal adaptation level)
2. Own desired position on tra it
3. Importance of tra it to self
4. Intensity of adaptation to tra it
5 . Internal source of request to complete task
Background
6 . Adults1 actual position on tra it (social adaptation level)
7. Adults1 desired position on trait
8 . Importance of trait to adults
9. Average importance of tra it to all subjects
10. Average importance of tra it to adults
11. External source of request to complete task
Focal
1 2 . Request to falsify: presence versus absence
13. Instructions as to direction of shift
Accordingly, seven types of hypotheses were distinguished--
those concerned with (a) position, (b) importance, (c) intensity,
(d) average importance, (e) source of request, (f) p resen ce/
absence of falsification, and (g) direction of shift.
Of the seven types of hypotheses that were tested, ^ five
received some degree of support. These included hypotheses 2, 3,
Hypotheses 2 through 8 .
105
4, 5, and 8 . In one or more of those which received support,
residual, background, and focal stimuli were each related signifi-
; cantly to the presence of differential falsification among individuals, !
apart from the predicted direction of the differences. Unlike the
2
residual and background stimuli, however, the focal stimuli
I showed no significant relationships to the falsification which had
; taken place in a direction predicted from the adaptation level theory.
The types of hypotheses receiving no support were the ones relating
to the source of request to falsify and to the presence versus absence
of a definite request to falsify.
In the remaining discussion, the fact that these hypotheses
could have been supported with respect to any one, or to all three
of the tra its, will be treated as follows. Hypotheses supported
with respect to one tra it will be said to have been given "slight
support" ; for those receiving support on two traits--"m o d erate
support" ; and for those with support on all three traits - -"consider
able support. " Because these three types of support could have
occurred in any or all of four contexts --(a) Own Actual, (b) Own
D esired, (c) Adults* Actual, and (d) Adults* D esired--the following
comments will, for the sake of clarity, be limited to context * a*
(Own Actual), unless otherwise specified. It is in this context that
the larg est number of significant, em pirical relationships were to
4
be found, and it is this context which --on non-em pirical grounds--
might be considered the one m ost directly relevant to and/or re p re
sentative of an individual* s adaptation level.
A conspicuous difference may be seen by comparing the three
types of hypotheses which related to indices of the adaptation level
_
Which were the ones least often represented.
3
In the case of Hypothesis 2 and Hypothesis 5.
4
In the case of the two hypotheses in which there had been
equal representation on the part of all four contexts.
106
per se ( hypotheses 2, 3, and 4 : position, intensity, and importance)
with the three types of hypotheses which did not refer directly to
I the adaptation level, but which instead tried to take into account ;
conditions which might impinge upon it, or bring about differential
changes in it ( hypotheses 6 , 7, and 8 : request to falsify or not
; falsify, source of request, and direction of shift ). All of the form er
' hypotheses received moderate to considerable support, whereas
two of the latter hypotheses- - 6 and 7--received no support, and the
third--num ber 8 --received slight support. We may say, from
this, that there appears to have been a greater degree of success
in predicting differential falsification on the basis of variables which
related to quantitative or quantified attributes of the adaptation
level--(position and intensity, prim arily)--than on the basis of
variables which related to qualitative conditions which impinged
upon the adaptation level (particularly the Falsification versus
non-Falsification structuring-of-task, and the Source of Request to
falsify). Although this generalization holds in the case of all three
tra its , the contrast in question, between quantitative attributes and
qualitative conditions, was considerably less sharp in the case of
M asc. -Fem ,
In connection with the two hypotheses which received no support,
it may be pointed out that, in order for subjects to become aware of
the (a) external versus the internal source of request, or of the (b)
falsification versus non-falsification structuring, it was necessary
for them to either read--on each of three occasions--a set of in stru c
tions approximately 250 words in length, or, in the case of * a*
(external versus internal source. ..) , to read a twenty-five word
sum m ary of these instructions. It is not improbable to suppose that
many subjects did not fully absorb the instructions, possibly as a
resu lt of their length, their relative complexity, or their relative
novelty. This may partially account for the lack of significant
relationships in connection with these two stimulus conditions.
5
But not in regard to the theoretically predicted direction.
Of the five residual stim uli and six background stimuli rep re
sented in this study, one representative of each of these classes
appeared to stand out strongly, in regard to its bearing upon
differential falsification, regardless of the tra it being falsified.
These were the intensity of adaptation to the given tra its, on the
part of individual subjects (Hypothesis 2), and the average, relative
importance of the tra its to the sample of subjects (Hypothesis 5),
respectively.
In comparison to these, one residual and one background fac-
tor showed considerably less of a capacity to discrim inate between
individuals, as to amount of shifting--again regardless of tra it
(Hypothesis 7). It would appeari from this, that an attempt to
generalize about the ability which the class of residual factors has
in regard to predicting the relative amount of falsification, in com-
7
parison to the ability possessed by background factors, would--
irrespective of context (in this case, traits)--b e overlooking such
g
apparent diversity within each of these classes of stimuli.
Apart from this overall diversity among the several background
as well as residual stimuli, there were a lim ited, but not sm all,
number of suggestions that the influence of particular background
and residual stimuli might still be unequal with respect to certain of
the attributes of an individual's adaptation level, or of the factors
which impinged upon it. In the case of the Importance of a tra it,
for example, a moderate amount of support was given to Hypothesis
4b, in connection with a residual factor (importance of tra it to indi
vidual) , whereas the opposite of support was given in connection
with an independently m easured background factor (importance of
tra it to adults).
^The Internal and External Sources of Request to falsify,
respectively.
7
At least, of those residual and background factors presently
studied.
Q
Factors other than sim ilarity in regard to range enter into
this statem ent. Both classes of stimuli accounted for roughly the
same proportion of significant relationships, in the case of hypothe
ses in which both had been represented.
T h e re w a s so m e in d ic a tio n th a t in d iv id u a ls w ho ( r e la tiv e to
j
: other individuals) were high (or low) shifters on one trait, would also
o
be high (or low) shifters on other tra its. In this particular sense,
some degree of generality may be said to have been indicated for
the phenomenon of falsification, at least as to the contexts in which
it might be manifested. It does not follow, however, that (1) there
: is only a single process of falsification--that the act of shifting
(regardless of [a] the tra it involved, or regardless of the [b]
; direction of shift) is always the product of the same factors; nor does
it follow that (2 ) all falsification will be amenable to prediction on
the basis of the same factors, apart from such differences in
object, or in direction of falsification.
The present evidence supports the view that the process of
falsification is not unitary, either in this theoretical sense (*1 ’,
above) or in this operational sense (*2f, above) of the term . It was
shown, for example, that (a) given residual as well as background
stim uli were differentially associated with the amount of shift on each
of the three tra its (Hypothesis 4b), and that (b) given residual as
well as background stimuli were significantly related, in the manner
predicted, to falsification which took place in the direction of the
high, but not the low end of a trait (Hypotheses 2b and 5).
It is interesting to note, in this same connection, that certain
links were found to exist between different types of factors and each
of the different traits: Variables which had reference to the self
appeared to be of particular significance to the tra it of R estraint,
whereas those which referred to adults showed certain close connec
tions with that of Ascendance. For example, (1) significant re la
tionships were found between the amount of falsification and the
amount of distance which existed between an individual’s Tolerance
Lim its and (a) his own actual position on R estraint, and (b) his
9
This evidence is based upon only the three traits presently
studied. Cf. intercorrelations among the criterion m easures.
(Table 29.)
109
Table 2 9 . — I n t e r c o r r e l a t i o n s o f th e C r it e r io n M easures
(N=213)
T r a it -
S c a le s R e s t r a in t A scendance M asc.-Fem . IRAM
R e s t r a in t • • .47 .27 .9 0
A scendance • ♦ « « .3 1 .7 0
M asc.-Fem . • • • • • • .5 8
C o s in e -p i e s t im a t e s o f th e t e t r a c h o r i c c o e f f i c i e n t o f
c o r r e l a t i o n . A l l c o r r e l a t i o n s a r e s i g n i f i c a n t beyond
th e .0 1 l e v e l , and a r e i n a p o s i t i v e d i r e c t i o n .
T ab le 3 0 . — S tan dard D e v ia tio n s o f Im portance o f T r a i t ,
P o s i t i o n on T r a i t , and I n t e n s i t y o f A d a p ta tio n
t o T r a it , f o r M ales Compared w ith F em a les, and
f o r Groups R eq u ested to F a l s i f y Compared w it h
Groups n o t R eq u ested to F a l s i f y
T r a it and Judgment
M ales
(N=127)
Fe
m ales
(N=86)
Not
R eq u ested R eq u ested
to t o
F a l s i f y F a l s i f y
(N=126) (N=87)
R e s t r a in t Impor .8 0 .7 9 .7 8
< * •
00
*
A scendance
ta n c e
to
.77 .87 .8 2 .8 0
M asc.-Fem . A d u lts .8 2 .79 .8 2 .7 9
R e s t r a in t Impor
.8 2 .7 6 .77 .8 1
A scendance
ta n c e
.8 2 .7 8 .8 1 .8 4
M asc.-Fem .
L v
S e l f
.8 0 .77 .8 5 .7 4
R e s t r a in t
A d u l t s '
1 .4 6 1 .4 6 1 .4 0 1 .5 3
A scendance A c tu a l 1 .5 4 1 .3 8 1 .4 6 1 .4 9
M asc.-Fem .
P o s i t i o n
1 .6 0 1 .8 0 1 .7 8 1 .6 0
R e s t r a in t
A d u lts 1
2 .1 0 2 .0 4
o
r-l
•
C M
2 .0 4
A scendance D e s ir e d 1 .8 0 1 .6 4 1 .7 5 1 .7 6
M asc.-Fem .
P o s i t i o n
1 .4 3 2 .5 1 2 .0 0 2 .0 3
R e s t r a in t
Own
1 .8 6 2 .1 4 1 .9 9 1 .9 8
A scendance A c tu a l 1 .9 4 2 .0 2 1 .9 9 2 .0 4
M asc.-Fem .
P o s i t i o n
1 .6 7 2 .2 4 1 .8 4 2 .0 6
R e s t r a in t
Own
2 .4 4 2 .0 8 2 .2 4 2 .3 8
A scendance D e s ir e d 1 .6 6 1 .7 1 1 .8 5 1 .6 1
M asc.-Fem .
P o s i t i o n
1 .4 2 2 .4 8 1 .9 1 2 .0 1
R e s t r a in t
I n t e n s i t y
1 .6 6 1 .6 2 1 .6 7 1 .6 1
A scendance o f Adap 1 .6 9 1 .6 6 1 .7 6 1 .6 1
M asc.-Fem .
t a t i o n
1 .7 3 1 .6 9 1 .7 4 1 .6 9
I l l
view of adults* actual position on Ascendance; (2) the tra it of (a)
R estraint was judged to be significantly more Important to the self
than it was to adults, whereas that of (b) Ascendance was consid
ered significantly more important to adults than it was to the self
(See Table 5); (3) the predicted relationships which existed between
the amount of falsification and the differing Positions on the tra it of
(a) R estraint, were strongest and most numerous in the context of
one’s own actual position on that trait, whereas those relating to
(b) Ascendance were both strongest and m ost numerous in the con
text of adults* actual position on the tra it (Hypotheses 2a and 2b;
See T ab lel7).
The distinctions that were drawn between Own Actual, Own
D esired, Adults’ Actual, and Adults' Desired position on a tra it
proved to be worthwhile, in the sense that, without them, an ov er
simplified picture would probably have emerged. A case in point
occurred in connection with Hypothesis 2b-1, in relation to the tra it
of R estraint. Here, the predicted relationship was supported in the
case of Own Actual position, but was not supported in the case of
Adults' Actual position; further, the opposite of this prediction
was supported in the case of Own Desired and Adults' Desired
positions. These stimuli, apparently, do not all play the same role
with respect to the process of change in adaptation level; nor do they
appear to operate in the same direction. ^ In still other contexts, ^
however, the sim ilarities equalled or even outweighed such differ
ences.
^ T h is latter point also applies in connection with the influence
of these stimuli upon the different tra its. Thus, in Hypothesis 2b-2,
the predicted relationship was found for the tra it of Ascendance,
with respect to Adults' Actual position on that trait; however, the
opposite of the predicted relationship was found for the tra it of
R estraint, with respect to Adults' Actual position on this trait.
"^For instance, those relating to Hypothesis 5a, with respect
to position on trait.
Still other, substantive findings w ere the resu lt of indices
which w ere derived from the relationships among these distinctions.
! These indices included (1) the direction of difference between one’s
own, and adults’, actual and d esired positions on tra its ; (2 ) the
absolute difference, in scale sc o re s, between all of these sam e four
! com binations of positions, taken together; (3) the absolute d iffe r
ence, in scale sc o re s, between individual combinations of position;
(4) the relativ e im portance of the tr a it to oneself as com pared with
its im portance to ad u lts. These indices--w hile not involved in the
p re se n t hypotheses - -now have reaso n to be p a rt of future predictions.
The type and amount of inform ation resulting from the sev eral
indices used in th is study, m ay be seen in the form of a "profile, "
in which the kinds of conditions associated with falsification on
p a rtic u la r tra its , appear in sum m ary form . Thus, for the tra it of
R e stra in t, the degree of falsification w ill tend to be, or w ill p ro b a
bly be, g re a te r in the case of the individual
1 . who believes him self to actually be m ore restrain ed (upper one-
half of the sam ple) than other individuals (low er one-half of the
sam ple) believe them selves to be (Hypothesis 2)12
2 . who has a le s s e r intensity of adaptation to the tra it of re s tra in t
than other individuals have (Hypothesis 3);
3 . who considers this tr a it to be relativ ely m ore im portant (to h im
self, personally) than he considers the rem aining tra its to be
(Hypothesis 4);
4 . who is shifting in the direction of the socially d esired region of
th is tr a it, ra th e r than in an opposite direction (Hypothesis 8);13
5 . who believes adults to actually be m ore restrain ed than he consid
e rs him self to be;
6 . who believes that adults want to be m ore restrain ed than he
considers him self to be;
7 . who believes adults to actually be m ore restrain ed than he
personally wants to be;
1 ?
This applies both in the case of shift which takes place in the
d irectio n of "the h ig h --m o re R estrain ed --en d of this tra it, and in th at
which occurs in the opposite direction.
13
R elative to the approxim ate midpoint of the tra it.
113
8 . who believes that there exist relatively sm all discrepancies
between Adults* Actual, Adults' D esired, Own Actual, and
Own D esired positions on the tra it of restrain t;
9 . who believes that there exists a relatively large difference
between Adults' Actual position on the tra it of restrain t, and
Adults' Desired position on this trait;
1 0 . who considers restra in t to be--relative to other tra its --a more
im portant tra it (to him self, personally) than he believes it to
be in the opinion of m ost adults;
1 1 . whose minimum tolerance lim it, for restrain t, is located
farth er from the point which he estim ates to be his Own Actual
position on this same trait; or, additionally, whose maximum
tolerance lim it on this tra it is located closer to this sam e,
actual position.
This lis t could be extended. Conceivably, sim ilar profiles
might be constructed for any m easurable trait; and, weights could
(in a more refined approach) be assigned to indicate the relative
degree of influence (or association) which each condition has upon
(with) the criterion m easure.
All three of the subject-variables that were involved in the
matching p ro cess--ag e, sex, and CEEB--w ere found to be related
to falsification. Fem ales, for example, shifted more than m ales on
an index which consisted of the total amount of shift given in the case
of all three tra its , combined. G reater shift on the p art of fem ales
has been found in other studies. (33, 46) As in these other studies,
this difference was not found with all of the tra its. For example,
in the present study, the fem ales shifted more than the m ales only
in the case of R estraint and Masc. -Fern. The adaptation level
approach sheds some light on this finding, in that the fem ales, for
instance, (a) were not found to have as great an intensity of adap-
14
tation to either of these tra its , and (b) were found to view--and
believed that adults view ed--these tra its as being relatively more
im portant than did the m ales. Since, at the same tim e, we also know
the kind of relationship which exists between (i) Intensity and (ii)
Im portance, on the one hand, and the amount of falsification, on the
other, it becomes possible to make predictions with respect to
subject-characteristics, or variables such as sex.
14
In both cases, p < . 05.
It is understandably difficult to classify variables such as age,
sex, and aptitude/achievement, simply as representing either
15
residual or background conditions. Yet--however categorized--
they m ust be dealt with: With respect to particular traits and
contexts, for example, each of these variables appeared to be as
strongly related to the present response bias as were some of
those which were closer to the main focus of investigation.
Further, the need to control for these variables was apparent even
in the p re-te st. Obviously, too, they are capable of contributing
to profiles such as the one shown above.
P arts of two hypotheses were directly contradicted. These
were Hypothesis 2b-2 and Hypothesis 5 (relative to Own Actual posi
tion, and the shifting which takes place in the direction of the low
end of a trait). Both of these predictions focused upon shift which
occurred in the direction of the low end of a trait. It had been a s
sumed that the same factors, and no others, would account for shift
which took place either toward the low or toward the high end of a
trait. But it was found that--w hereas individuals who were high on
a tra it did, as predicted, shift more toward the high end of a tr a it- -
those who were lower on the tra it did not, as predicted, shift more
_
In connection with the two latter variables, it is, for example,
probably not possible to decide this m atter on logical grounds alone,
without in the first place begging the question in term s of as sump- ,
tions which would have to be made regarding the nature of personality
and of cognition. In term s of em pirical reality, we are apparently
dealing with an interaction between all three classes of stimulus
conditions. For practical purposes, age, sex, and aptitude/
achievement may be said to function in a manner which comes
closest to fitting the broad definition of residual stimulus conditions.
This latter definition, in turn, contains im plicit assumptions about
the nature of "stability-am idst-change, " v iz ., that growth or
change--personal or otherw ise--is a manifestation of the "pooling"
(integration) of an indefinitely large number of stimulus events. (26) ;
This would seem to imply that anything which exists at any point in
time may be thought of as incorporating everything that it has been
influenced by in the past, and need not be looked at as an entirely
unique event.
toward the low end. If the low end of a trait is relatively distant
from the socially and personally desired optimum levels (for the
trait), as well as from the actual social adaptation level, it is pos
sible that subjects who consider themselves to be ’’low" on the tra it
already believe that they are so "far out" as to feel they hardly need
to shift at all, in order to carry out the instructions. Instead, they
may answer the questions relatively truthfully. This, of course, is
an ad hoc account, and must be pursued in its own right.
A sim ilar explanation may have relevance to another set of
findings. It was seen that, in regard to all hypotheses combined,
the fewest num ber of significant relationships had been found in
connection with the tra it of Masc. - Fern, - -apart from the question
of whether these relationships were in the predicted direction.
At the same tim e, Masc. -Fern, was the trait on which the present
sample of subjects were, on the average, actually located farthest
from the hypothetical midpoint ('5') : The subjects’ average degree
of possession of this trait w as--according to self-ratings- - 6 . 27 , on
a 10 point scale. ( See Table 8 .) Thus, there is a possibility that
many individuals - -particularly those already relatively high on the
trait--d id not shift very far in the direction of the requested "high"
point (’8 1 ) , and that they felt, instead, that the request would be
carried out even if they portrayed themselves in an honest light.
At the same tim e, the remaining individuals who were high on this
same tra it, but whose task had been that of shifting toward the "low"
point {’2 ’) --these subjects may have done, relatively speaking, a
considerable amount of shifting. An average difference of 1. 73
scale points existed between the actual and the requested locations
in the case of the form er individuals, whereas a difference of 4. 27
points existed in the case of the latter. If some of the individuals
who were actually fairly high on the tra it were doing relatively little
shifting, and other individuals, also high on the tra it, were doing
relatively much, then--regardless of whether the reverse applied in
the case of individuals who were fairly low on the trait--one would
not expect high degrees of positive relationship between the amount
of falsification and variables such as an individual's actual position
on the particular trait. Needless to say, it would not follow, from
these considerations, that a sm aller overall amount of falsification
had been taking place in connection with the tra it of Masc. - Fern,
than in connection with either or both of the remaining traits.
The present variables did leave considerable variance un
accounted for. N evertheless, they were, for the most p art, related
in a non-chance manner to the phenomenon or phenomena in ques
tion; and, in addition, the intercorrelations among these several
variables were not extremely high. Each such variable, in con
sequence, m erits further pursuing. Regardless of the factors that
could not be accounted for in this study, it appears, on balance,
as if a substantial amount of research could be conducted along the
present lines of endeavor, and could be carried out in term s of
variables which were sim ilar to several of those used in the present
study.
Support has been given to an approach in which falsification
is viewed as the product of an interplay between background, re sid
ual, and focal conditions. This view --together with the attendant,
operational or research requirem ents--is both more complex and
more cumbersome than that involved in any of the single-m otive
explanations. But it appears better suited to address the question of
individual differences in regard to falsification. The form er
explanations had been offered prim arily in answer to questions which
related to the existence or non-existence of the phenomena, in
particular contexts. It is not unreasonable to assume that one or
m ore of the gratification and the relief motives are operating to some
extent in m ost falsification situations; the extent to which these may
be operating has not been quantified with respect to specific situa
tions. Be that as it may, the evidence of the present study would
suggest that each of these m otives--to the extent that they are indeed
operating--w ill be likely to express them selves in differing degrees,
as a function of particular, perceived characteristics of tra its, and
of quantifiable attributes of individuals’ level of adaptation to those
tra its . Thus, further research could look into the relationships
117
which each of the different gratification and relief motives has to
the perceived characteristics of tra its (such as their relative im
portance) as well as to quantifiable attributes of individuals* level of
adaptation to traits (such as intensity of adaptation), for specific
situations. 3h these respects, the present approach might contribute
to research which is conceptualized in term s of a single motive.
In regard to Loevinger's approach, the present variables show
prom ise of supplying the needed operational links between indi
viduals' response to specific situations, aims, and objects (traits),
and the more general, or long-range (across-situation) trends in
their personality.
We are not able to indicate whether the present findings tend
to support or not support Loevinger's overall approach to response
bias, prim arily because we did not derive indices of the subjects'
level of social development, or level of self-aw areness. As an
illustration of this , it was found that there was a greater degree of
shift on the part of individuals whose Own Actual position on the
tra it of Ascendance was closest to that of the median position on
Ascendance for all individuals. ^ ( The same was found with
17
respect to [a] Own Desired position on Ascendance and [b] Adults*
18
Actual position on Masc. -Fern. , but was not found for any of the
remaining nine combinations of "position by tra it." ) According to
Loevinger's theory, the individuals who typically conform to social
stereotypes--irrespective of the nature of their own identities--m ay
also be expected to falsify more than those individuals who are less
conforming. In the present study, however, it was not known
whether those individuals who had come closest to the median trait-
positions were any more or less conforming than were the remaining
individuals. Also, it was not known whether their level of social
development or of self-aw areness--either in general or with respect
to the tra it of Ascendance--was closer to that of Stage 2 or Stage 3
of Loevinger's outline. Her theory would predict greater or lesser
■^Cosine-pi est. of r^, = .20; N= 213; p < .01.
17 18
p < .0 1 ; cosine-pi est. = .25. p < . 05 (cosine-pi est.).
118
falsification, depending upon the stage the individuals were in.
In. the case of response patterns other than that of falsifica
tion, it appears possible to use an approach sim ilar to the one
presently employed. In regard to individual differences which
might bear upon the likelihood and extent of gambling, for instance,
one could firs t distinguish--within each of the three broad classes
of stimulus conditions--a number of factors which, on rational
grounds, appeared as if they might relate to individuals1 adaptations
with respect to the contents--the perceived gains and risk s, for
exam ple--of the particular gambling situation, as well as of gam
bling situations generally.
3h this case, residual stimuli might include:
1 . the im portance, for the individual, of what he perceives to be
the greatest possible gain that might result from successful
gambling alone--its value, that is, in comparison to the value
of the perceived alternatives, or lesse r gains, that would
result from unsuccessful gambling, or from no attempt at
gambling;
2 . the individual’s actual and desired positions with respect to
the perceived "greatest gain"--that is, his nearness to or
distance from any such goal;
3. the individual’s intensity of adaptation to the present degree of
possession of that which might be obtained by means of gambling;
also, intensity of adaptation to the perceived alternatives of not
gambling, or of unsuccessful gambling;
4. the success-in-general which the individual perceives him self
to have as a gambler;
5. the average importance, to the individual, of succeeding as
compared with that of not succeeding.
Background stimuli might include;
6 . perceived penalties of gambling;
7. relative importance of the "greatest gain" to most adults;
8 . the perceived social importance of succeeding as compared with
not succeeding, in the present situation.
F o c a l s tim u li m ig h t in c lu d e ;
119
9 . the number of choices;
1 0 . the reversibility or irrev ersib ility of the decision to gamble;
1 1 . the amount of tim e allotted for making the decision to either
gamble or not gamble;
1 2 . the extent to which gambling is structured as being an
inescapable means to the desired end.
Variables sim ilar to these might also be employed in the case
of 11 impulsion, 11 as well as in that of "indecision or evasion. " (1 )
In each of th ese--as in the case of "acquiescence" - -the relationship
between an individual and a specified goal (which is perceived to be
involved in the given situation) could first be singled - out. This
relationship could then be described operationally, in term s of the
im portance which is attached by the individual to the goal, the status
or position (actual and desired) of the individual with respect to the
goal, the intensity of the individual*s adaptation to his present status
on the dimension along which the goal is to be found, and so forth.
(Such an approach would have to be supplemented with other v a ri
ables, to the extent that [a] several goals coexisted, and [b] the goal
or goals were either unclear or not conscious. ) These considera
tions relate mainly to the residual stimuli alone; thus, it is apparent
that such an approach is fairly complex, and would call for relatively
system atic research programming.
The relevance of an adaptation level approach to biases which
reflect individual differences in the definition of such relative term s
as "som etim es," or "frequently" -- " sem antics" --is apparent, and
has already been illustrated. (26) This study involved a legs
intricate--though still not simple - -approach, and accounted for a
large portion of the variance. To be sure, the possibility exists that
the phenomenon of "sem antics" is not of equal complexity with that of
"falsification. " Be that as it may, it would seem as if, in some
contexts, a somewhat less formidable degree of complexity may be
called for in inquiries which would attempt to pursue the problem of
bias, from the present fram e of reference.
Summary
120
The present study was set in the context of a growing need for
a rigorous, theoretical framework that might help system atize the
rapidly increasing amount of current research on response bias,
as well as help clarify the interpretation of particular biases.
Helson*s theory of adaptation level was selected as the fram e-of-
reference which might meet such needs. The present study consis
ted of an application of this theory to a "test case" of response
bias, viz. , the widely recognized, and relatively complex bias of
falsification.
In relation to personality questionnaires, a falsified item was
defined as one which an individual had consciously answered in such
a manner as to convey an im pression, or to obtain a score, which
would differ from that which would have resulted from his carrying
out an intention to portray an accurate picture of himself. This
definition included falsification for purposes of re se a rc h --
falsification-upon-request--as well as falsification for more obvious
form s of personal gain.
The general hypothesis was that variation in degree of in
accurate self-portrayal would be a function of specifiable focal,
background, and residual stimulus conditions; these referred to
the three classes of stimuli distinguished in adaptation level theory.
A number of variables, together with other, categorical dimensions
were chosen to represent these classes of stimuli, in the context of
falsification on bipolar tra its.
These included (a) position on tra it, (b) intensity of adap
tation to tra it, (c) relative importance of tra it (for individuals),
(d) average, relative importance of tra it (for sample of individuals),
(e) presence versus absence of definite request to falsify, (f) source
of request to complete the task, and (g) the direction of falsifica
tion. Seven kinds of specific hypotheses--in term s of which the
general hypothesis was to be evaluated--were then (1) concep
tualized in term s of these variables, and were (2) predicted, as to
121
direction of relationship, both on the basis of certain assumptions of
adaptation level theory, and on that of certain em pirical findings
resulting from it.
The study was conducted during April and May of 1962.
The subjects were 127 male and 86 female undergraduates, m ost of
whom were enrolled in different sections of an introductory psychol
ogy course, at the University of Southern California. They were
divided into eight groups, each of which received a different com
bination of task, type of instruction, or type of falsification -
structuring. A close matching was achieved for the variables of
age, sex, and College Entrance Examination Board score (or sub
stitutes thereof, with six per cent of the sample). Neither re s tric
tions nor other types of conditions were placed upon individuals, or
groups of individuals, as to eligibility to participate in the study.
The study was conducted unannounced; it took the place of a
regularly scheduled classroom lecture. Participation was voluntary.
The situation was presented as being illustrative of the kinds of
things the subjects were reading or hearing about in their course.
The participants were told that the general results would be discussed
with them, in class, at a later point in the sem ester. No other
rew ards were offered. One group of individuals believed that they
would earn additional course credits by participating in the study;
it was not possible to circumvent this pre-existing expectation.
These subjects, however, were evenly distributed among all eight
groups within the total sample. All subjects were told immediately
prior to the testing that the manner in which they responded to the
item s would have no bearing upon their grades in the course in
which they were enrolled. Anonymity prevailed, and confidentiality
was assured.
The criterion m easures employed in evaluating all hypotheses
were drawn from the R estraint, Ascendance, and Masculinity -
Femininity scales (each reduced in length) of the Guilford-Zim m er
man Temperament Survey. A pre-test--conducted with a sample
sim ilar to that used in the present study--had demonstrated the
122 1
feasibility of employing these scales, as well as that of the rem ain- ;
ing p a rt of the task: This p art consisted of fifteen ratings and two
rankings, designed to supply information relating to the first four
variables mentioned above.
The basic task, for all subjects, was that of portraying them
selves in a manner described by w ritten instructions, in connection
with particular tra its. This task was carried out three tim es--
separately for each tra it. An average of about forty-five minutes
was required to carry out all tasks, including those which involved
ratings and rankings. No time lim it was set, although most subjects
attempted to complete the total task within the regular class hour,
and were successful in doing so.
Satisfactory levels of reliability were obtained for the criterion
m easures, using the Kuder-Richardson form ula number 20.
Hypotheses relating to (a) position on tra it, (b) intensity of
adaptation to trait, (c) relative importance of trait (for individuals),
and (d) average, relative importance of tra it (for total sample of
individuals)--each of these received support in the case of two or
more tra its . The hypothesis which related to the direction of
falsification received some support in the case of one trait.
Specifically, the m ajor findings were:
1. Subjects who perceived them selves to be relatively high on
R estraint, or on Ascendance (those in the upper half of the
sam ple--that is, the subjects who, on the basis of self-ratings,
*possessed more of1 these traits)--th ese individuals differed
significantly, in the extent to which they shifted th eir scores,
from individuals who perceived them selves to be relatively
How* on those same traits. (Hypothesis supported. )
In addition, individuals who were high on R estraint were also
those most likely to shift their scores the farthest, in the
direction of the high end of that same trait (Hypothesis suppor
ted), but not, however, in the direction of the low end of that
same tra it (Opposite of hypothesis supported).
2. In the case of all three tra its , the amount of shift was propor
tionately less for individuals who manifested greater intensity
of adaptation to their perceived, actual positions on each
respective tra it. (Hypothesis supported. )
123
3. The amount of shift on R estraint, as well as on Ascendance,
was proportionately greater for individuals who considered
either of those tra its to be of relatively greater personal
importance than any of the rem aining traits. (Hypothesis
supported.) The same relationship--also hypothesized--
applied in the case of the estim ated, relative importance of
R estraint to most adults.
4. In regard to the relationship between an individual*s position
on a tra it and the amount of shift in the direction of the high
end of that same tra it, there was a significantly stronger
relationship between the trait which the total sample of
subjects had viewed as the one m ost important to them
personally (M asculinity-Femininity), and the amount of
shift bn that same tra it, as compared to the strength of
relationship which existed between the traits that had been
considered less im portant, and the amount of shift on these
latter tra its . (Hypothesis supported. ) The opposite of this
hypothesized relationship was found in the case of shift which
took place in the direction of the low end of a trait. Essentially
the same resu lts--m o st of which supported the hypothesis--
were found in connection with the amount of shift that had
occurred on tra its which the total sample of subjects considered
to be relatively more important to most adults (M asculinity-
Femininity, and Ascendance), as compared with the tra it
(R estraint; which they thought of as less im portant to m ost
adults.
5. Subjects who had been requested to portray them selves as
possessing relatively high or relatively low degrees of any of
the three tra its, without falsifying--these individuals shifted
neither more nor le s s, in the requested direction, than did
individuals who had been requested to falsely portray them selves
as possessing the same levels of the trait. (Hypothesis not
supported. )
6. Subjects whose requests to shift th eir scores had been structured
as having a common, external origin--these individuals shifted
neither more nor less than did individuals whose requests to
shift presum ably had represented their own, personal p re fe r
ences. (Hypothesis not supported. )
7. Subjects who had shifted their scores in the direction of the
pole, on the tra it of R estraint, in which were located both the
personally and the socially desired optimum levels of that same
trait--th ese individuals differed significantly, as to amount of
shift, from individuals who had shifted in the direction of the
opposite pole. (Hypothesis supported.) The same, hypothe
sized relationship was not found in the case of Ascendance,
nor in that of M asculinity-Femininity. (Hypothesis not
supported. )
124
In all, four of the seven specific hypotheses received a fair
amount of support, while a fifth hypothesis received a slight
amount of support. Residual, background, and--to a somewhat
le s se r extent--focal stim uli were each included among the
predictions which received support. Non-chance correlations
w ere typically low, in the case of all predictions.
The overall picture was thus one of partial, yet general
support for the main hypothesis.
Variables which referred to quantitative attributes of the
adaptation level per se--such as (1) intensity of adaptation,
(2) position on tra it, or (3) relative importance of tra it to
individuals - -predicted differential falsification better than such
dimensions as (4) falsification-structuring versus non-falsification-
structuring, or (5) source of request to complete the task--w hich,
in theory at least, referred more to qualitative conditions which
might impinge upon an adaptation level.
A number of indices were derived from relationships among
the different variables, and these were shown to bear several
significant relationships to the dependent m easures. When the
substantive information that had been provided by all of the indices
and variables was collated, it was seen to be possible to construct
a fairly detailed profile of conditions associated with falsification,
for given tra its.
Evidence was obtained which indicated that individuals who
falsified relatively much on any one tra it would tend to do the same
on other tra its. Even though this type of generality was suggested
in regard to the manner or contexts in which falsification was
expressed, there were, at the same tim e, a number of specific
suggestions that the process of falsification is not "unitary" --
is not the product of the same factors, in the case of all traits
and tasks.
125
A brief account was given of the manner in which an approach
sim ilar to that used in the case of falsification might be applied
to other sources of bias. Evidence from the present study provides
some basis for expecting that the type of approach which was
suggested, would be capable of bringing a substantial amount of
rigor--and research of a cumulative nature--into the general
area of bias.
BIBLIOGRAPHY
BIBLIOGRAPHY
Adams, G. S. Techniques of Minimizing or Capitalizing Upon
Response Tendencies in Structured Self-Report Inventories^
A Report to the International Psychology Congress, Bonn,
Germany, August, I960. Los Angeles , I960. ■
Bennett, G. K ., and Gordon, H. P. "Personality Test Scores
and Success in the Field of Nursing, " J. appl. Psychol. ,
1944, 28, 267-278.
Blake, R. R. , Helson, H. , and Mouton, J. S. "The Generality ;
of Conforming Behavior as a Function of Anchorage,
Difficulty of Task, and Amount of Social P re s s u re ,1 1
J. P e rs. , 1957, 25, 294-305.
Blake, R. R ., Rosenbaum, M ., and Duryea, R. A. "Gift-
giving as a Function of Group Standards, " Human
Relations, 1955, _ 8 , 61-73.
Bouvier, E .A ., P erry, N. C. , Michael, W. B ., and Hertzka,
A. F. "A Study of the E rro r in the Cosine-pi Approxima
tion to the Tetrachoric Coefficient of Correlation, "
Educ. psychol. M easmt. , 1954, 14, 690-699.
Christie, R. , Havel, J. , and Seidenberg, B. "Is the F Scale
Irreversible?" J. abn. soc. Psychol., 1958, 56, 143-
159.
Cronbach, L. J. "Response Sets and Test Validity, "
Educ. psychol. M easmt. , 1946, 6, 475-494.
C ross, O. H. "A Study of Faking on the Kuder Preference
Record," Educ. psychol. Measmt. , 1950, 10, 271-277.
Dicken, C. F. "Simulated Patterns on the California Psycho
logical Inventory," J. counsel. Psychol., I960, 7,
24-31.
Drasgow, J . , and Barnette, L. W. "F-K in a Motivated
Group," J. consult. Psychol., 1957, 21, 399-401.
Edwards, A. L. "Social D esirability and Probability of
Endorsement of Items in an Interpersonal Check L ist, "
J. abn. soc. P sychol., 1957, 55, 434-436.
___________. The Social D esirability Variable in Personality
Assessm ent and R esearch. New York: The Dryden
P ress. 1957.
127
128
13. Ekstrom , P. R. "An Attempt to Develop a Falsification Key
for the Guilford-Martin Personnel Inventory. " Unpub
lished M aster's thesis, The University of Southern
California, Los Angeles, 1950.
14. Ellis, A. "The Validity of Personality Questionnaires, "
Psychol. Bull. , 1946, 43, 385-440.
15. Fordyce, W. E ., and Lamphere, A. V. "The Control of Social
Desirability in a Structured Q -sort on Dependency-
Independency," Educ. psychol. M easm t., I960, 20,
103-110.
16. Goldfarb, J . , Jacobs, A ., and Levitan, S. "Variables
Determining the Ability to Estimate One*s Scores on
Objective T ests," J. psychol. Stud., I960, 11, 232-237.
17. Goldstein, H. "A Malingering Key for Mental T ests,"
Psychol. Bull. , 1945, 42, 104-118.
18. Green, R. F. "Does a Selection Situation Induce T esters to
Bias their Answers on Interest and Temperament T ests?"
Educ. psychol. Measmt. , 1951, 11, 503-515.
19. Guilford, J. P. Fundamental Statistics in Psychology and
Education. New York: McGraw-Hill Book Company,
Inc. 19^6.
20._ ___________. Psychometric Methods. New York: McGraw-
Hill Book Company, Inc. 1959.
21._ ___________. Personality. New York: McGraw-Hill Book
Company, Inc. 1959.
22._ ___ _ _ _ _ _ _ _ • "The Correlation of an Item with a Composite of
the Remaining Items in a Test, " Educ. psychol.
M easm t., 1953, 13, 87-93.
23. Guilford, J. P . , and Zimmerman, W. S. The Guilford-
Zimmerman Temperament Survey, Manual of Instructions
and Interpretations. Beverly Hills, California: Sheridan
Supply Company. 1949.
24. Hansen, A. V. "An Experimental Study of the Ability to
Simulate Humm-Wadsworth Temperament Perform ance. "
Unpublished M aster's thesis, The University of Southern
California, Los Angeles, 1942.
129
25. Heidbreder, E. "Self-ratings and P references,M J. abn.
soc. Psychol., 1930, 25, 62-74.
26. Helson, H. "Adaptation-Level Theory," in Psychology:
A Study of a Science. Koch, S ., Ed. Vol. I. New York:
McGraw-Hill Book Company, Inc. 1959.
27. Helson, H ., Blake, R. R ., and Mouton, J. S. "Petition-
signing as Adjustment to Situational and Personal Factors, "
J. soc. Psychol. , 1958, 48, 3-10.
28. Helson, H. , Blake, R. R. , Mouton, J. S. , and Olmstead, J. A.
"Attitudes as Adjustments to Stimulus, Background, and
Residual Factors, " J. abn. soc. Psychol., 1956, 52,
314-322.
29. Heron, A. "The Effects of Real-life Motivation on Question
naire Response," J. appl. Psychol. , 1956, 48, 65-68.
30. Herzberg, F. "Temperamant Measures in Industrial
Selection," J. appl. Psychol., 1954, 38, 81-84.
31. Jones, R. A. Personal communication.
32. Kelly, E. L ., Miles, C. C ., and Terman, L. C. "Ability to
Influence One*s Scores on a Pencil-and-Paper Test of
Personality," Character & P e r s ., 1935-36, 4, 206-215.
33. Kimber, J.A .M . "The Insight of College Students into the
Items of a Personality Test. " Unpublished Doctor*s
dissertation, The University of Southern California,
Los Angeles, 1950.
34. Kinder, J. S. "Through Our Own Looking-glass, " Sch. &
Soc., 1925, 22, 533-536.
35. Lindquist, E. F. Design and Analysis of Experiments in
Psychology and Education. Boston: Houghton Mifflin
Company. 1953.
36. Loevinger, J. "A Theory of Test Response," Proceedings
of the Invitational Conference on Problems of Testing.
Educational Testing Service, November 1, 1958.
37. . "Theory and Techniques of Assessm ent," in
Annual Review of Psychology. Farnsworth, P. R. , Ed.
Vol. X. Palo A lto: Annual Reviews, Inc. 1959.
130
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
Longstaff, H. P . , and Jurgen sen, C. E. "Fakability of the
Jurgensen Classification I n v e n t o r y , J . appl. P sychol.,
1953, 37, 86-89.
Longstreth, L. E. The Predictive Validity of the Guilford-
Zimmerman Temperament Survey. Working Paper #11.
Youth Studies Center, The University of Southern Califor
nia, Los Angeles. 1961.
Lunneberg, P. L. M The Relationship between Social D esira
bility Response Set and Anxiety M easures in Children. "
Unpublished Doctor*s dissertation, The University of
Texas, Austin, 1961.
Mais, R. D. "Fakability of the Classification Inventory Scored
for Self-confidence," J. appl. Psychol., 1951, 35,
172-174.
M arsh, C. J. "The Influence of Supplementary Verbal D irec
tions upon Results Obtained with Questionnaires, "
J. soc. Psychol. , 1945, 21, 275-280.
McGee, R. K. "The Relationship between Response Style and
Personality Variables: Acquiescence and Social O rienta
tion. 1 1 Unpublished Doctor1 s dissertation, Vanderbuilt
University, Nashville, 1961.
Michael, W. B. , Jones, R. A ., Cox, A ., Gershon, A .,
Hoover, M. , K atz., and Smith, D. "High School
Record and College Board Scores as Predictors of Success
in a Liberal A rts Program during the Freshman Year of
College," Educ. psychol. Measmt. , 19&2, 22, 399-400.
Mitzel, H. E ., Rabinowitz, W ., and O streicher, L. M.
"The Effects of Response Sets on the Validity of the
Minnesota Teacher Attitude Inventory, " Educ. psychol.
M easm t., 1956, 16, 501-515.
Noll, V. H. "Simulation by College Students of a Prescribed
P attern on a Personality Scale, " Educ. psychol.
M easm t., 1951, 1 1 _ , 478-488.
Palm er, T. B. "B rief Description of the D ry-run."
Unpublished P rogress Report. 1962. (Mimeographed. )
Quiggle, R. R. "The Detection and Measurement of Faking
on Personality Inventories." Unpublished M aster, s
thesis, The University of Southern California, Los
Angeles, 1956.
131
49. Rabinowitz, W, "Fakability of the Minnesota Teacher
Attitude Inventory," Educ. psychol. M easm t., 1954,
14, 657-664.
50. Rosen, E. "Self-appraisal, Personal Desirability, and
Perceived Social D esirability of Personality Traits, "
J. abn. soc. Psychol. , 1956, 52, 151-158.
51. Rosenzweig, S. "A Suggestion for Making Verbal Personality
T ests More V alid," Psychol. Rev. , 1934, 41, 400.
52. Ruch, F . L. "A Technique for Detecting Attempts to Fake
Perform ance on the Self-Inventory Type of Personality
T e st," in Studies in Personality. New York* McGraw-
Hill Book Company, Inc. 1942.
53. Sanford, N. , W ebster, H, , and Freedman, M. "Impulse
E xpression as a Variable of Personality," Psychol.
Monogr. , 1957, 71^(11), 21pp.
54. Steinmetz, H. C. "M easuring Ability to Fake Occupational
In tere st," J. appl. Psychol. , 1932, 16, 123-130.
55. Strieker, L. J. "Some Item C haracteristics that Evoke
Acquiescent and Social Desirability Response Sets on
Psychological Scales. " Unpublished Doctor,s d isserta
tion, New York University, New York, 1962.
56. Sundberg, N. D. , and Bachelis, W. D. "The Fakability of
Two M easures of Prejudice: The California F Scale and
Gough's P r Scale," J. abn. soc. Psychol., 1956, 52,
140-142.
57. Tables of the Cumulative Binomial Probability Distribution.
The Staff of the Computation Laboratory, Cambridge :
H arvard University P re ss. 1955.
58. van Wagenen, N. B. The Stability of Self-description Tests
of Personality Adjustment. New York; The author.
1935.
59. Wesman, A. G. "Faking Personality Test Scores in a Simula
ted Employment Situation, " J. appl. Psychol., 1952,
36, 112-113. ------------------
APPENDICES
133
A P P E N D IX A
COMMENTS REGARDING THE MATCHING PROCEDURE
Had every subject been present on the day that the study was
conducted in their class, a total of 275 subjects could be expected
to take part. The 213 individuals who comprised the final sample
thus represented 77. 5% of the maximum possible number of sub
jects. Information concerning age, sex, and Board scores of the
275 subjects was gathered prior to the study, and (with the aid of
graph-plots) was the basis for assigning all individuals to one or
another of the eight groups, in such a way as to approximately
equalize these groups with respect to all matching variables.
Shortly after the study proper had gotten underway, it became
apparent that a class attendance of 65-70% was not atypical for this
point in the school sem ester, and that if chance as well as unknown
factors relating to attendance were allowed to continue, a dispropor
tionate number of individuals - -as well as wide divergencies from
the hoped-for averages--m ight appear and p ersist in some of the
eight groups.
Thus, after data had been collected on approximately 100 sub
jects, the means and standard deviations for CEEB and age--
together with the m ale/fem ale ratios--w ere computed for the groups
to which these subjects belonged. If any of these groups were, for
instance, found to be falling considerably below the averages of many
of the other groups on any of the matching variables, a few individ
uals--w ho were part of the class next scheduled for testing, and who
were slightly higher than average on the matching variable--would
be reassigned to that group, in exchange for individuals who were
slightly lower on that same variable, and who previously had been
next scheduled for testing as part of a group which might have been
averaging somewhat higher than the remaining groups, on the v a r
iable in question. This operation tended to produce a convergence
in the averages of all eight groups; sim ilarly, sex ratios were con
tinuously checked and brought closer together by these reassignm ents
134
of a sm all proportion of subjects next scheduled for testing. The
same procedures were then carried out after each class of subjects
had been tested (except for the final class)--thus also ensuring that
an approximately equal number of individuals would be assigned to
each group.
135
A P P E N D IX B
PROCEDURES RELATING TO INDICES SUBSTITUTED
FOR COLLEGE ENTRANCE EXAMINATION
BOARD SCORES
The procedures used for the original matching ( N = 275 )
of the subjects for whom CEEB scores were not known, were based
upon the results of a study of 209 male and 233 female freshmen
who entered the University of Southern California in September of
I960. (44) From this study of college grade point average^- in
relation to the unweighted CEEB total score, it was possible to
derive the following equations for converting the form er m easures
into a common scale whose mean and standard deviation were those
of the CEEB total score in the present sample:
High school GPA (Xo) to CEEB total score (Xs) :
Xs * 297. 91X0 ♦ 98.96
College GPA (Xo) to CEEB total score (Xs) :
Xs = 266. 33Xo - 366.95
The formula for converting the CEEB total scores for the
275 subjects selected originally for the present study into CEEB
total scores for the i960 sample, weighted according to number of
males and fem ales, was:
Xs = 0.954X o * 53.83
If both the high school and college GPA were available, both
item s of information were used. The relative contributions of the
two were then based upon the square of the correlations between
^For the first two sem esters of college.
136
2
each of these m easures and the CEEB total score, from the
I960 study.
Subjects* scores on the English Classification Test were
already available in term s of standard scores, thus making possible
a relatively direct conversion to CEEB scores. The equivalent
CEEB score was provided by the U niversity's Testing Bureau in
the case of the subject for whom only the SCAT was available.
The I960 study was not needed when either the ECT or the SCAT
were available.
Separately for males and for fem ales.
A P P E N D IX C
137
COPY OF NOTE REQUESTING PARTICIPATION OF
SUBJECTS NOT PRESENT ON DAY THE
STUDY HAD BEEN CONDUCTED
Last week, members of the class took part in a small research
project as part of a study I have been working on. My research
design required that everybody take and complete the task, but,
unfortunately, about one-third of the class was not present.
So, the purpose of this note is to ask if those individuals who
were not present last week (or who did not complete the task) would
be able to arrange a time (within the next week or two) when they
could take part in the experiment. If so, please write a phone num
ber next to your name--on the attached alphabetical list--and I will
contact you within 3 days to arrange a time (about 40 - 50 minutes)
at your convenience.
I would greatly appreciate your cooperation.
Thank you,
(signed)
A P P E N D IX D
138
INTRODUCTORY REMARKS TO SUBJECTS
Instructor. - -The instructor began the class by stating that today,
instead of a regular lecture, there would be an experiment. . .one
which would illustrate some of the things which had been, or were
about to be, covered during the course. . . and that Mr. Palm er
(introducing the experim enter), who was a graduate student in
psychology, would now tell the class more about it.
Experim enter. - -As D r._________ indicated, this is a little experiment
of a psychological type. Basically, what's being studied are three
separate tra its of personality. I won't say anything more about this
at this p o in t--it's one of those things in which the less that's said
about it ahead of tim e, the more accurate the results are. I will be
back in a few weeks to tell you what it was about and to give you an
idea of the general results. The results for each individual person
will be confidential. The only information I would like you to put
on the booklet you will get is your birthdate--m onth, day, and y e ar--
and whether you are male or fem ale.
There are several different form s of this booklet. Each one
represents a special p art of the experiment. As a result, it's
im portant not to ask your neighbor about certain questions, because
the chances are that he will be working on something very different.
If you do have any questions, just raise your hand, or come up and
ask me. Needless to say, the way that you answer questions on the
booklet has nothing to do with your grades in the course you are
taking.
I'm going to give you each an alphabetized list that has on it
your name, and also a letter in front of your name. The letters
were randomly chosen, and they correspond to the form you will be
getting.
139
Read the instructions carefully. You’ll probably find them
pretty much self-explanatory.
It will seem like a lot to do at first, but you will be able to
finish without hurrying. Answer sheets are inside, though you
won’t have to use them right away. For the first part, you write
right on the booklet. Pencil or p en --it doesn’t m atter. You can
sta rt whenever you get the booklet. (The booklets are then passed
out. )
A P P E N D IX E
140
GZTS ITEMS WHICH DIFFERENTIATED SUBJECTS
AS TO AMOUNT OF SHIFT1
I. Amount of Shift on R estraint
A. R estraint items
GZTS
Book
let #
Exper.
Book
let #
pbir ir
142 1 . 122
92
9 . 192
27
10 . 180
7 14 . 216
77
18 . 237
17 21 . 175
112
24
.199
42
25 . 178
22 32 . 195
Upper 50. 02% versus lower 49.98%, as to amount of shift.
Subjects responding correctly to any item s listed, were those likelier
to shift to a greater degree. Differentiating response for any unkeyed
item , for any given task, can readily be surm ised by analogy to
response for keyed item.
2
Point b iserial r, corrected for item -total overlap by means
of the following formula: ,----
_ _ m rr J PI
fii tr
J n ~ 2 ^ * 1 ■fpt'
To attain the .01, . 05, and . 10 levels of significance with 211 d. f.,
this coefficient must equal or exceed . 175, . 134, and . 112, resp ec
tively.
141
GZTS
Book
let #
52
62
132
122
87
117
127
2
102
147
B. Other item s^
155
78
290
93
138
83
275
13
68
33
Exper.
Book- pbir ir
let #
33 .216
45 .264
51 .223
57 .191
61 .133
62 .220
63 .283
64 .206
65 .163
66 .226
6 .132 (M)
7 .133 (A)
11 .135 (M)
16 .188 (A)
20 . 128 (A)
22 .119 (A)
27 .117 (M)
44 . 198 (A)
41 .117 (A)
47 .200 (A)
3
Unkeyed with respect to the particular task. GZTS scales
to which these item s belong are indicated in the parentheses re la
ting to the respective item.
142
GZTS Exper.
Book- Book- pbi ir
let # let #
48 53 . 158 (A)
235 55 .174 (M)
II. Amount of Shift on Ascendance
A. Ascendance item s
148 93 .116
73 94 .156
143 130 .119
B. Other items
None
III. Amount of Shift on M asculinity-Femininity
A. M asculinity-Femininity items
230
152
.113
215
161 . 234
195
164
. 186
200 173
. 337
165 175 . 189
295 178
. 274
290 186
. 143
185 195
. 134
Other items
228
163 . 138 (T)
92 192 .1 4 8 (R)
Appendix F .—B ooklet number in GZTS, and Type o f S ca le on GZTS, o f T est-Item s
Appearing in Experim ental B ooklet
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
S ca le
in
G ZTS
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
S ca le
in
G ZT S
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
S cale
in
G ZTS
1 142 R 27 275 M 53 48 A
2 170 M 28 270 M 54 295 M
3 210 M 29 185 M 55 235 M
4 53 A 30 255 M 56 286 0
5 128 A 31 220 M 57 122 R
6 155 M 32 22 R 58 225 M
7 78 A 33 52 R 59 144 S
8 148 A 34 165 M 60 118 A
9 92 R 35 73 A 61 87 R
10 27 R 36 287 F 62
117 R
11 290 M 37 143 A 63 127 R
12 214 P 38 103 A 64 2 R
13 230 M 39 195 M 65 102 R
14 7 R 40 239 P 66 147 R
15 160 M 41 68 A 67 300 M
16 93 A 42 98 A 68 133 A
17 107 R 43 179 P 69 15 E
18 77 R 44 13 A 70 260 M
19 215 M 45 62 R 71 <52 R
20 138 A 46 38 A 72
65 E
21
17 R 47 33 A 73 133 A
22 83 A 48 88 A 74 38 A
23 28 A 49 200 M 75 292 F
24 112 R 50 261 0 76 33 A
25 42 R 51 132 R 77 107 R 4
26 156 0 52 264 P 78 224 P 0
A ppendix F . (C on tin u ed )
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
S ca le
in
G ZTS
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
Sea le
in
G ZTS
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
S ca le
in
G ZTS
79 160 M 106 215 M 133 93 A
80 132 R 107 103 A 134 2 R
81 77 R 108 290 M 135 300 M
82 92 R 109 112 R 136 78 A
83 68 A 110 226 0 137 17 R
84 42 R 111 62 R 138 102 R
85 195 M 112
170 M 139 28 A
86 128 A 113 48 A 140 13 A
87 270 M 114 255 M 141 142 R
88 127 R 115 260 M 142
225 M
89 106 G 116 83 A 143 124 S
90 34 S 117 235 M 144 148 A
91
185 M 118 122 R 145 143 A
92
118 A 119 95 E 146 78 A
93
148 A 120 7 R 147 128 A
94 73 A 121
87 R 148 282 F
95 88 A 122
210 M 149 7 R
96 142 R 123 230 M 150 73 A
97 155 M 124 53 A 151 147 R
98 275 M 125 130 E 152
230 M
99 220 M 126 27 R 153 260 M
100 147 R 127 295 M 154 127 R
101 200 M 128
225 M 155 118 A
102
169 P 129 117 R 156 22 R
103 208 T 130 143 A 157 117 R
104 138 A 131 165 M 158 112 R £
105 98 A 132 22 R 159 42 R ^
A ppendix F . (C on tin u ed )
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
S ca le i
in
G ZTS
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
Sea le
in
G ZTS
No. in
Exp.
B ooklet
No. in
G ZTS
B ooklet
S ca le
in
G ZTS
160 103 A 177 93 A 194 220 M
161 215 M 178 295 M
195 185 M
162 122 R 179 255 M 196 28 A
163 228 T 180 132 R 197 170 M
164 195 M 181 270 M 198 216 O
165 235 M 182 83 A 199 2 R
166 138 A 183 13 A 200 102 R
167 48 A 184 52 R 201 77 R
168 107 R 185 62 R 202
88 A
169 289 P 186 290 M 203 163 T
170 98 A 187 275 M 204 162 F
171 53 A 188 300 M 205 33 A
172 68 A 189 155 M 206
17 R
173 200 M 190 38 A 207 35 E
174 27 R 191 160 M 208 133 A
175 165 M 192 92 R
209 210 M
176 87 R 193 171 0 210 61 G
U 1
A P P E N D IX G
TEST FORMS I : INTRODUCTION, RATINGS,
AND RANKINGS
^The firs t part of the experimental booklet was identical for
all 213 subjects. It is reproduced on the following pages.
147
Please indicate: Date
of
Birth
Sex
We are studying the meaning and importance of three separate
traits of personality. These traits are *Restraint, 1 'A scendance,1
and 'M asculinity-Fem ininity.1 As traits which describe people,
think of these three as including various combinations of the
following characteristics:
R estrained: unhurried, persevering,
cautious, self-controlled.
Ascendant: assertive, socially bold,
outspoken, social leader.
Masc. Fem.
Masculine-Feminine : more 'tough-minded1 Vs. more 'tender-
minded*
less emotional Vs. more emotional
have different . . . have different
interests than interests than
females males
The more of these characteristics a person has, the more of
the tra it he would be said to have.
You will be asked to make a number of ratings, in some cases
this calls for a good deal of imagination and the use of your knowl
edge of people. For example, you will be asked to estimate where
it is that most adults stand on the traits just described; also, you
will be asked to introspect in order to estim ate how much or how
little of the trait you personally have.
L ater on, you will be asked a number of other questions about
such tra its . When making the ratings, and when answering the
later questions, bear in mind the trait-definitions given above.
148
We will use ’scales* to make the ratings just mentioned.
Here is an example of how this works : The scale you see below ,
re fe rs to ’degrees of possession’ of certain traits, e. g. , the tra it
of ’F rie n d lin ess.’ A ratine of *1’, for instance, would mean a
person had
relatively little oi mpared to other
people, while a ’9' would mean he had quite a lot of that tra it when
com pared with others.
Beginning on the next page, you are to answer six different
questions about each of three tra its. They can be answered by
means of rating scales such as the one above. The six questions
relate to
1. Where you think m ost adults actually stand on each trait.
2. Where you think m ost adults would like to stand on each trait.
3. Where you would estim ate yourself to actually stand on each
tra it.
4. Where you would like to stand on each trait.
5. Which of the tra its you think most adults consider the most
im portant, the next m ost im portant, and the least im portant.
6. Which of the tra its is m ost im portant to you personally, which
is next most im portant, and which least important.
Most of these questions will have three p arts to them (one for each
trait).
Be sure to answer every question, even if you are not 100% sure.
P lease begin, starting with the firs t question.
0 1 2 ~ ~ ^ 3 S K 7 h
Not at
all
9 10
Very
much
Friendly Friendly
Turn the page and begin
149
1. For each, of the tra its listed, check the point (or circle the
num ber) at which you think most adults would actually be
located.
R estraint
O '" 1 2 3 i 5 5 7 § 9 T O
Not at Very
all much
Restrained Restrained
Ascendance
0 I 2 3 4 5 6 7 8 9 10
Not at Very
all much
Ascendant Ascendant
Masculinity-
Femininity
Not^ at ^
all
Masculine
(Very
much
Feminine)
"3 4 5 5
Note: m ales
please answer
about males;
females please
answer about
females
10
Very
much
Masculine
(Not at
all
Feminine)
For each trait, please check (or circle) the point at which
you think most adults would like to be located; that is, where
they would feel the most socially comfortable or respected.
R estraint
T J 1 Z 3------4------5------6------7 S 9 HT
Not at Very
all much
Restrained Restrained
150
Ascendance
0 1
Not at
all
Ascendant
Mas culinity -
2 3 4 5 6 7 8 9 10
Very
much
Ascendant
F e mxninity q ^
2 3 4 5 6 7 8 9 10
Not at Very
all Note; males
much
Masculine
please answer
Masculine
(Very
about m ales;
(Not at
much
females please
all
Feminine)
answer about
Feminine)
females
3, Check or circle the point at which you estim ate yourself to
actually be located, on each trait.
R estraint
0 1 2 3 4 5 6 7 8 9 10
Not at Very
all much
R estrained Restrained
Ascendance
0 1 2 3 4 5 6 7 8 9 10
Not at Very
all much
A s c endant Ascendant
151
Mas culinity -
Femininity
0 I
Not at
all
Masculine
(Very
much
Feminine)
Note: males
please answer
about m ales;
females please
answer about
females
) T o
Very
much
Masculine
(Not at
all
Feminine)
4. This tim e, for each trait, check or circle where you person
ally would want to be located; that is, indicate how much or
how little of each trait you would prefer to ‘have. 1
R estraint
0 1 2 3 4 5 6 7 8 9 10
N ot at Ve ry
all much
Restrained Restrained
Ascendance
0 1 2 3 4 5 6 7 8 9 10
Not at Very
all much
Ascendant Ascendant
Mas culinity -
Femininity
0 1
Not at
all
Masculine
(Very
much
Feminine)
Note: males
please answer
about males;
females please
answer about
females
) 10
Very
much
Ma!sculine
(Not at
all
Feminine)
152
5. Now rank each of the traits relative to one another, as to how
im portant you think most adults consider th em --i. e. , how
im portant they consider it to have the 'right amount* of each.
Place a * 1 * next to the most im portant one, a '2' next to the
second most important, and a '3* next to the one least important
to most adults.
R estraint
Ascendance
Note: if a m ale, answer about males;
if a fem ale, answer about fem ales.
Do this only for the Masc. -Fern, trait.
Masculinity
(for males)
or _
Femininity
(for females)
6. Do the same thing--i. e. , rank each tra it--b u t this time as to
how important you personally consider each one to be, for your
self. '1* for the most im portant, '2* for the next most im por
tant, etc.
R e str aint ______
Ascendance
Masculinity
(for m ales) Note: if a m ale, answer about males;
or if a fem ale, answer about fem ales.
. . Do this only for the Masc. -Fern, trait.
Femininity
(for females)
153
7. One additional rating, before going on to the main part of the
questionnaire: Would you indicate (via the 10 point scales you
have been using) the TWO points on each trait (Restraint,
Ascendance, M asculinity-Femininity) below and above which
you would probably feel quite uncomfortable or anxious--that is,
if you actually possessed those degrees of each tra it (for exam
ple, a * 3 * and a ’9 * on one of the traits, or a , 11 and a * 6 * on one
of the tra its , etc. ). You m ight think of these two points as being
minimum and maximum *tolerance lim its. 1 - ,0I is the low end
tz *10* is the high end of the scale.
R estraint and
Ascendance and
Masculinity
(if male)
or and
Femininity
(if female)
Go on to the next page
A P P E N D IX H
154
TEST FORMS II : INSTRUCTIONS FOR DIFFERENT
EXPERIMENTAL GROUPS
Following the ratings and rankings (Appendix G), all subjects
answered three sets of questions, in connection with R estraint,
Ascendance, and Mas culinity-Fem ininity, respectively. All eight
groups of subjects received differing written instructions for these
three tasks. A complete set of these task-instructions, for group K,
appears on the following six pages. The subsequent six pages contain
the complete set of instructions for group T. These two sets of
instructions illustrate those for the remaining six groups as well,
since these latter instructions run parallel with the form er, in
accordance with the paradigm presented in Table 13. For purposes
of ready reference, however, the task-instructions for groups K and
T will then be followed by the firs t page of the task-instructions--
for the tra it of R estraint--for groups L, M, R, S, W, and Z, resp ec
tively.
155
Now for the main part of the task, which is in three parts.
Each part has 70 short questions.
P art I
In answering questions #1 through #70, imagine yourself to be
a person who is in the
(a) following situation, and who has the
(b) following trait :
In connection with some kind of experiment, you are being given a
Uest1 of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain traits of
personality are being studied, and, judging from the test item s, you
think one of those traits is that of *Restraint. *
Nowthen, imagine yourself as being an '81 on the trait of R estraint--
in other words, say you are actually a fairly restrained person.
However, on the test you try to convey a picture of yourself as being
a * 2 * (fairly un-Re strained) on that same trait. Ordinarily, you
would rarely falsify things or m isrepresent yourself, particularly in
such a test situation. But, be that as it may, you are now answering
as a * 2 * because you would probably feel somewhat unhappy at the
thought th at--if you answered honestly--the person who was giving
the test would learn that you are actually a fairly restrained person--
and you generally do not want people to know this fact. This just
happens to be the way you feel; and because of this reason, you would
probably answer no differently even if you knew the researcher
personally (which you do not) and even if you knew what he was after
or what answers (if any) he might prefer (which you are also unaware
of), or knew if he was interested in your particular score. -In short,
you try to seem like a * 2 * for purely personal reasons alone.
Go on to th e n e x t p ag e
156
To Repeat:
1. You are actually an *8* (fairly Restrained)
2. But you want to appear to be a * 2 * (fairly un-Re strained)
3. Because, for personal reasons, you don*t want to indicate to
the researcher where you actually stand on this trait.
Please answer questions 1 through 70 in light of this purpose and
situation.
Use the answer sheets provided. Be sure to answer every question,
even if you are not 100% sure of what the best answer would be.
Turn the page and begin
P a r t II
In a n s w e rin g q u e s tio n s #71 th ro u g h # 1 4 0 , im a g in e y o u r s e lf to b e
a p e r s o n w ho is in th e
(a) following situation, and who has the
(b) following tra it :
In connection with some kind of experiment, you are being given a
*test* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain traits of
personality are being studied, and, judging from the test item s, you
think one of those traits is that of ’Ascendance, *
Nowthen, imagine yourself as being an * 8 * on the tra it of Ascendance -
in other words, say you are actually a fairly ascendant person.
However, on the test you try to convey a picture of yourself as being
a * 2 * (fairly How1 on the tra it of Ascendance). Ordinarily, you would
rarely falsify things or m isrepresent yourself, particularly in such a
te st situation. But, be that as it may, you are now answering as a
*2* because you would probably feel somewhat unhappy at the thought
th a t--if you answered honestly--the person who was giving the test
would learn that you are actually a fairly ascendant person--and you
generally do not want people to know this fact. This just happens to
be the way you feel; and because of this reason, you would probably
answer no differently even if you knew the research er personally
(which you do not) and even if you knew what he was after or what
answers (if any) he might prefer (which you are also unaware of), or
knew if he was interested in your particular score. -In short, you try
to seem like a * 2 * for purely personal reasons alone.
Go on to the next page
N o te ; Y o u r a n s w e rs do n o t h av e to b e c o n s is te n t w ith any e a r l i e r
a n s w e r s , b e c a u s e o n ly th e p r e s e n t t r a i t (A sc e n d a n c e ) is
im p o r ta n t now .
To Repeat:
1. You are actually an * 8 * (fairly Ascendant)
2. But you want to appear to be a * 2 * (fairly ^ow1 on Ascendance)
3. Because, for personal reasons, you don*t want to indicate to
the research er where you actually stand on this trait.
Please answer questions 71 through 140 in light of this purpose and
situation.
B e s u r e to a n s w e r e v e r y q u e s tio n , e v e n if you a r e n o t 100% s u r e
o f w h at th e b e s t a n s w e r w o u ld b e .
T u rn th e p ag e an d b e g in
P a r t III
159
In a n s w e rin g q u e s tio n s #141 th ro u g h #2 1 0 , im a g in e y o u r s e lf to be
a p e r s o n w ho is in th e
(a) following situation, and who has the
(b) following tra it :
In connection with some kind of experiment, you are being given a
*test* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain traits of
personality are being studied, and, judging from the test item s, you
think one of those traits is that of * Mas culinity-Femininity. 1
Nowthen, imagine yourself as being an 1 81 on the trait of Mas culinity -
Fem ininity--in other words, say you are actually (if a male) fairly
•high* on Masculinity (or, if a female, fairly *low* on Femininity).
However, on the te st you try to convey a picture of yourself as being
a *2J (fairly How1 on Masculinity, if a male: fairly ^ ig h 1 on
Femininity, if a fem ale). Ordinarily, you would rarely falsify things
or m isrepresent yourself, particularly in such a test situation. But,
be that as it may, you are now answering as a '2 1 because you would
probably feel somewhat unhappy at the thought th at--if you answered
honestly--the person who was giving the test would learn where you
stand on this trait--a n d you generally do not want people to know this
fact. This just happens to be the way you feel; and because of this
reason, you would probably answer no differently even if you knew
the research er personally (which you do not) and even if you knew
what he was after or what answers (if any) he might prefer (which you
are also unaware of), or knew if he was interested in your particular
score. -In short, you try to seem like a * 2 * for purely personal
reasons alone.
Go on to th e n e x t p ag e
N o te : Y o u r a n s w e r s do n o t h a v e to b e c o n s is te n t w ith an y e a r l i e r
a n s w e r s , b e c a u s e o n ly th e p r e s e n t t r a i t (M a s c u lin ity -
F e m in in ity ) i s im p o r ta n t n o w .
To Repeat:
1. You are actually an ’8 * (if male: fairly ‘high* on Masculinity;
if fem ale: fairly tlowl on Femininity)
2. But you want to appear to be a t 2 1 (if male: fairly 'low1 on
Masculinity; if female:
fairly ^igh* on Femininity)
3. Because, for personal reasons, you don*t want to indicate to
the research er where you actually stand on this trait.
Please answer questions 141 through 210 in light of this purpose and
situation.
Be sure to answer every question, even if you are not 100% sure
of what the best answer would be.
Turn the page and begin
161
N o w f o r tlie m a in p a r t o f th e t a s k , w h ic h i s in t h r e e p a r t s .
E a c h p a r t h a s 70 s h o r t q u e s ti o n s .
P a rt I
In answ ering questions #1 through #70, im agine yourself to be
a p erso n who is in the
(a) following situation, and who has the
(b) following tr a it :
In connection w ith som e kind of experim ent, you are being given a
*test* of p erso n ality . You do not exactly know what the re s e a rc h is
fo r and, as fa r as you can tell, you ju st happen to have been one of
the people picked as a subject. Your guess is that certain tr a its of
p erso n ality are being studied, and, judging fro m the te s t ite m s, you
think one of those tr a its is th at of ’R estrain t. 1
Nowthen, im agine yourself as being an t81 on the tr a it of R estrain t - -
in other w ords, say you are actually a fairly re stra in e d p erso n .
W hile taking the above-m entioned te s t, this is also the picture you
tr y to convey of yourself: you answ er accu rately , in o rd er to appear
as you actually are on this tr a it- - a n *8* (fa irly R estrain ed ).
You a re trying to do that because you have an idea that the perso n
giving the te s t p re fe rs that the subjects (yourself included) answ er
th a t w a y --it seem s so from c ertain re m a rk s (possibly hints) he has
m ade. O rdinarily, you would ra re ly falsify things or m isre p re se n t
y o u rself, p a rtic u la rly in such a te s t situation. But, be th at as it m ay,
you a re now answ ering as an *8* so as to satisfy the research e r* s
probable p refe ren c e, even though you don*t know why he m ight p re fe r
an *8, * and even though you do not know him personally, o r know if
he is in te reste d in your p a rtic u la r sco re. -In short, you try to seem
like an * 8 * for p u rely ’im personal* reasons alone.
Go o n to th e n e x t p a g e
To Repeat:
162
1. You are actually an '8* (fairly Restrained)
2. And you want to appear to be an * 8 * (fairly Restrained)
3. Because that’s what the research er appears to prefer.
Please answer questions 1 through 70 in light of this purpose and
situation.
Use the answer sheets provided. Be sure to answer every question,
even if you are not 100% sure of what the best answer would be.
Turn the page and begin
163
P a r t II
In a n s w e rin g q u e s tio n s #71 th ro u g h #1 4 0 , im a g in e y o u r s e lf to b e
a p e r s o n w ho i s in th e
(a) following situation, and who has the
(b) following tra it :
In connection with some kind of experiment, you are being given a
•test1 of personality. You do not exactly know what the research is for
and, as far as you can tell, you just happen to have been one of the
people picked as a subject. Your guess is that certain traits of p e r
sonality are being studied, and, judging from the test item s, you
think one of those traits is that of ’Ascendance,1
Nowthen, imagine yourself as being an ’8 * on the tra it of Ascendance--
in other words, say you are actually a fairly ascendant person.
While taking the above-mentioned test, this is also the picture you
try to convey of yourself: you answer accurately, in order to appear
as you actually are on this tra it--a n * 8 * (fairly Ascendant).
You are trying to do that because you have an idea that the person
giving the test prefers that the subjects (yourself included) answer
that w ay--it seem s so from certain rem arks (possibly hints) he has
made. Ordinarily, you would rarely falsify things or m isrepresent
yourself, particularly in such a test situation. But, be that as it may,
you are now answering as an ’8* so as to satisfy the research er’s
probable preference, even though you don’t know why he might prefer
an '8 1 , and even though you do not know him personally, or know if
he is interested in your particular score. -In short, you try to seem
like an ’8’ for purely ’im personal1 reasons alone.
Go on to the next page
164
Note: Your answ ers do not have to be consistent with any earlier
answ ers, because only the present tra it (Ascendance) is
im portant now.
To Repeat:
1. You are actually an '8* (fairly Ascendant)
2. And you want to appear to be an * 8 * (fairly Ascendant)
3. Because that*s what the research er appears to prefer.
Please answer questions 71 through 140 in light of this purpose and
situation.
B e s u r e to a n s w e r e v e ry q u e s tio n , ev en if y o u a r e n o t 100% s u r e
o f w h a t th e b e s t a n s w e r w ould b e .
T u rn th e p a g e and b e g in
P a r t III
165
In a n s w e rin g q u e s tio n s #141 th ro u g h # 2 1 0 , im a g in e y o u r s e lf to b e
a p e r s o n w ho is in th e
(a) following situation, and who has the
(b) following tra it :
In connection with some kind of experiment, you are being given a
’test* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain traits of
personality are being studied, and, judging from the test item s, you
think one of those tra its is that of ’M asculinity-Femininity. 1
Nowthen, imagine yourself as being an ^8^ on the trait of Masculinity -
F e mininity - - in other words, say you are actually (if a male) fairly
'high* on Masculinity (or, if a fem ale, fairly *low* on Femininity).
While taking the above-mentioned te st, this is also the picture you
try to convey of yourself: you answer accurately, in order to appear
as you actually are on this tra it--a n ^ (fairly high on Masculinity,
if a male; fairly llowl on Femininity, if a fem ale). You are trying to
do that because you have an idea that the person giving the test p re
fers that the subjects (yourself included) answer that w ay--it seems
so from certain rem arks (possibly hints) he has made. Ordinarily,
you would rarely falsify things or m isrepresent yourself, particularly
in such a test situation. But, be that as it may, you are now answ er
ing as an * 8 * so as to satisfy the resea rch e r’s probable preference,
even though you don’t know why he might prefer an *8*, and even
though you don’t know him personally, or know if he is interested in
your particular score. -In short, you try to seem like an * 8 * for
purely ’impersonal* reasons alone.
Go on to th e n e x t p ag e
N o te : Y our a n s w e rs do n o t hav e to be c o n s is te n t w ith any e a r l i e r
a n s w e r s , b e c a u s e o n ly th e p r e s e n t t r a i t (M a s c u lin ity -
F e m in in ity ) i s im p o rta n t now .
To Repeat:
1. You are actually an * 8 * (if male: fairly 'high* on Masculinity;
if female: fairly 'low* on Femininity)
2. And you want to appear to be an '8 ' (if male: fairly 'high* on
Masculinity: if female:
fairly 'low* on Femininity^
3. Because that's what the researcher appears to prefer.
Please answer questions 141 through 210 in light of this purpose and
situation.
B e s u re to a n s w e r e v e ry q u e s tio n , ev en if you a r e n o t 100% s u r e
o f w h at th e .b e s t a n s w e r w o u ld b e .
T u rn th e p a g e and b eg in
167
N ow f o r th e m a in p a r t of th e ta s k , w h ich i s in th r e e p a r t s .
E a c h p a r t h a s 70 s h o r t q u e s tio n s .
P art I
through #70, imagine yourself to be
following situation, and who has the
following tra it :
In connection with some kind of experiment, you are being given a
•test* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain traits of
personality are being studied, and, judging from the test item s, you
think one of those tra its is that of ‘R estraint. *
Nowthen, imagine yourself as being an * 8 * on the tra it of R estrain t--
in other words, say you are actually a fairly restrained person.
However, on the te st you try to convey a picture of yourself as being
a * 2 * (fairly un-Re strained) on that same tra it. You are trying to
do that because you have an idea that the person giving the test p re
fers that the subjects (yourself included) answer that w ay--it seems
so from certain rem arks (possibly hints) he has made. Ordinarily,
you would rarely falsify things or m isrepresent yourself, particularly
in such a test situation. But, be that as it may, you are now answ er
ing as a * 2 * so as to satisfy the researcher*s probable preference,
even though you don't know why he might prefer a *2*, and even
though you do not know him personally, or know if he is interested in
your particular score, -In short, you try to seem like a *2* for purely
•impersonal* reasons alone.
In answering questions #1
a person who is in the
(a)
(b)
Go on to th e n e x t p a g e
168
N ow f o r th e m a in p a r t o f th e ta s k , w h ic h is in th r e e p a r t s .
E a c h p a r t h a s 70 s h o r t q u e s tio n s .
P a rt I
In answering questions #1 through #70, imagine yourself to be
a person who is in the
(a) following situation, and who has the
(b) following tra it :
In connection with some kind of experim ent, you are being given a
’test* of personality. You do not exactly know what the research is
for and, as fa r as you can tell, you ju st happen to have been one of
the people picked as a subject. Your guess is that certain tra its of
personality are being studied, and, judging from the te st item s, you
think one of those traits is that of *Restraint. 1
Nowthen, imagine yourself as being a *2* on the trait of R estrain t--
in other w ords, say you are actually a fairly un-restrained nerson.
However, on the test you try to convey a picture of y o u rselfa s being
an *81 (fairly R estrained) on that same trait. Ordinarily, you would
ra re ly falsify things or m isrepresent yourself, particularly in such a
test situation. But, be that as it may, you are now answering as an
*8 * because you would probably feel somewhat unhappy at the thought
th a t--if you answered honestly--the person who was giving the test
would learn that you are actually a fairly un-restrained person--and
you generally do not want people to know this fact. This just happens
to be the way you feel; and because of this reason, you would probably
answer no differently even if you knew the research er personally
(which you do not) and even if you knew what he was after or what
answers (if any) he might prefer (which you are also unaware of), or
knew if he was interested in your particular score. -In short, you try
to seem like an * 8 * for purely personal reasons alone.
Go on to th e n e x t p a g e
169
N ow f o r th e m a in p a r t o f th e ta s k , w h ich is in th r e e p a r t s .
E a c h p a r t h a s 70 s h o rt q u e s tio n s .
P a rt I
In answering questions #1 through #70, imagine yourself to be
a person who is in the
(a) following situation, and who has the
(b) following trait :
In connection with some kind of experiment, you are being given a
’test* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain traits of
personality are being studied, and, judging from the test item s, you
think one of those traits is that of 'R estraint. *
Nowthen, imagine yourself as being a * 2 * on the tra it of R estraint- -
in other words, say you are actually a fairly un-restrained person.
However, on the test you try to convey a picture of yourself as being
an (fairly R estrained) on that same trait. You are trying to
do that because you have an idea that the person giving the test p re
fers that the subjects (yourself included) answer that w ay--it seems
so from certain rem arks (possibly hints) he has made. Ordinarily,
you would rarely falsify things or m isrepresent yourself, particularly
in such a test situation. But, be that as it may, you are now answ er
ing as an * 8 * so as to satisfy the research er's probable preference,
even though you don't know why he might prefer an *8*, and even
though you do not know him personally, or know if he is interested in
your particular score. -In short, you try to seem like an *8* for
purely 'impersonal* reasons alone.
Go on to th e n e x t p a g e
170
N ow f o r th e m a in p a r t o f th e t a s k , w h ic h is in th r e e p a r t s .
E a c h p a r t h a s 70 s h o r t q u e s tio n s .
P a rt I
In answering questions #1 through #70, imagine yourself to be
a person who is in the
(a) following situation, and who has the
(b) following tra it :
In connection with some kind of experiment, you are being given a
Hest* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain tra its of
personality are being studied, and, judging from the te st item s, you
think one of those tra its is that of *Restraint. *
Nowthen, imagine yourself as being an * 8 * on the tra it of R estrain t--
in other words, say you are actually a fairly restrained person.
While taking the above-mentioned test, this is also the picture you
try to convey of yourself: you answer accurately, in order to appear
as you actually are on this tra it--a n (fairly R estrained).
You are doing this for purely personal reasons alone--not to satisfy
anyone else--and apart from the fact that, ordinarily, you rarely
would falsify things or m isrepresent yourself, particularly in such a
test situation. Your reason for trying to appear as an *8* is that you
want the person who is giving the test to learn that you are actually a
fairly restrained person--and you generally want people to know this
fact. This just happens to be the way you feel; and because of this
reason, you would probably answer no differently even if you knew the
research er personally (which you do not) and even if you knew what he
was after or what answers (if any) he might prefer (which you are also
unaware of), or knew if he was interested in your particular score.
-In short, you try to seem like an *8* for purely personal reasons
alone.
Go on to th e n e x t p a g e
171
N ow f o r th e m a in p a r t o f th e ta s k , w h ic h is in th r e e p a r t s .
E a c h p a r t h a s 70 s h o r t q u e s tio n s .
P a rt I
In answering questions #1 through #70, imagine
a person who is in the
(a) following situation,
(b) following tra it :
In connection with some kind of experiment, you are being given a
*test* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain tra its of
personality are being studied, and, judging from the test item s, you
think one of those traits is that of *R estraint.'
Nowthen, imagine yourself as being a ^2* on the tra it of R estraint--
in other words, say you are actually a fairly un-restrained person.
While taking the above-mentioned test, this is also the picture you
try to convey of yourself: you answer accurately, in order to appear
as you actually are on this tra it--a * 2 * (fairly un-Re strained).
You are doing this for purely personal reasons alone--not to satisfy
anyone else--and apart from the fact that, ordinarily, you rarely
would falsify things or m isrepresent yourself, particularly in such a
te s t situation. Your reason for trying to appear as a *2* is that you
want the person who is giving the test to learn that you are actually a
fairly un-restrained person--and you generally want people to know
this fact. This just happens to be the way you feel; and because of
this reason, you would probably answer no differently even if you knew
the research er personally (which you do not) and even if you knew
what he was after or what answers (if any) he might prefer (which you
are also unaware of), of knew if he was interested in your particular
score. -In short, you try to seem like a * 2 * for purely personal
reasons alone.
yourself to be
and who has the
Go on to th e n e x t p ag e
172
N ow f o r th e m a in p a r t of th e t a s k , w h ich is in th r e e p a r t s .
E a c h p a r t h a s 70 s h o r t q u e s tio n s .
P a rt I
In answering questions #1 through #70, imagine yourself to be
a person who is in the
(a) following situation, and who has the
(b) following tra it :
In connection with some kind of experiment, you are being given a
•test* of personality. You do not exactly know what the research is
for and, as far as you can tell, you just happen to have been one of
the people picked as a subject. Your guess is that certain traits of
personality are being studied, and, judging from the test item s, you
think one of those traits is that of * R estraint. *
Nowthen, imagine yourself as being a !2* on the trait of R estraint- -
in other words, say you are actually a fairly un-restrained person.
While taking the above-mentioned test, this is also the picture you
try to convey of yourself: you answer accurately, in order to appear
as you actually are on this tra it--a * 2 * (fairly un-Restrained).
You are trying to do that because you have an idea that the person
giving the test p refers that the subjects (yourself included) answer
that w ay--it seem s so from certain rem arks (possibly hints) he has
m ade. Ordinarily, you would rarely falsify things or m isrepresent
yourself, particularly in such a test situation. But, be that as it may,
you are now answering as a * 2 * so as to satisfy the researcher*s
probable preference, even though you don’t know why he might prefer
a '2*, and even though you do not know him personally, or know if he
is interested in your particular score. -In short, you try to seem like
a *2* for purely ’impersonal* reasons alone.
Go on to th e n e x t p a g e

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

On The Relationship Between Anxiety And Aggression In Nine-Year-Old Boys

PDF

Intellect After Lobotomy In Schizophrenia: A Factor-Analytic Study

PDF

The Implications Of Cultural Anthropology For The Question: What Is The Basis Of Moral Obligation?

PDF

A Preliminary Study Of The Dimensions Of Future Time Perspective

PDF

A Rotational Approach To Psychological Invariance

PDF

A Methodological Investigation Of Affect Response Bias

PDF

The Development Of Attitude Scales To Predict Accident Repeater And Moving Violator Drivers

PDF

Semantic-Memory And Creative (Divergent-Production) Abilities Of Senior-High-School Students

PDF

A Factor-Analytic Study Of Military Leadership

PDF

A Factor Analysis Of The Figural-Evaluation Abilities

PDF

A Factor Analysis Of The Symbolic-Evaluation Abilities

PDF

Teaching Machines: An Investigation Of Constructed Versus Multiple-Choicemethods Of Response

PDF

Creativity In Children: A Study Of The Relationship Between Temperament Factors And Aptitude Factors Involved In The Creative Ability Of Seventh Grade Children With Suggestions For A Theory Of C...

PDF

A Multidimensional Scaling Of Mood Expressions

PDF

Validity Concomitants Of Various Scoring Procedures Which Attenuate The Effects Of Response Sets And Chance

PDF

Factorial Stability As A Function Of Analytical Rotational Method, Type Of Simple Structure, And Size Of Sample

PDF

Human Performance As A Function Of The Joint Effects Of Drive And Incentive Motivation

PDF

Figural And Symbolic Divergent-Production Abilities In Adults And Adolescents

PDF

Prediction Of First Grade Reading Achievement From Selected Structure-Of-Intellect Factors

PDF

A Study Of Criteria Of Social Perception And Some Related Variables

Asset Metadata

Creator Palmer, Theodore Bernard (author)

Core Title An Application Of Adaptation Level Theory To The Response Bias Of Falsification

Degree Doctor of Philosophy

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,psychology, general

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Michael, William B. (committee chair), Guilford, Joy P. (committee member), Werkmeister, William H. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-310292

Unique identifier UC11359094

Identifier 6403103.pdf (filename),usctheses-c18-310292 (legacy record id)

Legacy Identifier 6403103.pdf

Dmrecord 310292

Document Type Dissertation

Rights Palmer, Theodore Bernard

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA