CROSS-CULTURAL ANALYSIS OF LANGUAGE
IN THE ASSESSMENT OF PERSONALITY VARIABLES
ALBERT F. INCLAN
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Albert F. Inclan
I am eternally indebted to the loyalty and faith my Doctoral
Committee has had in me and in what I was attempting to do in this
study. I would like to express my gratitude to Dr. Janet Larsen for
her total confidence in me and for not allowing me to turn back on
those occasions when I felt I had to; to Dr. Mary McCaulley for
making me fully realize the impact of what I was doing and for giving
me, not only support, but valuable information about the Myers-Briggs
Type Indicator that only a person of her calibre could offer; and to
Dr. Rod McDavis for many insightful suggestions for this study, and
for his support whenever it was needed. The success of this research
was also made possible by the backtranslators and the many subjects
who gave so freely of their time. Without them this study would not
have been possible.
I would also like to express a sincere 'thank you' to my mother
who has been patiently waiting for this day to arrive and who never
once doubted that it would; and to those I have lived with and gotten
close to over the years for their patience when I had none.
This dissertation is a product of many hearts and minds working
together. It is my sincere hope that the results will be of benefit
to many more.
TABLE OF CONTENTS
ABSTRACT ....................................................... vi
ONE INTRODUCTION .................................... 1
The Problem ..................................... 4
Purpose of the Study ............................ 7
Need for the Study .............................. 7
Definition of Terms ............................. 13
Organization of the Study ....................... 13
TWO REVIEW OF THE LITERATURE ....................... 15
Theories and Methods of Translations.............. 15
Emic and Etic Analysis ...................... 17
Associative Method .......................... 18
Cultural subjectivity ..................... 20
Implications for communication
and bilingualism......................... 22
Implications of theory for
translations ........................... 24
Transformational Generative Theory
of Language .............................. 24
Universal Patterns of Human Thought ............. 25
Cultural Difference Issues in Translations.... 26
Theoretical Base.............................. 31
Strategies for Insuring Equivalence of
Translation ............................... 34
Assessment of Equivalence ................... 40
Culture Fairness ............................... 41
Myers-Briggs Type Indicator ..................... 50
Intercorrelation of MBTI Scores............... 53
Validity ...................................... 56
Introduction ................................... 60
Research Questions .............................. 60
Procedures ..................................... 61
Phase I: The Translation ..................... 61
Phase II: The Administration................... 65
Population.. .............................. 65
Collection of data......................... 67
Phase III: Data Analysis...................... 68
FOUR ANALYSIS OF THE RESULTS ............................70
Research Question 1 ................................72
Research Question 2 ................................75
Research Question 3 ................................77
Research Question 4 ..................................78
FIVE DISCUSSION, LIMITATIONS AND CONCLUSIONS ............81
Limitations of the Study ............................82
A JUNGIAN TYPES ......................................87
B INTERCORRELATIONS OF MBTI TYPE CATEGORIES............89
C PRODUCT-MOMENT CORRELATIONS OF MBTI CONTINUOUS
SCORES AND OTHER PERSONALITY TESTS IN
MEDICAL STUDENT SAMPLES .......................... 91
D LETTER OF PERMISSION FROM TEST PUBLISHER.............96
E CONTACT LETTER TO INITIATE STUDY ...................98
F COVER LETTER FOR TEST PACKET......................100
G TYPE TABLES ENGLISH/SPANISH VERSION
MYERS/BRIGGS TYPE INDICATOR, FORM G..............102
REFERENCES................... ...... ... .............. ...............104
BIOGRAPHICAL SKETCH ...............................................111
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
CROSS-CULTURAL ANALYSIS OF LANGUAGE
IN THE ASSESSMENT OF PERSONALITY VARIABLES
Albert F. Inclan
Chairman: Dr. Janet Larsen
Major Department: Counselor Education
The purpose of this study was to develop and validate a Spanish
version of an English language personality questionnaire that would
be faithful to the original while taking into account differences in
values and attitudes of the two cultures. Specifically, the study
translated and acculturated the Myers-Briggs Type Indicator into
the Spanish language within the intent of the authors of the Myers-
Briggs Type Indicator and Jung's theory of personality.
It was most important to address the intent of the written
message in the questionnaire. This aspect had to be considered in
order to help in the selection of the words being used in the target
version of the instrument.
The translation was conducted in several phases and combined
several theories of translation. Analyses of the deep meaning of
language were used, as opposed to simply a linguistic analysis of
The Spanish version was given to a sample of 209 bilingual
adult men and women. The results indicate that the instrument is
very comparable in what it measures in the original English version.
By all indications the instrument is measuring the same universal
characteristics and in very much the same way as the original MBTI.
"From a psychological point of view, communicating is largely
a matter of knowing what themes are important to people and addressing
those themes in ways that are in accord with the subjective meaning
people attach to them" (Szalay & Deese, 1978, p. vii).
Differences among people were noted almost 2,000 years ago by
the Greek physician, Galen. He made the first clear distinctions
among four temperaments in man, and coined them: sanguine, phlegmatic,
choleric, and melancholic (Fordham, 1979). Although the terms are
commonly used, they are not widely recognized within the present day
psychological interpretations. It was because of a need to understand
human differences that, years later, C. G. Jung analyzed people in
a novel way and postulated that they habitually take either an ex-
traverted or introverted attitude depending on whether their flow
of interests is directed outwardly or inwardly, respectively
(Jung, 1923/1971). These distinctions helped pave the way for an
entirely unique outlook on personality. Jung felt that the flow of
interests, which he saw as being universal in nature, was expressed,
subjectively, by means of language. If his theory is accepted, the
manner in which people perceive and make their perceptions known to
others is the basis of communication.
Persons' concepts of the world often color their perceptions
and determine their reactions to the environment. The way people
perceive the world around them determines what words they choose in
order to verbally express their attitudes and experiences. Words,
then, become the medium which transfer the message from the person
having the experience to the realm of recognition of those with whom
they are communicating. The proper choice of words is, therefore,
crucial to convey the exact interpretation of messages given to
others. They represent a medium through which the listener or reader
can identify with the experiences being shared. By looking at words
in this perspective it is possible to recognize that the deep struc-
ture of language is within the inner recesses of the person. Feelings
and emotions are really being tapped by the spoken or written word.
The way each person responds, internally, to stimuli before putting
the response into words is affected by many variables such as age, sex,
educational background, experience and health. Also, the cultural
background will be reflected in the way an individual responds to
certain stimuli, whether they are written, oral, or visual. Conse-
quently, it is important to consider the element of culture in construc-
ting any test that evaluates attitudes and values, and personality
The problem of culture is compounded in the translation and/or
adaptation of any psychological measure from one language into another.
Not only must the original version reflect the culture as accurately
as possible in the language in which it was developed, but subse-
quently the new version must give full attention to the culture, as
well as the language, of the countries where it will be used.
A translation is not merely a linguistic undertaking with only
words being viewed and analyzed; it is a part of a much larger domain,
namely, that of communication (Neubert, 1969). This concept points
out that words in any language are something more than just written
symbols which represent an entity or idea; they are conveyors of
images which are in turn, reflections of each person's subjective
world. And, in order to tap into each person's world, a translator
must keep in mind as many of the cultural and social aspects, not
only of the work being translated but also of the person who will be
interpreting the translated work.
In order to bridge the gaps between one language and its trans-
lated form it becomes necessary to tap into what has been coined
"Grand Theory," which is a form of linguistic meta-theory or universal
grammar of human language. This language is the very basis of human
nature, that which knows no cultural boundaries (Werner & Campbell,1970).
In an attempt to bring forth the importance of the true, complete
meaning of language and to attempt to determine specifically what a
native speaker knows about his language, the transformational "Generative
Theory of Language" came into light (Chomsky, 1965). In this method
the translator uses projection in order to paraphrase the meanings
of the statements in question. By doing so, it becomes possible to
explore, not just superficial characteristics, but the inner, deep
structure of language.
Recent comparative studies in psycholinguistics are making it
clear that although languages do exhibit a uniqueness in superficial
phonology, grammar and semantics which render them mutually unintel-
ligible, at a deeper level they exhibit certain universal charac-
teristics which render them mutually translatable (Osgood, May, &
Miron, 1975). It is important, therefore, to understand the intent
of the communication, as opposed to simply the words that comprise
it, in order to obtain a test which can be validated or standardized
in a second culture.
Assertions have been made that many tests are unfair to culturally
different persons (Anastasi, 1968). It is because of this criticism
that the long-term concern to develop "culture-fair" tests has become
an increasingly important topic in the field of measurement. Culture
fairness and all its implications are topics of heated debate in the
field of measurement today. For some there will never be a perfectly
culture-fair test. For others there is still the hope that by per-
fecting the existing instruments and becoming more aware of what is
involved in culture fairness, that ideal situation may someday be
Because the United States traditionally has been a melting pot
nation which absorbs people from all parts of the world, there is a
critical need to understand differences in persons from many different
cultures. When any specific group increases dramatically, American
institutions and governmental agencies must be in a position to react
adequately to the needs of the newcomers.
One group of immigrants to the United States has been increasing
rapidly. According to the 1980 census (U.S. Department of Commerce, 1981)
there has been a 61% increase of Hispanic people moving to the United
States over the past 10 years, bringing the total Hispanic population
to 14.6 million people (Olmedo, 1981). Coupled with these augmenting
figures, there is evidence that the Hispanic population is increasing
at a faster rate than its American counterpart (Macias, 1977). The
Mexican-American population has been increasing steadily over the
years, many entering illegally as farm workers. This group of people
has experienced many problems, especially in the area of employment.
Mexican-Americans have been exploited and many Chicanos have gained
entry into the American labor force only because of their willingness
to work for substandard wages.
Many of the Cuban immigrants of the 50's and 60's have come from
the upper class strata of Cuban society. Now, and without warning, they
find themselves fighting for a place on the lower echelons of the American
social system. As if leaving one's homeland as an adult is not crucial
enough in itself, there is the added adjustment to a new social system
as well. As a result of these and many other situations involving
cross-cultural adaptation, a wave of new and diverse problems is
emerging in many phases of education and industry with their repercussions
being felt across many fields, including mental health.
In helping immigrants make the transition from their country to
the United States, agencies find themselves coping with people who
have personality disturbances, lack of education, physical illness,
and possible criminal behavior. For these reasons, it is important
to find expedient and effective ways of identifying problems as early
as possible so that prevention or intervention can be implemented.
When immigrants seek counseling it is not only because of the
problems that cause their American counterparts to seek help, but
also because of adaptation problems. These problems can center
around environmental issues, changes in careers, language problems,
and changes in values and mores. It becomes apparent that American
counselors working with immigrants must take into account complex
patterns of differences.
The first set of differences are cultural in nature, i.e., those
aspects which set immigrants, in general, apart from their American
counterparts. These areas would include such aspects as social class
structure, value systems, work habits, and more basically, living
The second set relates to interpersonal differences. These are
the nuances which make some people compatible with some and not
others, with some types of work and not others. It is in this second
area of differences that counseling can be especially useful, for it
is here that counselors can exercise their expertise in helping people
to become more effective and fulfilled people, regardless of their
culture. To do so, it would be necessary to identify those differ-
ences which are universal in nature, those traits which are typical
of each specific culture, and those traits specific to the individual.
A counselor would need to understand the immigrant's cultural, universal,
and unique attitudes and values through an evaluation of personality.
A questionnaire such as the Myers-Briggs Type Indicator (MBTI) might
help in the process of understanding the dynamics of personality in-
cluding the values, preferences, and interests of the immigrant.
Purpose of the Study
The purpose of the study was to develop and validate a Spanish
version of an English language personality questionnaire that would
be faithful to the original while taking into account differences
in values and attitudes of the two cultures. Specifically, the study
translated the MBTI into Spanish within the intent of the authors of
the Myers-Briggs Type Indicator and Jung's theory of personality.
Need for the Study
Basing his concepts on the premise that people prefer one of two
attitudes, either extraversion (E) or introversion (I), Carl G. Jung
(1923/1971) postulated that persons operate within these attitudes
using their most developed function. Other differences between extra-
verts and introverts were attributed to what Jung (1923/1971) called
"four functions of consciousness." These functions were a matter of
personal choice, one person may prefer thinking as a means of making
a judgment while someone else may choose feeling; one person may want
to experience his surroundings through his senses while another may
look into the deeper series of possibilities in his surroundings
thereby relying more heavily on his intuitive processes.
Sensing and intuition (S and N) are the basic functions of perception
and thinking and feeling (T and F) are the functions of judgment and
evaluation. If one of the perceptive functions (S or N) takes control
as the primary or dominant function, one of the judgment functions
(thinking or feeling) will be secondary or auxiliary; or if one of
the judgment functions (thinking or feeling) becomes dominant, then
the secondary or auxiliary will be one of the perceptive functions,
namely sensing or intuition (Jung, 1923/1971). Preferences of these
attitudes and functions are the basis of Jung's Typology, thus, the
term preference type, or simply type.
The attitudes of E and I, and the functions S, N, T and F are
assumed in the theory to be universal. The reference is to life, in
general, and in its broadest sense. This includes every aspect of
what makes a person's life an entity in itself, i.e., what makes it
the individualized subject it is. Thus, the very essence of a person
is being considered when these functions are experienced but they
are relating to more general, universal, aspects of life. Perception,
then, marks the universal aspect of this theory, but how a person
perceives marks the intimate, subjective way people experience things.
This combination allows for a theory so universal in nature that
people from every culture are touched and evaluated in the same ways.
It is the difference in perception and judgment among peoples that
provides the diversity of culture throughout the world. The theory
postulates 16 dynamic combinations of perception and judgment, called
the 16 types. All 16 are expected to appear in all cultures, but the
proportions of each type are expected to vary from culture to culture.
Jung's original premise, and his definition of function was: "It is
a particular form of psychic activity that remains the same in principle
under varying conditions. From the energy standpoint a function is
a manifestation of LIBIDO (q.v.), which remains constant in principle,
in much the same way as physical force can be considered a specific form
or manifestation of physical energy" (Jung, 1923/1971, p. 436).
This initial manifestation, according to Jung, is directed
outwardly extravertedd) or inwardly, into oneself (introverted).
These two general attitudes are then subdivided with a four-fold
classification of the functions of thinking (T), feeling (F), sen-
sation (S), and intuition (N). This allows for a total of eight
variants, namely, thinking, feeling, sensation and intuition in
either the extraverted or introverted form. Every person uses all
four functions, but the relative importance of each function, and
the attitude in which each function is used is different for each of
the types which comprise the Jungian typology (McCaulley, 1978).
By definition in the theory, one of the four major functions
(S,N,T,or F) becomes dominant in the person. This dominance is inborn
and as reenforcement takes place this dominant function becomes more
rewarding to the person. As the person matures, an auxiliary function
begins to emerge which complements the dominant function. For example,
if the dominant function provides mature perception (if it is S or
N), the auxiliary will provide mature judgement (T or F). When both
the dominant and auxiliary functions become differentiated, the
person achieves a balance (McCaulley, 1981).
It was through an interest in personality differences and the
development of her own typology from biographies that Katherine
C. Briggs first came in contact with Jung's "Psychological Types"
(Myers, 1979). Realizing that her ideas and Jung's were channeled
in the same direction, she and her daughter, Isabel Briggs Myers, did
further studies on Jung's work on psychological types. Ultimately
these led to the development of the Myers-Briggs Type Indicator
(MBTI), a questionnaire designed to permit a person to indicate
preferences in a non-threatening manner.
The Myers-Briggs Type Indicator, hereafter referred to as the
MBTI, was published solely as a research tool in 1962 after 20 years
of development and research. Many studies were conducted in order
to establish its validity and reliability. These studies led to
its publication, for general use, in 1975.
The MBTI can enter people's lives when they are still in high
school,for the purpose of helping them determine a preference for
a college education or for a business or vocational course upon
graduation (Myers, 1962). If persons choose to go to college, their
choice of major could be influenced by MBTI results. Having a
better understanding of themselves and where their preferences lie
could serve a useful function in determining which careers are com-
patible with those choices. As adults, people often experience
mid-life career changes and in this area the Indicator is also useful
in helping people match their preferences at that stage of their
development with career choices.
In education the most obvious preferences are between sensing
and intuition. Sensing persons focus their attention on the world
as it can be perceived through the concrete reality of the five
senses. They are "doers" and would rather engage in doing something
than listen to another person speak. Intuitive people, however,
rather assimilate as much as possible through language, either oral
or written, and incorporate these ideas into the subconscious process
In the world of work the Indicator is serving as an instrument
which facilitates decisions regarding the matching of successful
proven types with vocational goals. Also, if there are communication
problems between two people on a job, a knowledge of their type can
often assist in pinpointing the possible reasons for the difficulties.
As an example, extraverts generally prefer variety and action in
their work while introverts prefer a quiet setting and the opportunity
to concentrate. A thinking type may often hurt other's feelings
without knowing it whereas the corresponding feeling type is generally
aware of this and will do whatever possible to please others. A sensing
person is usually patient with routine details while an intuitive
person often becomes impatient under these circumstances. If these
differences are known to the employee and/or employer they can lead
to constructive uses of differences instead of their being a liability.
In the psychotherapeutic realm the Indicator is finding wide
usage. In individual counseling, people's understanding of their
basic preferences can often help clarify goals and reasons for problems
with other people. It might also help individuals discover preferences
in others that compliment their own, thus affording them a source
from which to further their own growth and development. In marriage
counseling, identifying the type needs of both partners is useful
in helping to resolve marital conflicts. A clearer understanding of
type preferences can help each person bring about those changes that
are necessary to live in better agreement with the immediate environment
The Myers-Briggs Type Indicator can be especially useful for
people from other cultures who are experiencing a cross-cultural
transition. It may point out their needs and desires in a very non-
threatening manner. This will involve an understanding of the process
required to adapt to their new environment. In order to accomplish
this in an ideal manner, the Indicator should be available in the
subject's native tongue.
Growing demands on the American society by the Spanish-speaking
immigrants to this country, make it necessary to find more effective
ways of helping this population. In order to bridge the gap between
cultures and be in a better position to empathize as well as understand
their points of reference, American counselors need to develop ways
of making their communications with their counselees as universal
as possible, so that in discussing problems and issues with them
they can genuinely say "our" problems because of the common element
of "humanness" which they share, and not alienate the counselee
by referring to "your" problem as a Hispanic. It is hoped that this
study will help move the counseling profession in that direction,
for the appreciation of differences among cultures of the world can
provide a valuable source of information for constructive progress
in the ways people relate to each other.
Definition of Terms
Typology is the description of personality wherein people's
behavior is described on the basis of differences in the way they
prefer to use perception and judgment (Jung, 1923/1971).
Personality tests are tests which measure the emotional, motivational,
interpersonal, and attitudinal characteristics of people as distinguished
from their abilities (Anastasi, 1968).
Immigrant is any person born and raised in a country other than the
United States who is presently living in this country.
Language facility in this study will be interpreted as the ability
of a bilingual person to read at the 6th grade level in the English
language as well as in his/her native tongue.
English form is the original standardized version of the Myers-Briggs
Type Indicator, Form G.
Spanish form is the Spanish version of the Myers-Briggs Type
Indicator, Form G. The development of this version will be the focus
of this study.
Organization of the Study
The remainder of this study will be presented in four Chapters.
Chapter Two, the review of the literature, will be subdivided into
three sections and will pertain to translation theories and some of
the problems encountered in translations. The methods and procedures
used in the study will be presented in Chapter Three. The findings of
the study will be presented in Chapter Four. This will incorporate
the intercorrelational results as well as the split-half reliabili-
ties and the results of the item analysis.
A summary, discussion of the results, recommendations and
conclusions will be presented in Chapter Five. Included within
the discussion section will be incorporated the limitations of
REVIEW OF THE LITERATURE
This chapter will present a review of the literature. There will
be three major areas considered. In the first there will be a review
of the theories and methods of translation. This section will be
followed by an overview of culture-fairness as an important element
of any assessment tool. The third, and last section, will consist
of a brief review of the Myers-Briggs Type Indicator, a personality
questionnaire based on Jungian theory.
Theories and Methods of Translations
In expressing ourselves in a language other than our own it
becomes very apparent that language has many nuances which can either
help or hinder the expression of our ideas at any given point in
time. When a person says, I am at a loss for words," he or she is
expressing that at that point in time there are concepts or feelings
present that cannot be put into words. They may be feelings which
are overwhelming, not understood, or simply, abstractions. Likewise,
when persons express themselves in their native tongue the repertoire of
words to express these feelings always seems greater than when
they are trying to express the same feelings in a foreign language.
Sometimes a person finds words to express those feelings which do
not translate linguistically, but do translate emotionally into the
There is no one correct translation of a sentence into another
language. For every sentence in the source language there are many
possible appropriate sentences in the-target language (Quine, 1964).
This is why translations are something more than linguistic interpre-
tations in another language. The transfer from one language into
another is something more than a transfer of words. Instead, it
represents a transfer of feelings, ideas, emotions, and values which
are all determined by the reference point of the reader and which is
influenced by the person's individual world. Because of this, literal
translations of written words are not truly translations of the in-
The person perceiving a certain verbal stimulus is doing so
from a very unique reference point which is affected by variables
such as his sex, race, cultural background, upbringing, i.e., his
individuality. The immediate environment is seen as something which
is either close and intimate ( a person's ingroup), or something
which is more distant (outgroup). For different people in diverse
cultures these ingroups and outgroups take on different meanings.
In the United States, for example, an ingroup consists of family,
close friends, and fellow countrymen. In Greece, however, the
ingroup consists of family, close friends and visitors, but excludes
other Greeks, who, in that culture, are considered members of the
These concepts are important to consider because they not only
help explain the degree of social distance which a person allows
others but it also affects the way a person would respond to a
subjective word-stimulus (Triandis, Vassiliou, & Nassiakou, 1968).
In translations, then, those terms which are influenced by the
person's culture will be responded to differently from culture to
culture. Only when these cultural nuances can be bridged will the
responses have the universality so desirable in most instruments
Emic and Etic Analysis
In order to gain this understanding across cultures it becomes
necessary to identify those aspects of the language which are related
to a specific culture as opposed to those other aspects which are
more universal in nature. Emic analysis refers to those variables
which are reflective of any one given culture, being meaningful,
therefore, to the members of that culture; etic analysis, on the
other hand, deals with the analysis of these variables in such a
way that generalizations across cultures can be made (Brislin, 1980).
Emic and etic data, therefore, do not constitute a rigid dichotomy
of bits of data, but often present the same data from two points
of view (Pike, 1966).
A summary of the two approaches and the distinctions between
the two was provided by Berry (1980):
Emic Approach Etic Approach
Studies behavior from within Studies behavior from outsic
the system. the system.
Examines only one culture. Examines many cultures,
Structure discovered by the
Criteria are relative to
Structure created by the
Criteria are considered
absolute or universal.
Etics, then, refer to culture-free items such as fire, moon, and
sun. Emics are those terms which are more abstract, and more sub-
jective in their definitions such as love, fairness, and honesty.
By definition, it is impossible to translate perfectly an emic
concept (Triandis, 1976).
It is possible, however, to define an emic concept in terms
of etic attributes. There is considerable evidence now that some
basic dimensions of social and cognitive behavior are universal.
These can be thought of as the etic dimensions that can be used
to define emic concepts. The great advantage of discovering such
universal, etic dimensions, is that it is possible to use them as
the framework for comparisons. The emic dimensions that are related
in reliable ways to etic dimensions can be compared using the etic
dimensions as a bridge (Triandis, 1976). These concepts are found
among several theories of translation, among them the theory of
association and the transformational theory of language.
A linguist will look at a word and define its meaning. The pro-
ponents of the associative theory, however, will describe a word in
a Gestalt way, its meaning embedded in a matrix of ideas, and not as
an isolated element in itself (Szalay & Deese, 1978).
Galton (1890) was the first to analyze associations empirically
and years later, in 1924, Freud also came to the conclusion that the
association method reveals the content of minds in a way that proposi-
tional language does not (Szalay & Deese, 1978).
In associative data we derive a better understanding of certain
concepts because of their association with other concepts. Citizens
of Colombia who are likely to say "polite" for "educated" are telling
us something about their attitudes toward education, perhaps that
only educated people are polite. In so doing, they incidentally
reveal a great deal about their culture that they might otherwise
have been unable to express (Szalay & Oeese, 1978).
The associative theory revolves around the concept of psycholog-
ical meaning, or the person's subjective perception and affective
reactions to segments of language. In analyzing groups of words
from an associative standpoint the assumption is never made that
the stimulus words are always exact translations (Kluckhohn, 1954).
This research theory focuses on equivalents; i.e., stimuli
that do not necessarily mean exactly the same but that allow the
tapping of English and foreign responses at comparable places in
the two representational systems. When exact translations are
lacking, selections are made of words closest in meaning, namely
the same topic (Szalay & Deese, 1978).
The content of the associations are also reflected in the
accuracy of the corresponding interpretation. Studies have been
conducted comparing the associations in English to English stimuli,
associations in a speaker's native language to English stimuli,
native words as responses to native stimuli, and English responses
to native stimuli. The results reflected that words referring to
concrete, manipulatable objects were more likely to be more alike
than words referring to abstract states or emotions (Kolers, 1963).
Not only is the content important in the associative method;
it is also relevant to know the order of response to the original
stimulus. Generally speaking, the earlier the response, the more
salient the component of meaning revealed by the response (Szalay
& Deese, 1978). The interpretation of these associations is not
a fixed, concise procedure. Due to its subjectivity it is important
that the researcher be familiar with the culture or realm from
which the responses are coming. The boundaries of a foreign lan-
guage must also be recognized and accepted, for usually, when responding
in a secondary language, the person's repertoire of responses will
be more restricted than if it were his native tongue ( Szalay
& Deese, 1978).
In order to penetrate the system from which a person is responding,
methods of free association have often proven useful. The system in
question is not a linguistic one; rather it is a means through
which the world is translated into an internalized subjective rep-
resentation (Szalay & Deese, 1978).
Among psychologists, Osgood (1964) and Triandis and his colleagues
(Triandis, 1964; Triandis, Vassiliou & Nassiakou, 1968) have been
concerned with the idea of "subjective culture."
Osgood defines "meaning" as "that process or state in the behavior
of a sign-using organism which is assumed to be a necessary conse-
quence of the reception of sign-stimuli and a necessary antecedent
for the production of sign-responses (Osgood, 1964, p. 9). Therefore,
meaning, like emotion, is a relational or process concept (Osgood,
Suci,& Tannenbaum, 1975).
Among linguists and philosophers many would say that two people
must first agree on the meaning of a sign before they are in a
position to disagree on their diverse reactions to it. As an ex-
ample, a person may find thunder (object) challenging and exciting
while someone else may see it as something extremely frightening,
but before they can communicate about this state of affairs they
must first agree on the referent of the linguistic sign "thunder"
in their common language (Osgood, Suci,& Tannenbaum, 1975).
In like manner,the pattern of stimulation which is a sign or
indicator is never identical with the behavior pattern which is the
significate. The word "hammer" is not the same stimulus as the
object it represents. The former is a pattern of sound waves; the
latter, depending on its mode of contact with the organism, is a
combination of visual, tactual, proprioreceptive and other stimula-
tions. In spite of this, the sign (hammer) does come to elicit
behaviors which are in some manner relevant to the significate
(hammer), something which is not shared by an infinite number of
other stimulus patterns that are not signs of this object (Osgood,
Suci,& Tannenbaum, 1975).
In dealing with abstractions in this manner the cultural milieu
often plays an important role in the deep meaning which the stimulus
elicits. For example, the American politician who uses the concept
'future' with his own connotations of good, strong, and active,
might fail to communicate this intention when he addresses a Finnish
audience, where "future" is seen as being both good and strong, but
also passive in nature (Osgood, May,& Miron, 1975).
These observations suggest that it may be difficult, or impossible,
to bridge cultural gaps in analyzing the written word. However, recent
comparative studies in psycholinguistics are making it clear that
although languages do exhibit a uniqueness in superficial phonology,
grammar and semantics which render them mutually unintelligible, at
a deeper level they exhibit certain universal characteristics which
render them mutually translatable (Osgood, May, & Miron, 1975).
It is important, therefore, to analyze the meaning of the
ideas, as opposed to simply the words that comprise it, in order to
develop a test which can be validated or standardized in a second
culture. In the process of standardization across diverse cultures
it may be necessary, purposefully, to modify the instructions and
procedures, whereas in within-culture research the instructions and
procedures may be identical (Osgood, May, & Miron, 1975).
Implications for communication and bilingualism
A basic misconception applicable to bilinguals is that in
speaking a second language it is assumed they can also be tested
in a second language (Padilla, 1979). In timed tests, the reading,
writing, and speaking fluency of persons taking the test is of utmost
importance. Experimentally, fluency has been studied in the ana-
lysis of reaction times. The results indicate that, in the case of
bilinguals, a delayed reaction time is an indication of less fluency
in the second language for it indicates a translation back into
the native tongue may be occurring before the response is provided
(Osgood, May, & Miron, 1975).
Bilingualism may tend to present problems of complete compre-
hension because of language differences that may not be recognized
in the second language. However, recent research indicates that
bilinguals possess what researchers are coining "cognitive flexibility,"
namely the problem-solving ability to see the problem and its possible
solutions from more than one viewpoint, thus, developing richer,
more creative ways of analyzing and solving the problem. The research
also implies that these bilinguals then have at least two ways of
looking at their world, enhancing their awareness and understanding
of their surroundings (Lambert, 1974].
From the standpoint of understanding, bilingualism presents
persons with a more ample repertoire of envisioning their surroundings.
However, in translating into a second language this asset may become
a liability; the translator must be aware of these differences in
perception. In order to develop an adequate translation he/she must
go deeper into the original language, probing not the linguistic
equivalents of the material in question, but rather the true meaning
being conveyed by the author Of the statement. Nida refers to the
term "dynamic equivalence" and defines it as the translation of a
culture symbol in the source language into a culture symbol in the
target language which elicits the same functional response (Nida, 1964).
Implications of theory for translations
The awareness of this problem has led to an analysis of language
in a more complete way. The surface structure is still very important
but without an understanding of its deep structure the translation
of an instrument into an equivalent in a second language would be
meaningless. In the case of the Myers-Briggs Type Indicator, for
example, the questions being asked are the elements of the surface
structure which are intended to measure Jung's theory which is the
Transformational Generative Theory of Language
In an attempt to bring forth the importance of the true, complete
meaning of language and in an attempt to determine specifically what
a native speaker knows about his/her language, the Transformational
Generative Theory of Language came into being (Chomsky, 1965]. In
this method the translator uses projection in order to paraphrase
the meanings of the statements in question. By doing this it becomes
possible to explore, not just superficial characteristics, but the
inner, deep structure of the language, which specifies the unique
shared structure closest to the semantics of a set of sentences which
are syntactic paraphrases of each other (Werner & Campbell, 1970).
These syntactic paraphrases are those which retain the same
lexical items rather than substituting synonyms and are essential in
the decomposition of complex sentences into their constituent simple
sentences (Chomsky, 1966). In addition to syntactic paraphrasing,
transformational theory recognizes lexical paraphrasing, whereby
synonyms become the medium through which one component is broken
down into its parts. This latter method is of primary importance
in componential semantic analysis, and it is based on the native
speaker's knowledge that certain words are indeed equivalent in
meaning to certain other words or phrases. This type of paraphrasing
is also of importance in translations. The longer paraphrase of
a term (e.g., "devoid of moisture" for "dry"), is more readily
translatable because of its specificity (Goodenough, 1956).
Sometimes this form of paraphrasing is only partial. In such
cases a more generic term of the source language is lacking in the
target language and a more specific term of the target language has
to be substituted. Conversely, often a more general term has to be
used in lieu of a specific term in the source language (Werner &
Campbell, 1970). In order to bridge the gaps between one language
and its translated form it becomes necessary to tap into what has
been coined "Grand Theory", i.e., a form of linguistic meta-theory
or universal grammar of human language. This language is the very
basis of human nature, that which knows no cultural boundaries
(Werner & Campbell, 1970).
Universal Patterns of Human Thought
There are also some basic thinking patterns which are universal
in nature. For example, the disposition of human beings to think
in contrasts or opposites is recognized (Deese, 1965; Lyons, 1963;
Ogden, 1967). The act of contrasting is something which is primitive
in man's thinking processes, and as Greenberg (1966) pointed out,
the pairing of adjectives into opposites is universal among languages.
Cultural Difference Issues in Translations
In spite of the fact that there are some universals among lan-
guages there are still cultural differences among cultures which
need to be recognized when translating into a target language.
Research has been conducted in the area of social psychology to de-
termine the social distance factor across cultures. These studies
analyze the relationship between characteristics such as race, na-
tionality, or religion, and the way typical members of different
cultures react to them (Triandis, 1976).
Although the concept of "social distance" is an abstract one,
studies by Thurstone in the 1930's developed strategies for obtaining
a value for each statement along the social distance scale, thus a
numerical value which was measurable could be given to the concept
(Triandis, 1976). Attempts to translate the Edwards Preference
Scale, for example, have met with difficulties which are culturally
related. Berrien (1966) first looked at the preferences on the test
and had them scaled for social desirability in Japan. Using this
method he reassembled these new values into paired items unlike the
original ones used. If the purpose was to make differentiations
among the Japanese, this procedure would have been correct, but in
comparing the United States and Japan, it was wrong. If the procedure
had been executed flawlessly, there would have been no value or
need differences between the cultures and within each culture, the
mean of each value or need would have been the same (Kikuchi & Gordon,
Another researcher faced further complications when trans-
lating the instrument into Japanese. The items pertaining to
heterosexual interest were too crude for Japanese sensibilities.
These items had to be reworded in order to make them more acceptable
in that culture. Under these circumstances the question arises
once again: Is it fair to compare the Japanese results with those
of Americans? In the extended answer to this question, the re-
searcher suggested using a bicultural construct validation technique
but failed to consider collecting new data in the United States using
the new Japanese version (Berrien, 1967).
Other methodological problems may arise in translating tests
into target languages. For example, direct translations do not or-
dinarily yield technically equivalent forms because the domains
sampled by the different language versions may have little in common
with each other, and the translated items may exhibit psychometric
projectives substantially different from those of the original
instrument. Additionally, since often the test remains culture-
bound, the interpretation of the resulting scores may be inaccurate
Verbal test items are usually more independent of culture than those
which are nonverbal. In the Thematic Apperception Test, for example, the
pictorial content of the instrument is very culture-bound. Although
efforts to remedy this have been attempted in the form of new pictures
which are more culture-specific, they have not proven successful, for
the possibility remains that the differences in responses are due to
the differences in the pictures ODoob, 1965; Lindzey, 1961).
These differences on either end of the translation, regardless
of whether the original stimulus is verbal or nonverbal, have been
recognized. If the aim of the translation is loyalty of meaning and
equal familiarity and colloquialness in both languages, the translation
is said to be symmetrical or decentered; if there is dominance in
one language, and it is usually in the source language, the translation
has been coined asymmetrical or unicentered; in this instance, a
product which may have been colloquial or familiar in its original
language is translated, approximately, into something unnatural,
even exotic, in the investigator's language (Werner & Campbell,
Other ways of looking at translations have been reported, focusing
more on the cultures in question than with language itself. In
"culturally ipsatized measures" the same instrument is used in the
societies being studied. The recording and interpretation of the
resulting samples of behavior are judged relative to others in that
society, rather than relative to some universal standard of inter-
pretation. If the recording of the responses are completed in an
identical manner with identical stimuli, the measurements are said
to be culturally universal. Sometimes the indicators are altered
to make them culturally appropriate but the-original scoring is
retained. These have been coined "culturally modified instruments."
In the last instance, or "culturally specific measures," the maximum
phenomenal variability is attained, with the objective of achieving
the maximum conceptual uniformity (Straus, 1969).
One of the most dynamic classifications of translations, however,
is one which combines the elements just considered with the element
of feeling tone of the message being conveyed. The first division
considered in this classification is known as "pragmatic translation,"
in which the message is translated with an interest in the accuracy
of the information that was meant to be conveyed in the source lan-
guage form. The "aesthetic-poetic" translation takes into account
the affect, emotions and feelings of an original language version as
well as any information contained in the message. In "ethnographic
translations," the cultural context of the source and second lan-
guage versions is considered. This would be similar to the "culturally
modified instruments" of the previous classification. Lastly, the
"linguistic translation" concerns itself with the equivalent meanings
of the constituent morphemes of the second language (Casagrande, 1954).
In addition to the above considerations, other cultural nuances
must be kept in mind in completing any translation. For example,
in Spanish there are five ways of saying "you"--the familiar singular
form "td"; the familiar plural "vosotros" (masculine) and "vosotras"
(feminine); and the formal "usted" and "ustedes". "Usted" comes from
the archaic construction "vuestra merced," meaning "your grace."
The feeling tone of the material being translated must be kept present
for with one form of "you" the mode of addressing the reader is an
informal tone whereas in the other it is a most formal one (Berlitz,
The Spanish language also faces the problem of being composed
of many dialects. Sometimes the translation may be adequate but
because of a diversity of dialects in the target language it might
not be understood (Padilla & Ruiz, 1973). Regardless of culture
or language, an important aspect of any translation is whether the
word being translated is a single entity in itself, or whether it
is part of a more complex structure. Researchers seem to agree on
the fact that in order to achieve a translation which is thorough,
the word being translated must be offered in context and not as a
single word (Chapanis, 1965; Osgood, May, & Miron, 1975; Seleskovitch,
Often, emphasis has been placed upon "situation context," i.e.,
that translation is not merely a linguistic undertaking with only
words being viewed and analyzed; it is a part of a much larger domain,
namely, that of communication (Neubert, 1969). This concept points
out, once again, that words in any language are something more than
just written symbols which represent an entity or idea. They are
conveyors of images which are in turn, reflections of each person's
subjective world. And, in order to tap into each person's world,
a translator must keep in mind as many of the cultural and social
aspects, not only of the work being translated, but also of the
person who will be interpreting the translated work. By doing this
it is then possible to tap into the intent of the message and not
just into the words that comprise it.
Situation context then leads to another area of concern,
namely problems of content. In translating the Greek classics, for
example, the translator must decide whether to preserve as much as
possible of the unique features of ancient Greece and thus attempt
to transport the reader back in time to the place or places where
the communication originally took place, or to provide a new cultural
setting for the corresponding cognitive content (Nida, 1976).
This concept also has its limitations as posed in the Sapir-
Whorf Hypothesis; namely that human beings who are speaking different
languages do not live in the same "real" world with different labels
attached. In actuality they live in different worlds altogether.
Language, in this instance, is nothing more than a filter of reality,
molding our perceptions of the universe around us (Werner & Campbell,
1970). It is important, therefore, for the researcher to have adequate
training in understanding a social system that is alien, complex,
and culturally different in order to collect valid data (Irvine, 1968).
In a recent work Nida presents a classification of theories
of translation which addresses a majority of these concerns, and
bases it on the focus of attention of the translation. First, philo-
logical theory concerns itself with literary texts. In this approach,
instead of just treating the form in which the text was first composed,
consideration is given to the structures in the source and receptor
languages and an attempt is made to evaluate their equivalences.
This includes, to a certain extent, a concern for deep structures.
Levy is recognized as a pioneer in this field for his concern over
sound linguistic principles was a major element in the foundation
of his literary translating theory. As early as 1931 comparative
studies of literary tests were being formulated but these were
attempts at comparisons and nothing else (Belloc, 1931).
In 1958 protests were voiced against this theory. Language
was not being considered strongly and it was felt that it should
be. That marked the beginning of a shift towards a concern for
the language which comprised the literary texts in question, and not
just over the texts themselves (Fedorov, 1958, 1968).
Although traditionally it was the philosophers and logicians
who were concerned with the nature of meaning, growing interest by
linguists over these issues led to the birth of what Nida has coined
"linguistic theories of translation" (Nida, 1964). The shift in
focus was also prompted, in part, by Fedorov's protests. As a
result, instead of looking at literacy genres and stylistic features,
the emphasis was to analyze the linguistic structures of source and
receptor texts. This shift in focus also brought with it a more
pronounced interest in the deep structure of language because
language cannot be discussed as though verbal communication occurs
in a cultural vacuum.
The last category recognized by Nida is called "sociolinguistic
theories of translation." In this model the translator is compelled
to take language performance as seriously as language competence.
This theory also recognizes the individual's changing emotional
state and needs (Nida, 1976). Due to the nature of this theory and
because it tends to be more explicit than the rest it has been de-
termined that in any good translation there will be approximately
50% redundancy in language. Therefore the results will be somewhat
longer than the original (Colby, 1958).
In spite of this seemingly thorough classification, there are
still some concepts that cannot be directly translated: puns, meta-
language, certain types of literary allusions, and some sociolin-
guistic dialects. At best, the resulting translation of items such
as these would only be paraphrases of the original text (House, 1973).
A further extension of the classification would include the process
of interpretation: the focus of attention in this case is directed
toward the ideas expressed in live utterances and the language
equivalents are not even attempted. In this respect the functions
of interpretation and code-switching are absolute opposites:the
former deals with ideas, while in the latter case words are regrouped
and analyzed, shifting the emphasis to the language itself.
Since interpretation deals with ideas, this introduces a new
element into, and becomes a part of, the code-switching chain. Lan-
guage X symbols lead to an interpretation of their intended meaning,
which in turn leads to a conveyance of the intended meaning (as
perceived by the interpreter) in language Y symbols. Therefore,
translation is seen as an attempt to introduce linguistic equiva-
lents, whereas interpretation aims at integral communication of
meaning (Seleskovitch, 1976). The translator is the person who
will determine which theory or theories are applicable to the trans-
lation in question. Thus, in a way, it becomes a very subjective
Strategies for Insuring Equivalence of Translation
Among the methods of translation, Brislin's (1976) classification
seems to be thorough and well accepted. He recognizes four distinct
approaches to any translation effort, and recommends using, not
one, but a combination of the approaches, depending on the need.
In the first, or back translation, the researcher prepares
material in one language and asks a bilingual to translate into a
second, or target, language. A second bilingual then translates
the material back into the original, or source language indepen-
dently. At that point the researcher has two source language
forms and even if he/she does not know the target language, can
make a sound judgement about the quality of the translation.
Studies report that in order for the back translation to be
optimally effective the passages in the source language should be:
1) simple sentences; 2) repeated nouns as opposed to the use of
pronouns; 3) devoid of metaphors and colloquialisms; 4) devoid of
passive tense; 5) devoid of hypothetical phrasings or subjunctive
mood (Werner & Campbell, 1970).
In back translation a limitation must be recognized. In order
for the translated form to be effective the source language may have
to be revised, based on informal comparisons of the meanings of
the source and back-translated versions (Werner & Campbell, 1970;
Fink, 1963). Research conducted in translations of European lan-
guages supports this concept (Bass, 1968; Jacobson, 1954).
Another consideration in the use of this method is that a perfect
translation, as evidenced by a sound back translation, is not always
a good adaptation. Sometimes the problems are as much bicultural
as they are bilingual, and only bilingual informants would be in a
position to notice bicultural differences (Osgood, May, & Miron, 1975).
Generally speaking, however, the back translation technique
seems capable of bridging cultural gaps, as evidenced by successful
research using this method in a diversity of cultural settings: in
Laos in an attitude survey (Fink, 1963); with the Navaho Indians in
the United States (Werner & Campbell, 1970); and in a French study
relating to governmental work (Sinaiko, 1963). Although it has been
strongly recommended that back translation methodology be used in
research that requires instruments in more than one language (Cortese
& Smyth, 1979), it has also been recognized by Brislin and others
that a researcher cannot depend solely on this technique (Gough,1968).
Brislin outlines seven steps in the use of this technique:
1) Write an English form which is translatable.
2) Secure competent translators.
3) Instruct one bilingual person to translate from the source
to the target language.
4) Have several raters examine the original, target, and/or
back-translated versions for errors that lend to differences
in meaning. If errors are found, step three must be repeated,
changing the original English when necessary, the process
known as "decentering" (Brislin, 1970).
5) When no meaning errors are found, pretest the translated
materials on target language speaking people.
6) Administer the materials to bilingual subjects, some who see
the English version, some who see the translation, and some
who see both. The responses should be similar across groups,
as assessed by means, standard deviations, and correlation
7) Report experience using the different criteria for equivalence
The most important step in this procedure is the decentering
process. In this process the source and target languages are equally
important in the translation procedure, i.e., they both contribute
to the final set of questions, and both are open for revision. It
should be noted, however, that in the case of translated tests, if
the original version is revised, this would indicate a need for a
restandardization of the original instrument. As a result, in the
case of measurements it would seem more advisable to revise the target
version and not the source whenever possible.
The decentering process is related to emics and etics. Those
concepts that "survive" the translation-backtranslation procedure
would be etic concepts since in order for them to survive the terms
must be readily available in both languages. By the same token,
those concepts which do not have equivalents in the second language
would be "lost"; those would be the emic concepts. The emic and etic
concepts would then be interrelated through statistical techniques
The second method of translation recognized by Brislin in his
classification is the "bilingual technique," whereby bilinguals take
the same test, or different groups take different halves of a test,
in the two languages that they know. Items in which there is a
discrepancy in the responses are easily identifiable. This technique
is advisable because of its precision and for the potential use of
sophisticated statistical analysis of the results, such as split-half
reliability assessment. The major criticism of this method, however,
is that the respondents, i.e., bilinguals, comprise an atypical
group within the population.
For some researchers this method represents an optimal way of
achieving equivalence of translation (Schachter, 1954). Used as a
formal statistical approach, the goal of this procedure should not
be to achieve identity on the item-by-item level, but rather equi-
valence of the Gestalt, i.e., of the means and variances, plus
appropriate correlations between scores on the two forms (Werner &
If the bilingual technique is used, careful consideration must
be given to the content of the source version. Chapanis suggests
three rules to follow in writing the material to be translated.
First, use as small a vocabulary as possible and make sure
that the vocabulary being used is known to all the communicators.
Secondly, the use of familiar, as opposed to unfamiliar, words should
be encouraged. The reason for this is that, generally speaking,
familiar words have a wider choice of target versions than those
that are not. Studies by Nida (1964) and Spilka (1968) stressed this
aspect of the translations, stating that the familiar usage of words
provided an easier access to the appropriate choice in the target
language than those words which, because of their less frequent
usage, were more difficult to interpret. Third, use as much context
as possible in the interpretation of the translated forms. Whenever
there are difficult words under consideration they should appear
in sentences, if at all possible (Chapanis, 1965).
In an investigation of the bilingual technique, Prince and
Mombour recommended a five-step procedure for obtaining optimal
results with this method (Prince & Mombour, 1967).
The primary consideration under this approach is a careful
a priori translation; it is suggested that words be chosen that would
even have the same frequency of usage in both languages. The obtained
translation is then tested on bilingual individuals; the researcher
randomly assigns bilinguals to two groups. One group is asked the
first half of the questions in the source language, and the second
half in the target language; this order is then reversed for the
second group. Next, the translation of items eliciting discrepant
response frequencies should be segregated and revisions of these
translated items should be attempted. The trial runs should be
continued until comparable frequencies are obtained. Only then
will operational equivalence have been achieved.
In the fourth step the authors recommend discarding those test
items which continue to have discrepancies in the responses. Finally,
a survey should be carried out in order to rule out any possibility
that any salient differences between the two instruments are not
due to any faulty translation procedures (Prince & Mombour, 1967).
A final method of translation, as outlined by Brislin, is what
is commonly referred to as the "committee technique." In this
approach the translation is carried out by a group of people, and
errors in the translation are caught by other committee members. This
approach has been criticized because in the event that members are
not open with each other the ultimate result of the translation would
be useless. In spite of the problem, this approach coupled with
one or more of the others has proven successful in a number of
studies (Bass, 1968; Hudson, Barakat, & LaForge, 1959; Jacobson, 1954).
In conducting pretest procedures, the translated form should be
field tested to insure tht people will understand the material
which they will be expected to respond (Brislin, 1970).
In the "random probe technique" the interviewer selects a random
sample of items on a questionnaire: and asks a probing question about
each of them, such as: "What do you mean?" If the response to the
question is bizarre, suspicions would arise concerning the conveyance
of the intent of the question (Schuman, 1966).
In a study conducted in Nigeria the interviewees were asked to
rate the 27 items of the questionnaire, the purpose of which was to
determine the clarity of the question to the people. It was added
that questions highest in rated clarity were highest in question
reliability as measured by two interviewers questioning the res-
pondent (Mitchell, 1966).
The protesting technique has also been successfully used in
conjunction with other methods (Hudson, et al., 1959; Kandel, Lesser,
Roberts, & Weiss, 1968).
Assessment of Equivalence
A difficulty present in the assessment of translation studies
is the lack of an ultimate criterion of translation quality (Spilka,
1968). This problem may have been suggested four years earlier when
Nida stated that one of the requirements of translation is to pro-
duce a similar response on the part of readers of a source and target
version of the material in question (Brislin, 1970). Other problem
areas have dealt with the difficulty of source language materials
and the negative effects they produce in translation (Treisman, 1965).
For most researchers, equivalence of meaning is regarded as
the most crucial element of any translation (Catford, 1965; Nida,
1964). Some criteria for translation equivalence has been suggested
by Miller and Beebe-Center (1956) and by Macnamara (1967). The former
research team reasoned that if people could perform bodily movements
after having heard either source or target language instructions,
and if the results of those bodily movements criterion were similar
across all people, then the source and its translation must be
equivalent (Miller & Beebe-Center, 1956).
Brislin also showed concerns regarding equivalence, and offered
several criteria for equivalence. First, he suggested that mono-
lingual raters examine the original and the back-translated forms
of a passage. As they do this they are instructed to write down
any errors they feel would lead to differences in meaning if the
two forms were administered. The same procedure is then held, en-
couraging the use of bilingual raters who look at both sides of the
translation, and write down errors as outlined in the first procedure.
The last consideration offered by Brislin is that the subjects should
be able to answer questions about the target language version and
back-translated version equally well (Brislin, 1970).
The theories and methods of translation are numerous. The under-
lying common thread, however, is the accuracy of the interpretation.
If the person doing the interpretation or interviewing is not adept
at what he/she is doing, the most thorough translation could prove
to be useless.
Psychological testing has been endorsed by many, yet refuted
by many others. From its early beginnings, many saw testing as an
aid to classify people in order to give them the treatment and/or
training which would best suit their particular situation. To others,
however, tests were used as labeling devices and, as such, were
instruments which could be used to prevent access to educational
opportunities or jobs. The responsibility assumed by the tester
was great and as years passed there was more resentment towards
those who used them and toward the field in general. Beginning
in the early 1950's this unrest seems to have escalated due, partly,
to the fact that testing itself has expanded dramatically since
The criticisms against testing stem from one of four sources;
first, there has been mounting concern regarding invasion of privacy
and confidentiality of test results. A second source of criticism
is the misuse of tests, either as diagnostic tools or as a placement
agent. A third source is derived from misinterpretation of test
results and misconceptions about the nature and purpose of tests
(Anastasi, 1968). The fourth source refers to the assertions being
made that tests are unfair to culturally disadvantaged persons.
It is because of this criticism and because of fairly recent legis-
lation, that the long term concern to develop culture-fair tests
has become increasingly important in the field of measurement.
A major reason for the increased interest in culture-fair
tests is the noticeable change in foreign population distribution
in the United States. A shift began in the early 1960's with the
heavy migration of Cubans to southern Florida, and more recently
there has been a migration of Haitians to this country. The use of
tests with these populations could be of great importance to coun-
selors and mental health workers.
Tyler and Wolf (1974) have defined culture-fair:
Fair use states that a common qualifying score may
be used with two groups if the regression line based
on one group does not systematically over or under-
predict criterion performance in the other. However,
it is shown that when the two groups differ appreciably
in mean test score, the above procedure, which is
"fair" to individual members of the group scoring
lower on the test, is "unfair" to the lower group
as a whole in the sense that the proportion qualified
on the test will be smaller, relative to the higher
scoring groups, than the proportion that will reach
any specified level of criterion performance. (p. 35)
This means, in essence, that when test results of two culturally
different populations are compared, the results of both should fall
along a common slope on a regression line, and any differences or
deviations from this norm should not be significant. If the two
groups are culturally different and the differences are indeed sig-
nificant, the cultural bias could be the reason for the discrepancy.
Summarizing, thus far, if a test is indeed culturally fair, the
results obtained from a testing situation with a group of Germans,
for example, would be as valid as the results obtained for a group
of native Americans who have taken the same test.
As early as 1923 there were documented situations where ethnic
groups form southern Europe were being unfairly characterized as
subaverage on the basis of test results. Researchers, in the belief
that these discrepancies were due to cultural influences and not
intellectual differences, began analyzing the cultural influences
on test results. Although the undertaking was unsurmountable for
the period and the efforts were discontinued, it marked the beginning
of a new area of concern for mental measurements (Eells, Davis,
Havighurst, Herrick, & Tyler, 1951).
As the years went by, more discrepancies in test scores between
populations were observed. One area of concern that was being
documented heavily was the area of mental retardation. The number
of children classified as mentally retarded and being placed in
special education programs was growing at a rapid pace. Initially,
the distinction was not being made, but after some time it was
determined that in many of the cases those children who were clas-
sified as mentally retarded were only culturally different and the
measures of their intelligence were not taking into consideration
those cultural differences (Anastasi, 1968).
As the incidents increased, several of them were handled in the
courts. As a result, and because of growing concern over tests
and their uses, P.L.94-142 was enacted. This law clearly mandated
that tests and other evaluational procedures would be non-discriminatory
in nature (Bailey & Hardin, 1980). As a result of P.L. 94-142 a
new concentrated effort to eliminate as much of the cultural bias
from testing as possible began to take place. This effort and concern
has grown in importance over the years and is still a major factor
of concern today in the area of measurement.
The process of eliminating culturally biased questions was not
an easy one. As an example, a question dealing with tying one's
shoe laces would be culturally loaded for it may be presented in
a society where shoes are more of a luxury item than a necessity.
The issue in question then would be to deal -with something as universal
as possible. Cattell (1979) explained it well: Sun, moon, and stars
belong to everyone across the earth, and if we could pose sufficiently
complex relational puzzles among them we could approach culture fair
conditions" (p. 7).
Since it is virtually impossible to deal with absolutes in
such situations, attention has been turned to eliminate as many of
the cultural parameters as possible. This meant eliminating and/or
changing elements in tests that would be definite marks of any one
given culture. The first such element that was considered a definite
bias in testing was language. It was felt that the person taking
the test would be greatly influenced in his reaction to the written
word by past experiences, including the cultural milieu in which
he/she was raised.
Although the Army Alpha and Army Beta had been published for
years, it took some time before a surge of interest in nonverbal
instruments took place. This added interest seemed to grow as a
result of the introduction of the nonverbal elements in the Army
Beta. In some instances it was possible for the tests to be cons-
tructed in a non-verbal manner and still measure what they were
intended to measure, but in other cases this ideal could not be
achieved, and has not been achieved to this day (Anastasi, 1968).
There are three important nonverbal tools in use today. The
first of these came into existence when the mandates of P.L. 94-142
were still very new, and at the time of its inception was seen as
the perfect instrument to carry out the mandates of that law. It
was called SOMPA (System of Multi-Pluralistic Assessment), and as
the name implies, this is not one but a series of nine tests which
is intended for children ages five through eleven. The test is
non-language in nature and attempts to determine estimated learning
potential (ELP) (Figueroa, 1979).
The SOMPA was seen as a new way of conceptualizing the assessment
process while at the same time taking into account underlying social
and political assumptions. It was also a tool which combined liberal
and humanistic values with factual empiricism. As such, the system
made it possible to see children form a biological, social, cultural,
and educational perspective (Nuttall, 1979).
The SOMPA was standardized on a sample comprised of 700 blacks,
700 English-speaking caucasians, and 700 Spanish surnamed individuals.
The sampling procedure appeared accurate but the drawback, and one
for which SOMPA has been criticized, was that the Hispanic group was
made up entirely of Mexicans. Other Hispanic subcultures such as
Puerto Ricans and Cubans were not included in the sample. In spite
of this criticism, however, the instrument has shown great potential
and the introduction of the estimated learning potential concept
into the field has been widely acclaimed (Nuttall, 1979).
Another intelligence test which is also considered culture-
fair and which is an individual performance test is the Leiter
International Performance Scale. This test was developed through
several years of use with different ethnic groups in Hawaii, and
included groups of elementary and high school students. The test
was later tested in Africa and further testing and revisions were
made in the United States (Anastasi, 1968]. In the Leiter test there
is an almost complete elimination of instructions, either spoken
or pantomine. The test consists of a response frame and cards
containing printed pictures which are then placed appropriately
within the response frame (Anastasi, 1968).
A third instrument which has the distinction of being consi-
dered culture-fair is the Culture-Fair Intelligence Test developed
by Cattell and published by the Institute for Personality and
Ability Testing (IPAT). This test is based on the idea that general
intelligence is a matter of seeing relationships in the things
with which we have to deal, and that this ability to see relationships
is measurable by means of simple diagrams. For a test to be usable
in different cultures the pictures or forms represented should be
of fairly universal objects (Thorndike & Hagen, 1955).
Although all three of the above mentioned tests are widely
used in spite of some criticisms, they still take into account that
the results are measuring intelligence and nothing else. This in
itself does not necessarily indicate that a person may need special
services in his/her education.
In order to further substantiate proper school placement in
as fair a way as possible, adaptive behavior scales are being used
with increasing frequency. These scales attempt to measure the
individual's ability to cope with the natural and social demands of
the environment. They are being advocated frequently by educators
today and it appears that their use tends to reduce the placement
of minority children in special classes. The use of dual criteria
(intelligence tests and adaptive behavior scales) is assumed to
result in a more valid diagnosis of mental retardation (Bailey
& Hardin, 1980).
Among the measurements of adaptive behavior can be found the
Vineland Social Maturity Scale which was designed to identify ab-
normal behavior and development as measured by sets of age related
developmental milestones. Unfortunately, however, this instrument
was normed on white, middle class populations, ignoring the varia-
tions of a culturally diverse population. As a result, the instru-
ment is not considered culture-fair and is not widely used (Bailey
& Hardin, 1980).
The Adaptive Behavior Inventory for Children (ABIC), however,
took into consideration cultural variations in its standardization
and is therefore, more widely accepted than the Vineland. Research
has also shown that the use of the ABIC has eliminated the over
representation of minority children in some school systems (Bailey
& Hardin, 1980).
There is another aspect to be considered in analyzing culture-
fair tests. In many instances an instrument cannot be nonverbal
in its presentation. It must be presented in such a way that its
interpretation across cultures is also fair even though it is being
presented verbally. In these instances the task at hand is much
more complex, for word interpretations across cultures vary greatly.
The problem becomes one of being able to somehow measure the affective
qualities of human communication, something which is much more sub-
jective. Charles Osgood in his Measurement of Meaning stresses the
need to be able to determine what the word means to the reader of
the written word in order to make objective evaluations of his/her
performance on the test.
In a more recent work, Szalay and Deese introduce the idea of the
associative theory of word analysis. In this theory they present the
idea that word associations often tap deep into a person's belief
system, regardless of culture. These associations are important
tools in looking at the person's world perception and how it affects
his/her behavior (Szalay & Deese, 1978).
The associative theory will hopefully be of significance as
an attempt is made to not only translate the Myers-Briggs Type In-
dicator into the Spanish language, but to construct a Spanish version
of the tool. In the proposed research, the instrument being adapted
is the MBTI. Because of cultural possibilities found in the Spanish
language, and keeping in mind that this new form will also have to
be as culture-fair as its original counterpart, it will be important
to focus on those items which tend to be culture-specific in making
Culture-fair and all its implications are topics of heated
debate in the field of measurements today. For some there will
never be a perfectly culture-fair test. For others there is still
the hope that by perfecting the existing instruments and becoming
more aware of what is involved in culture-fairness that ideal test
may someday be achieved.
Myers-Briggs Type Indicator
The Myers-Briggs Type Indicator, or MBTI as it is more commonly
known, is an instrument which came into existence out of a need to
measure the personality preferences advocated in Jungian Typology.
At the beginning of this century, Katherine C. Briggs developed,
based on her own interests in the area of personality differences,
her own typology. In 1921, however, with the publication of Jung's
"Psychological Types," Briggs became very aware that Jung's ideas
and her own were very compatible except for the fact that Jung's
thoughts seemed more complete and complex than her own.
Briggs studied the Jungian text thoroughly and then taught the
obtained information to her daughter, Isabel Briggs Myers. For the
subsequent twenty years Myers and Briggs were involved in "type
watching," and became convinced of the strength and credibility to
Jung's theory in understanding and predicting human behavior.Observation
of the American people during and after World War II led Isabel
Myers to feel that it was both unfortunate and unnecessary for many
Americans to be working at jobs which did not lend themselves to
the preference of the employee.
In order to help those who were in such positions and to help
prevent the further spread of such incidents, Myers felt the need
to develop some method by which the. Jungian preferences could be
indicated and then acted upon. As a result, in 1942, she began to
develop the first questions that would, years later, lead to the
Myers-Briggs Type Indicator.
For fifteen years Myers developed and tested questions and
scales. Large samples of subjects took the preliminary forms of
the questionnaire, and by 1962 Form F of the MBTI was made available
through the Educational Testing Service. This version was published
for research purposes only, and consisted of a total of 166 questions,
many of which would not enter in the determination of the ultimate
profile but, instead, were part of the questionnaire for use in future
By 1975 the instrument had been tested widely enough to warrant
publication of the final form of the instrument for professional
use. In 1977, however, a shorter form (Form G), consisting of 126
questions was developed and instituted. This form eliminated a
great many of the research items presented in Form F.
The MBTI derives from Jung's three basic variables, namely,
extraversion-introversion (EI), sensing-intuition (SN), and thinking-
feeling (TF). To these variables is added a fourth one, implicit in
Jung's theory but only made explicit by Isabel Myers, judgment-
According to the Manual, the distinctions among the four are as follows:
Index Preference as between Affects individual's choices as to
El Extraversion or Intro- Whether to direct perception and
version judgment upon environment or world
SN Sensing or Intuition Which of these two kinds of per-
ception to rely on.
TF Thinking or Feeling Which of these two kinds of
judgment to rely on.
JP Judgment or Perception Whether to use judging or per-
ceptive functions for dealing with
These four indices, explains Myers (1962), are indices of direc-
tion of preferences and not scales to measure traits. The dynamic
interrelationship of the four indices result in a total of 16 Jungian
types, which Myers described (Appendix A). The items of each index
offer "forced" choices involving the preference at issue. Responses
pointing in opposite directions bear separate weights of 0, 1 or 2,
enabling the evidence in each direction to be separately summed. This
device permits (a) control of the effect of omission, and (b) an
item-by-item correction for social desirability, undistorted by
Persons with more points for E than for I, for example, are
classified as extraverts. MBTI scoring transforms points for each
preference into a preference score. The preference score is made
up of a letter to denote direction of the preference and a number
to denote strength of preference. For example, E29, N15, etc.
Since the score is always the difference between the two directions,
the result will always be shown in terms of one or the other, not
both. Preference scores may be converted by a linear transformation
into continuous scores which run in both directions from the center.
For preferences of E, S, T or J, the numerical portion of the pre-
ference score is subtracted from 100; for I, N, F or P the numerical
portion is added to 100. Thus, E 29 is expressed as a continuous
score of 71; N 15 is expressed as a continuous score of 115. Myers
(1962) attempted to make the division points on each preference as
accurate as possible in order to a) classify people with maximum
accuracy, and b) using these classifications, to obtain evidence
as to whether the preferences are indeed dichotomous in nature (Myers,
1962, p. 89).
Intercorrelations of MBTI Scores
The intercorrelation of the type categories as well as of the
continuous scores have been explored. The results of these studies
indicate that the type categories as well as the continuous scores
are independent of each other with the exception of the S-N and J-P
indices which tend to be correlated (Appendix B) (Carlyn, 1977;
McCaulley, 1978; Myers, 1962; Stricker & Ross, 1963; Webb, 1964).
As was the case with the studies of intercorrelation, the relia-
bility studies regarding the Indicator have also been conducted per-
taining to the type-category as well as the continuous scores. In
separate studies, Myers (1962) and Webb (1964) estimated split-half
reliabilities by calculating phi correlation coefficients and
applying the Spearman-Brown formula. The shortcoming of this type
of research, however, lies in the fact that it is known for under-
estimating factors fairly frequently (Carlyn, 1977).
The recommendation of the author (Myers, 1962) was to calculate
tetrachoric correlation coefficients and estimate split-half relia-
bilities based on those figures. The results would then be used in
the Spearman-Brown prophecy formula. The use of the tetrachoric
correlation assumes a normal distribution of scores. The distribution
of the Indicator scores, however, tend to be platykurtic (Myers, 1962).
A compromise between these two estimates has been achieved, and for
any given sample it is reasonably certain that the actual category-
type reliabilities will fall somewhere between the lower underestimated
estimates derived with the phi coefficients, and the higher often over-
estimated estimates derived by means of the tetraphoric r.
Studies conducted by Hoffman (1974), Myers (1962), and Webb
(1964), have reported phi coefficients ranging from .55 to .65
(E-I), .64 to .73 (S-N), .43 to .75 (T-F), and .58 to .84 (J-P).
The reports of the tetrachoric r system for determining the internal
consistency figures yield coefficients of .70 to .81 (E-I), .82 to
.92 (S-N), .66 to .90 (T-F), and .76 to .84 (J-P) (Carlyn, 1977).
The research to determine the internal consistency of the continuous
scores has been less diversified and because of this, has yielded
ranges of scores which are not as broad. Myers (1962) developed
a split-half procedure involving Pearson product-moment correlations;
Webb (1964) used a similar procedure in his studies and Stricker
and Ross (1963) used Cronbach's Coefficient Alpha. The ranges
resulting from these studies were: .76 to .82 (EI), .75 to .87 (SN),
.69 to .86 (TF) and .80 to .84 (JP).
Several methods have been explored to help explain the internal
consistency of type-category scores. Stricker and Ross (1963) and
Webb (1964) computed lower-bound estimates of reliability. These
estimates, however, do not have much meaning to further researchers
without corresponding upper limits. Test-retest data were computed
by Levy, Murphy and Carlson (1972), Stalcup (1968), and Stricker and
Ross (1964a) using college students as subjects, and by Wright (1966)
using a sample of elementary school teachers. The results of these
studies were presented as percentage of agreement figures and ranged
from a low of 0% change in 4 areas to a high of 90% 'no change' in
T-F and J-P typing in the Wright study. These last figures may help
confirm the theory that type seems to become more stable as a person
grows older and becomes stabilized in his/her occupation (Carlyn,
1977; McCaulley, 1978).
The continuous scores have also yielded high correlation
coefficients when testing for stability. In studies by Levy, Murphy
and Carlson (1972), and Stricker and Ross (1964a), these coefficients
ranged from .48 on the T-F scale to .83 on the E-I scale. Additional
recent research results are presented in Appendix B. These results
are all consistent with the results of earlier research on the subject.
Research has also been conducted to determine the correlations
between type and scholastic achievement. The results of this re-
search indicated that there were low correlatons between type and
cases where the subject being tested was an underachiever (Myers, 1962).
The purpose of the research done in this area has been to
determine how well the MBTI measures its intended target, namely,
Jung's typology. The most extensive research done to date on the
content validity of the Indicator was done by Myers (1962) in de-
vising the instrument. Stricker and Ross (1964a) concluded that the
S-N and T-F scales seem largely consistent with their conceptual
definitions but the E-I and J-P scales may not be consistent with
what Myers postulated in the manual.
Self-report studies were conducted by Bradway (1964) using 28
Jungian analysts as her subjects. In her study Bradway asked the
subjects to classify themselves in each of three dichotomies: E-I,
S-N, and T-F. Comparisons were taken made with MBTI typing. There
was a remarkable 100% agreement on E-I typing, 68% agreement on
S-N, 61% agreement on the T-F scale, and 43% agreement on all
In another study, Stricker and Ross (1964b) compared continuous
scores on the MBTI with scores obtained on the Gray-Wheelwright
Questionnaire, another measure of Jungian type. The sample for
this study was comprised of 47 male college students. The resulting
correlations were as follows: E-I scales exhibited a .79 correlation,
S-N scales were .58, and the T-F scales were .60. The Gray-Wheelwright
does not recognize the J-P dichotomy.
The ability to predict major choice in college with success has
been researched in several studies. In a study in 1967, Goldschmid
was able to predict college majors for two groups of undergraduates,
finding that the Indicator scales had moderate predictive validity.
Using a sample of 1709 entering freshmen, Conary (1966) speculated
that certain personality types within his sample would be more likely
to choose certain college curricula in their freshman year and would
be more likely than certain other types to make good grades. These
predictions were later substantiated.
In an attempt to measure its construct validity, the MBTI has
been correlated with the constructs of a variety of other instruments.
Saunders (1960) made a comparison between the constructs of the
Indicator and those of the Allport-Vernon-Lindzey Study of Values
(AVL), another instrument concerned with the measurement of types.
Using factor analysis the results indicated that both instruments
were measuring similar constructs. In further research, in a study
conducted with 507 freshmen engineering students at Cornell, and using
the Personality Research Inventory for comparative purposes, all
25 scales of the PRI correlated significantly with one or more of the
type indices, with 49 of the r's being significant at the .01 level
Extensive research has been conducted with the MBTI in order to
determine preference in aptitude, achievement and career choice
and the manner in which these preferences correlate with the scales
of other instruments, such as the 16 Personality Factor Test, the
Omnibus Personality Inventory, the Opinion, Attitude and Interest
Survey, to cite a few. The Myers Longitudinal Medical Sample was
conducted over a 12-year period and tested the predictive validity
of the instrument by offering the MBTI, a second time, to a group
of people who had taken the Indicator twelve years earlier, as
freshmen in medical school. The results of these studies are shown
in Appendix C (McCaulley, 1978).
Summarizing, the levels of internal consistency and test-retest
reliability have proven adequate in the MBTI. The research associated
with its validity indicates that the instrument is significantly
correlated regarding type preferences and those characteristics
which are similarly measured by other instruments; there is also
significance in the correlation of type with career choices, indica-
ting that certain types seem better suited for certain careers.
The MBTI is becoming widely accepted as a helping tool in coun-
seling. It is an aid for those who work in the helping professions
to better understand the direction and many facets of their client's
life. The self-report questionnaire helps people understand them-
selves better as well. It also promotes a better understanding of
a person's relationship to others, thus leading to increased well-
being and advancement in daily activities. The differences which
the MBTI measures are those found in all persons, and it does not
concern itself with pathological issues.
The reasons for choosing the MBTI as the instrument to be trans-
lated in this study are both theoretical and practical. In his theory,
Jung postulated that these preferences are universal in nature. As
such, they are found in all cultures. The MBTI, as a tool which
measures these universal characteristics, has been found to have a
wide range of practical applications which are of value in this
culture. When faced with the problem and challenge of helping
Spanish-American immigrants find a new beginning in the United States,
the Spanish version of the instrument should be of special value.
In summary, a number of theories of translation have been explored
prior to embarking on this study. Several methods of translation
were then analyzed as a way to help determine which ones would be
relevant for this study. An analysis of culture-fairness followed
as a way of illustrating some of the cultural nuances which could
possibly affect the outcome of the study. The final section of the
chapter dealt with an overview of the Myers-Briggs Type Indicator
with special emphasis being placed on its validity and reliability
and the reasons it was chosen as the instrument to be translated
in this particular study.
The research questions will be presented in the first section
of this Chapter. In the second section the three phases that comprised
the study will be presented, namely, the translation, the administration
which will include a description of the sample, and the data analysis.
1. Is the Spanish translation of the MBTI an adequate representation
of the original English version?
a. Are intercorrelations of continuous scores for EI, SN, TF
and JP preferences comparable to those reported for the
b. Are the statistics for the English-Spanish administration
comparable to test-retest data in the literature for
English-English testing, in
1) Percentage of agreement in assignment to categories
EI, SN, TF or JP?
2) Percentage of persons agreeing on 4, 3, 2, 1, or 0
2. Are modifications in the psychometrics of the MBTI necessary
for the Spanish version to represent the intent of the English
version? Specifically, how do item weights for responses of the
subjects of this study reporting both English and Spanish compare
to the published item weights of the English version?
3. Are differences between responses to the English and Spanish
forms in the directions expected from actual differences in
the two cultures?
The study was conducted in three phases. Phase I included the
development and pilot study of the Spanish form of the MBTI, Form G;
Phase II consisted of the administration of a reading sample, followed
by the administration of the MBTI, Form G, in English and Spanish;
and Phase III consisted of scoring and analyzing the items as well
as the resulting data.
Phase I: The Translation
Phase I included the development of a translation of the MBTI
which took into consideration not only the language, but also the
cultural and social values of Spanish-speaking cultures. For the
purposes of this study, the English version of the Myers-Briggs
Type Indicator, Form G, was used. This form was chosen over the
older Form F because it is shorter, newer, and the only differences
between the two instruments are to be found in the unscored items,
i.e., those that are included in the test strictly for research
purposes, and the order in which the items are presented. (Permission
was secured from Consulting Psychologists Press in Palo Alto, CA,
to develop a translation of Form G of the instrument, Appendix D).
A first consideration in completing the translation was the form
of address to be used in the Spanish form. In the Spanish language
"you" can be interpreted as "tii", in the case of the informal way
of addressing someone, or as "usted" in the event the approach is
more formal. For the sake of eliminating a possible insult or dis-
comfort to elders in a Spanish culture, the formal "usted" was used
for the purposes of this study.
A second consideration taken into account was that within the
Spanish language there are many dialects. Most, if not all, of the
dialects share similarities with the other dialects. For the purposes
of this study a concentrated effort was made to use the Latin American
dialects, avoiding regionalisms which may be typical of any group.
A third consideration taken into account was the use of sexist
language. In English, the adjectives used in the word pairs section
of the MBTI are non-sexist. In the translation, however, this feature
is lost, and in Spanish, most adjectives will have either an 'o' or
an 'a' ending depending on whether the noun they modify is masculine
or feminine. For the purposes of this study, an accepted rule in
Spanish grammar was used, namely, that in such situations, the male
gender will apply.
A fourth consideration came from the fact that responses to the
questions are the result of self-evaluations. Due to the subjectivity
of the questions, the responses encompass not only the general factors
to be considered across cultures, but also the factors which apply
to the way each person reacts to those elements. For example, regardless
of a person's nationality--Spanish, Portuguese, Italian, German, French,
or English--the reaction to a greeting or salutation is usually a
corresponding 'hello' in the appropriate language. As the salutation
and reaction are merely conventional responses and the response
itself is fairly universal in nature, the response is not affected
by the person's emotions to any great extent.
But what happens to people, internally, when they hear the
word 'affection'? The reaction to this word is a very subjective
one. In some cultures, 'affection' is openly discussed; in others,
it is something which is only discussed under a given set of
circumstances; and in still other cultures, it is a word which
men avoid lest their masculinity come under attack.
In the first step of Phase I, the researcher made an informal
linguistic translation of the MBTI and the translation was self-
administered. It was discovered that there were many linguistic
and cultural nuances which seemed to have been overlooked. Secondly,
the translated form was given to two bilingual subjects for their
review. The same questions arose again concerning the interpre-
tation of some of the test items. It became apparent that the
translation was merely a linguistic one and did not measure the
feelings and emotions intended by the authors of the instrument.
At this point it was determined that further analysis of the
instrument and its intent would be in order.
Phase I included the development of a translation of the MBTI
which took into consideration not only the language, but also the
cultural and social values of Spanish-speaking cultures.
The next step consisted of an analysis of the English version
of the instrument. The purpose of this was to determine the di-
chotomies which the author of the Indicator intended to tap in
each of the 126 questionnaire items. This facilitated the selection
of the Spanish word/words which approximated the same dichotomy in
the Spanish language.
Then, the Indicator was translated into Spanish, with a deliberate
attempt being made to preserve the format of the English version as
much as possible. The original draft of the Spanish form incorporated
as many word choices as possible for those items where there might
be more than one choice in the translation. On completion, the
translated version was compared to the English version. Based upon
the dichotomy being tapped in the original form, the word or words
in Spanish which best tapped that dichotomy, in Spanish, were chosen.
Next, the Spanish version was presented to a linguist. The
purpose of this was to determine the reading level of the new version
In the following step, the Spanish version was presented to
three bilingual people for any reactions that they may have had to
the choice of words, the nature of the questions, and to determine
if there were any questions that may have been offensive, because of
cultural differences. The people chosen for this step were adults
who were not familiar with the Indicator, who were fluent in both
languages, and who were from three different Spanish-speaking cultures.
An informal interview followed, the purpose of which was to
determine the person's affective reaction to some of the questions
that could have been seen as being culturally laden or which could
have been interpreted in more than one way.
The completion of this step led to additional necessary refine-
ments. As an example, a concentrated effort was made to create a
more informal tone in some of the items which one of the interviewees
indicated she found to be very personal. Some other suggestions could
not be followed for they would have involved the use of regionalisms
of the culture in question, something the author has made a deliberate
attempt to avoid.
Once these refinements were made, the Spanish version was sent
to five bilingual people in five different Spanish-speaking cultures:
Puerto Rico, Cuba, Uruguay, Mexico, and Peri. These people were
asked to back-translate the Spanish version into the English language
of the original version of the MBTI. The back-translations were then
compiled and further revisions were made on the basis of the comments
received. Once these refinements were made, the Spanish form was
ready for reproduction.
Phase II: The Administration
The population for this study was a sample of bilingual subjects, who
had a minimum of a sixth grade reading level in both English and Spanish.
The length of stay in the United States for the members of the sample
varied, but a concentrated effort was made to obtain representative
samples from diversified cultural and socioeconomic backgrounds as
well as a variety of residence time in this country. The sample also
consisted of members of several Hispanic subgroups. The sample was
obtained from several resource areas: Hyattsville, MD; Orlando, FL;
West Palm Beach, FL; Jacksonville, FL; and Gainesville, FL.
Groups of bilingual subjects were gathered with the assistance
of key people who were contacted in order to help with the project.
The sample consisted of 209 subjects who met the outlined criteria
The second phase of the study included the administration of a
sample reading passage, at approximately the sixth grade reading
level in English and in Spanish in order to assess reading competency.
Next, both forms of the MBTI were administered to the sample population.
Administration of the reading evaluation and of the MBTI was
done in an alternate fashion, namely, giving the English form first
to half the sample, and last to the second half and performing the
same procedure with the Spanish form. In order to avoid placing any
of the subjects in an embarrassing situation, the English and
Spanish forms of the MBTI were administered to all the subjects
and only at the time of scoring and recording the data were those
subjects eliminated from the study who did not achieve the required
sixth grade reading level on either or both the reading tests.
A letter was sent to the directors of prospective organizations
from which samples for the study were obtained (Appendix E). In the
case of individuals, a letter or telephone contact was made in order
to document the subject's willingness to participate in the study.
A packet of the tests was given or mailed to each participant.
Each packet consisted of a brief cover letter (Appendix F), a demo-
graphic questionnaire, and an English and Spanish form of the
reading sample and the Myers-Briggs Type Indicator, Form G. In addition,
for those being mailed, a stamped, self-addressed envelope was
included for the return of the materials to the researcher.
Collection of data
Because all directions were given on the instruments, the ad-
ministration of the forms did not require supervision. Demographic
data such as race, sex, age, and time in the United States were
obtained in the form of a short questionnaire that was stapled to
the front page of the instrument. These forms were numbered consecu-
tively and corresponded with the numbers assigned to each packet,
thus assuring confidentiality after the questionnaires had been ad-
ministered. In order to alternate the administration format, the
odd numbered packets had the English form of the reading sample first,
followed by the Spanish form. Next was the English form of the MBTI,
ending with the Spanish form of the MBTI. The even numbered packets
began with the Spanish form of the reading sample, followed by the
English form of the reading sample. Next was the Spanish form of the
MBTI, ending with the English form of the Indicator. The accompanying
instructions encouraged the participants to take the instruments in
the order in which they were presented.
The reading samples which were given took no more than 20 minutes
to complete. The two forms of the Indicator could be completed in
approximately one hour. The completed forms were computer-scored
at the Center for Applications of Psychological Type, Inc., in
Prior to beginning Phase III of the study, the reading samples
were scored. Those who had not scored at least at the 6th grade reading
level on both reading samples, were eliminated from the study. In
order for a person to have passed this reading test, they must have
scored correctly on three out of the four questions being presented
at the end of each reading sample.
Phase III: Data Analysis
The next step consisted in scoring the MBTI. An optical scanning
procedure was used by a recognized scoring program at the Center for
Application for Psychological Type, Inc., in Gainesville, Florida.
In the next step, using a correlation matrix of all recorded
scores, the full scores as well as the scores of the phrase questions
and word pairs were analyzed. This procedure yielded the continuous
scores, but also helped analyze those scores with the greatest
similarities and differences. This procedure was done for the total
sample, as well as for males and females, separately.
The next step in Phase III was a tabulation of the numbers and
percentage of the subjects remaining the same in their reports of
El, SN, TF, and JP for English and Spanish. Closely associated,
the next step was a tabulation of numbers and percentages of the
sample which remained the same on all four, three, two, one, or
no preferences in the English and Spanish forms. This tabulation
also broke down the changes by strength of preference.
In the final step, an item analysis was conducted, first on the
English form, then on the Spanish form. The formula used by Myers (1962)
was applied to both forms. This formula takes into account social
desirability of items and their ability to predict total score for
each preference. The item analysis was done separately for males and
females, and for A and B answers on each item. The goal of the item
analysis was to lead to an index of prediction for each response.
The items which resulted in a O, 1 or 2 weight were then
compared with Myers' published scoring keys for Americans taking the
questionnaire in English, i.e., the normative population. This indicated
whether the sample bilingual population were answering the questions
in the same way as the norm.
The final step consisted of a comparison of those items where
there had been a shift in the item weights between the two versions.
This was conducted in an attempt to determine whether any cultural
factors or differences in social desirability between the two cultures
could have affected the results.
ANALYSIS OF THE RESULTS
The purpose of this study was to develop and validate a Spanish
version of an English language personality questionnaire that would
remain faithful to the original version, while at the same time
taking into account cultural differences between the populations in
question. More specifically, the study was an attempt to develop
a Spanish version of the Myers-Briggs Type Indicator (Form G) that
would consider the intent of the authors of the English form and of
Jung's theory of personality. The development of the Spanish version
of the MBTI which was used in this study was conducted in several
Initially, a translation was done by the researcher. It was
noted, however, during a self-administration of the questionnaire,
that the questionnaire being analyzed was simply a linguistic trans-
lation and nothing else.
Several revisions were made to the instrument, each one keeping
in mind the target population of the translation as well as the
source from where it was being derived.It was then sent to three
bilingual people who were asked to read it and react to it, i.e.,
to the degree of intimacy incurred in any.of the questions, any
word or words which were either offensive or unclear, etc.
Informal interviews were then conducted with the three people,
and based on their suggestions and recommendations, another revision
was completed. The revision was then sent to five bilingual people
from five different Spanish cultures. They were asked to back-translate
the instrument into English, and were given the freedom to add any
suggestions they would like to make to improve the way in which the
ideas were being conveyed. Once the back-translations were received,
another revision took place and the final instrument was reproduced
and made ready for sampling.
Table 1 presents the distribution of the sample for this study
which was comprised of 209 bilingual men and women who had successfully
completed a reading test which determined their ability to read and
understand the English and Spanish languages at a sixth grade level
N = 209
Sex N %
Males 84 40.19
Females 125 59.81
Country of Origin
Central American Countries
South American Countries
Time Lived in the United States
10 years or less
Over 20 years
Occupation N %
Student (High School) 44 21.1
Student (College) 41 19.6
Blue collar worker 13 6.2
White collar/Professional 66 31.6
Housewife 13 6.2
Self-employed 1 0.5
Unemployed 2 1.0
Clerical worker 16 7.7
Retired 7 3.3
Not given 6 2.9
The type distribution that resulted from this sample is presented in
An analysis of this distribution indicates that the number of
extraverts and introverts in the sample was almost equal, a fact
which is common among the college population of the United States.
The Sensing/Intuitive ratio for the sample is approximately 2:1
which is slightly below the expected ratio in the United States (3:1).
The Thinking/Feeling ratio was also approximately 2:1. There was an
unusually large number of Thinking females in the sample, something
which is not usually expected in the sample distributions in the United
States. The Judging/Perceiving ratio was approximately 55/45, which
is typical of the distribution of American samples.
Research Question 1
Is the Spanish translation of the MBTI an adequate representation
of the original English version?
a. Are intercorrelations of continuous scores for EI, SN,
TF and JP preferences comparable to those reported for the
Table 2 indicates the intercorrelations obtained when the English and
Spanish versions of the MBTI, Form G, were administered to 209 adults.
INTERCORRELATIONS OF CONTINUOUS SCORES (MBTI FORM G)
EI/SN EI/TF EI/JP SN/TF SN/JP TF/JP
English -.17 -.19 -.10 .17 .56 .31
Spanish -.22 -.20 -.13 .14 .54 .25
English -.10 -.13 -.14 .03 .54 .33
Spanish -.08 -.08 -.09 .11 .54 .37
English -.23 -.19 -.00 .21 .43 .30
Spanish -.32 -.25 -.08 .19 .38 .18
Table 3 presents the intercorrelations for data compiled in the
literature. These data were made available by the Center for Applications
of Psychological Type. They include 21 samples from the MBTI Data Bank;
15 samples of 8-12th grade students; 14 samples of college students,
and-10 samples of students in professional programs.
INTERCORRELATIONS OF CONTINUOUS SCORES (MBTI, FORM G)
EI/SN EI/TF EI/JP SN/TF SN/JP TF/JP
Low -.27 -.26 -.21 -.20 -.03 -.15
Median -.03 -.04 .01 .06 .30 .14
High .17 .20 .18 .25 .60 .34
MBTI DATA BANK
Males -.03 -.03 .02 .08 .20 .15
Females -.05 -.02 -.03 .01 .28 .11
b. Are the statistics for the English-Spanish administration
comparable to test-retest data in the literature for English-English
1) Percentage of agreement in assignment to categories EI,
SN, TF or JP?
2) Percentage of persons agreeing on 4, 3, 2, 1, or 0 preferences?
For purposes of comparison the administration of the second form
(the Spanish form in half of the questionnaire packets, and the
English form in the alternating half) was considered the retest situa-
tion in this study. The scoring of these questionnaires was based
on the 1977 weights, the ones presently in use with the English form.
Table 4 illustrates the results obtained in this study. Table 5 pre-
sents comparative data from the literature for changes in test-retest
situations. These data were compiled and supplied by the Center for
Applications of Psychological Type in Gainesville, Florida.
In the present study the T-F scales showed the greatest number
of changes (12%). This was also the index where 1977 weights changed
toward T in comparison with 1962 weights.
The category labeled "All 4 Preferences the Same" shows that 157
of the subjects or 75% of the sample taking the Indicator did not
deviate in the retest situation in any of the four dichotomies. It was
suspected that it was possible that some of the subjects had simply
duplicated the answers onto the second answer sheet and, therefore,
had only answered the questionnaire once. An analysis was done of
the continuous scores of all those subjects who were in that category.
This analysis showed that most of the scores were not alike, as had
been suspected, but had simply not shifted enough to change the resulting
Table 6 illustrates the correlations of the continuous scores
between the English and Spanish versions of the MBTI. The correlations
were obtained for the total sample, as well as for males and females
COMPARISON OF ORIGINAL AND RETEST PREFERENCE CATEGORIES
(ENGLISH/SPANISH USING 1977 PUBLISHED WEIGHTS)
Agreement on E-I Typing
Agreement on S-N Typing
Agreement on T-F Typing
Agreement on J-P Typing
All 4 Preferences the Same
Change in 1 Preference
Change in 2 Preferences
Change in 3 Preferences
Change in 4 Preferences
COMPARISON OF ORIGINAL AND RETEST
RANGE OF COMPARISONS REPORTED
Agreement on E-I Typing
Agreement on S-N Typing
Agreement on T-F Typing
Agreement on J-P Typing
All 4 Preferences the Same
Change in 1 Preference
Change in 2 Preferences
Change in 3 Preferences
Change in 4 Preferences
IN THE LITERATURE
CORRELATIONS OF CONTINUOUS SCORES
ENGLISH/SPANISH VERSIONS OF MBTI
Research Question 2
Are modifications in the psychometrics of the MBTI necessary
for the Spanish version to represent the intent of the English version?
In all but one instance the percentages of agreement were higher than
those reported in the literature. Therefore, no modifications were
The internal consistency of the instrument was determined by
the split-half method, and was within acceptable limits. The results
of that analysis are presented in Table 7.
INTERNAL CONSISTENCY OF CONTINUOUS SCORES BASED ON PRODUCT
MOMENT CORRELATIONS OF SPLIT-HALF CONTINUOUS SCORES WITH
SPEARMAN-BROWN PROPHECY FORMULA CORRECTIONS
Sex N El SN
English Form M 84 .81 .84
F 125 .74 .83
Total 209 .77 .83
Spanish Form M 84 .83 .83
F 125 .79 .81
Total 209 .81 .81
COMPARATIVE SAMPLES FROM THE
PROVIDED BY THE CENTER FOR APPLICATIONS
15 High School M Low .43 .51
Samples High .85 .86
F Low .70 .41
High .87 .85
10 College M Low .56 .75
Samples High .87 .90
F Low .78 .80
High .86 .87
OF PSYCHOLOGICAL TYPE
A comparison of the split-half scores between the results
obtained in the study with the results as shown in the literature
indicate that on the Spanish administration of the instrument all
the split-half correlations were in accord with those reported in
the literature. Most of these correlations approximated the highest
recorded correlation in the literature. In the English administration
there was only one category that was not in accord with the literature.
Research Question 3
How do item weights for responses of the subjects of this study
reporting both English and Spanish compare to the published item
weights of the English version?
One of the elements that must be considered in weighting responses
in this instrument is expected social response in any culture. The
social desirability of any question can vary greatly from culture to
culture. MBTI item weights were designed to take social desirability
into account. For this reason, item weights for the translation
were compared with published item weights.
All the items of the Indicator are forced choice questions. The
item analyses are separate for males and females, and each choice is
analyzed separately. During the development of the instrument, each
choice discriminated. As cultural changes have affected social desira-
bility of choices, some choices on the Indicator do not carry weights
at the present time.
The prediction formula used to calculate item weights leads to
a decrease in weights if there are too many subjects who answer a
question which was not intended for their preference, e.g., if too
many E's give the I answer to a question, the item weight is decreased,
making the response less discriminating between E's and I's.
In this study, some answers which are not now weighted in the
current scoring for the English version discriminated at a 1-point or
even a 2-point level. Some 1-point items discriminated at the 2-point
level in the Spanish version, and some 2-point items discriminated at the
1-point level or not al all. These distinctions are presented in Table 8.
COMPARISON OF SPANISH VERSION WEIGHTS TO PUBLISHED ENGLISH WEIGHTS
1 2 3 4
N % N % N %
El 41 22 54 8 20 11 27
SN 51 27 53 15 29 9 18
TF 44 22 50 11 25 11 25
JP 48 21 44 21 44 6 13
EI 41 19 46 13 32 9 22
SN 51 28 55 14 27 9 18
TF 45 24 53 11 24 10 22
JP 48 30 63 14 29 4 8
EI 82 41 50 31 26 20 24
SN 102 55 54 29 28 18 18
TF 89 46 52 22 25 21 24
JP 96 51 53 35 36 10 10
Column 1 Number of possible responses
Column 2 Item weights equal to published weights
Column 3 Item weights greater than published weights
Column 4 Item weights less than published weights
A review of Table 8 indicates that in the great majority of the
cases, the weights which resulted in the administration of the
Spanish version in this study were as discriminating, if not more so,
than those weights in the literature for the English version. This
statement is reflected in an analysis of Columns 2 and 3 of the table.
Column 4 represents the percentage of items which are less discriminating
in the Spanish version than in the original English form.
Research Question 4
Are differences between responses to the English and Spanish
forms in the directions expected from actual differences in the two
A review of the previous data indicates that in this study items
discriminated, as in English, on all four scales and for both sexes.
However, there were a number of items which discriminated better in
this study while others did not discriminate quite so well as the
original English version. For approximately half of the total sample
the published weights as they are being used at the present time with
the English version are consistent with those found in the study for
the Spanish-speaking subjects.
On the J-P scale there seems to be an indication that the English
weights are much more discriminating than are the obtained ones in the
Spanish version of the questionnaire. One possible explanation for
this could be that Latin Americans living in this country are either
striving for more order in their lives or are being forced into this
form of behavior in this society. This would hold especially true
for those who have been born and raised outside the United States and
have recently come to this country to live.
There have also been differences on the T-F scales in this study.
This may be due to the fact that there were an unusually large number
of thinking types among the females in the sample. In order to
either substantiate or negate this premise, the study should be re-
plicated with a larger sample. Also, Latin American men find it very
difficult to either discuss or admit to personal feelings. This may
have been a factor in choosing the opposite choice as a way of
avoiding dealing with the issue while they were answering the question-
There was also a shift of the E-I scale in the sense that the
English weights were not discriminating on the Spanish form in 21 out
of the possible'82 responses on that scale. According to the statistics
that have been presented, the predominant way to live in the United
States is toward the extravertive end of the scale. It is conceivable,
from a cultural standpoint, that many Latin Americans who have come
to this country have not totally adapted to this preferred way of
In summarizing, the percentage of agreement of preference scores
as well as the percentage of agreement of types were high. The internal
consistency scores are congruent with what is indicated in the literature
for the English version; and the test-retest data are equal if not
better, in some cases, that what is reported in the literature.
As a result, it is felt that the translation is indeed a valid
one which, when administered and scored using existing keys, would
yield fairly accurate results. It is recommended, once again, that
the study be replicated with a larger sample in order to make even
finer discrimination which may in turn lead to a different set of
weights for the Spanish version.
DISCUSSION, LIMITATIONS, AND CONCLUSIONS
The purpose of this study was to develop a Spanish version of
an English language questionnaire that would remain faithful to the
original version, while at the same time taking into account cul-
tural differences between the populations in question.
In this study and due, possibly, to the universality of the
original instrument, the weights which are presently in effect for
the original English version of the MBTI are fairly accurate in
the determination of type in the Spanish version. There are some
discriminating factors, however, between the two instruments. Because
of differences in social desirability between the two cultures, re-
weighting the scales to account for this in Spanish cultures should
yield even stronger correlations between the two instruments.
In addition to high intercorrelations, there is added support
to the universality of what is being assessed. This is based on the
fact that the split-half reliabilities obtained in this study are as
high, if not higher, than those offered in the literature on the
The ideas, feelings, and emotions being tapped are consistent
with the original form. Therefore, the same questions are being
posed to both populations. The discrepancies being presented as a
result of the research lie in the weight of the items themselves.
It is in this area that culture is playing a most important role. As
has been noted earlier, several of the "feeling" questions are being
responded to consistently and in accord with the present published
weights by the females in the sample, but in the case of the males
there seem to be elements of the "machismo" in Latin American countries
which are affecting the weights of those same responses by the men
in the sample.
In other instances, especially on the El scale, there seems to
be an indication that because of an incomplete acculturation process,
some of the native Spanish-speaking subjects are tending toward an
introverted way of thinking in a culture alien to their own, while
showing tendencies toward extraversion when assessed within their
Limitations of the Study
There are diverse cultures in Spanish-speaking countries. The
research sample for this study did not include large enough numbers
in each culture to enable the researcher to arrive at any definite
conclusions about each of the cultures in question. The next step
in this study would be to enlarge the size and diversification of
the samples within each subculture. In other words, it would be
necessary to increase the sample in each country being assessed in
order to arrive at any definite conclusions about that particular
country. In doing this, a conscious attempt should be made to con-
tinue in the heterogeneous vein of the original study sothat
people from as many walks of life as possible can be represented in
Secondly, and as a result of the item analysis of the first
study, refinements in the instrument should be considered. These
refinements could be, primarily, in two areas.
In the first, a careful analysis of those items where the
weights are in discrepancy with the published weights in the original
version should be studied. The purpose of this will be to identify
any need for improvements in the translation.
The second type of refinement to be considered would be in the
realm of the weights themselves. The discriminatory weights may
indeed be different in Spanish cultures in which case a new scale
of weights may be indicated for the Spanish version. Furthermore, it
may be discovered that within the Spanish version there may be a
need for several weighting systems, depending on the subculture in
question. By increasing the size of the sample the determination
could be made as to whether the weights of the original instrument
must be changed or not. Also, and due to the diversity of subcultures,
further research is encouraged in order to determine whether norms
for members of each subculture living in this country, as opposed
to their own, are necessary.
Another recommendation is that the study be conducted using a
more informal "tl" form of the Spanish version. This would indicate
if the formal approach of the original Spanish form used in the study
had any effects on the end results.
Additionally, while the above mentioned studies are taking place,
the translation of accompanying literature should be undertaken. In
addition to the MBTI Manual, translations of Myers' "Introduction
to Type," as well as her more recent book, "Gifts Differing," would
be valuable to the user of the Spanish version of the questionnaire.
This study has helped prove that although the languages and
cultures of the world are indeed different, there are still some
very basic elements of being human which people from all walks of
life share. The abilities to love, fear, hate, feel compassion, and
perceive, in general, are common to us all. Language is the bridge
between those elements of the human mind and how the people in the
environment react to them. The importance, therefore, lies not in
the fact that these are universal elements, but rather in how they
are expressed and perceived so that the underlying message of the
emotion can be perceived by the recipient in the manner intended
by the provider.
The importance of the clarity of the message's intent can be
seen in many areas. In assessments, for example, a misinterpretation
of an intended message could lead to a missed diagnosis of a problem,
barring the client from receiving proper services. In counseling,
if the therapist cannot empathize with the feeling behind the message
being conveyed by clients, it can be very difficult to help them deal
with their problems. On a more basic level, individuals must be in
tune to their own feelings so that they can be in a better position
to act and react to those around them.
It cannot be stressed enough that what is important are not
the words conveying the message, but the conglomerate of feelings
that led to the development of those words. It is only through a
deeper, more universal system of analyzing man's behavior, that
language may someday be seen as a more complete way of describing
ourselves and what we are feeling to those around us.
It is hoped that through this study the reader can now feel
that with the proper choice of words messages can be perceived in
the manner in which they were intended. This is why translations must
be seen in a much broader scope--it is necessary to go below the
surface of words into the deeper, more complex realm of meaning
before complete and accurate translations can ever be achieved.
, .- Ih I-
C C- h0I Cii
* "* C s i
a. 4.'aC 1
a -a 4.
C ~' -J a C -
2 as 0.34.
- L 0
INTERCORRELATIONS OF MBTI TYPE CATEGORIES
INTERCORRELATIONS OF MBTI TYPE CATEGORIES
MBTI Type Categories
EI/SN EI/TF EI/JP SN/TF SN/JP TF/JP
422 Freshmen at
Emory (Webb, 1964) 0.04 -0.03 -.01 -.05 .99** .08
300 Male Freshmen at
L.I.U. (Stricker & Ross,
1963) .08 -0.04 .14* .07 .23** .09
184 Female Freshmen at
(Stricker & Ross, 1963) .03 .05 .13 .02 .31** .05
NOTE: All reported correlation coefficients in this table are phi
coefficients, which tend to yield low estimates of true values
PRODUCT-MOMENT CORRELATIONS OF MBTI CONTINUOUS SCORES
AND OTHER PERSONALITY TESTS IN MEDICAL STUDENT SAMPLES
Product-Moment Correlations of MBTI Continuous Scores and Oth
SCALE GROUP N
16 PERSONALITY F
C. Stable, faces reality
(ligh rgo Streith)
G. Proper, Rule-Bound
N. Shrewd, Worldly
er Personality Tests in Medical Student Samples
El SN TF JP STRONGEST
-18 -24 01 -22 -
-43"* -07 -05 -13 E -
-22" -01 31"' -02 E F
-40"* -08 16"* -18" E F J
-35"* -01 19"* -07 E F
04 35" 22 19 N -
01 25" -02 -05 N
-01 10 -07 01 -
05 14" 08 03 .- I
00 10* -01 02 -
-30* -22 -12 -14 .
-23" -17 -18 -11 E -
-32-* -16* 00 -12 E -
-28-" -11 -06 -08 E -
-31*' 02 -06 09* E -
-18 -18 -18 22 -
-08 25" -06 32*" N P
-12 25" -06 10 N -
-21"* 23** -06 20"* E N P
-32"' 29"' -18"* 18*** E N F P
-30* 09 03 10 -
-36-' 33"' 28"* 34" E K F P
.50"* 03 17 13 -
-56"' 12* 20"* 21-- E- F P
-59"* 13"* 05 14" E N P
-11 -39"* -22 -54"'- S J
-12 -44-* -24" -50*** S T J
-22" -32"* -22" -57"*' E S T J
-17** -36-" -11* -48** E S J
-11* -38"* -06 -55- S J
-53"* 03 -09 01 E -
-67-* 09 -09 01 E -
-63"* 03 00 -14 ( -
-64"* 13"* 03 04 E N -
-76"* 13-" -03 00 E N -
05 17 29* -01 .
-04 25" 30"' 08 N F -
20* 31"' 36"* 33"' N F P
06 31"' 37"* 25- N F P
-01 33** 25ZS 19"* N F P
01 12 -08 09 -
19* 09 04 24** p
17* 01 -19 00 -
13" 03 06 09 I -
09* 04 -02 09* P
06 31* 18 21 -
09 30'- 19* 20" N P
25" 34"* 17* 40"** N P
26"* 37"* 07 21"' I N P
02 36"* -03 15" N P
22 -16 -06 -31 -
01 -31"* -29" -32 S T J
-03 -1; -25" -28 T J
-09 -10* -19"* -13" -S T -
14** -27** -07 24" I S P
27* 22 11 21 -
33"* -06 22* 06 I -
25" -10 06 02 I -
20** -11 12 -10 -
23"* -07 16" -03 I F
Table 2 continued
SCALE GROUP N El SN TF JP IRLATIONSHIP
16 PERSONALITY FACTOR TEST
Primary Factors (Continued)
(figh trgic Tension
(mare reuroec Trendl
(lore Le derhip Potencial)
(CrmCaive personal ew)
66 -02 32" -23 22
122 00 25* -40~ 12
149 -13 28'* -10 13
484 09 31** -15** 15-
645 02 27" -25m 21"**
66 41** 02 -03 12
122 52* 00 00 12
149 44"* 22" -03 21"
484 41- 11* -06 12*
645 41"* 14** -11* 12"
66 -05 -43** -21 -41"
122 -11 -19* -19* -36"M
149 -01 -16* -24** 41"*
484 -10* -15"* -17"* -27*"
64! -07 -29*** -17** -40***
66 32" 26* 17 21
122 22* 02 30** 06
149 28** -03 04 02
484 23"' -03 14" -02
645 26** -02 09* 00
66 -51" 03 -OS 00
484 -70M 13" 12* 10*
645 -74** 12" 03 05
66 35" 23 16 17
484 35"* -01 15"** 04
645 34"- -04 14** 06
66 -18 13 -17 23
484 -17"' -21"' -34"' -08
645 -30- -06 -28"* OI
66 -10 33". -12 32"
484 29"* 45*" -10* 36"'
645 -10* 46** -23** 35"*
66 37" 16 22 12
484 44*** -01 15** -04
645 48** -04 18"* 00
66 -44** -35" -26* -37"
484 -54** -09 -13" -15*-
645 -58** -15"- -16"* -26-
66 12 50o" 19 39"
484 43** 37"* 02 25"*
645 16"* 41** -09* 25"*
OMNIBUS PERSONALITY INVENTORY
TI Thinking Introversion
TO Theoretical Orientation
(Artise c seanicivity)
(L beral, needs indepsodence)
RO Religious Orientation
(Skeptical, reject orthodoxy)
484 -10* 53"' 10* 11*
648 -06 59"' 04 12"
484 05 41"* -16** 07
648 -04 49"' -15*** 10
484 -05 46"* 29- 21-*
648 -11* 48*" 22** 17"*
484 03 .60" 13" 57***
648 -08 62"' 12" 53"*
484 11' 48*' 09 37"**
648 08 47* 01 33*
484 16** 20m" -15"* 26-*
648 09* 19"* -15** 25"*'
- NT -
N T P
X T P
S T J
- ST J
I N -
I F -
E N -
E S -
E T -
- N P
- F P
- N F P
- N F P
- N P
oiqHIBUS PERSOKALITY INVENTORY
Table 2 continued
SCALE GROUP N El S fF JP STRONGEST
OMNIBUS PERSONALITY INVENTORY
SE Social Extraversion 0 484 -75"m 10* 06 -02 E -
(Lke being wlth people) E 648 -77"' 07 01 -02 E -
IE Impulse Expression 0 484 -08 33** 07 42* N P
(lapulaJiv, Ia-inativ.) E 648 -20"* 29-" -06 34"* E N P
PI Personal Integration D 484 -38** -10' -08 -15* E J
(Not alienaetd or dAsturbed) E 648 -37"* -0S -09' -16"* E J
AL Anxiety Level D 484 -20** -07 -19** -06 E T -
S(Dnial of aniety or worry) E 648 -23-* -02 -19" -02 E T -
Am Altruism D 484 -44** 21m 25"* -06 EN F -
(Affillaive. care for others) E 648 -39"* 27**" 29"m 01 E N F .
PC Practical Outlook 0 484 -06 -67- -21** -41* S T J
(LUke applied, concrete) E 648 -03 -64"** -12" -32"' S T J
MF Masculir ty-Femininity D 484 10* -31' -35 -12* S T
(masculine, less social E 648 15** -32"* -36-"' 11 I S T
RI Response bias D 484 -26** 00 -11* -30"' E J
(Good iapression) E 648 -26"* 06 -09* -32"' E J
IOC Intellectual Disposition D 484 -03 -65* -11* -32"* S J
(Ierning is for practical use) E 648 07 -65- -04 -27"** S J
OPINION, ATTITUDE AND INTEREST SURVEY
Opinion and Attitude Scales
Set for True
484 -34** -36-" -05
658 -30** -42"- -05
46 00 -29* -04
484 06 30." 14"
658 01 35"' 04
46 09 36* 07
484 -49*** 26** 15**
658 -53"' 26*- 05
46 -55** 03 -01
484 48** -04 -25**
658 49** -04 -21"*
46 25 13 -25
484 05 -18** -02
658 11* -18** 08
46 20 -04 02
484 17*** 03 05 -25"*
658 14** -02 04 -36"*
46 17 12 21 -28Z
484 18** 35** 05 30**
658 20"* 27'* -06 27"'
46 03 33* -04 46***
484 -22"* 43"* 00 33***
658 -27" 46*** -08 32'*
46 -27 33* 10 36*
484 -44"** -12* 14"* -13"
658 -37*** -15"' 23** -13**'
46 -43" 06 24 -03
484 -47** 01 -18-* -12*
658 -47"** 01 -15*" -12**
46 -54"* -04 -20 -24
484 03 -09 -35" 09
658 04 -03 -33"* 02
46 09 -13 -56" -15
484 06 -14" 01 -10*
658 08 -12" 02 -05
46 08 19 -11 -11
484 32"' 05 -08 12*
658 25"* 08 -05 14-
46 21 06 -01 -14
484 23*' -02 02 17"'
658 26** -03 01 21*
46 -01 04 -14 45"
E S -
E S J
N F P
E N F
I T -
I N P
E F J
E S F J
E T -
E T T J
- T -