The Stereotype as a Research Tool Author(s): Irwin Deutscher Source: Social Forces, Vol. 37, No. 1 (Oct., 1958), pp. 55-60 Published by: University of North Carolina Press Stable URL: http://www.jstor.org/stable/2573780 Accessed: 25/11/2010 08:00 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=uncpress . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. University of North Carolina Press is collaborating with JSTOR to digitize, preserve and extend access to Social Forces. http://www.jstor.org
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
The Stereotype as a Research ToolAuthor(s): Irwin DeutscherSource: Social Forces, Vol. 37, No. 1 (Oct., 1958), pp. 55-60Published by: University of North Carolina PressStable URL: http://www.jstor.org/stable/2573780
Accessed: 25/11/2010 08:00
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
In this paper we have begun to specify some of the
social dynamics involved in such a system of
"rights and duties." We have done so in a manner
which, we expect, will lend itself to empirical
study. We have not concernedourselves with the
methodological aspects involved in the implemen-
tation and testing of the ideas presented. This
must await subsequent work.2Ibid., p. 406.
THE STEREOTYPEAS A RESEARCH TOOL*
IRWIN DEUTSCHER
Community tudies Incorporated
T IS possible or symbolswhichevokestereo-
types to be received through any of the fivesenses: the smell of garlic, the clasp of a
calloused hand, the taste of curry, the sight of
excessive make-up, the sound of a familiar phono-
graph record. On the one hand these may be
highly individualized symbols, operating uniquely
on a person as a result of some configurationof his
own past experiences. On the other hand, stereo-
types are often evoked by verbal symbols: a
whore, an angel of mercy, a Sunday school teacher,
a housewife. To the layman, these verbal symbols
mobilize a complexof ideas and images in much the
same manner that a sound concept does for the
scientist. Strong feelings about issues, personal
values, attitudes, motives, etc., when internalized
by an individual, become condensed into some
sort of stereotyped shorthand reference.The social
scientist, who participates in or otherwise under-
stands the subtleties of a culture, can learn to
identify these stereotypes and the verbal folk-
symbols which evoke them. Once understood,
these stereotypes can be exploited for research
purposes.'
THE TRADITIONAL VIEW OF OBJECTIVITY, BIAS,
AND LOADED QUESTIONS
The possibility of such exploitation is grudgingly
acknowledged by some of the better texts on
methodology, but only after careful and detailed
instruction concerning the efficacyof "objectivity"
and freedom from "bias." Such acknowledgments
may be made parenthetically:
Is the wording biased?Is it emotionally loaded orslanted towarda particularkind of answer?Does itemploystereotypes?Does it containprestige-carryingnames?Does it employ superlative erms whichpush
the answeroneway or the other?(If suchelementsofbias are present, are they there intentionally-anddoes the research urpose ustifytheir nclusion?)2
The expected answer to the rhetorical questions
up to theparenthesis s an emphatic "no." But it is
the parenthetical question which will be em-
phasized in the present paper; it implies a "yes"
answerand when such an answer can be given then
a "yes" answer becomes permissible for all the
preceding questions. If not parenthetical, the
grudging acknowledgment may be made by foot-
note. For example,after stringentwarnings against
"leading questions" in his text, one author buries
this statement in just such a manner: "Leading
questions may, of course, have value if they are
* The research used for illustrative purposes in this
paper was made possible by a grant to Community
Studies, Inc., from the American Nurses' Association.
I am especially indebted to my colleagues Howard S.
Becker, Dan C. Lortie, Thomas A. McPartland, Peter
Kong-ming New, Warren A. Peterson, and Julius Roth
for their constructive criticisms of this paper in an
earlier form.I For a penetrating analysis of the nature of stereo-
types see Maurice N. Richter, Jr., "The Conceptual
Mechanism of Stereotyping," American SociologicaZ
Review, 21 (October 1956), pp. 568-571. An example of
a rewarding functional analysis of a stereotype is pro-vided by Alvin W. Gouldner, "Red Tape as a Social
Problem," in Robert K. Merton et at. (eds.), Readerin
Bureaucarcy (Glencoe, Ill.: The Free Press, 1952), pp.
410-418.
2 Arthur Kornhauser, "Constructing Questionnaires
and Interview Schedules," in Marie Jahoda, Morton
Deutsch, and Stuart Cook (eds.), Research Methods in
Social Relations, Part II, Selected Techniques (NewYork: The Dryden Press, 1951), p. 448.
However, if one is looking for a distorted response,
then loading becomes an advantageous methodo-
logical device. Becausestereotypes are by definition
distortions, it should be apparent that purposeful
distortion is necessary in the identification of
stereotypical images. A final form of grudging
acknowledgment of the possibility of research
value in stereotypes comes in the form of an
afterthought: "Is the question content biased orloaded in one direction-without accompanying
questions to balance the emphasis?"5In the dis-
cussion which follows it will be seen that when a
careful effort is made to maintain just such a
balance in bias or loading, the results can be both
reliable and valid.
Rather than bemoan popular stereotyping and
attempt by complicated means to circumvent it,
the phenomenon can be exploited as an entree to
the feelings, attitudes, images, and latent re-
sponses, obscured within the respondent. In a
sense, the principleis the same as that employed in
the use of psychological projective tests, and, like
projective devices, the common stereotype can be
especially useful in research concerned with self-
conceptionsand conceptionsof others. In the latter
case, if the symbols which evoke the stereotype of
the relevant "others" can be identified, then the
respondent'sattitudes toward and images of those
others can be determined. As Litwak has so
clearly demonstrated, "The purpose of the in-
vestigator defines the bias."6No question is ever
inherently biased. As a result of his arguments,Litwak concludes that, "No longer do we say a
given type of question is 'good' or 'bad.' The
search for the 'perfect' question which is not
loaded, double-barreled,or vague becomes trivial
under this new formulation." Among other things,
Litwak now asks "For what purpose is each of
these questions useful?"
AN ILLUSTRATIVE CASE OF PURPOSEFUL BIAS
As an example of how such a process works when
applied empirically, we will examine a study, the
purpose of which was to learn something of the
nature of public images of the nurse.7 Differences
in evaluations of nurses and nursing by different
publics within a metropolitan area were to be de-
termined quantitatively (by means of a scaling
device) and interpretedwith the aid of qualitative
analysis of responsesto an open-ended question at
the end of the interview. A probability sample of
householdsin the metropolitan area, in conjunction
with a random sample of persons 18 years of ageand over within the households, resulted in nearly
one thousand complete, personally administered
interviews."
On the basis of informal conversations about
nurses, with people at bus stops, parties, cafeterias,
bars, and the like, and more formalinterviews with
such groups as unemployed laborers loitering
around their union hall and Junior League ladies
at their monthly meeting, the investigators
familiarized themselves with the way in which
people talk about nurses. It was determined that
there were four recurrent themes in the evaluative
comments about nurses; these themes were em-
ployed as the major components of an evaluative
scale which was designed to reveal differentials in
public images of the nurse. In brief, these four
components are as follows:
(1) A moral evaluiation,e.g., "Nursesare easy makes,"or "Theyareabovereproach ndremindme of [Catho-lic] sisters; they are like angels in their white uni-
forms.
(2) A social-class evaluation, e.g., "They come from
deprived backgrounds-not very good homes-andcan't affordto go to college,"or "Nursesarea better
class of people who have had good upbringingand
knowhow to talk to people."
I George A Lungberg, Social Research (New York:
Longmans, Green, and Co., 1942), p. 193, n. 11.
4 Leon Festinger and Daniel Katz, ResearchMethiods
in the Behavioral Sciences (New York: The Dryden
Press, 1952), p. 347 (italics mine).5Kornhauser, op. cit., p. 440.
6 Eugene Litwak, "A Classification of Biased Ques-
tions," American Journal of Sociology, 62 (September
1956), pp. 182-186.
7 Public Images of theNurse, Part II of A Study of the
Registered Nurse in a Metropolitan Community
(Kansas City, Mo.: Community Studies, Inc., August
1955).
8 The size of the selected sample was 1,200 individuals
with an anticipated loss of approximately 20 percent.
Actual loss was 20.2 percent (the refusal rate was 11.3
percent; the remaining 8.9 percent loss is attributable
(3) Evaluations based on self- or family-identification,
e.g., "I thinkthat any womanwho tries canfind betterwork than that; I certainlywouldn'tchoose to be anurse,"or "There s nothing I wouldrathersee than
to have my daughtergrowup to be a nurse."
(4) Evaluative comparisons wuith ther women's occupa-tions, e.g., "Teachings a much betterjob than beinga nurse,"or "Nurses have made somethingof them-
selves; heyaren'tsatisfiedwithjust clerkingn a storeorbeingan office lunkyorsomethingike that."
Four arbitrary scalesweredevised to elicit value-
laden stereotypes of the nurse.9 To get at the
moral evaluation we asked, "Tell me which of the
following you consider closest to your idea of a
nurse? (1) a saint, (2) a Sunday school teacher, (3)
a housewife, (4) a waitress, (5) a loose woman."
The social-class dimension was includedby asking:"Imagine that you could say all Americans are
upper class, middle class, or lower class. Tell me
where you would put most nurses?" The five
answers provided ranged from (1) upper class to
(5) lower class. The self- or family-identification
dimensionwas obtained by supplying the familiar
range of alternatives from (1) I would be very
happy, to (5) I would strongly disapprove, for the
question, "If you had a daughter, which one of
these most closely approachesthe way you would
feel about her being a nurse?"An evaluative com-
parison with other women's occupations was in-
cluded in the following form: "Imagine that you
are going to workvery close to a person for several
years. You do not know anything about the
person except that you have been given a list
which tells you the person's past occupation. You
must select one with only this knowledge. Which
one type would you select to work around?" The
alternative responsesare (1) a private secretary to
an executive, (2) a saleslady in a department
store, (3) a nurse in a doctor's office, (4) a lady
taxicab driver.There is no "out" provided for any of these
questions by a "don't know" alternative. Either
the respondentanswersthe questionor he does not.
It is true that some people resented being pushed
into selecting one of several value-laden cate-
gories. Resistance appeared most frequently
among the best educated and the most poorly
educated groups, although, for the latter group,
"resistance" may not be an accurate description.
The most poorly educated group seemed not somuch to resist stereotyping as to be unaccustomed
to thought processes involving generalization and
conceptualization. They were unable to cope with
the categories provided; the better educated group
appeared unwilling to do so.10The reaction of some
respondents who complained that the questions
didn't "make sense" is equally understandable
when we recall that the questions were designed to
get at a stereotype through an associational
pattern-to trigger a sequence of ideas which
would lead to the revelation of the respondent's
subjective image; the fact that he does not under-
stand the purpose of a question may result in its
apparent senselessness to him. Although this
could be avoided by asking only survey-type
questions with an obvious purpose, students of
attitude and opinion are learning that it is not
always advantageous for a respondent to be aware
of their purposes. Experiences with such pro-
jective devices as Rorschach, T.A.T., sentence
completion, and word-association tests (none of
which "make sense" to the respondents), indicate
that the richest and most meaningful results maybe derived from the application of more subtle
tactics.
We asked "leading" questions and we asked
"loaded" questions, because we were seeking
neither superficial information nor to test the re-
spondent's knowledge. We literally desired to
"lead" the respondent into revealing his "loaded"
feelings, rather than to obtain simperingcliches or
permit devious evasions into the realm of "don't
know." We asked "what do you feel?" not "what
do you know?" In spite of scattered resistance, avast majority of the respondentswere able to select
one response over the others and in doing so, re-
vealed something of their own value systems.
AN EMPIRICAL TEST
One way to test the proposition that stereotypes
can be employed in this manner is to observe the
9The somewhat tenuous assumption that the scale
items approximate an equal distance from one another
was made. The use of Analysis of Variance is of course
dependent on the correctness of that assumption. The
only vindication for the assumption lies in the fact that
(as is pointed out below) independent techniques not
requiring such an assumption culminate in results whichclosely approximate those obtained by analyzing the
scales.
10 For a discussion of the kinds of resistance displayed
by a sample of physicians who received this same ques-
tionnaire by mail, see this writer's "Physicians' Reac-
tionsto
aMailed Questionnaire: A Study in 'Resisten-
tialism,' " The Public Opinion Quarterly, 20 (Fall