Professional Documents
Culture Documents
net/publication/230034577
CITATIONS READS
564 18,794
2 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Maryon F. King on 07 October 2017.
Top of text
Top of CT
Social Desirability Bias:
A Neglected Aspect of
Validity Testing
Maryon F. King and Gordon C. Bruner
Southern Illinois University
ABSTRACT
Top of text
1988). Among the most widely used measurement techniques in mar- Base of text
keting is the multi-item summated scale. In order to be effective in the
generation of interpretable research data, such scales must, at a mini-
mum, be unidimensional, reliable, and internally valid. Most research-
ers would readily acknowledge their desire for, and commitment to, the
implementation of valid measures; yet previous reviews of scale usage
in the marketing literature suggest that little effort is typically ex-
pended to assess scale validity (Bruner, 1993; Peter, 1981; Peter &
Churchill, 1986).
One potentially important threat to the validity of research employ-
ing multi-item scales is social desirability bias. The pervasive tendency
of individuals to present themselves in the most favorable manner rel-
ative to prevailing social norms and mores has threatened to compro-
mise research findings in the social sciences for more than 50 years.
Provision of socially desirable responses in self-report data may lead to
spurious correlations between variables as well as the suppression or
moderation of relationships between the constructs of interest (Ganster,
Hennesey, & Luthans, 1983). To the extent that the effects of social
desirability bias represent a source of influence which obscures mea-
surement of the primary relationships under investigation, validity may
be compromised (Malhotra, 1988). Hence, it is incumbent upon re-
searchers to identify situations in which data may be systematically
biased toward respondents’ perceptions of what is socially acceptable,
to determine the extent to which this represents contamination of the
data, and to implement the most appropriate methods of control.
Social-desirability bias (SDB) is considered to be one of the most com-
mon and pervasive sources of bias affecting the validity of experimental
and survey research findings in psychology and the social sciences (Ned-
erhof, 1985; Paulhus, 1991; Peltier & Walsh, 1990). Although identifi-
cation and control of SDB has presented an ongoing challenge for re-
searchers in most social science disciplines (cf. Fisher, 1993), only a
handful of articles published in marketing research journals have re-
ported attempts to systematically identify contamination from this
source. Considering the frequent use of self-report measures in mar-
keting research, as well as the social and normative implications inher-
ent in many marketing investigations, the low incidence of routine test-
ing for this potential confound may indicate that marketing researchers
have underestimated the importance of SDB in interpretation and ap-
plication of research findings.
Clearly, a prima fascia case may be made to argue the importance of
identifying and controlling for SDB in marketing research. The purpose
of this article is to facilitate systematic consideration of this issue in
future research activities by exploring the nature of social desirability
bias within the context of marketing. After reviewing the historical de-
velopment of the SDB construct, situations in which SDB may pose a short
significant threat and methods for managing SDB are examined. The standard
Top of text
few occasions on which tests for SDB have been conducted in the mar- Base of text
keting discipline are discussed, and the role of social desirability bias
within the marketing domain is evaluated. Finally, recommendations
for improving the quality of psychometric measures employed in future
marketing research are identified.
1
Although SD bias is the focus of this article, it is not the only methodological check needed to
assess the internal validity of measures. Admittedly, social-desirability bias is but one of many
response biases that may affect behavioral measures (cf. Paulhus, 1991). However, SD bias has
proven to be particularly troublesome in behavioral sciences research and has received more
attention than any other potential source of bias. Moreover, it is a particularly salient concern short
when assessing the validity of consumer-behavior measures. standard
Top of text
bada, Rubenfeld, & White, 1983; Dicken, 1963; Ganster et al., 1983). Base of text
Investigators have variously addressed the question of data contami-
nation from both the perspective of SDB as a behavioral correlate, as in
studies on personality (Edwards, 1957) and psychopathy (Meehl &
Hathaway, 1946), as well as a threat to the validity of the scales mea-
suring constructs of primary interest, as in studies of the MMPI (Robi-
nette, 1991) and the Chapman scales (Peltier & Walsh, 1990). Of par-
ticular importance to marketing researchers are studies published in
the psychology literature that have found significant relationships be-
tween the SDB scales and attitudes (Lenski & Leggett, 1960), attitude
change (Bagozzi, 1985; Buckhout, 1965a, 1966; Goldsmith, 1989; Green-
baum, 1966; Miller, Doob, Butler, & Marlowe), consumer satisfaction
(Sabourin, Bourgeois, Gendreau, & Morval 1989), dyadic interactions
(Buckhout, 1965b), innovativeness (Goldsmith, 1987), and risk-taking
behavior (Kogan, 1964; Kogan & Wallach, 1967). Hence, it is reasonable
to expect that many areas of marketing, particularly topics in consumer
behavior, would subject to social desirability bias.
Today, SDB is considered to be one of the most common and pervasive
sources of bias affecting the validity of experimental and survey re-
search findings in psychology (Nederhof, 1985; Paulhus, 1991; Peltier &
Walsh, 1990). In fact, the vast majority of research investigations of
attitudes, opinions, and personality that have been reported in psy-
chology journals in the last four decades have simultaneously admin-
istered social-desirability scale items as part of the validation process.
A recently published monograph on scale development for applied social
research recommends routinely testing for SDB during scale construc-
tion (DeVellis, 1991). The widespread use of SD scales is based on the
belief of psychology researchers that it is “essential to discriminate be-
tween the effects of item content and the needs of subjects to present
themselves in a socially desirable (or undesirable) light” (Crowne &
Marlowe, 1960, p. 21).
Top of text
with SDB (Paulhus, 1991). If a significant correlation is found, regres- Base of text
sion analysis may be used to determine the proportion of variance at-
tributable to SDB, with an ensuing correction for this distortion by use
of either partial correlations (McNemar, 1969) or canonical correlation
analysis (Greenblatt, Mozdzierz, & Murphy, 1984).
In more recent years, challenges regarding the unidimensionality of
the MCSDS have been voiced (cf. Ballard et al., 1988; Paulhus 1984,
1986; Robinson & Shaver, 1973). Low intercorrelations among the orig-
inal MCSDS and several of its short forms prompted factor analysis of
the various instruments, revealing two primary factors (Borkenau &
Ostendorf, 1989). Zerbe and Paulhus (1987) have concluded that these
two factors include self-deception, “the conscious tendency to see oneself
in a favorable light,” and impression management, “the conscious pre-
sentation of a false front, such as deliberately falsifying test responses
to create favorable impressions” (p. 253). They argue that the former
dispositional tendency is a relatively invariant personality trait and
thus should not be considered a contaminant per se, whereas the latter
factor, impression management, is the culprit in confounding research
data. To assist in individually assessing the effects of these two factors,
Paulhus (1992) has offered the Balanced Inventory of Desirable Re-
sponding (BIDR), a 40-item scale that reportedly offers minimal corre-
lation between impression management and self-deception. Despite its
theoretical advantages over the MCSDS, this scale has not been widely
adopted or reported in published research (cf. Moorman & Podsakoff,
1992). Hence, there is little empirical evidence to date that this scale
represents a significant advance in identifying and interpreting the ef-
fects of socially desirable responding.2
Other scales that have been used to identify SDB include the Martin-
Larsen Approval Motivation Scale (Larsen, Martin. Ettinger, & Nelson,
1976), a desirability scale based on attitude opinion items that is par-
ticularly intended for use in survey research (Schuessler, Hittle, & Car-
dascia, 1978); the Other-Deception Questionnaire (ODQ) and the Self-
Deception Questionnaire (SDQ), both developed by Sackeim and Gur
(1978, 1979); and (as previously noted) the Balanced Inventory of De-
sirable Responding (Paulhus, 1992). Although these scales have not
been tested or validated as extensively as the MCSDS, administration
of any SD scale represents an effort to detect and measure SDB in col-
lected data. Selection of an appropriate test instrument depends, in
part, on the projected role of SDB. That is, if SDB is considered to be a
contaminant that confounds responses to target variable scales, the in-
vestigator might choose the MCSDS, Other-Deception Questionnaire or
the impression management items from the BIDR. Conversely, if the
2
Other measures developed expressly to test for SD bias will be discussed later in this article. At
this juncture, the primary focus will remain on the Marlowe– Crowne SD scale (and its varia-
tions) as well as the BIDR, because these were the scales primarily employed in the marketing short
research discussed herein. standard
Top of text
tendency to respond in socially desirable ways is considered within the Base of text
research context to be a variable of theoretical interest, the Martin-
Larsen Approval Motivation Scale, the Self-Deception Questionnaire
(SDQ), or the self-deception items from the BIDR may be more appro-
priate.
Administration of a scale to detect patterns of socially desirable re-
sponses has several practical applications in theoretical research. First,
pretesting of new content scales while the scales are in their develop-
mental stages permits detection of items that are significantly corre-
lated with SDB measures, providing the opportunity to modify or elim-
inate content items prior to scale administration. Second, simultaneous
administration of a scale to detect SDB in conjunction with target con-
tent scales may reveal the presence of individual differences in response
styles. This type of systematic response bias can confound content with
style, and interfere with the accurate assessment of content variables.
Identification of SDB as a behavioral covariate can afford the investi-
gator the opportunity to accommodate the findings (e.g., partial out the
effects of the SD). (For a more complete review of this subject, see Gans-
ter, Hennessey, & Luthans, 1983.) Hence, determination of SDB is an
equally important validity test for both newly constructed and estab-
lished measures (Fraboni & Cooper, 1989). Finally, administration of
SD scale items can assist in identification of situational demands that
may evoke socially desirable responding. For example, enlightened a
priori evaluation of testing methods (e.g., computerized testing, ques-
tion format), research contexts (e.g., degree of anonymity), interviewer
characteristics, and other issues involved in research design will serve
to minimize SDB contamination in the target study. These issues will
be revisited later in this article.
Top of text
Marketing’s long-standing preoccupation with self-reports of poten- Base of text
tially sensitive topics related to information search patterns, product
attitudes, purchase behaviors, and persuasion, coupled with the fact
that much of this research involves survey data collection techniques,
would suggest that SDB testing is generally a good idea. Moreover, SDB
testing would seem not only to be warranted, but mandatory within the
rapidly growing body of literature in ethics and ethical behavior in mar-
keting situations (Randall & Fernandes, 1991). Research involving self-
reports of the behavioral aspects of consumers or marketers in conjunc-
tion with multi-item, summated scales should be suspect with respect
to SDB, and hence it is appropriate to investigate for SDB effects.
The twin concerns of conducting research involving sensitive topics
that could arguably elicit SDB as well as continued interest in producing
and refining valid scales suggests that efforts should be redoubled on
this front. However, before recommendations can be made regarding the
measurement and control of SDB, past efforts in this regard must be
assessed. Accordingly, the following review of efforts to isolate the ef-
fects of SDB in the marketing literature focuses on answering three
broad questions: How have marketers viewed the role of SDB in their
discipline? How and why have they tested for its presence in past stud-
ies? What impact has evidence of SDB had on research practices and
data interpretation?
3
The research reviewed here includes only those studies that employed SD testing in conjunction
with the use of multiitem, summated scales. It is also possible that potential SD bias was as-
sessed during the construction of some multi-item scales, but the authors chose not to report the short
findings. standard
MAR
WILEJ
Table 1. Summary of Articles Reporting Social Desirability Bias (JCR, JMR, JM, JAMS: 1980–1997)
Correlation
Coefficients
Type of Use (Content/SDB Source of Content
Author(s) Source Reported SDB Scale Scales) Scales Interpretation
Carlson & JCR (1988) Behavioral bias 19 item version Not reported Several existing scale Correlation was statistically
Grossbart of MCSDS items summative in- significant; authors con-
(⫽0.71) dices of general so- cluded that the magni-
cialization tendencies tude was too small to rep-
resent a relevant effect
Childers et JCR (1985) Scale construc- MCSDS ⫽ ⫺0.03 to 0.20 VVIQ (Marks 1973) Authors report that correla-
al. tion tion coefficients indicate
that no SDB is present
Fisher JCR (1993) Situational Reynolds (13 ⫽ ⫺0.49 to 0.10 New scale; multiple Use of indirect questioning
analysis items) items measuring per- techniques reduces so-
⫽0.69) sonal and normative cially desirable respond-
outcomes ing
LEFT INTERACTIVE
Friedman & JCR (1987) Scale construc- MCSDS ⫽ ⫺0.02 to 0.06 New scale; multi-item No significant SDB noted
Churchill tion measures of satisfac-
tion, compliance and
Base of RF
Base of text
Top of text
standard
short
SOCIAL DESIRABILITY RESPONSE BIAS
MAR
Raju JCR (1980) Scale construc- MCSDS None reported Arousal seeking tend- Several items eliciting SDB
tion ency (Robinson & were eliminated
Shaver 1973)
Richins JCR (1983) Scale construc- MCSDS ⫽ ⫺0.28 to 0.13 New and purified scales; Correlations were statisti-
WILEJ
tion multi-item measures cally significant; author
of assertiveness and concluded that the mag-
aggressiveness nitude was too small to
represent a relevant ef-
fect
Richins & JCR (1992) Behavioral bias Ballard et al. ⫽ ⫺0.12 to New and existing scale Authors report that low in-
Dawson short form ⫺0.03 items that were previ- significant correlations
(10 items) ously purified (Ri- suggest no significant
chins & Dawson SDB
1990)
Saxe & JCR (1982) Scale construc- MCSDS None reported New scale; multi-item Authors report that low in-
Weitz tion selling orientation significant correlations
scale suggest no significant
SDB
Unger & JCR (1983) Scale construc- MCSDS None reported New scale; multi-item Some evidence of the pres-
Kernan tion measures of leisure ence of SDB; interpreted
determinants as situation specific
Westbrook JCR (1980) Behavioral bias MCSDS ⫽ ⫺0.13 to 12 Several preexisting No significant SDB noted
RIGHT INTERACTIVE
scales
Westbrook JCR (1987) Behavioral bias MCSDS None reported DES-II (Izard 1977) No significant SDB noted;
pilot study scale was purified by orig-
inal author
87
Base of RF
Base of text
Top of text
standard
short
MAR WILEJ LEFT INTERACTIVE
Top of text
that individuals conducting research in consumer behavior may not be Base of text
aware that such a scale exists or under what circumstances adminis-
tration of such an instrument could be beneficial. Furthermore, there
may be some confusion regarding interpretation of results obtained from
administration of the MCSDS. A review and discussion of the few in-
stances in which testing for SDB has been conducted will illustrate con-
ditions under which marketers have previously identified this as a po-
tential threat to the validity of the research, and how SDB instruments
have been administered and results interpreted. An overview of each
study is presented below; a summary of articles is presented in Table 1.
In each of these studies an SD scale was self-administered, along with
scales measuring constructs of theoretical interest. All studies incor-
porated items from the MCSDS; in some cases an abbreviated version
of the MCSDS was employed in order to prevent task overload. The
BIDR scale was administered in only one instance. As previously noted,
a significant correlation between scores on the SD scale and a target
construct scale suggests that the data are confounded because of the
respondent’s desire to answer in socially desirable ways, thereby seri-
ously weakening the validity of the findings. Conversely, a low corre-
lation suggests that the measure is relatively free of SDB.
A collective evaluation of these studies provides insight into how mar-
keters have historically viewed the role of SDB in their discipline. SDB
testing was performed for three distinct reasons across these studies.
First, some studies sought to confirm the discriminant validity of pro-
posed items for summated measures during the construction of their
scales. In this case, authors were attempting to eliminate any items that
were identified as potential sources of contamination via an SD response
bias in order to purify their measures. Second, some studies used the
MCSDS responses as a correlate to their primary scale of interest. Here,
testing for SDB was performed in conjunction with a previously devel-
oped primary scale (rather than one in the developmental stages) in
order to identify behavioral tendencies of the subjects that might con-
taminate the study’s findings. Finally, two recent studies recognized the
broader importance of SDB across marketing contexts. Rather than ad-
ministering social-desirability scale items as part of a limited validation
process for a single content instrument, these investigations have em-
braced SDB as their core concern, engaging in situational analysis to
highlight and evaluate broad methodological and conceptual applica-
tions of this construct. In the ensuing discussion, studies are grouped
into categories by the intended role of SDB testing and presented in
chronological order.
Top of text
optimum stimulation levels and exploratory behavior, and personality Base of text
and demographic traits (Raju, 1980). The MCSDS was administered to
prescreen 90 statements to ensure selection of items with low correla-
tion with social desirability. Although no actual correlation coefficients
or significance levels were reported, several items eliciting SDB were
eliminated, leaving a pool of items that exhibited little or no SDB.
The MCSDS was administered during only one stage of a large study
reported by Unger and Kernan (1983). The study examined dimensions
of the subjective leisure experience in six different scenarios. During the
initial construction of a six-dimensional scale of leisure, the MCSDS was
included as one of several measures of reliability and internal consist-
ency. As in the previous study reviewed, no correlation coefficients or
significance levels were reported. Unlike the previous study, however,
the authors indicated that there was some evidence that SDB might be
present when using the scales. However, the authors chose to interpret
the bias as being situation specific and not necessarily inherent in the
scales themselves.
Saxe and Weitz (1982) administered the MCSDS to salespeople rep-
resenting a wide variety of sales positions. The purpose of this study
was to construct a scale for measuring a salesperson’s customer orien-
tation. Each item on the primary scale of interest was correlated with
the MCSDS scores. Although no actual correlation coefficients or sig-
nificance levels were reported, the authors indicated that all correla-
tions were small and insignificant and concluded that the items were
not contaminated by a SD factor.
The construction of two new scales, one for assertiveness and one for
aggressiveness, was the focus of a study by Richins (1983). SDB was
evaluated in two ways. First, during initial screening of proposed Lik-
ert-type scale items, those which appeared to have a strong potential
for eliciting a socially desirable response were removed. Respondents
completed both the assertiveness and aggressiveness scales along with
a shortened version of the MCSDS. The correlations between the SD
scale and assertiveness and aggressiveness were 0.13 (p ⬎ .10) and
⫺0.28 (p ⬍ .01), respectively. Although the latter correlation is statisti-
cally significant, the author concluded that the magnitude was too small
to represent a relevant effect.
Childers, Houston, and Heckler (1985) examined several measures of
visual and verbal mental imagery. The MCSDS was administered in
order to test the discriminant validity of the primary scale of interest
during the scale development process. Unlike the studies summarized
earlier, Childers et al. (1985) reported coefficient alphas for each of two
groups of subjects. Although prior studies have suggested that individ-
ual difference measures may be correlated with social desirability
(White, Sheehan, & Ashton, 1977), none of the scales proposed by Chil-
ders et al. were correlated with the social-desirability measure. short
The final instance of the administration of the MCSDS during scale standard
Top of text
construction was the only time it was employed in marketing outside Base of text
the arena of consumer behavior. The purpose of a study by Friedman
and Churchill (1987) was to examine how social power behaviors can be
effectively employed by physicians. Again, the MCSDS was used during
the scale construction phase of the study in order to assess the discrim-
inant validity of the dependent measures of satisfaction and compliance.
Individual, pairwise correlation coefficients were reported, with all val-
ues evidencing insignificant correlations ranging from negative to very
low positive levels.
Top of text
Carlson and Grossbart (1988) investigated the relationship between Base of text
parental styles and their effects on the manner in which children are
socialized with respect to consumption behavior. The primary research
instrument included 11 summative indices extracted from prior studies
predominantly reported in psychological journals. A shortened (19 item)
version of the MCSDS was used to identify any SD behavioral response
tendencies that might exist in the subject pool. Both alpha and beta
coefficients were reported, as well as some factor scores from the cluster
analysis indicating the presence, direction, and strength of the relation-
ship between SDB and group classification. Although some significant
but weak relationships between SD and factor scores existed, the au-
thors concluded that SD did not have a major influence on group clas-
sification.
Finally, Richins and Dawson (1992) incorporated a short form of the
MCSDS as a covariate in a study of the materialism construct. The
authors reported that their 18-item materialism scale was uncorrelated
with SDB items. This finding is in contrast to later evidence presented
by Mick (1996) that SDB negatively correlates with the materialism
scale, systematically lowering materialism scores. This contradiction in
findings will be more fully explored in the discussion section.
Top of text
explanatory power of this bias is “neither dramatic nor comprehensive” Base of text
(p. 116). The author cautions, however, that “. . . it would be unfor-
tunate if these results invited continued complacency about SDR among
consumer researchers . . .” (p. 116).
This research also addressed two measurement-related issues. First,
the original MCSDS was administered in conjunction with materialism
and compulsive buying instruments. As previously noted, in contrast to
the Richins and Dawson study (1992), materialism was significantly
related to socially desirable responding, as were other content variables
of interest (self-esteem and compulsive buying). Next, the newer BIDR
scale was incorporated into a study of consequences of materialism (i.e.,
neuroticism, self-actualization, and impulse buying). To date, this ap-
pears to be the only instance in which this reconceptualized scale has
been administered in published marketing research. Although the
BIDR scale was administered primarily to identify covariation of SDB
with the content variables, it was also expressly selected for use to pro-
vide evidence of its discriminant validity in partialling out the effects
of the self-deception and impression management SDB components. The
data indicate that impression management influenced the measure-
ment of several content variables; however, only two of the theoretically
proposed relationships appeared to be artifacts of this response bias.
Discussion of Findings
Although the preceding review is limited in terms of numbers of studies,
a wide range of topical areas in marketing research is represented. Of
the six studies involving scale construction, half provided evidence of
significant correlations between target construct items and the MCSDS.
Of these three studies, only one author chose to use evidence of an SD
bias to eliminate confounded scale items (Raju, 1980); in both of the
other studies (Richins, 1983; Unger & Kernan, 1983) the authors did
not modify their scales even though a significant relationship between
their content scales and SDB scales was noted. (This issue will be re-
visited shortly.) Although Westbrook (1987) found no significant rela-
tionship between the MCSDS and the DES-II, it should be noted that
the DES-II scale was purified with the use of the MCSDS during its
original construction. Hence, four instances of significant relationships
between construct scales and social-desirability scales are evident in
seven studies, providing compelling evidence for further investigation
of potential contamination due to response bias during marketing scale
construction.
Although less evidence of a response bias was apparent in the five
research studies involving the administration of the MCSDS (or one of
its shortened forms) as a behavioral covariate, two of these studies are
worthy of closer examination. First, Carlson and Grossbart (1988) iden- short
standard
Top of text
tified significant contamination effects, but suggested that elicited SDB Base of text
did not have a major influence on group classification. Even though the
content scales were initially developed in psychology (where they were
presumably covaried with an SD scale during development), the signif-
icant relationships identified suggest that residual bias may be obscur-
ing accurate assessment of the content variables. More perplexing is the
discrepancy between the findings of the Richins and Dawson (1992)
study and Mick’s (1996) study in assessing the potential response bias
contamination of the materialism scale. Identical content and SDB in-
struments were administered in both studies, with dramatically differ-
ing results. Although it has been suggested that “the norms guiding the
acceptability of materialism (when the data was collected in the 1980s)
may have been quite different from the norms of the 1990s” (Mick, 1996,
p. 108), other variables such as data collection methods or subject pop-
ulations should be scrutinized to identify any context or situational ef-
fects present in these studies. Other situational factors that could have
produced this effect are discussed later.
It is troublesome that in the majority of cases where SDB was re-
ported at significant levels the authors chose to disregard this potential
threat to the discriminant validity of the content scales. Although in-
terpreting the conceptual relevance of the bias remains in the domain
of the investigator (and those reviewing his or her work), every observed
correlation between the two instruments leaves the data open to at least
two possible explanations. Regardless of whether SDB is viewed as a
contaminant in self-assessment (the historical perspective) or as a sta-
ble individual difference variable (e.g., Paulhus’s self-deception factor),
it should be identified and controlled or explained. Furthermore, it has
been suggested that reliance upon differences in mean SD scale scores
or simple correlations between SD and content scales may result in un-
derestimating the extent of contamination (Zerbe & Paulhus, 1987).
When evidence of such a systematic bias is evident, it is paramount to
acknowledge that the bias exists and to take steps to partition out its
effects, rather than relying on post hoc explanations. Otherwise, inter-
pretation of results remains open to question.
Regardless of how the effect of SDB has been interpreted or con-
trolled, more than half of the studies examined herein have provided
evidence of response bias. Hence, it is reasonable to expect that other
studies in these areas of research could be biased by the subject’s desire
to respond in a socially desirable manner. In fact, it has recently been
suggested that “although the routinization of checking for social desir-
ability bias is much greater in psychology, it has been identified as being
potentially more problematic in marketing research than in psychology”
(Nunnally & Bernstein, 1994, p. 315). To encourage and facilitate sys-
tematic assessment of this response bias in future research, several is-
sues involved in controlling SDB are examined in the ensuing section. short
standard
Top of text
IDENTIFYING AND CONTROLLING THE INFLUENCE Base of text
OF SDB
Top of text
Scale Selection. The use of preexisting content scales that have been Base of text
examined for evidence of SDB may aid in reducing the likelihood of
encountering a response bias. Although selection of such scales does not
guarantee freedom from SDB in ensuing studies (e.g., the Richins &
Dawson vs. Mick controversy), there is evidence that such a strategy
may be beneficial (e.g., Westbrook’s 1987 application of the DES-II
scale).
Top of text
convince subjects that the equipment can determine whether or not the Base of text
subject is telling the truth (Jones & Sigall, 1971). In this rather extreme
and costly measure, subjects are hooked up to electronic equipment that
reputedly can physiologically assess their attitudes and determine if
their responses are consistent with their feelings. A simpler, albeit less
effective method is to warn subjects that the test instrument contains
methods for detecting faking. Additionally, the interviewer selected can
strongly affect the amount of SD bias evoked, because results of an in-
terview are more likely to be biased when subjects and interviewers are
similar with respect to social distance (e.g., Dohrenwend, Colombotos,
& Dohrenwend, 1968). These are but a few of the many available meth-
ods for reducing elicitation of socially desirable responding; a more com-
plete review of demand-reduction techniques is presented by Paulhus
(1991).
Top of text
tion toward social objects” (Greenwald & Banaji, 1995, p. 8). Because Base of text
implicit attitudes are under the control of “automatically activated eval-
uation,” the IAT has the potential to measure socially significant auto-
matic associations without relying on the veracity or objectivity of the
respondent. Because the IAT may have the ability to “resist self-pres-
entational forces that can mask personally or socially undesirable eval-
uative associations” (Greenwald et al., in press), it may prove be a valu-
able addition to the repertoire of tools in future investigations of socially
sensitive topics.
Top of text
interest in its own right. Prediction and treatment of socially desirable Base of text
responding within an experimental design will largely depend upon
theoretical evaluation of SDB relative to the nomological network of
interest. However, emerging interest in the dimensionality of SDB may
contribute to future research in this regard. Recent identification and
clarification of the self-deception and impression-management dimen-
sions of SDB have substantially contributed to the conceptual develop-
ment of this issue. Although development of the BIDR scale has signif-
icantly advanced knowledge, much investigative work remains to refine
the conceptual dimensions as well as to construct and test appropriate
scales.
Although there may be valid reasons why social-desirability bias
should not be examined in some cases, such cases would seem, from this
vantage, to be the exception rather than the rule. It is evident that social
desirability influences the responses of consumers to many kinds of
measures; to be unaware of the presence, direction, and extent of such
a bias represents a highly risky endeavor. Instead of ignoring the pos-
sibility of a confounding response bias in a scale, testing should be rou-
tinely performed in order to purify the measure and aid in its future
usage. Therefore, until such time when testing for SD bias is routine,
the findings in much of the marketing research literature will have to
be accepted with caution.
REFERENCES
Top of text
Borkenau, P., & Ostendorf, F. (1989). Descriptive consistency and social desir- Base of text
ability in self- and peer-reports. European Journal of Personality, 3, 31 – 45.
Brinberg, D., & Hirschman, E. C. (1987). Multiple orientations for the conduct
of marketing research: An analysis of the academic/practitioner distinction.
Journal of Marketing, 50, 161 – 173.
Bruner, G. C. II. (1993). A census of multi-item scales used in marketing re-
search (Technical Report No. 9305). Carbondale, IL: Southern Illinois Uni-
versity, Office of Scale Research.
Buckhout, R. (1965a). Need for social approval and attitude change. Journal of
Psychology, 60, 123 – 128.
Buckhout, R. (1965b). Need for social approval and dyadic verbal behavior.
Psychological Reports, 16, 1013 – 1016.
Buckhout, R. (1966). Changes in heart rate accompanying attitude change.
Journal of Personality and Social Psychology, 4, 695 – 699.
Campbell, D. R., & Fiske, D. W. (1959). Convergent and discriminant validation
by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81 – 105.
Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental
designs for research. Boston: Houghton Mifflin.
Carlson, L., & Grossbart, S. (1988). Parental style and consumer socialization
of children. Journal of Consumer Research, 15, 77 – 94.
Childers, T. L., Houston, M. J., & Heckler, S. E. (1985). Measurement of indi-
vidual differences in visual versus verbal information processing. Journal of
Consumer Research, 12, 125 – 134.
Churchill, G. A., Jr. (1979). A paradigm for developing better measures of mar-
keting constructs. Journal of Marketing Research, 16, 64 – 73.
Cofer, C. N., Chance, J., & Judson, A. J. (1949). A study of malingering on the
MMPI. Journal of Psychology, 27, 491 – 499.
Conger, A. J. (1974). A revised definition for suppressor variables: A guide to
their identification and interpretation. Educational and Psychological Mea-
surement, 34, 35 – 46.
Cook, T. D., & Campbell, D T. (1979). Quasi-experimentation: Design and anal-
ysis issues for field settings. Boston: Houghton Mifflin.
Crino, M. D., Svoboda, M., Rubenfeld, S., & White, M. C. (1983). Data on the
Marlowe – Crowne and Edwards social desirability scales. Psychological Re-
ports, 53, 963 – 968.
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability inde-
pendent of psychopathology. Journal of Consulting Psychology, 24, 349 – 354.
Crowne, D. P., & Marlowe, D. (1964). The approval motive. New York: John
Wiley & Sons.
De Vellis, R. F. (1991). Scale development: Theory and applications. London:
Sage Publications.
Dicken, C. (1963). Good impression, social desirability and acquaintance as
suppressor variables. Educational and Psychological Measurement, 23, 699 –
720.
Dohrenwend, B. S., Colombotos, J., & Dorhenwend, B. P. (1968). Social distance
and interviewer effects. Public Opinion Quarterly, 32, 410 – 422.
Edwards, A. L. (1957). The social desirability variable in personality assess-
ment and research. New York: Dryden Press.
Edwards, A. L. (1970). The measurement of personality traits by scales and short
inventories. New York: Dryden Press. standard
Top of text
Edwards, A. L., Diers, C. J., & Walker, J. N. (1962). Response sets and factor Base of text
loadings on sixty-one personality scales. Journal of Applied Psychology, 46,
220 – 225.
Fisher, R. J. (1993). Social desirability bias and the validity of indirect ques-
tioning. Journal of Consumer Research, 20, 303 – 315.
Fraboni, M., & Cooper, D. (1989). Further validation of three short forms of the
Marlowe – Crowne scale of social desirability. Psychological Reports, 65, 595 –
600.
Friedman, M. L., & Churchill, G. A. (1987). Using consumer perceptions and a
contingency approach to improve health care delivery. Journal of Consumer
Research, 13, 492 – 510.
Ganster, D. C., Hennessey, H. W., & Luthans, F. (1983). Social desirability
response effects: Three different models. Academy of Management Journal,
26, 955 – 966.
Goldsmith, R. E. (1987). Self-monitoring and innovativeness. Psychological Re-
ports, 60, 1017 – 1018.
Goldsmith, R. E. (1989). A validity study of the affect intensity measure. Jour-
nal of Social Behavior and Personality, 4, 133 – 140.
Goode, W. J., & Hart, P. K. (1952). Methods in social science. New York: Mc-
Graw-Hill.
Greenbaum, C. W. (1966). Effect of situational and personality variables on
improvisation and attitude change. Journal of Personality and Social Psy-
chology, 4, 260 – 269.
Greenblatt, R. L., Mozdzierz, G. J., & Murphy, T. J. (1984). Content and re-
sponse-style in the construct validation of self-report inventories: A canonical
analysis. Journal of Clinical Psychology, 40, 1414 – 1420.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (in press). Measuring
individual differences in implicit cognition: The implicit association test.
Journal of Personality and Social Psychology.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes,
self-esteem, and stereotypes. Journal of Personality and Social Psychology,
102, 4 – 27.
Greenwald, H. J. Satow, Y. (1970). A short social desirability scale. Psycholog-
ical Reports, 27, 131 – 135.
Humm, D. G., & Wadsworth, G. W., Jr. (1939). The Humm-Wadsworth tem-
perament scale. Los Angeles: Doncaster G. Humm.
Hunt, S. D. (1990). Truth in marketing theory and research. Journal of Mar-
keting, 54, 1 – 15.
Izard, C. E. (1977). Human emotion. New York: Plenum Press.
Jacoby, J. (1978). Consumer research: A state of the art review. Journal of
Marketing, 42, 87 – 96.
Jones, E. E., & Sigall, H. (1971). The bogus pipeline: A new paradigm for mea-
suring affect and attitude. Psychological Bulletin, 76, 349 – 364.
Kogan, N. (1964). Risk taking: A study in cognition and personality. New York:
Holt, Rinehart & Winston.
Kogan, N., & Wallach, M. A. (1967). Group risk taking as a function of members’
anxiety and defensiveness levels. Journal of Personality, 35, 50 – 63.
Larsen, K. S., Martin, H. J., Ettinger, R. H., & Nelson, J. (1976). Approval
seeking, social cost and aggression: A scale and some dynamics. Journal of short
Psychology, 94, 3 – 11. standard
Top of text
Lenski, G. E., & Leggett, J. C. (1960). Caste, class, and deference in the research Base of text
interview. American Journal of Sociology, 65, 463 – 467.
Malhotra, N. K. (1988). Some observations on the state of the art in marketing
research. Journal of the Academy of Marketing Science, 16, 4 – 24.
Marks, D. F. (1973). Visual imagery differences in the recall of pictures. British
Journal of Psychology, 64, 17 – 24.
McNemar, Q. (1969). Psychological statistics. New York: John Wiley & Sons.
Meehl, P. E., & Hathaway, S. R. (1946). The k factor as a suppressor variable
in the MMPI. Journal of Applied Psychology, 30, 525 – 564.
Mick, D. G. (1996). Are studies of dark side variables confounded by socially
desirable responding? The case of materialism. Journal of Consumer Re-
search, 23, 106 – 119.
Miller, N., Doob, A. N., Butler, D. C., & Marlowe, D. (1965). The tendency to
agree: Situational determinants and social desirability. Journal of Experi-
mental Research in Personality, 1, 78 – 83.
Moore, E. M., Bearden, W.O., & Teel, J. E. (1985). Use of labeling and assertions
of dependency in appeals for consumer support. Journal of Consumer Re-
search, 12, 90 – 96.
Moorman, R. H., & Podsakoff, P. M. (1992). A meta-analytic and empirical test
of the potential confounding effect of social desirability response sets in or-
ganizational behavior research Journal of Occupational and Organizational
Psychology, 56, 131 – 149.
Muncy, J. A., & Fisk, R. P. (1987). Cognitive relativism and the practice of
marketing science. Journal of Marketing, 51, 20 – 33.
Nederhof, A. J. (1985). Methods of coping with social desirability bias: A review.
European Journal of Social Psychology, 15, 263 – 280.
Nunnally, J. C. (1967). Psychometric theory. New York: McGraw-Hill.
Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory. New York:
McGraw-Hill
Orne, M. T. (1969). Demand characteristics and the concept of quasi-controls.
In R. Rosenthal & R. L. Rosnow (Eds.), Artifact in behavioral research
(pp.143 – 179). New York: Academic Press.
Park, C. W., & Lessig, V. P. (1977). Students and housewives: Differences in
susceptibility to reference group influence. Journal of Consumer Research,
4, 102 – 110.
Paulhus, D. L. (1991). Measurement and control of response bias. In John P.
Robinson et al. (Eds.), Measures of personality and social psychological at-
titudes, (Vol. 1, pp. 17 – 59.). New York: Academic Press.
Paulhus, D. L. (1992). Assessing self-deception and impression management in
self-reports: The balanced inventory of desirable responding (Reference man-
ual, version 6). Vancouver, BC: University of British Columbia.
Paulhus, D. L. (1984). Two-component models of socially desirability respond-
ing. Journal of Personality and Social Psychology, 46, 598 – 609.
Paulhus, D. L. (1986). Self-deception and impression management in test re-
sponses. In A. Angleitner & J.S. Wiggins (Eds.), Personality assessment via
questionnaire (pp. 143 – 165). New York: Springer.
Peltier, B. D., & Walsh, J. A. (1990). An investigation of response bias in the
Chapman Scale. Educational and Psychological Measurement, 50, 803 – 815.
Peter, P. J. (1979). Reliability: A review of psychometric basics and recent mar- short
keting practices. Journal of Marketing Research, 16, 6 – 17. standard
Top of text
Peter, P. J. (1981). Construct validity: A review of basic issues of marketing Base of text
and practices. Journal of Marketing Research, 18, 133 – 145.
Peter, P. J., & Churchill, G. A., Jr. (1986). Relationships among research design
choices and psychometric properties of rating scales: A meta-analysis. Jour-
nal of Marketing Research, 23, 1 – 10.
Raju, P. S. (1980). Optimum stimulation level: Its relationship to personality,
demographics, and exploratory behavior. Journal of Consumer Research, 7,
272 – 282.
Randall, D. M., & Fernandes, M. F. (1991). The social desirability response bias
in ethics research. Journal of Business Ethics, 10, 805 – 817.
Rentz, J. O. (1988). An exploratory study of the generalizability of selected
marketing measures. Journal of the Academy of Marketing Science, 16, 141 –
150.
Reynolds, W. M. (1982). Development of reliable and valid short forms of the
Marlowe – Crowne social desirability scale. Journal of Clinical Phsychology,
38, 119 – 125.
Richins, M. L. (1983). An analysis of consumer interaction styles in the mar-
ketplace. Journal of Consumer Research, 10, 73 – 82.
Richins, M. L., & Dawson, S. (1990). Measuring material values: A preliminary
report of scale development. In M. E. Goldberg, G. Gorn, & R. W. Pollay (Eds.)
Advances in consumer research (Vol. 17, pp. 169 – 175). Provo, UT: Associa-
tion for Consumer Research.
Richins, M. L., & Dawson, S. (1992). A consumer values orientation for mate-
rialism and its measurement: Scale development and validation. Journal of
Consumer Research, 19, 303 – 316.
Robinette, R. L. (1991). The relationship between the Marlowe – Crowne Form
c and the validity scales of the MMPI. Journal of Clinical Psychology, 47,
396 – 399.
Robinson, J. P., & Shaver. P. R. (1973). Measures of social psychological atti-
tudes. Ann Arbor, MI: University of Michigan, Institute for Social Research.
Sabourin, S., Bourgeois, L., Gendreau, P., & Morval, M. (1989). Self-deception,
impression management, and consumer satisfaction with mental health
treatment. Psychological Assessment, 1, 126 – 229.
Sackeim, H. A., & Gur, R. C. (1978). Self-deception, self-confrontation and con-
sciousness. In G. E. Schwartz & D. Shapero (Eds.) Consciousness and self-
regulation: Advances in research and theory (Vol. 2, pp. 139 – 197). New York:
Plenum.
Sackeim, H. A., & Gur, R. C. (1979). Self-deception, other-deception and self-
reported psychopathology. Journal of Consulting and Clinical Psychology, 47,
213 – 215.
Saxe, R., & Weitz, B, A. (1982). The SOCO scale: A measure of the customer
orientation of salespeople. Journal of Marketing Research, 14, 343 – 351.
Schuessler, K. F., Hittle, D., & Cardascia, J. (1978). Measuring responding
desirability with attitude-opinion items. Social Psychology, 41, 224 – 235.
Skipper, R., & Hyman, M. R. (1987). Evaluating and improving argument-cen-
tered works in marketing. Journal of Marketing, 51, 60 – 75.
Skipper, R., & Hyman, M. R. (1990). Marketing and logical thought. Journal
of Marketing, 54, 89 – 92.
Strahan, R., & Gerbasi, K. C. (1972). Short, homogenous versions of the Mar- short
standard
Top of text
lowe-Crowne social desirability scale. Journal of Clinical Psychology, 28, Base of text
191 – 193.
Sudman, S., & Bradburn, N. M. (1974). Response effects in surveys. Chicago:
Aldine.
Thomas, K. W., & Kilmann, R. H. (1975). The social desirability variable in
organizational research: An alternative explanation for reported findings.
Academy of Management Journal, 18, 741 – 752.
Unger, L. S., & Kernan, J. B. (1983). On the meaning of leisure: An investi-
gation of some determinants of the subjective experience. Journal of Con-
sumer Research, 9, 381 – 391.
Westbrook, R. A. (1980). Intrapersonal affective influences on consumer satis-
faction with products. Journal of Consumer Research, 7, 49 – 54.
Westbrook, R. A. (1987). Product/consumption based affective responses and
postpurchase processes. Journal of Marketing Research, 24, 258 – 270.
White, K., Sheehan, P. W., & Ashton, R. (1977). Imagery assessment: A survey
of self-report measures. Journal of Mental Imagery, 1, 145 – 170.
Wiggins, J. S. (1959). Interrelationships among MMPI measures of dissimu-
lation under standard and social desirability instructions. Journal of Con-
sulting Psychology, 23, 419 – 427.
Zerbe, W. J., & Paulhus, D. L. (1987). Socially desirable responding in orga-
nizational behavior: A reconception. Academy of Management Review, 12,
250 – 264.
Correspondence regarding this article should be sent to: Maryon F. King, De-
partment of Marketing, College of Business and Administration, Southern Il-
linois University, Rehn Hall, Room 229, Mail: Code:4629, Carbondale, IL
62901-4629.
short
standard