Socially Desirable Responding

Abstract

This article investigated the dimensionality, measurement invariance, and cross-cultural variations of social desirability. A total of 3,471 university students from 20 countries completed an adapted version of the Marlowe–Crowne scale. A two-dimensional structure was revealed in the pooled sample, distinguishing enhancement (endorsement of positive self-description) and denial (rejection of negative self-description). The factor structure was supported in most countries; medium-sized item bias was found in two denial items. In a multilevel analysis, we found that (a) there was more cross-cultural variation in denial than enhancement; (b) females tended to score higher on enhancement whereas males tended to score higher on denial; (c) the Human Development Index, an indicator of country socioeconomic development, was the best (negative) predictor of denial; and (d) both enhancement and denial seemed to be associated with country-level values and personality pertinent to “fitting in.” We conclude that social desirability has a positive and a negative impression management dimension that are meaningfully associated with country-level characteristics, and we argue that social desirability is better interpreted as culturally regulated response amplification.

Keywords

social desirability cultures values personality multilevel analysis

Socially desirable responding (SDR) refers to the tendency of respondents to reply in a manner that will be viewed favorably by others (Paulhus, 1991). SDR can challenge the validity of psychological measures. Respondents high on SDR tend to respond according to how they think people in their immediate environment would like them to react, whereas the interpretation of psychological measures is based on responses that are not contaminated by SDR. In this line of thinking, SDR is a nuisance factor that should be minimized, through a careful research design or statistical corrections (e.g., Nederhof, 1985). In another interpretation, SDR is more about substance than style (Schwartz, Verkasalo, Antonovsky, & Sagiv, 1997); it is then considered part and parcel of the psychological makeup of individuals that reflects culturally preferred ways of communication associated with various other cultural characteristics (van Hemert, van de Vijver, Poortinga, & Georgas, 2002). Here, SDR reflects genuine individual and cultural differences, representing effective and truthful self-presentation. The accurate measurement of this construct is a prerequisite for resolving the nuisance versus substance interpretation of SDR (Leite & Beretvas, 2005). However, the steadily accumulating empirical evidence about probably the most widely used SDR instrument, the Marlowe–Crowne scale (Crowne & Marlowe, 1960), has not produced clear conclusions as to its dimensionality and cross-cultural equivalence (e.g., Li & Reb, 2009; Verardi et al., 2010). Not surprisingly, the debate of the nature of SDR is continuing and a clear psychological meaning of SDR has not been adequately established. In our view, cross-cultural evidence could help to examine the stability of its dimensionality and provide important information about the nuisance versus substance discussion. With a shortened and adapted version of the Marlowe–Crowne scale, the present study examines the factor structure of SDR across 20 countries and the associations of SDR with country-level characteristics.

Dimensionality of the Marlowe–Crowne Scale

Studies of SDR do not reveal the same number of factors, and different instruments yield very different factor structures (Paulhus, 2002). The Marlowe–Crowne scale, consisting of 33 descriptions of highly desirable but rare and highly undesirable but common behaviors, measures respondents’ tendency to present themselves in a positive light (Crowne & Marlowe, 1960). Initially conceptualized as unidimensional (Crowne & Marlowe, 1964), this scale has been suggested to be multidimensional, although there is no convergence on the number of factors (Barger, 2002; Loo & Loewen, 2004). For instance, Verardi et al. (2010) administered a shortened version of this scale in eight African countries and Switzerland, where they distinguished achievement and international relationship; neither scale reached scalar invariance.

Millham (1974) and Ramanaiah, Schill, and Leung (1977) found a two-dimensional structure of the Marlowe–Crowne scale: enhancement (i.e., the tendency to attribute socially desirable characteristics to oneself) and denial (i.e., the tendency to deny undesirable characteristics). Such a distinction is in line with the two basic self-presentation motives: looking good and avoiding looking bad (Schütz, 1998). The two dimensions were found to have differential validity in predicting scales in the Minnesota Multiphasic Personality Inventory. According to Paulhus (1991) and Ventimiglia and MacDonald (2012), the Marlowe–Crowne scale taps mainly impression management. Using the Balanced Inventory of Desirable Responding, an alternative two-dimensional structure of SDR was proposed by Paulhus (1984). He differentiated impression management (i.e., deliberate self-presentation to an audience) and self-deception (i.e., favorably biased but honestly held self-descriptions). The validity and utility of these two dimensions are not always supported in cross-cultural contexts (e.g., Helmes & Holden, 2003; Li & Bagger, 2006). Moreover, Paulhus and Reid (1991) reported that the distinction between enhancement and denial was more salient than that between impression management and self-deception.

Individual- and Country-Level Variations of SDR

At the individual level, education and socioeconomic status have been found to be negatively related to SDR (e.g., Uziel, 2010). Both males and females have the tendency to attribute socially desirable characteristics to themselves (Press & Townsley, 1998), with females often reporting higher SDR than males (Barger, 2002). There were no gender differences found in subdimensions of SDR such as enhancement and denial (Ramanaiah et al., 1977). Given the inconclusive findings regarding gender, we do not specify a directional hypothesis about gender differences but explore these across cultural contexts.

Response styles in general have been found to be related to cultural values and personality traits (e.g., He, Bartram, Inceoglu, & van de Vijver, 2014; Smith, 2004). Lalwani, Shavitt, and Johnson (2006) reported that impression management, the main dimension tapped by the Marlowe–Crowne scale, was higher among collectivists than individualists. Schwartz et al. (1997) found a similar positive association between SDR and value types emphasizing social harmony in Finland and Israel. Trimble (1997) reported a positive association of SDR with intrinsic religiosity. Musek (2007) argued that SDR was positively related to the general factor of personality, a combination of the Big Five traits. At the country level, SDR was reported to be negatively associated with country affluence and individualism (Johnson & van de Vijver, 2003). Van Hemert et al. (2002) studied the Lie Scale from the Eysenck Personality Questionnaire (which is traditionally associated with SDR) in a cross-cultural meta-analysis, in which they confirmed the associations of the Lie Scale with affluence and individualism, and they reported a positive association with embeddedness measured with the Schwartz Value Survey. They also found a positive correlation with emotional stability and a negative one with extroversion at the country level. If SDR indeed reflects valid individual and cultural differences, we expect that the aggregated values, beliefs, and personality traits shared by individuals in each country are associated with SDR at the country level.

The Present Study

It has been argued that it is difficult, if not impossible, to find scalar invariance of SDR across cultures (e.g., Smith, 2009), given that what is considered desirable varies from culture to culture. Comparing students from Singapore and the United States, Li and Reb (2009) found weak support for the cross-cultural invariance of SDR in a multigroup confirmatory factor analysis; a similar conclusion was reached in a nine-country study (Verardi et al., 2010). With large-scale cross-cultural data, it is common to find nonequivalence, and the underlying reasons are often unclear (Byrne & van de Vijver, 2010). It could be due to misspecification of the constructs in a few countries, accumulated small (even inconsequential) differences in parameters, or a combination of both. Some researchers argue that measurement invariance constraints in multigroup confirmatory factor analysis (i.e., invariance of loadings and intercepts) may be overly restrictive and that we need to allow for psychologically inconsequential variation in these parameters, as done in Bayesian Structural Equation Modeling (Muthen & Asparouhov, 2012). We wanted to avoid these fit problems and did not want to use Bayesian Structural Equation Modeling given the lack of experience with its usage in empirical projects. Therefore, we resorted to an exploratory factor analysis approach (Costello & Osborne, 2005; Helmes & Holden, 2003) to study the factor structure of SDR with an adapted scale. The equivalence of the structure in different cultures was checked by means of calculations of Tucker’s phi which is the congruence index of two sets of factor solutions (van de Vijver & Leung, 1997), followed by a differential item functioning (DIF) analysis to tease out items not suitable for cross-cultural comparisons.

We reasoned that one cause for the poor replicability of the factor structure of the Marlowe–Crowne scale is ambiguity in some items and potential inapplicability of some items in different cultures or with different populations. For instance, the original item “I never make a long trip without checking the safety of my car” does not apply to most people in less developed countries or to university students who do not own a car. Moreover, the wording in some items is redundant and outdated (e.g., “I don’t find it particularly difficult to get along with loud mouthed, obnoxious people”), which poses challenges in precise translation to other languages. Given that some original items had limited discriminatory ability, various shortened versions of this scale have been proposed and validated (e.g., Reynolds, 1982; Strahan & Gerbasi, 1972). Similar to these previous studies, we used shortened and adapted items in the present study. Moreover, we aimed at maximizing cross-cultural comparability by adapting items.

Finally, to better understand the underlying mechanism of SDR, we studied the individual- and country-level correlates of SDR measured with this adapted scale in a multilevel design, taking into consideration data dependency at both levels. Specifically, we explored gender differences of SDR in the cross-cultural contexts and replicated and extended the study of the effects of country affluence, values, beliefs and personality traits on individual SDR.

Method

Participants

Participants were 3,471 university students with an age range of 17 to 35 in 20 countries. The mean age of these participants was 21.59 years (SD = 3.38). Thirty-one percent of the respondents were males. The sample size per country ranged from 95 to 389. The demographics are presented in Table 1.

Table 1.

Demographics of the Participants.

Country	Sample size	Mean age (SD)	Percentage of males	Language	Collection mode
Bulgaria	194	20.53 (2.27)	23	Bulgarian	1
China	374	21.12 (2.45)	48	Chinese	1 and 2
France	389	19.05 (1.73)	20	French	2
Germany	102	23.95 (3.01)	24	German	1
Greece	167	25.60 (4.67)	32	Greek	1
Indonesia	150	19.93 (1.08)	17	English	2
Israel	98	27.70 (3.92)	36	Hebrew	1
Italy	220	21.27 (0.63)	36	Italian	2
Kenya	157	22.04 (2.63)	39	English	2
Mexico	131	21.50 (3.94)	18	Spanish	1
Netherlands	199	19.85 (2.34)	21	Dutch	2
New Zealand	153	18.69 (1.91)	29	English	1
Portugal	117	26.16 (4.73)	23	Portuguese	1
Romania	193	22.53 (2.85)	17	Romanian	1
Singapore	148	22.50 (2.09)	47	English	1
South Africa	166	19.68 (1.60)	27	English	1
Spain	106	20.77 (3.80)	36	Spanish	1
Togo	201	22.48 (2.59)	50	French	2
Turkey	95	23.49 (3.03)	22	Turkish	1
United States	111	23.40 (3.48)	16	English	1

Note. Collection mode: 1 = online; 2 = paper and pencil.

Measures

The social desirability scale used in the present study was a shortened and simplified version of the Marlowe–Crowne scale. We selected items from the original scale with two criteria: Items should not have ambiguous meaning and items should be appropriate in different cultural contexts. In addition, we simplified the original items to improve the cross-cultural comparability with translated versions. For instance, the original item “there has been times when I was quite jealous of the good fortune of others” was rephrased as “I am jealous of others with good fortune.” There were 15 items in total, of which 9 items were worded as desirable attributes or behaviors (e.g., “I help others in trouble”), and 6 items were worded as undesirable attributes or behaviors (e.g., “I gossip”). All items were formulated affirmatively to avoid artifacts from using negation (item keying). The wording comparison of the original items and the adapted items is presented in Table 2.

Table 2.

Factor Structure of the 15-Item Social Desirability Scale With the Pooled Sample.

Original Marlowe–Crowne items	Items adapted and used in the current study	Enhancement	Denial
Before voting I thoroughly investigate the qualifications of all the candidates.	I think about my options before I make a choice	.53	−.06
I never hesitate to go out of my way to help someone in trouble	I help others in trouble	.58	−.16
It is sometimes hard for me to go on with my work if I am not encouraged	I continue with my work if I am motivated	.53	.04
On occasion I have had doubts about my ability to succeed in life	I have doubts about my ability to succeed in life	−.04	.43
I sometimes feel resentful when I don’t get my way	I am satisfied when I get my way	.50	.34
I am always careful about my manner of dress	I am careful about my way of dressing	.47	.04
I like to gossip at times	I gossip	.22	.59
No matter who I’m talking to, I’m always a good listener	I am a good listener	.55	−.10
I sometimes try to get even rather that forgive and forget	I forgive others for their wrongdoings	.39	−.19
When I don’t know something I don’t at all mind admitting it	I admit when I do not know something	.47	−.18
At times I have really insisted on having things my own way	I do things my way	.42	.19
I would never think of letting someone else be punished for my wrongdoings	I let someone else be punished for my wrongdoings	−.17	.56
There have been times when I was quite jealous of the good fortune of others	I am jealous of others with good fortune	−.01	.65
I am sometimes irritated by people who ask favors of me	I am irritated by people who ask favors	−.06	.53
I have never deliberately said something that hurt someone’s feelings	I say things that hurt others’ feelings	−.07	.57

To obtain sufficient psychometric details and allow more nuances in responses to this shortened scale, a 5-point Likert-type response format ranging from 1 = strongly disagree to 5 = strongly agree was used. The original dichotomized (i.e., true vs. false) and the Likert-type scale responses to the Marlowe–Crowne scale have been shown to be comparable (Domínguez Espinosa, Salas Menotti, & Reyes-Lagunes, 2008), and support was found that a shortened version with the Likert- type scale format did not result in substantial loss of information (Thunholm, 2001). Participants also provided demographic information such as their age and gender.

Country affluence was measured by the Human Development Index (HDI). The HDI is a composite measure of the average achievements in a country in three basic dimensions of human development: a long and healthy life, access to knowledge, and a high standard of living; it has a score range from 0 to 1, with a higher value indicating a higher level of development (United Nations, 2012). Data for all 20 countries were available (denoted as N = 20).

Scores of country values, beliefs, and personality were extracted from various large-scale international studies. Scores of Hofstede’s four cultural value dimensions (N = 17), measured from working adults, were taken from Hofstede (2009). Scores of Schwartz’s seven culture-level value dimensions (N = 17), measured from teachers and students, were taken from Schwartz (2009). The five belief dimensions in Social Axioms (N = 15) with students samples from Leung and Bond (2004) were also included. Scores of Agreeableness, Conscientiousness, Emotion Stability, Extroversion, and Openness (N = 13) were taken from the Big Five Inventory (Schmitt, Allik, McCrae, & Benet-Martínez, 2007). The respondents were mainly university students and in a few countries community-based samples.

Procedures

Administration procedures varied across countries. For countries where English is not the native language, the scale was translated, checked, and administered in the native language. In China, Mexico, and Singapore, the scale was administered on its own; in France, Togo, and the Unites States, the scale was administered as part of a bigger study on education and career development; in all the other countries, the scale was administered as part of a bigger study on youth identity and well-being. In France, Indonesia, Italy, Kenya, the Netherlands, Togo, and the United States, data were collected with paper and pencil, whereas in the other countries an online survey was administered. In China, both data collection modes were used: 274 respondents filled out the survey online and 100 with paper and pencil. The participation of all the students was voluntary.

Results

We describe the findings in three parts. First, we report the cross-cultural equivalence of the social desirability scale between the two data collection modes (within the Chinese sample). Second, we describe the cross-cultural equivalence and the item bias across countries. Finally, we report the multilevel analysis of SDR addressing the effects of gender and country characteristics.

Mode Effects

To account for possible administration bias caused by different data collection modes (Dwight & Feigelson, 2000), we first compared the factor structure and item means of the scale between the online survey and the paper-and-pencil sample within China. Principal component analysis with direct Oblimin rotation in either sample supported a two-factor solution: with all the positively worded items loading on the first factor (i.e., enhancement) and all the negatively worded items on the second factor (i.e., denial). Structural equivalence was evaluated with Tucker’s phi (above .90 as acceptable and above .95 as excellent; van de Vijver & Leung, 1997). The values of Tucker’s phi for the two factors were 1.00 and 1.00 across modes, pointing to excellent structure invariance. Independent sample t tests on the 15 items between the collection modes showed no significant mean differences at p < .01. We then computed a DIF analysis using a multiple regression approach. Each enhancement item was predicted with the total score on the enhancement factor, administration mode, and the interaction of the scale score and the administration mode, and each denial item was predicted with the total score on the denial factor, administration mode, and their interaction. The effect of the administration mode indicated uniform bias and that of the interactions indicated non-uniform bias, both of which were evaluated by Cohen’s f² when adding each predictor in multiple regressions (Cohen, 1988). Items with f² values larger than .15 (lower bound of medium effect size) were flagged as having non-negligible DIF. In the current analysis, no items were found to display a medium or large effect size; the values of Cohen’s f² ranged from 0 to .01. To summarize, the results indicated that the social desirability data collected in these two modes measure the same constructs and can be compared directly.

Structural Equivalence Across Countries

Rather than computing all pairwise comparisons between countries, we chose for an approach in which we compared the factor structure of each country with the pooled sample (i.e., combining all participants and correcting for mean score differences on items by computing a weighted covariance matrix). Principal component analysis of the 15 items with direct Oblimin rotation was performed with the pooled sample (Table 2) and in each country. Two factors were extracted based on the scree plot, explaining 17% and 13% of the variance in the pooled sample, respectively. The first five eigenvalues in the pooled solution were 2.56, 1.90, 1.25, 0.99, and 0.95, respectively. All the positively worded items loaded on the enhancement factor and all the negatively worded items loaded on the denial factor. The explained variance of enhancement ranged from 15% to 21% and of denial from 11% to 15% in different countries. Similar to previous studies on the shortened Marlowe–Crowne scale (Verardi et al., 2010), the amount of variance explained in these factors was not high, possibly because response styles including SDR are broad, general tendencies, which can be difficult to capture wholly in questionnaire items. Structural equivalence was checked though comparing each country’s factor solution with the pooled solution using Tucker’s phi (Table 3). On average, the values of Tucker’s phi for the two factors were .92 and .92. All the other countries showed acceptable structural invariance except Bulgaria, which might be due to the fact that we sampled students from the National Sports Academy who were not on an academic track as students majoring in social sciences in other countries. It was likely that the low convergence of the factor structure in Bulgaria resulted from the lack of familiarity in such survey tasks. We excluded Bulgaria in the following analyses.

Table 3.

Values of Tucker’ Phi of the Factor Solutions Between Each Country and the Pooled Sample.

Country	Enhancement	Denial
Bulgaria	.50	.77
China	.98	.98
France	.98	.94
Germany	.96	.95
Greece	.91	.91
Indonesia	.85	.90
Israel	.88	.88
Italy	.98	.96
Kenya	.95	.94
Mexico	.97	.94
Netherlands	.96	.93
New Zealand	.96	.91
Portugal	.95	.89
Romania	.96	.87
Singapore	.92	.94
South Africa	.95	.94
Spain	.88	.93
Togo	.97	.95
Turkey	.93	.91
United States	.95	.92

DIF Analysis

Item bias across countries was examined using multiple regression analyses. We took South Africa as the reference group because the adapted scale was first tested and validated in English in South Africa (van de Vijver & Meiring, 2011). Specifically, each of the nine enhancement items was regressed on three blocks of variables: the deviance scale score of the enhancement factor, 18 country dummy variables, and 18 interactions between the country dummy variables and the deviance score of the enhancement factor. The same procedure was applied to the six denial items, which were predicted by the deviance score of the denial factor, the 18 country dummy variables, and the interactions between the two. To correct for the large number of predictors in the blocks, values of adjusted R² were used to calculate Cohen’s f². Two denial items were found to have uniform bias with medium effect sizes (Table 4). On the item “I am irritated by people who ask favors,” China, France, Germany, Indonesia, Israel, Italy, Mexico, New Zealand, Singapore, Togo, Turkey, and the United States showed higher intercepts. On the item “I say things that hurt others’ feelings,” China, France, Portugal, Romania, and Togo showed higher intercepts whereas Greece had a lower intercept. These DIF effects could be due to different connotations of the words “irritated” and “hurt” in the various languages. These two items were excluded from the following analyses.

Table 4.

Effect Sizes in Regression Analyses: Uniform and Non-Uniform Bias Detection.

	Uniform Bias f²	Non-Uniform Bias f²
Enhancement Items
I think about my options before I make a choice	.03	.01
I help others in trouble	.04	.00
I continue with my work if I am motivated	.04	.00
I am satisfied when I get my way	.10	.00
I am careful about my way of dressing	.10	.00
I am a good listener	.05	.00
I forgive others for their wrongdoings	.04	.01
I admit when I do not know something	.02	.00
I do things my way	.08	.00
Denial Items
I have doubts about my ability to succeed in life	.07	.00
I gossip	.10	.00
I let someone else be punished for my wrongdoings	.09	.01
I am jealous of others with good fortune	.14	.00
I am irritated by people who ask favors	.16	.00
I say things that hurt others’ feelings	.20	.00

The values of Cronbach’s alpha of the final enhancement (nine items) and denial scale (four items) were .62 and .54, respectively. The low reliability values were not unexpected; Beretvas, Meyers, and Leite (2002) in a reliability generalization study of the Marlowe–Crowne Social Desirability Scale reported that the estimated reliability of this scale was .53. Scale scores of the two subscales were calculated for each individual and aggregated to country level. The scores of the denial items were reverse coded, thus a higher score on this subscale representing a higher tendency to deny negative self-descriptions. The two subscales were weakly correlated at the individual level, r(3274) = .09, p < .01; their correlation at country level was nonsignificant, r(17) = −.10, p = .69, which could be due to the limited number of observations. The country scores of enhancement and denial are presented in Table 5.

Table 5.

Country Scores of Enhancement and Denial Across 19 Countries.

Country	Enhancement	Denial
China	3.83	2.57
France	3.98	2.55
Germany	4.07	2.50
Greece	3.83	2.64
Indonesia	4.07	2.12
Israel	4.03	2.43
Italy	3.83	2.62
Kenya	4.06	3.01
Mexico	4.11	2.56
Netherlands	3.99	2.44
New Zealand	3.95	2.05
Portugal	4.14	2.88
Romania	4.29	2.64
Singapore	4.27	1.98
South Africa	4.16	2.80
Spain	4.09	2.38
Togo	4.00	3.22
Turkey	4.07	2.63
The United States	4.12	2.45

Multilevel Analysis

Before applying the multilevel analysis, we correlated the country-level scores of enhancement and denial with affluence, values, beliefs, and personality traits. Due to the small sample sizes at the country level and skewed distributions of some variables, we resorted to bootstrapping and the significance level of the correlations was determined in 1,000 bootstrap samples (Table 6). Enhancement at country level was positively related to embeddedness and religiosity, and denial was negatively associated with HDI and positively associated with uncertainty avoidance, harmony, agreeableness, conscientiousness, and openness.

Table 6.

Country-Level Correlations of Enhancement and Denial With Affluence, Values, Beliefs, and Personality.

Country-level correlation	Enhancement	Denial
Human Development Index (N = 19)	−.07	−.60^a
Hofstede Values (N = 17)
Power distance	.25	.15
Individualism	−.24	.02
Masculinity	−.15	.13
Uncertainty avoidance	−.14	.60^a
Schwartz Values (N = 17)
Harmony	−.18	.34^a
Embeddedness	.37^a	−.13
Hierarchy	.03	−.21
Mastery	−.34	.20
Affective autonomy	−.20	−.12
Intellectual autonomy	−.28	.17
Egalitarianism	−.19	.18
Social Axioms (N = 15)
Social cynicism	.03	.23
Reward for application	.30	−.26
Social complexity	−.06	−.07
Fate control	.11	−.09
Religiosity	.30^a	−.22
Big Five Personality (BFI; N = 13)
Agreeableness	−.08	.55^a
Conscientiousness	.09	.60^a
Emotion stability	.38	−.03
Extroversion	.14	−.37
Openness	−.22	.32^a

Significance level established with 95% CI based on 1,000 bootstrap samples.

We tested the effects of gender and country characteristics on enhancement and denial in a multilevel design with HLM Version 6 (Raudenbush & Bryk, 2002). The intraclass correlation coefficient was 9% for enhancement and 17% for denial, suggesting sufficient variations at country level to conduct multilevel analyses (van de Vijver & Poortinga, 2002). In accordance with Enders and Tofighi (2007), we centered the country-level predictors (converted to the standardized z scores) on the grand mean. We entered one predictor per analysis. All multilevel analyses employed a random intercept and a fixed slope (Table 7).

Table 7.

Coefficients From Multilevel Analyses.

Predictor	Enhancement	Denial
Gender (male)	−.08**	.03**
Human Development Index (N = 19)	−.01	−.15**
Hofstede Values (N = 17)
Power distance	.03	.04
Individualism	−.03	.01
Masculinity	−.03	.04
Uncertainty avoidance	−.02	.13**
Schwartz Values (N = 17)
Harmony	−.03	.08
Embeddedness	.05	−.03
Hierarchy	.00	−.05
Mastery	−.05	.05
Affective autonomy	−.03	−.03
Intellectual autonomy	−.04	.04
Egalitarianism	−.03	.04
Social Axioms (N = 15)
Social cynicism	.00	.06
Reward for application	.04	−.06
Social complexity	−.01	−.02
Fate control	.02	−.02
Religiosity	.04	−.05
Big Five Personality (BFI; N = 13)
Agreeableness	.00	.12*
Conscientiousness	.03	.19*
Emotion stability	.06	−.02
Extroversion	.02	−.10
Openness	−.02	.15

Note. N stands for the number of countries in the analysis.

p < .10. **p < .01.

We first checked the effects of differences in data collection modes (online vs. paper and pencil) on enhancement and denial and found no significant differences. Compared with females, males scored lower on enhancement and higher on denial, indicating that there were differential effects of gender on the subdimensions of SDR. HDI was negatively associated with denial; its association with enhancement was nonsignificant yet in the expected (negative) direction.

Caution is needed in the interpretation of the country-level results, because the number of countries available for analyses with cultural values and personality traits were smaller (e.g., in most cases these data were not available for Kenya and Togo), which limited the cross-cultural variations that we could study. Albeit nonsignificant, power distance, hierarchy, and emotion stability showed positive associations, and individualism and autonomy showed negative associations with enhancement, which well replicated the findings from van Hemert et al. (2002). Denial was predicted by agreeableness and conscientiousness. The combined evidence from the country-level correlations suggests that both enhancement and denial were related to cultural values and personality traits pertaining to “fitting in” and the two dimensions might be related to different aspects of “fitting in.”

Discussion

We studied the factor structure, structural equivalence, and cross-country variations of a shortened and simplified Marlowe–Crowne Social Desirability Scale among university students in 20 countries. Our findings supported a two-dimensional structure of SDR, distinguishing the endorsement of positive self-description (enhancement) and the avoidance of negative self-description (denial). The structure was largely invariant across countries. There were gender differences in the two dimensions: Enhancement was stronger among females and denial among males. A similar finding was reported by Sutton and Farrall (2005). It seems that there is a general difference in impression management in which females make more efforts to create a positive impression. There were more cross-country variations in denial than enhancement, and HDI was the most significant predictor for denial, which is in line with previous studies (Johnson & van de Vijver, 2003).

There has been much debate on the factor structure of SDR. Our study confirmed that SDR is a multidimensional construct. The distinction of enhancement and denial that we found in multiple countries is unlikely to be an artifact of item wordings, because all the items are formulated as affirmation of either positive or negative traits, emotions, and behaviors (Paulhus & Reid, 1991). The two aspects of SDR seem to be triggered by similar cultural mechanisms (i.e., fitting in). Given the small value of intraclass coefficients and the few significant predictors for enhancement, it seems that attributing positive traits to oneself is rather universal and is not much under cross-culturally differential control; however, denial has more cross-cultural variations, as people in countries low in affluence and high in agreeableness and conscientiousness seem to have a higher tendency to deny negative self-descriptions. These country-level correlates were similar to those found in a general response style (with SDR and extreme response style as positive indicators and acquiescent and midpoint response styles as negative indicators; He & van de Vijver, 2013; He, van de Vijver, Domínguez Espinosa, & Mui, 2014), suggesting that SDR, as part of a general response style factor, can be interpreted as a means of response amplification motivated by “fitting in.”

Our study has a few limitations. We used data of SDR from various bigger projects with different administration modes; thus, we did not have data on other constructs available in all countries that could be used to study the convergent and divergent validity of SDR at the individual level. The university student sample may not be equally representative in each culture. In particular, access to higher education in less developed countries (e.g., Togo) is largely restricted to elites who do not necessarily reflect the values of the general population. In Bulgaria, students from the non-academic track were sampled, which caused some incomparability with other countries. Future efforts should ensure the comparability of samples across cultures and replicate the study with different conditions that may motivate or demotivate respondents to present themselves in a positive light (e.g., employment selection process). Yet we confirmed an equivalent structure of SDR within 19 countries and meaningful country-level correlates. Our study has important implications for cross-cultural research. First, we found that SDR measured by the adapted Marlowe–Crowne scale has a positive and a negative component that are weakly related to one another but that do not show the same gender differences. Second, we find some systematic cross-cultural differences in enhancement and denial, which provides a piece to the puzzle as to whether SDR is a nuisance or substance. We argue that SDR has at least some substantive meaning (McCrae & Costa, 1983) as the two dimensions, especially the denial dimension, are influenced by country affluence, cultural values, and personality traits pertinent to “fitting in.” In such a case, removing the effects of SDR can erroneously eliminate valid variations between individuals and cultures.

Footnotes

Authors’ Notes

An earlier version of the article was discussed in the 2013 Writing Week of the Early Researcher Union of the European Association of Developmental Psychology.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We acknowledge the support from the European Association of Developmental Psychology.

Author Biographies

Jia He is a PhD researcher in Tilburg University, the Netherlands. She obtained her MA degree in Intercultural Communication from Shanghai International Studies University, China. Her current research includes the psychological meaning of survey response styles, values, social desirability, and other methodological aspects of cross-cultural studies. She is also interested in modern research methods such as structural equation modeling and multilevel analysis.

Fons J. R. van de Vijver has in a chair in cross-cultural psychology at Tilburg University in the Netherlands. He has (co-)authored more than 400 publications, mainly in the domain of cross-cultural psychology. The main topics in his research involve bias and equivalence, psychological acculturation and multiculturalism, cognitive similarities and differences, response styles, translations, and adaptations. He is president-elect of the International Association for Cross-Cultural Psychology.

Alejandra Dominguez Espinosa is the current coordinator of the PhD program and the leading researcher of the Culture, Emotions and Gender Laboratory at the Iberoamerican University, Mexico City. Her current areas of research involve impression management strategies and acculturation.

Amina Abubakar is a visiting research fellow at Tilburg University in the Netherlands. Her main research topics include test development in cross-cultural context and child development in adverse conditions.

Radosveta Dimitrova is a COFAS Marie Curie Fellow at Stockholm University, Sweden. She holds a PhD in developmental psychology from Trieste University, Italy (2009 Best Doctoral Thesis Award, Italian Psychological Association), and a PhD in cross-cultural psychology from Tilburg University, the Netherlands (2012 SECC Outstanding Dissertation Award, Society for Research in Child Development). She is the president of the Early Researchers Union (ERU) of the European Association of Developmental Psychology (EADP).

Byron G. Adams is an industrial psychologist from Johannesburg, South Africa. He lectures industrial psychology at the University of Johannesburg and will soon start a post-doctoral fellowship at Tilburg University. His broad research interests include personality, psychological assessment, transformation, and organizational behavior. He has a more specific interest in identity, and its associations with well-being and intergroup relations. His objective is to broaden theoretical arguments and to provide empirical evidence for a model of identity that integrates personal, social, and relational dimensions.

Arzu Aydinli is a PhD student in social and cross-cultural psychology and is enrolled in a joint doctorate program between Tilburg University (the Netherlands) and Koç University (Istanbul, Turkey). Her current research interest revolves around identifying motivational pathways leading to helping and volunteering, from a cross-cultural perspective. Moreover, she is interested in studying acculturation, ethnic identity and life satisfaction among bicultural adolescents.

Kokou Atitsogbe is a licensed psychologist in career counseling at the National Center for Career Information and Vocational Guidance in Lomé (Togo).

Itziar Alonso-Arbiol is an associate professor at the Department of Personality, and Psychological Assessment and Treatments of the University of Basque Country, Spain. Her main research topics include psychological assessment, attachment, and other emotion processes in relational and cross-cultural contexts.

Magdalena Bobowik has graduated in psychology at Warsaw University and completed her PhD in 2013 at the University of the Basque Country. Currently, she works as a post-doctoral researcher at the University of the Basque Country. Her main research interests cover social identity and intergroup relations, psychology of immigration, cross-cultural psychology, political psychology, and positive psychology.

Ronald Fischer is an associate professor in Victoria University of Wellington, New Zealand. His research is broadly situated in cultural and applied social psychology.

Venzislav Jordanov has received his PhD in sports psychology at the Bulgarian National Sports Academy, Department of Psychology, Pedagogy and Sociology in 2012. Currently, he is a lecturer at the University of World and National Economy (UWNE) in Sofia, Bulgaria. His main research interests revolve around topics of sport psychology, basketball, and cross-cultural psychology. Since 2009, he has been actively involved in the study of identity, acculturation, and well-being of youth and young adults in Bulgaria.

Stefanos Mastrotheodoros is a PhD student in clinical psychology in the University of Athens, Greece. In his PhD project, he studies attachment styles, identity development, and adaptation in late adolescence, using a longitudinal design. He has worked in many research projects of resilient immigrant youth adaptation in Greece. He is a clinical psychologist and has also been trained in cognitive therapy.

Félix Neto is a professor of psychology at the University of Porto, Portugal. His research interests include cross-cultural psychology and social psychology (especially migration, subjective well-being, and forgiveness). He has published 18 books and about 250 scientific articles.

Yael J. Ponizovsky is a social worker (MSW) and a PhD student at the School of Social Work and Social Welfare, Hebrew University of Jerusalem, Israel. She is specializing in the study of immigrant children and adolescents’ well-being and adjustment from a resilience perspective. Moreover, she is interested in the study of children, youth, and families coping with change in the context of cultural encounters.

Jochen Reb is an associate professor of organizational behavior and human resources at the Lee Kong Chian School of Business, Singapore Management University. His research focuses on two main areas: judgment and decision making in organizations, and the role of mindfulness in organizational contexts such as leadership and performance. His work has been published in leading academic journals such as the Journal of Applied Psychology, Journal of Management, Organizational Behavior and Human Decision Processes, Personality and Social Psychology Bulletin, and Personnel Psychology. He currently serves on the editorial board of Organizational Behavior and Human Decision Processes, Journal of Management, and Journal of Business and Psychology.

Samantha Sim is currently a PhD student at Singapore Management University, Lee Kong Chian School of Business, organizational behavior and human resource department. Her research interests include compassion at the workplace, organizational justice, and cross-cultural differences.

Laurent Sovet is a PhD candidate in career counseling psychology. His research is focused on the interplay between career choices, subjective well-being, and cross-cultural differences.

Delia Stefenel obtained her doctorate at University of Bucharest, Romania, with a dissertation on intercultural communication and acculturation. She is working as teaching assistant at “Lucian Blaga” University of Sibiu, Romania. Her main research interests include migrants’ acculturation, psychological adjustment and entrepreneurship, identity, cross-cultural communication.

Angela O. Suryani gained her doctoral degree at Department of Cross-Cultural Psychology at Tilburg University, and now she is working at Faculty of Psychology, Atma Jaya Catholic University in Jakarta, Indonesia. Her studies are about leadership in Indonesia involving a mixed-methods, indigenous and comparative cross-cultural perspectives, and psychological test construction/adaptation.

Ergyul Tair is an associate professor at the Department of Psychology of the Institute for Population and Human Studies by Bulgarian Academy of Sciences, Bulgaria. The main topics in her research included different personality traits, psychological well-being and acculturation, and motives and values in cross-cultural contexts.

Arnaud Villieux is an associate professor in occupational psychology in Rouen University. His current research involves well-being and personality measurement in organizational contexts.

References

Barger

S. D.

(2002). The Marlowe-Crowne affair: Short forms, psychometric structure, and social desirability. Journal of Personality Assessment, 79, 286-305. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1207/s15327752jpa7902_11

Beretvas

S. N.

Meyers

J. L.

Leite

W. L.

(2002). A reliability generalization study of the Marlowe-Crowne Social Desirability Scale. Educational and Psychological Measurement, 62, 570-589. doi:10.1177/0013164402062004003

Byrne

B. M.

van de Vijver

F. J. R.

(2010). Testing for measurement and structural equivalence in large-scale cross-cultural studies: Addressing the issue of nonequivalence. International Journal of Testing, 10, 107-132. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1080/15305051003637306

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

Costello

A. B.

Osborne

J. W.

(2005). Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis. Practical Assessment, Research and Evaluation, 10, 1-9.

Crowne

D. P.

Marlowe

(1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354.

Crowne

D. P.

Marlowe

(1964). The approval motive: studies in evaluative dependence. New York, NY: John Wiley.

Domínguez Espinosa

Salas Menotti

Reyes-Lagunes

(2008). Validez Concurrente de la Escala de Deseabilidad Social de Domínguez utilizando la Escala de Deseabilidad Social de Marlowe-Crowne [Concurrent validity of the Scale of Social Desirability of Domínguez using the scale of social desirability of Marlowe-Crowne]. Revista Iberoamericana de Diagnóstico y Evaluación Psicológica, 25, 125-139.

Dwight

S. A.

Feigelson

M. E.

(2000). A quantitative review of the effect of computerized testing on the measurement of social desirability. Educational and Psychological Measurement, 60, 340-360. doi:10.1177/00131640021970583

10.

Enders

C. K.

Tofighi

(2007). Centering predictor variables in cross-sectional multilevel models: A new look at an old issue. Psychological Methods, 12, 121-138. doi:10.1037/1082-989x.12.2.121

11.

Bartram

Inceoglu

van de Vijver

F. J. R.

(2014). Response styles and personality traits: A multilevel analysis. Journal of Cross-Cultural Psychology, 45, 1028-1045. doi:10.1177/0022022114534773

12.

van de Vijver

F. J. R.

(2013). A general response style factor: Evidence from a multi-ethnic study in the Netherlands. Personality and Individual Differences, 55, 794-800. doi:10.1016/j.paid.2013.06.017

13.

van de Vijver

F. J. R.

Domínguez Espinosa

Mui

P. H. C.

(2014). Acquiescent, extreme, and midpoint response styles: A multilevel study. International Journal of Cross-Cultural Management. doi:10.1177/1470595814541424

14.

Helmes

Holden

R. R.

(2003). The construct of social desirability: One or two dimensions? Personality and Individual Differences, 34, 1015-1023. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/S0191-8869(02)00086-7

15.

Hofstede

(2009). Dimension data matrix. Retrieved from http://www.geerthofstede.eu/dimension-data-matrix

16.

Johnson

T. P.

van de Vijver

F. J. R

. (2003). Social desirability in cross-cultural research. In Harkness

J. A.

van de Vijver

F. J. R.

Mohler

P. P.

(Eds.), Cross-cultural survey methods (pp. 195-204). Hoboken, NJ: John Wiley.

17.

Lalwani

A. K.

Shavitt

Johnson

(2006). What is the relation between cultural orientation and socially desirable responding? Journal of Personality and Social Psychology, 90, 165-178. doi:10.1037/0022-3514.90.1.165

18.

Leite

W. L.

Beretvas

S. N.

(2005). Validation of scores on the Marlowe-Crowne Social Desirability Scale and the Balanced Inventory of Desirable Responding. Educational and Psychological Measurement, 65, 140-154. doi:10.1177/0013164404267285

19.

Leung

Bond

M. H.

(2004). Social Axioms: A model for social beliefs in multicultural perspective. Advances in Experimental Social Psychology, 36, 119-197. doi:10.1016/s0065-2601(04)36003-x

20.

Bagger

(2006). Using the BIDR to distinguish the effects of impression management and self-deception on the criterion validity of personality measures: A meta-analysis. International Journal of Selection and Assessment, 14, 131-141. doi:10.1111/j.1468-2389.2006.00339.x

21.

Reb

(2009). A cross-nations, cross-cultures, and cross-conditions analysis on the equivalence of the Balanced Inventory of Desirable Responding. Journal of Cross-Cultural Psychology, 40, 214-233. doi:10.1177/0022022108328819

22.

Loo

Loewen

(2004). Confirmatory factor analyses of scores from full and short versions of the Marlowe–Crowne Social Desirability Scale. Journal of Applied Social Psychology, 34, 2343-2352. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1111/j.1559-1816.2004.tb01980.x

23.

McCrae

R. R.

Costa

P. T.

(1983). Social desirability scales: More substance than style. Journal of Consulting and Clinical Psychology, 51, 882-888. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1037/0022-006x.51.6.882

24.

Millham

(1974). Two components of need for approval score and their relationship to cheating following success and failure. Journal of Research in Personality, 8, 378-392. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/0092-6566(74)90028-2

25.

Musek

(2007). A general factor of personality: Evidence for the Big One in the five-factor model. Journal of Research in Personality, 41, 1213-1233. doi:10.1016/j.jrp.2007.02.003

26.

Muthen

Asparouhov

(2012). Bayesian SEM: A more flexible representation of substantive theory. Psychological Methods, 17, 313-335. doi:10.1037/a0026802

27.

Nederhof

A. J.

(1985). Methods of coping with social desirability bias: A review. European Journal of Social Psychology, 15, 263-280. doi:10.1002/ejsp.2420150303

28.

Paulhus

D. L.

(1984). Two-component models of socially desirable responding. Journal of Personality and Social Psychology, 46, 598-609. doi:10.1037/0022-3514.46.3.598

29.

Paulhus

D. L.

(1991). Measurement and control of response biases. In Robinson

Shaver

Wrightsman

(Eds.), Measures of personality and social psychological attitudes (Vol. 1, pp. 17-59). San Diego, CA: Academic Press.

30.

Paulhus

D. L.

(2002). Socially desirable responding: The evolution of a construct. In Braun

H. I.

Jackson

D. N.

Wiley

D. E.

(Eds.), The role of constructs in psychological and educational measurement (pp. 49-69). Mahwah, NJ: Lawrence Erlbaum.

31.

Paulhus

D. L.

Reid

D. B.

(1991). Enhancement and denial in socially desirable responding. Journal of Personality and Social Psychology, 60, 307-317. doi:10.1037/0022-3514.60.2.307

32.

Press

J. E.

Townsley

(1998). Wives’ and husbands’ housework reporting: Gender, class, and social desirability. Gender and Society, 12, 188-218. doi:10.2307/190530

33.

Ramanaiah

N. V.

Schill

Leung

L. S.

(1977). A test of the hypothesis about the two-dimensional nature of the Marlowe-Crowne Social Desirability scale. Journal of Research in Personality, 11, 251-259. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/0092-6566(77)90022-8

34.

Raudenbush

S. W.

Bryk

A. S.

(2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Newbury Park, CA: Sage.

35.

Reynolds

W. M.

(1982). Development of reliable and valid short forms of the Marlowe-Crowne social desirability scale. Journal of Clinical Psychology, 38, 119-125.

36.

Schmitt

D. P.

Allik

McCrae

R. R.

Benet-Martínez

(2007). The geographic distribution of Big Five personality traits. Journal of Cross-Cultural Psychology, 38, 173-212. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1177/0022022106297299

37.

Schütz

(1998). Assertive, offensive, protective, and defensive styles of self-presentation: A taxonomy. The Journal of Psychology, 132, 611-628. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1080/00223989809599293

38.

Schwartz

S. H.

(2009). Culture matters: National value cultures, sources, and consequences. In Wyer

R. S

Chiu

C.-y.

Hong

Y.-y.

(Eds.), Understanding culture: Theory, research, and application (pp. 127-150). New York, NY: Psychology Press.

39.

Schwartz

S. H.

Verkasalo

Antonovsky

Sagiv

(1997). Value priorities and social desirability: Much substance, some style. British Journal of Social Psychology, 36, 3-18. doi:10.1111/j.2044-8309.1997.tb01115.x

40.

Smith

P. B.

(2004). Acquiescent response bias as an aspect of cultural communication style. Journal of Cross-Cultural Psychology, 35, 50-61. doi:10.1177/0022022103260380

41.

Smith

P. B.

(2009). On finding improved ways of characterizing national cultures. In Wyer

R. S

Chiu

C.-y.

Hong

Y.-y.

(Eds.), Understanding culture: Theory, research, and application (pp. 151-161). New York, NY: Psychology Press.

42.

Strahan

Gerbasi

K. C.

(1972). Short, homogeneous versions of the Marlow-Crowne Social Desirability Scale. Journal of clinical psychology, 28, 190-193.

43.

Sutton

R. M.

Farrall

(2005). Gender, socially desirable responding and the fear of crime: Are women really more anxious about crime? British Journal of Criminology, 45, 212-224. doi:10.1093/bjc/azh084

44.

Thunholm

(2001). Social desirability in personality testing of military officers. Military Psychology, 13, 223-234. doi:10.1207/S15327876MP1304_3

45.

Trimble

D. E.

(1997). The Religious Orientation Scale: Review and meta-analysis of social desirability effects. Educational and Psychological Measurement, 57, 970-986. doi:10.1177/0013164497057006007

46.

United Nations. (2012). UN data. Human Development Index. Retrieved from http://data.un.org/DataMartInfo.aspx#15

47.

Uziel

(2010). Rethinking social desirability scales: From impression management to interpersonally oriented self-control. Perspectives on Psychological Science, 5, 243-262. doi:10.1177/1745691610369465

48.

van de Vijver

F. J. R.

Leung

(1997). Methods and data analysis of comparative research. Thousand Oaks, CA: Sage.

49.

van de Vijver

F. J. R.

Meiring

(2011). Social desirability among Blacks and Whites in South Africa. Cross-Cultural Psychology Symposium, Tilburg, The Netherlands, Unpublished paper.

50.

van de Vijver

F. J. R.

Poortinga

Y. H.

(2002). Structural equivalence in multilevel research. Journal of Cross-Cultural Psychology, 33, 141-156. doi:10.1177/0022022102033002002

51.

van Hemert

D. A.

van de Vijver

F. J. R.

Poortinga

Y. H.

Georgas

(2002). Structural and functional equivalence of the Eysenck Personality Questionnaire within and between countries. Personality and Individual Differences, 33, 1229-1249. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/s0191-8869(02)00007-7

52.

Ventimiglia

MacDonald

D. A.

(2012). An examination of the factorial dimensionality of the Marlowe Crowne Social Desirability Scale. Personality and Individual Differences, 52, 487-491. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/j.paid.2011.11.016

53.

Verardi

Dahourou

Ah-Kion

Bhowon

Tseung

C. N.

Amoussou-Yeye

Rossier

(2010). Psychometric properties of the Marlowe-Crowne Social Desirability Scale in eight African countries and Switzerland. Journal of Cross-Cultural Psychology, 41, 19-34. doi:10.1177/0022022109348918