Method

Results

In regard to the NARQ reliability coefficients, the results were similar to the ones found in other Spanish speaking countries. In comparison with the results from ^{Doroszuk et al. (2019}) in Chile, Spain and Colombia, the Cronbach’s Alpha coefficients for the complete scales were similar to the Venezuelan sample. Specifically, narcissistic admiration showed an Alpha value of .823 and narcissistic rivalry of .803. These values show that the measurements made with both scales have an adequate reliability, also, these results are similar to the ones reported by ^{Doroszuk et al. (2019)}. Descriptive and reliability estimates for all instruments are included in Table 1.

Table 1: Reliability coefficients and descriptive statistics for general strategies and subcomponents of NARQ, Big-Five Personality and Self-Esteem 

^Notes: α = Cronbach’s Alpha; M = arithmetic mean; SD = standard deviation; As = asymmetry; K = kurtosis. All statistics were calculated with a N = 159 for all variables.

Subsequently, the Factor program (^{Lorenzo-Seva & Ferrando, 2013}) was used in order to carry out an exploratory factor analysis; specifically, a polychoric correlation matrix was used, given the ordinal level of measurement present in psychological tests with Likert type items (^{Lloret-Segura, Ferreres-Traver, Hernández-Baeza & Tomás-Marco, 2014}). Including the 18 items from the questionnaire, both the KMO (0.858) and the Bartlett’s sphericity test (χ² ₍₁₅₃₎ = 1751.9; p < .001) pointed out the adequacy of the data to carry out a factorial analysis. The number of factors to be extracted was estimated through a parallel analysis (^{Horn, 1965}), only two factors showed a percentage of explained variance superior to the 95^th percentile of the average random explained variance, achieving between these two factors 54.8 % of total variance.

For extracting these factors, the robust unweighted least squares (RULS) method was used, in conjunction with a Robust Promin factor rotation method (^{Lorenzo-Seva & Ferrando, 2018}), following the recommendations from ^{Lloret-Segura et al. (2014}). The resulting factorial structure matched the expected results (see Table 2), given that the items 4, 6, 9, 10, 11, 12, 13, 14 and 17 corresponded to the first factor, narcissistic rivalry; while the remaining items 1, 2, 3, 5, 7, 8, 15, 16 and 18 corresponded to the second factor, narcissistic admiration. It’s important to remark that two items showed high factorial loadings with both factors: item 4, from rivalry, and item 18, from admiration. The first one presented positive and very similar factorial loadings, although slightly higher in its proper factor; the second item presented opposed factorial loadings in each factor, although the positive correlation was maintained with its proper factor.

Table 2: Rotated factorial loadings matrix for NARQ 

^Note: Values equal to or below .30 were omitted.

In regard to the correlation between variables, it was found that the rivalry strategy is negatively associated with the extraversion and agreeableness traits and self-esteem. In the case of the admiration strategy, this one showed positive associations with extraversion, openness, affability and responsibility traits and self-esteem. In most cases, these correlations were low, except the association between rivalry and affability (r = -.45), which is moderate (see Table 3).

Table 3: Pearson product-moment correlations for age, NARQ, Self-Esteem and Big-Five Personality 

^Notes: All correlations were calculated with N = 159. *The correlation is significant at the .05 level (two-tailed). ** The correlation is significant at the .01 level (one-tailed).

As to the differences generated by sex, there were not statistically significant variations in the general scales of rivalry and admiration; nonetheless, there were differences in the striving for supremacy and devaluation subscales, both which correspond to the general scale of narcissistic rivalry; the first one is significant at the 5 % level and the second one at the 10 % level (see Table 4). For both subscales, the men showed a higher average score, although, this difference size is low. It is important to remark that the statistical power of the test is also low (1-β < 0.8), due to the discrepancy between the number of participants from both sexes, so these results must be interpreted with caution, and expanded upon in balanced samples.

Table 4: Descriptive statistics for general strategies and subcomponents of NARQ, with a sex mean difference test 

^Notes: M (CI95 %) = arithmetic mean with confidence intervals at 95%; SD = standard deviation; t = independent samples Student’s t test; df= degrees of freedom; Sig. = statistical significance; d = Cohen’s d effect size coefficient; 1-β = statistical power. a Feminine n = 118, Masculine n = 41. b Equality of variance not assumed, by the Levene test.

Comparing these results obtained from the Venezuelan sample with the ones from Spain, Chile and Colombia (^{Doroszuk et al., 2019}), the following differences were found: in regard to narcissistic admiration, there were significant differences between these countries, although the effect size was low (η ² = 0.06). Specifically, the Venezuelan sample showed higher levels of admiration than the Spanish one, but lower than the Colombian sample. In comparison with de Chilean results, there were no differences with the ones obtained in Venezuela (see Table 5).

For narcissistic rivalry, again significant differences were found, although these were low (η ² = 0.016). Nonetheless, the pattern for this variable was different: the data from the Venezuelan sample only showed differences with the ones from Colombia, given that the Colombians presented lower levels of rivalry. With the Chilean and Spanish samples, there were no significant differences when compared with the Venezuelan one.

Table 5: Descriptive statistics and ANOVA test for narcissistic rivalry and admiration between Spanish, Chilean, Colombian and Venezuelan samples 

^Notes:M (CI95 %) = arithmetic mean with confidence intervals at 95 % level; SD = standard deviation; N = sample size; F = Fisher’s F test for independent samples; df = degrees of freedom (numerator/denominator); Sig. = two-tailed statistical significance; η ² = etha-squared effect size coefficient. ^a The data from Spain, Chile and Colombia were obtained from ^{Doroszuk et al. (2019}).

Discussion

The principal research objective was to carry out a first approximation to prove the psychometric adequacy of the Spanish version of the Narcissistic Admiration and Rivalry Questionnaire (NARQ), developed by ^{Doroszuk et al. (2019}), applied to a sample of Venezuelan college students.

In regard to the reliability indexes of the questionnaire, it can be stated that both Rivalry and Admiration scales provide reliable measures, given .80 criteria for Cronbach’s Alpha coefficient; however, the majority of the subscales from the questionnaire didn’t achieve a minimum reliability of .70. These differences could be attributed to the reduced number of items in each subscale, that, when summed, allow for an increased reliability in each general scale (^{DeVon et al., 2007}). Specifically, each subscale is composed of three items, for a combined total of nine items for each scale. Therefore, the results from Rivalry and Admiration scales can be considered reliable; although, the subscales should be analyzed with caution.

The factorial analyses carried out for Admiration and Rivalry scales adjusted to the results reported by ^{Back et al. (2013}), albeit the specific techniques used for the exploratory factor analysis were different, in order to adjust them to the recommended practices for this type of variables (^{Lloret-Segura et al., 2014}). Despite these differences, all analyzed items charged in their respective factors of narcissistic rivalry and admiration.

When delving into the associations between the resulting scores and the variables that, according to the literature, have presented specific correlation patterns, these results can be considered adequate. Specifically, self-esteem has been a variable usually associated with narcissism (^{Geukes et al., 2017}); despite this, there are contradictory results in regard to the direction of the association between these. In this particular research, the results obtained demonstrate that two narcissistic strategies showed opposed correlations with self-esteem: one positive, for narcissistic admiration, and another negative, for narcissistic rivalry, just as stated by ^{Doroszuk et al. (2019}) in their research.

The most remarkable difference was observed in the association between the big-five personality traits and the narcissistic strategies, according to the original reports of ^{Back et al. (2013}), given the correlation pattern between agreeableness and narcissistic admiration. Originally, these authors found that the relationship between both variables was not significant; however, in this research it was found a positive and significant association. Despite this particular discrepancy, the association patterns were similar to the ones expected, given that narcissistic rivalry correlated with introversion and antagonism, while the narcissistic admiration did it with extroversion and openness to the experience; similar results to the ones obtained by ^{Leckelt et al. (2017}).

Another difference found consisted in the role of sex in regard to rivalry and admiration; but it must be taken into consideration that even in the literature theses discrepancies appeared to be present. In their intercultural research, ^{Doroszuk et al. (2019}) analyzed the differences between men and women, both in rivalry and in narcissistic admiration. They found that in Chile, Colombia and Spain, men showed higher levels of rivalry than women. But only in Spain this pattern was repeated with admiration, given that men showed higher levels than women. In the other two countries no differences were found.

There are cultural differences in regard to the predominance of men and women in one particular narcissistic strategy, given that usually men show a higher use of the rivalry strategy (^{Doroszuk et al., 2019}). In this Venezuelan sample, this particular pattern was no observed in the general strategies; however, delving in the subcomponents of rivalry, the usual pattern present in other countries was identified. Specifically, men presented a higher devaluation and striving for supremacy (both elements from narcissistic rivalry), when compared to women. It’s important to take into consideration that the test power for sex comparisons was below the ideal levels, so, it’s possible that the reduced sample size may be producing a false negative result (^{Peña, 2017}). Therefore, the Venezuela pattern seems to be similar to the ones found in other countries, once the narcissistic rivalry components are analyzed; however it is necessary to expand on these results recurring to a bigger simple size and balancing the number of group members, achieving in this way an adequate test power, in order to identify possible differences produced by sex.

Conclusion

The Narcissistic Admiration and Rivalry Questionnaire is a measurement instrument that is getting more present in the narcissistic personality research. Being a short instrument, it allows to obtain relevant information in a short time and with relative easiness. Its utility lies in the fact that it facilitates the simultaneous measurement of two narcissistic strategies present in the literature and which influences determine the type of relationships with other people and are usually associated with particular social consequences in short and long-terms.

As a sample of its actual relevance, there are multiple language adaptations of this questionnaire. This line of research has even reached Spanish speaking regions, which now have a Spanish adaptation. Despite this, some Latin-American countries do not have a specific study to assess the adequacy of the instrument to their population. This research aimed to contribute the Venezuelan population with a preliminary study that allowed the identification of the adequacy of the NARQ, in regard to reliability and validity standards for scientific and applied research. This objective is considered accomplished, therefore, this can be considered a starting point for the inclusion of the NARQ as an available tool for researchers and practitioners interested in the study of the narcissistic personality, as well as future research that further expand in the study of its psychometric properties in the Venezuelan population.