**3. Results**

#### **3.1 Preliminary analysis**

The presence of important independent variables, sample size, missing data, range, extreme scores, independence of scores, normality of variables, linearity, multicollinearity, normality of residuals, and homogeneity of the variance of the residuals (homoscedasticity) was examined. Multivariate extreme scores were checked by examining the Mahalanobis distance. All of these factors and assumptions were respected.

#### **3.2 Moderation analysis**

To avoid potential multicollinearity issues, all continuous independent variables were centered. Next, we multiplied together the independent variable image distorting style with the moderator NU to create the interaction term. Model 1, without the interaction term, was significant F (2, 29) = 28.961, p < 0.001. Model 2, with the interaction term, was also significant F (3, 28) =18.750, p < 0.001. However, model 2 with the interaction term between image distorting style and NU did not account for significantly more variance than model 1. The R squared of our interaction term was found to be insignificant ΔR2 = 0.001, p = 0.744, indicating that there is no interaction. These results do not support the hypothesis that the association between image distorting style and RA is potentially moderated by the different levels of NU.

### **3.3 Partial mediation analysis**

A proposed partial mediation model was tested to verify the hypothesis that the association between image distorting style and RA is not mediated by NU. The determinant

#### **Figure 1.**

*a. Total effect of the image distorting-reactive aggression relationship. The beta shown in this figure is standardized. b. Mediation of the image distorting-reactive aggression relationship through negative urgency. The betas shown in this figure are all standardized.*

of the correlation matrix is 0.145, which is greater than 0.0001 and suggests an absence of multicollinearity. Looking at the model (see **Figure 1a**), the results indicated that the total effect was significant (β = 0.692, p < 0.001, (95% CI [0.562, 0.822])). By observing the model with the inclusion of the NU variable as a mediating variable (see **Figure 1b**), the results demonstrated that the direct link between image distorting style and RA turned out to be nonsignificant (β = 0.202, p = 0.119, (95% CI [−0.052, 0.455])). The results also indicated a significant direct relationship between image distorting style and NU (β = 0.750, p < 0.001, (95%CI [0.635, 0.865]). In addition, a direct link between NU and RA was found (β =0.654, p = p < 0.001, (95% CI [0.401,0.907]). The results indicated a significant indirect relationship between image distorting style and RA via NU (β = 0.491 (95% CI [0.257, 0.724]), p = 0.001. The results suggest that NU fully mediates the association between image distorting style and RA.

An effect size of 0.563 was obtained for NU and 0.666 for RA, which is considered large effects according to [59]. This indicated that 56.3% of the average NU variance is explained by the variables in our model, while 66.6% of the variance in RA is explained by other variables in our model.

We tested the proposed partial mediation model by adding the variables age and education as covariates. None of our covariates was found to be significantly related to RA. The results obtained by performing a mediation test with the addition of one of the covariates are not essentially different from those obtained without the addition of the covariates.

#### **4. Discussion**

The objective of this study was to explain the nature of the relationship between image distorting style and reactive aggression, more specifically to demonstrate whether, in nonclinical participants, the relationship between image distorting defensive style and reactive aggression is moderated, not mediated by the negative

#### *Negative Urgency and Its Role in the Association between Image Distorting Defensive Style… DOI: http://dx.doi.org/10.5772/intechopen.107531*

urgency impulsivity trait. Contrary to the hypothesis that was proposed, the results of the moderation analysis were not able to demonstrate that NU acts as a moderator in the relationship between image distorting style and RA.

Results obtained from the proposed partial mediation model are the significant indirect link and the non-significant direct link between image distorting style and RA (see **Figure 1b**), even after controlling for age and number of years of education. This suggests that NU mediates not partially but fully the relationship between image distorting style and RA. Thus, the more a person distorts the image of oneself or others, the more he/she acts impulsively with NU, and the more he/she will react aggressively in response to provocations or frustration. This is in accordance with the results of Gagnon and Rochat [34], demonstrating a similar mediation role with hostile attribution biases, instead of image distorting style. This makes sense when considering HAB as a distortion of reality, since it arises from a person's subjective interpretation of it, based on a given context and can be a distorted, or misinterpreted reflection of reality. In Gagnon et al. [19], a significant correlation has been found between the event-related N400 in a nonhostile condition and image distorting style, namely between hostile intent attribution and image distorting style. It is plausible that hostile intent attribution/hostile attribution biases are conceptually closely related to image distorting style, given the projective components of hostility. Consequently, it is possible to believe that the mediation role of NU obtained in the actual paper is an extension of the one found between hostile attribution biases and RA. The direct link between image distorting and NU was found significant, which suggests that the more an individual makes use of a defense mechanism distorting the image of oneself or others, the more this individual will have a disposition to have reckless behaviors during negative affective state, which is consistent with studies that have found indirect results on this relationship [19, 40, 41]. The direct link between NU and RA was found to be significant, supporting findings found in the literature, suggesting that the more individuals have a disposition to initiate reckless behaviors when resenting negative affective state, the more aggressively they will react to provocation [7, 31–34].

The total effect was found to be significant, suggesting that image distorting has a positive influence on RA when not controlling for the relationship between the mediator NU and RA, which is in agreement with the results of Gagnon et al. [19]. A surprising result is the complete mediating role of NU in the relationship between image distorting style and RA (given the lack of knowledge in the literature on the mediating role of NU in this relationship). We believe that other variables may have a role in the proposed causal chain, for instance, other types of impulsivity such as state impulsivity [60]. Taken together, our results suggest that when individuals put themselves in a defensive state leading to distorted images of self and others, they become more emotionally impulsive, leading to RA. In other words, the use of defense mechanisms associated with image distorting style could be the starting point in triggering RA when NU mediates this relationship. Without NU, the individual could simply distort their reality, but when NU is involved, RA is more likely to translate at the behavioral level. This is plausible on a theoretical level, given the significant mediator effect of NU in the relationship between image distorting style and RA is conceptually close to the one found between hostile attribution biases and RA [34], hostile attribution biases being conceived, such as a kind of projection defense mechanism or a defense distorting the image of others. Globally, two paths may elucidate the relationship between image distorting style and NU in the endorsement of RA behaviors. The first possible path might be that the tendency to make use of image distorting style increases negative effect, which in turn increases the difficulty to

control negative effect, and ultimately facilitates RA behaviors. A second path might be that image distorting style may activates dysfunctional beliefs (e.g., categorizing someone as bad), which may reduce the capacity of an individual to control his/her perception of other individuals, then facilitating the enactment of RA behaviors via an oversimplified and subjective perception [40].

Our study has several limitations that should be taken into consideration. It is paramount to consider that longitudinal designs are greatly preferable for testing if an effect is stable across time and if there is a proof of temporal precedence [51]. The actual study used a cross-sectional design, so we must therefore interpret our results with caution because the cross-sectional design does not allow us to really evaluate the proposed causal chain. Thus, we recommend for future studies to develop a protocol with a longitudinal design. Another limitation to consider is the small sample of the actual study, which might have restrained sufficient statistical power for examining particular hypotheses. The data used for our analyses are secondary, so it is not possible to make causal inferences from our obtained statistical results. We, therefore, recommend setting up a protocol with the specific aim of verifying the hypotheses proposed in this study. We also recommend using tools other than self-report questionnaires, as participants may demonstrate social desirability when completing this questionnaire or may simply be unaware of the defense mechanisms they are using. Qualitative interviews and projective tests would make it possible to evaluate the use of defense mechanisms of image distorting style of the participant from another perspective (for instance the Thematic Apperception Test or the Rorschach Inkblot Test).

#### **5. Conclusion**

The present study investigated the role of NU in the relationship between image distorting style and RA. Taken together, the results obtained in this preliminary study suggest that NU plays a crucial role in the relationship between image distorting style and RA and possesses important theoretical and clinical implications for enabling clinicians to better intervene at the source of aggressive behaviors, which according to our results, could be found as much in dysfunctional beliefs or the use of defense mechanisms with the function of distorting reality (the use of image distorting defense mechanisms being known to harm the therapeutic alliance [61]) as in certain facets of impulsivity such as NU. Our results may provide a better explanation of the internal factors that could increase RA and an opportunity to prevent aggressive behaviors based on this finding. Thus, the results of the present study make it possible to better understand the complexity of the link, which unites the image distorting style and RA.

#### **Acknowledgements**

This study was supported by a research grant to JG and Pierre Jolicoeur (PJ) from the Social Sciences and Humanities Research Council of Canada SSHRC (no. 430-2014-00729), the Canada Foundation for Innovation (201412), and the Canada Research Chairs (950-212663) program awards to PJ.

*Negative Urgency and Its Role in the Association between Image Distorting Defensive Style… DOI: http://dx.doi.org/10.5772/intechopen.107531*
