Introduction

Aggression is a term applied to a wide range of behaviors that may appear to be interrelated, while in close analysis, they are proven to be entirely different. In previous research focusing on aggression, there has been substantial controversy around the alternative definitions of this behavior; however, a consensus has been reached on two aspects of the definition. On one hand, aggressive behavior is defined as the infliction of harm or injury on another individual or organism. On the other hand, another definition is needed as to consider the intent of an aggressive individual to exclude accidental harm or aversive stimulation administered for beneficial reasons (1).

The definition proposed by Baron covers both these aspects (2); accordingly, aggression is defined as any form of behavior seeking to harm or injure another living being who is motivated to avoid such treatment. The operational definition used in the current study describes aggression as a behavior consisting of three dimensions: anger-temper, inroad-insult and obstinacy-malice. These dimensions were resulted from the analysis of test scores.

The classic attachment theory was first proposed by Bowlby, which presents a comprehensive definition for aggression in adult relationships (3). This theory has been widely used in recent years to explain the nature of relationships during adulthood (4). According to this theory, attachment is defined as the emotional bond between infant and caregiver, who is typically the mother (5). In this regard, Bowlby's seminal work on attachment, separation and childhood loss has resulted in a theoretical basis for the accurate evaluation of adult attachment patterns (3). Furthermore, descriptions by Ainsworth, Blehar, Waters and Wall are classified into three basic attachment patterns of secure, ambivalent and avoidant, which were used to develop other measures for equivalent attachment patterns in the adult population (6).

In general, children with secure attachment tend to become visibly upset when their caregivers leave and are happy when they return. When frightened, these children seek comfort from the parent or caregiver. Contact initiated by a parent is readily accepted by securely attached children, and they cherish the presence of a parent with positive behavior (7).

Children with ambivalent attachment tend to be extremely suspicious of strangers. These children display considerable distress when separated from their parents or caregivers and may not be reassured or comforted by their return. In some cases, the child might passively reject the parent by refusing comfort or may openly exhibit aggression towards the parent or caregiver (8).

Children with avoidant attachment tend to ignore their parents and caregivers. This avoidance often becomes especially pronounced after a period of absence. These children might not reject attention from parents; however, they neither seek comfort nor contact from their caregivers. Moreover, children with avoidant attachment may show no preference between contact with their parents and a complete stranger (8).

In this regard, Bowlby emphasized the quality of mother-infant relationship since it laid the foundation for later working models and a set of internalized beliefs to form future relationships and social interactions (3). Attachment theorists have postulated that experiences from early attachments in life largely influence social cognitions in children in their social and physical perception of the world, as well as their interactions with other people during adulthood (5, 9).

Children with secure attachment are believed to form a working model that perceives oneself as valuable and the caregiver as supportive and responsive. In this style of attachment, consistent responsiveness of the caregiver to the emotional needs of the child helps build a positive model of the world and people as safe and approachable.

On the other hand, early parent-child relationships marked by anger, hostility, mistrust and insecurity (i.e., insecure attachment) may foster an internal working model based on which future relationships seem untrustworthy as well (10). Consequently, the child is likely to have maladaptive views of self and others and may attribute the behavior of other people to negative intentions. Insecure attachment leads to unfavorable outcomes in children, including aggression, delinquency, substance abuse and emotional disturbances (11).

Several authors have attempted to expand their research on attachment to different stages of life, while the classic attachment theory has not been investigated equally at different ages. Moreover, developmental research has highlighted the pivotal role of mother-infant attachment in psychological and behavioral outcomes. According to the literature, insecure attachment styles are positively linked to indices of psychological distress, such as negative affectivity, depressive symptoms, eating pathology, dysfunctional sexual beliefs, emotional distress, anxiety and general distress symptoms (12-17). Also, insecure attachment is associated with interpersonal difficulties and greater hostility towards other people (18).

Existing body of literature is indicative of a significant correlation between the quality of parent-child attachment and outcomes such as aggression (19, 20). Few studies have evaluated the relationship between attachment styles and aggression in Iran. Findings of the current paper could be used to investigate the efficacy of attachment theory in social and cultural structure of the Iranian community and collect empirical evidence on this theory.

In the present study, we extended the attachment theory to explain the occurrence of aggression in adult relationships. To do so, adult attachment patterns and related interpersonal problems were compared and evaluated among individuals with distinctive experiences of aggression. This study mainly aimed to investigate the role of mother-infant attachment in the severity of aggression.

Materials and Methods

This descriptive correlational study was conducted on 750 master’s students of Bu-Ali Sina University in Hamadan, Iran in 2013. Participants were selected by cluster random sampling, and sample size of the study was determined to be 159 students using Morgan’s table. After eliminating distorted questionnaires, 150 participants remained in the study (21). Participants included 75 male students aged 23-29 years (mean: 23.3±6.6) and 75 female students aged 22-26 years (mean: 22.8±1.5). Selected students were educated during September 2012-May 2013. Out of faculties including chemistry, arts and architecture, engineering, and economics and social sciences, three were selected randomly. In addition, three classes were selected randomly from each faculty, and the students were enrolled in this study.

Adult Attachment Styles (AAS): In this study, we used the Persian version of AAS designed by Ainsworth et al. This measure is widely used to evaluate the level of attachment based on the classification of parent-child attachment styles into three types of secure, ambivalent and avoidant (5, 6). It is noteworthy that Mayseless added another category to this definition in order to cover disorganized or disoriented types of mother-infant attachment (22). In the current study, we used the Persian translation of AAS, which has been normalized by Besharat (23). Internal consistency of secure, ambivalent and avoidant subscales of attachment was estimated at 0.74, 0.69 and 0.71 for female participants, and 0.73, 0.72 and 0.71 for male participants, respectively. Regarding the reliability of AAS, coefficient alpha was calculated to be 0.70. Moreover, test-retest correlation was estimated at 0.72 after one week, which was indicative of adequate validity (23).

Ahvaz Aggression Inventory (AAI) :AAI was first developed by Zahedi Far, Najarian and Shokrkon using factor analysis in order to evaluate aggression on single samples of college students (24). This measure is a pen and paper self-report questionnaire consisting of 30 items scored on a four-point scale (never, rarely, sometimes and always). Questionnaires had three subsets, and high scores were indicative of high levels of aggression. In their study, Zahedi Far et al. used Eysenck Personality Questionnaire, Minnesota Multiphasic Personality Inventory and Buss-Durkee Validity Inventory to evaluate the reliability coefficient of AAI and reported all the obtained coefficients to be significant (P<0.001) (24). Reliability coefficient of this scale is estimated at the acceptable level of 0.84.

Written informed consent was obtained from all participants, and the study protocol was officially approved by the university. Afterwards, we initiated the tests, and students were fully instructed about questionnaires and rating scales. During the study, participants were encouraged to enquire about the research project and their possible concerns. Moreover, participants were asked to select the scale which could best describe their emotions. All participants received the same instructions about completing the questionnaires. Also, objectives of the study were explained to the selected students who met the inclusion criteria and had consent to participate in the investigation.

Inclusion criteria of this study were as follows: 1) consent for participation; 2) lack of trauma or pregnancy within the past six months; 3) no history of mental disorders and use of psychiatric drugs and 4) no experience of divorce within the past year. Participants were assured of confidentiality terms, and provided data were only used for research purposes. All participants completed AAS, AAI and demographic questionnaires. Data were analyzed using Pearson correlation and regression analysis. Data analysis was performed using SPSS statistical package.

Results

Mean of aggression scores and their correlation with different attachment styles are shown in Table 1.

As can be seen in Table 2, Pearson coefficients were indicative of a negative correlation between secure attachment and aggression scores (P<0.05, r=-0.43). Accordingly, secure mother-infant attachment was observed to have a significant negative correlation with all subscales of aggression.

On the other hand, avoidant attachment style had a significant positive correlation with total score of aggression (P<0.05, r=0.24), and ambivalent mother-infant attachment had a significant positive correlation with all subscales of aggression, with the exception of inroad-insult subscale.

Results of regression analysis indicated that aggression scores were significantly different between the three attachment styles (secure, avoidant and ambivalent) (P<0.05, β=0.29). Considering R², it was observed that attachment styles could predict a 0.27 variance of aggression among the students (R²=0.27). Moreover, regression coefficients could determine the effects of different attachment styles on the level of aggression. These findings are presented in Table 3.

As can be seen in Table 3, secure attachment had a significant negative effect on aggression (P<0.001, β=-0.44), so that with a one-unit increase in the standard deviation of the predicting variable (secure attachment style), standard deviation of the dependent variable (aggression) decreased by 0.44 units.

According to our findings, ambivalentattachment had a significant positive effect on aggression (P<0.001, β=0.29), so that with a one-unit increase in the standard deviation of the predicting variable (ambivalent attachment style), standard deviation of aggression increased by 0.29 units. However, avoidant attachment style was found to have no significant effect on the overall score of aggression (P<0.05, β=0.09).

According to the information in Table 3, there is no significant correlation between sexuality and level of aggression. In other words, mean scores of aggression were not significantly different between male and female students.

Discussion

The results obtained in the present study are consistent with previous findings in the literature. Accordingly, there is a negative correlation between secure attachment and aggression scores. In addition, secure attachment was observed to have a significant negative correlation with all subscales of aggression. Also, there was a significant positive correlation between avoidant attachment style and total scores of aggression, while ambivalent attachment had a significant positive correlation with all subscales of aggression, with the exception of inroad-insult subscale.

Attachment styles could be significant predictors for aggression variance. In the present study, we used regression coefficients to determine the effect of attachment styles on the score of aggression. According to our findings, secure attachment has a significant negative effect on the level of aggression, whereas ambivalentattachment has a significant positive effect on aggression. However, avoidant attachment style has no significant effect on the total score of aggression. These findings are in line with the results obtained by previous studies.