The growing juvenile crime problem is one of the most important issues facing around the globe. Juvenile crime stems from a complex array of causes including a lack of adult supervision for youth, a lack of strong role models, and limited opportunities in addition to negative conditions associated with poverty, abusive backgrounds, and a host of other reasons. Juvenile crime often varies depending on the region; community and neighborhood. Statistics provide evidence that it is a serious problem nationwide. Juvenile delinquency is defined in terms of acts or infractions which are prohibited in the statues of the individual states. Juvenile delinquents are youth who commit one or more of these infractions with the most common upper age limit of 17 years [1]. According to Juvenile justice system Ordinance of Pakistan [2] a juvenile is a lawbreaker under 18 years of age who can be sent to a juvenile jail upon conviction. Although small percentages (approximately 16% to 23%) of delinquents are serious chronic offenders [3,4] they account for about 50% to 60% of all juvenile offenses and about 75% of all violent juvenile offenses [5]. The study of emotional expression has a long history, which dates back to the 1870s with scientific investigation undergone by Charles Darwin [6]; His work emphasized the biological utility of emotional expression. Thus it contributed to the development of an evolutionary-expressive approach to emotion, which suggests that emotion exists because it contributes to survival [7].

Barriga et al. [8] categorize four main cognitive distortions in delinquents: self-centered, blaming others, minimizing/mislabeling, and assuming the worst. Facial expressions of emotions provide a means of communicating and obtaining information about one's social environment [9,10] and provide cues to interpersonal exchanges [11]. Bandura [12] pointed out that the ability to read the signs of emotions have important adaptive value in guiding actions toward others. Selective impairments in the recognition of both sad and fearful expressions have been found in children with psychopathic tendencies [13]. Children and adolescents with learning disabilities [14,15] and attention-deficit hyperactivity disorder have also shown deficits in interpretation of emotions in facial expressions. Deficits in reading emotion in facial expressions have also been found to be related to lower levels of social competence in children [16]. Cooley, Triemer [17] explored that facial affect recognition might interact with skills such as cognitive understanding of social interactions and expressing empathy.

Carr and Lutjemeier [18] conducted research on facial affect recognition, empathy and delinquency among youth offenders who had been involved with multiple and severe offenses. Additionally, a link was found between the negative expressions, fearful and sad, angry and sad. Felleman, et al. [19] found 4-5 year old children as well as adults are reasonably accurate in recognizing children's facial expressions of several basic emotions (i.e., happiness, sadness, anger, and neutrality). A recent research conducted on misrecognition of emotions expression [20] Delinquents were less accurate in the recognition of facial expressions that conveyed disgust than were control participants. The delinquents misrecognized the facial expressions of disgust as anger more frequently than did controls. These results suggest that one of the underpinnings of delinquency might be impaired recognition of emotional facial expressions, with a specific bias toward interpreting disgusted expressions as hostile angry expressions.

The adolescents are the important segment of society and those who develop delinquent behavior become a threat to themselves and society; there is need to understand the mechanism working behind the delinquent behavior. The research to explore the differences of emotions and their interpretation is scarce therefore, the current research is designed to identify the selective attention and recognition patterns in delinquents compare to non-delinquents. There is hardly any research on emotional recognition of delinquents done in Pakistani cultural context; specifically our study focuses the following issues. The objectives of the study were to compare the emotion recognition through facial expression in delinquents and non-delinquents youth to determine the selective attention of emotion recognition in terms of positive (happiness, surprise and negative emotions (anger, fear, disgust, sadness) and to determine the role of Socio-economic status and educational level in recognition of facial expressions. On the basis of the previous studies and objectives of the present research it was assumed that the delinquents and non-delinquents will have differences in the emotion recognition ability. And it was also assumed that the recognition of negative emotional expressions (anger, fear, disgust, sadness) and distress will be high in juvenile delinquents than positive emotions as compared to non-delinquents. Further it was assumed that the SES, education level and family size will be an important determinant in emotion recognition as well as delinquency trends.

Inclusion criteria

The Juveniles respondents for experimental group were selected on the basis of the conduct disorder criteria of Diagnostic Statistical Manual-IV-TR (2000) which includes showing the behavior of aggression to people and animals, destruction of property, deceitfulness or theft and serious violation of rules. The adolescents who were not fulfilling this criterion were not included in the final sample of experimental group. Whereas the adolescents included in the control group were the youth who did not meet the criteria of DSM-IV-TR (2000). The project was approved by IRB and all ethical considerations were taken into account while designing and implementing it. Debriefing/ counseling sessions was arranged for the subjects.

Inclusion criteria

The Juveniles respondents for experimental group were selected on the basis of the conduct disorder criteria of Diagnostic Statistical Manual-IV-TR (2000) which includes showing the behavior of aggression to people and animals, destruction of property, deceitfulness or theft and serious violation of rules. The adolescents who were not fulfilling this criterion were not included in the final sample of experimental group. Whereas the adolescents included in the control group were the youth who did not meet the criteria of DSM-IV-TR (2000). The project was approved by IRB and all ethical considerations were taken into account while designing and implementing it. Debriefing/ counseling sessions was arranged for the subjects.

Experimental Task [21]:

The ability to recognize facial emotional expressions was measured with the help of the Facially Expressed Emotion Labeling (FEEL) test [21], which was translated into Urdu for the current study. This is a newly developed computer program in which color photographs of neutral faces followed by the same faces expressing a certain basic emotion appeared on the screen, and subjects had to identify the emotion shown. The pictures used in the FEEL test were taken from the Japanese and Caucasian Facial Expressions of Emotion series. The 6 displayed basic emotions included: anger, sadness, fear, disgust, happiness, and surprise. The task entailed viewing 42 pictures 7 pictures for each emotion) presented on a computer screen. Presentation of the stimulus and selection of the emotion scene proceeded in the same way for all 42 pictures. The FEEL test has shown to have high reliability (Cronbach’s coefficient of a=0.77). The FEEL test showed the highest reliability (Cronbach´s coefficient of r=.77) in research conducted by all tests for emotion recognition published so far [21,22].

Procedure

Data collection for this study proved to be a difficult task as it took almost a year to get permission from the jail authorities to collect data for the current project. The researcher team visited the adolescents unit of district jail Lahore to be familiar with the prisoners. The research team was introduced as a part of NGO which was already in contact with the adolescents in jail. After developing the appropriate level of rapport the participants were informed about the objectives of the study and the procedure of the experiment participants had to go through, they were assured about the confidentiality of information sought from them. Informed consent form was given to the participants before proceeding with the interview information and feel task. The data for the current study were collected into four steps. In the first step subject were selected by matching on DSM_IV Criteria for conduct disorder both under trial and under punishment. Participants who did not fulfill the criteria of conduct disorder were not included in the next step of the study. In the second step a semi-structured interview was conducted with participants, which included the personal data, childhood information’s, crime history, information related to marital and sexual life, hobbies and religious information. In the third step after their diagnosis as delinquents they were tested on Feel Task in order to assess their emotional recognition. The debriefing sessions were conducted with the adolescents after the task accomplishment while telling stories and showing the different movies clips. How did they find it, either interesting or boring so that they may be relaxed by the imprinted effects of threatening word stimuli presented to them? Adolescents who remained unable to complete the task had not been included in the final sample.

The data of the current study were analyzed with SPSS (Version10) while employing appropriate statistical techniques like 2x6 Between-Within ANOVA, t-test, and MANOVA used to determine the main effects, interactions, differences and comparisons. First of all psychometric properties for the measures used in the current research were determined for the sample. The alpha coefficient computed for the FEEL task was also significant (a=70). Table 1 is showing demographic characteristics of the sample included in the current research which was comprised of Experimental group (Delinquents) and Control group (Non-delinquents) (Table 1).

Table 1: Demographic Characteristics of Sample

To find out the differences between delinquents and non-delinquents in the ability of emotion recognition the emotion recognition, FEEL test [21] was used. Moreover differences in both groups on different types of six emotions i.e., Fear, Anger, Disgust, Sadness, Surprise, and Happiness were explored by using 2 x 6 Between-Within ANOVA (Table 2).

Table 2: 2x6 Between-within ANOVA for Emotion Recognition Ability Task Accuracy Scores for Delinquents (n=50) and Non delinquents Group (n=50) *** p<.001.

The result of 2 x 6 Between-Within ANOVA indicated non-significant main effect for emotions recognition across groups of delinquents and non-delinquents, F (1,98)=.097, p=ns. Within subjects effects were highly significant for types of emotions F (5,490)=31.72, p<.0001.Highly significant interaction of emotion recognition with the groups (delinquent and non-delinquents), F=(5, 490)=16.24, p<.0001. Data were divided for positive (surprise and happiness) and negative emotions (fear, anger, disgust and sadness) and negative emotion between delinquents and non-delinquents (Table 3).

Table 3: 2x6 Between-within ANOVA of Positive and Negative Emotions scores Delinquents (n=50) and Non delinquents Groups (n=50).

2 x 6 between-within ANOVA indicated highly significant results for positive and negative emotion for within group (1, 98)=196.78, p<.0001 and interaction of two categories across groups (1, 98)=14.28, p<.0001. When data were further broken down for positive and negative emotions categories of FEEL Task and difference across experimental and control groups were determined while computing t-test (Table 4).

Table 4: Difference between Positive and Negative Emotions of Experimental Group (n=50) and Control Group (n=50) on emotion recognition task. df=98,<.001.

The results in Table 4 indicate highly significant differences in recognition of positive and negative emotions between delinquents and non-delinquents, t (98)=3.55, p<.0001.The delinquent group scored high (M=17.06) on recognition of negative emotions than non-delinquents (M=15.40). Whereas non-delinquents scored high on positive emotions (M=10.22) than delinquents (M=8.06). These differences are depicted in Figure 1.

Figure 1: Differences of Positive and Negative Emotions across Delinquents (n=50) and Non delinquents (n=50)

Additionally t-test was used to find out the differences in positive emotions (happiness and surprise) and negative emotions (fear. anger, disgust, sadness) between delinquents and non-delinquents.

Table 5 indicates the differences in recognition of emotions like anger, fear and surprise between delinquents and non-delinquents. The delinquents scored high on anger (M=5.4) than non-delinquents (M=4.0). In the same way delinquents scored high in recognition of fear (M=5.3) than non-delinquents (M=3.6). Results also indicates that non-delinquent group was significantly high on the emotion recognition of surprise (M=4.5) than that of delinquents group (M=1.9). There were highly significant differences between delinquents group and non-delinquents group on the emotion recognition of anger, t (98)=3.65, p<.0001, surprise, t (98)=6.53, p<.0001. However on emotion recognition of sadness, disgust and happiness, statistically no significant differences were found between delinquents and non-delinquents. Overall the table shows that there were no significant differences between the delinquents and non-delinquents group on the total emotion recognition task. These differences are clearly depicted in Figure 2.

Figure 2: Mean scores Differences of Delinquents and nonDelinquents on Emotions Recognition Task

Table 5: Difference between Delinquents (n=50) and Non delinquents Groups (n=50) on FEEL Task Accuracy Scores (N=100) df=98.

Figure 1 is showing that respondents from experimental group scored high on anger, happiness, and fear whereas scored low on sadness, surprise. Both groups were same on the emotion recognition of disgust. However the difference on total FEEL score between delinquents (experimental group) and non-delinquents (control group) was very small. Data was further broken down for positive and negative emotions categories of FEEL.

Delinquency Level as Determined by DSM-IV Criterion

In order to find out the effect of delinquency level in emotion recognition the data were broken down into categories of delinquency according to the criterion of DSM-IV-TR (2000) to see how many of the experimental and control were falling into mild, moderate or severe categories of delinquency (Table 6).

Table 6: Mean scores for each Level of Delinquency is Determined by DSM-IV Criterion on 6 Emotions of FEEL Task for Experimental and Control Group.

Table 6 shows that most of the subjects (n=43) in the experimental group were fulfilling the criteria of severe delinquency trends and only 7 subjects were falling in the moderate category. In the same way the table is showing that 43 adolescents in the control group are falling in the category of non-delinquency and only 7 seven people in the category of mild delinquency.

The data in the Table 6 indicates that the adolescents in experimental group with the moderate delinquency level scored high in accuracy for total FEEL Task (M=27.43) than the sever level delinquents (M=25.00). In control group adolescents with mild level of delinquency scored low on FEEL Task (M=23.29) than the delinquents having no delinquency trends (M=26.33). The table also indicates that respondents within experimental group had highest score on anger (M=5.44), happiness (M=6.14) fear (M=4.32) than the control group with low level of delinquency had lower scores on anger (M=4.02), fear (M=5.72), happiness (M=5.72). The table also indicates that respondents in experimental group were accurate on identifying anger (M=6), happiness (6.4). The least accurate recognition for experimental group was on surprise (M=1.92). For the control group, the most accurately recognized emotion was happiness (M=5.72). Below Figures 3 and 4 show these differences.

Figure 3: Mean scores differences of moderate and severe levels of Delinquents determined by DSM-IV Criterion

Figure 4: Means Differences of Non- delinquents for mild and no delinquency as Determined by DSM-IV Criterion

The figure exhibits that the adolescents with moderate level of delinquency were more accurate in the recognition of emotions than the adolescents with the severe level of delinquency.

The figure exhibits that in the control group the adolescents with no level of delinquency were more accurate in the recognition of emotions than the adolescents with the mild level of delinquency. 2X6 ANOVA was computed to determine the effect of delinquency level on emotion recognition task (FEEL) with six emotions (anger, fear. disgust, sadness, surprise and happiness). Table 7 presents Mixed ANOVA for Emotions x Level of Delinquency as Determined by DSM-IV- TR Criteria. The Table 7 shows that within subject there was no statistically significant difference among the four delinquency levels (mild, moderate, severe, no), F (3,99)=.886, p<.46 in emotion recognition whereas between subject difference among no, mild, moderate and severe delinquents at six emotions (anger, fear, disgust, surprise, sadness, happiness) was significant F (5,99)=15.81, p<.0001. Interaction between level of delinquency and types of emotions was found to be highly significant, F (15, 99)=5,68, p<.0001.

Table 7: Mixed ANOVA for Emotions x Level of Delinquency as Determined by DSM-IV Criteria. *p<.0001.

Finally we conducted the multivariate analysis in order to see the effect of the SES, education level and family size in emotion recognition as well as delinquency trends. The results indicate that education has significant effect on delinquency (1,99)=2.75, p<.05 and interaction effect of family income and education on total feel score F (1,99)=2.49, p<.05. Rest of the independent and interaction effects remain non-significant.

The current research was designed to identify the selective attention and emotion recognition patterns in delinquents as compared to non-delinquents. It inquires the recognition patterns of delinquents and non-delinquents on six universal emotions like anger, fear, disgust, sadness, surprise and happiness. It also targets to explore the difference in Psychopathic trends of Delinquents and Non-delinquents. Moreover, the associations and outcomes and interaction of demographic variables with other study variables like group, socioeconomic status, education, and family size were explored along with the basic in depth information were obtained by conducting in depth interview which was comprised of information related to childhood, marital and sexual life, social relationships crimes, hobbies and religious affiliation. Although scarce research is available on the emotion recognition of delinquents but some studies have shown that delinquent adolescents have scored lower than non-delinquent adolescents on cognitive aspect of empathy, such how they identify with others emotions [23]. The rationale behind this study was to understand the mechanism working behind the delinquent behavior. There is hardly any research on emotional recognition of delinquents done in Pakistani cultural context; specifically this study focuses on different objectives. The first objective of the research was to compare the emotion recognition through facial expression in delinquents and non-delinquent’s youth. The results showed the differences of delinquents and non-delinquents group in the total emotion recognition ability were not significant across the groups. While the both groups had highly significant differences on different types (anger, fear, disgust, sadness, surprise and happiness) of emotions and interaction between both groups and different types of emotions were highly significant. It means that both group delinquents and non-delinquents were significantly different from each other in the recognition of emotion ability. These findings are consistence with the findings of Kaplan and Arbuthnot [23]. They found that delinquent adolescents have scored low than non-delinquent adolescents in the identification of others emotions. There were significant differences in the recognition of emotions like anger, fear and surprise. The delinquents scored high on anger and fear than non-delinquents. The non-delinquent group was significantly high on the emotion recognition of surprise than that of delinquents group. There were highly significant differences between delinquents group and non-delinquents group on the emotion recognition of anger, surprise. However on emotion recognition of sadness, disgust and happiness, statistically no significant differences were found between delinquents and non-delinquents. Overall the results show that there were no significant differences between the delinquents and non-delinquents group on the total emotion recognition task. The findings by Carr and Lutjemeir [18] support a link between facial affect recognition and delinquency. They explored more youth offenders self-reported to have had been involved in acts to theft and acts of overall delinquent behavior, the more accurate they were at identifying the emotion angry in adult faces. A possible explanation for this findings might be that youth offenders are severe at recognizing the expression angry in adult faces because their high level of exposure to this expression. The research done by Blair and Frith [24] also provide a link among facial affect recognition and delinquency. Another research done by Blair [25] also has clear evidence that that response to fearful and sad expressions is attenuated in reduced regulation of social behavior.

The second objective of the study was that the tendency to recognition of negative emotions like anger, fear, disgust and sadness will be high in delinquents than the positive emotions like happiness and surprise. The results indicate the non-significant main effect for the group (delinquents and non-delinquents) in recognition of positive and negative emotions. There were no significant differences in the recognition of positive (happiness and surprise) and negative (anger, fear, disgust, sadness) emotion ability across the groups. However within subject there were highly significant differences in the recognition ability of positive and negative emotions of both delinquents and non-delinquent. The results showed that the delinquents (experimental group) were high on the recognition of negative emotions like anger and fear than the non-delinquents (control group). The non-delinquents were higher on the positive emotions like happiness and surprise. It means experimental group and control group in the recognition of negative and positive emotions were entirely different. These findings are consistent with the prior researches [26-28]. A possible explanation for perceiving the negative emotion more might be that from the very childhood these children have faced these emotions frequently.

Moreover, in order to find out the effect of delinquency level in emotion recognition the data were broken down into categories of delinquency according to the criterion of DSM-IV-TR (2000) to see how many of the experimental and control were falling into mild, moderate or severe categories of delinquency.

The results shows that most of the subjects (n=43) in the experimental group were fulfilling the criteria of severe delinquency trends and only 7 subjects were falling in the moderate category. In the same way 43 adolescents in the control group were falling in the category of non-delinquency and only seven people in the category of mild delinquency. 2X6 ANOVA was computed to determine the effect of delinquency level on emotion recognition task (FEEL) with six emotions (anger, fear. disgust, sadness, surprise and happiness). The results revealed that within subject there was no statistically significant difference among the four delinquency levels (mild, moderate, severe, no) in emotion recognition whereas between subject difference among no, mild, moderate and severe delinquents at six emotions (anger, fear, disgust, surprise, sadness, happiness) was significant. Interaction between level of delinquency and types of emotions was found to be highly significant these findings are consistent with the Frick and White [29].

Further results revealed that the adolescents in experimental group with the moderate delinquency level scored high in accuracy for total FEEL than the severe level delinquents. In control group adolescents with mild level of delinquency scored low on FEEL Task than the delinquents having no delinquency trends [30-33].

The results also indicates that respondents within experimental group had highest score on anger, happiness, fear than the control group with low level of delinquency had lower scores on anger, fear, happiness. The respondents in experimental group were accurate on identifying anger, happiness. The least accurate recognition for experimental group was on surprise. For the control group, the most accurately recognized emotion was happiness [34,35].

The further results showed that adolescents in the experimental groups with moderate level of delinquency were more accurate in the emotion recognition than the adolescents with the severe delinquency level. In the same way the adolescents in the control group with mild delinquency level were low in recognition of emotions than the delinquents with level of no delinquency. The adolescents with the severe level of delinquency were high on the recognition of anger, disgust, fear and happiness than the adolescents having the mild and no level of delinquency. Adolescents with mild and no level of delinquency were high on the recognition of sadness and surprise. The adolescents with the moderate and severe level of delinquency were higher on the total emotion recognition task than the adolescents having mild and no level of delinquency [18,23].

Multivariate analysis of variance was conducted to compute the independent and interaction effects of group (experimental and control), education. The results indicate that education has significant effect on delinquency and interaction effect of family income and education on total feels score. Rest of the independent and interaction effects remain non-significant.

Conclusively, both group delinquents and non-delinquents were significantly different from each other in the recognition of different types of emotion. However there was no difference on the total emotion recognition ability between both groups. The delinquents scored high on anger and fear than non-delinquents. The non-delinquent group was significantly high on the emotion recognition of surprise than that of delinquents group. There were highly significant differences between delinquents group and non-delinquents group on the emotion recognition of anger, surprise. However on emotion recognition of sadness, disgust and happiness, statistically no significant differences were found between delinquents and non-delinquents. The delinquents (experimental group) were high on the recognition of negative emotions like anger and fear than the non-delinquents (control group). The non-delinquents were higher on the positive emotions like happiness and surprise. Different levels of delinquency effects the ability to identify different types of emotions. The adolescents in experimental group with the moderate delinquency level scored high in accuracy for total FEEL than sever level delinquents. In control group adolescents with mild level of delinquency scored low on FEEL Task than the delinquents having no delinquency trends. Respondents within experimental group had highest score on anger, happiness, and fear than the control group with low level of delinquency had lower scores on anger, fear, and happiness.