Rasch Analysis, Dimensionality, and Scoring of Aggressive Behavior Inventory for Junior High-School Students

Aggressive behavior often appears in junior high-school students who are entering adolescent development and requires serious attention from many parties. To measure these behavioral tendencies, many assessment instruments have been developed but none of them use Item Response Theory. This study aims to develop and test the validity of the instrument of aggressive behavior of junior high-school students, see the level of suitability of the item and see the level of difficulty of the item and be able to see variable maps of the ability of the person to answer and the ability of items to reveal aggressive behavior. This study used a sample of 360 students with 47 items. The analysis technique used is Rasch analysis to test the reliability, person validity, item validity, and rating validity. The analysis showed that overall the inventory developed is valid and reliable (person reliability is 0.89 and item reliability is 0.98.) The rating scale results using the Andrich Threshold Value indicate that the five choices given are valid for respondents. It can be concluded that the inventory behavior aggressive is valid internally to assess aggressive behavior of junior high-school students and can be used by teachers, school counselors in identifying aggressive behavior.


Introduction
The rapid development of science and technology makes teenagers often experience many challenges in achieving their developmental tasks, especially in the social environment. Teenagers who have failed in developing their sense of identity, will lose direction and have a negative impact on their development and often cause problems and behave negatively [1], [2]one of them is a tendency to behave aggressively [3]- [6].
Aggressive behavior is any form of action to hurt or hurt others, both physically, such as damaging or hurting other people [7], [8]or mentally hurt [9]- [11]. Aggressive behavior shows a negative effect on the effectiveness of learning and needs serious treatment by the counselor so that this behavior does not develop in a worse direction, especially in the achievement of developmental tasks [12]. The results found 69% of teenage respondents had had a fight, which was dominated by male teenage students [13]. While other studies show that 56.66% of respondents have participated in brawls [14].
Education has an important role in building human ecological beliefs, understanding and behavior. The emergence of various natural damages, disasters, floods, landslides and other environmental crises is assessed due to human activities outside proportional limits, religious and cultural values [36]. So that education services in schools can be optimal [37], there needs to be identification of students' aggressive behavior.
The problem is that until now there is no instrument that can be used to measure the aggressive behavior of junior high school students who are tested using the item response theory. The development of this instrument uses various concepts of aggressiveness, including verbal aggression [38]direct and indirect aggression [39],proactive aggression with reactive [40], or instrumental aggression  Trianto (2012).

Materials and Methods
The research sample consisted of 360 studentsin 8 junior high schools (public and private) West Sumatra. The research data were analyzed using the Rasch model using statistical analysis of suitability [57]- [59].Statistical analysis of suitability using MNSQ outfit parameters with ideal range (+0.5 to + 1.5), ZSTD outfit with ideal range (-2.0 to +2.0) to find the suitability of items and people, detect measurement biases, item strengths and weaknesses, and the level of difficulty of the items from the ability of the person to answer and the ability of items to reveal aggressive behavior [57].
The instrument development steps use the Oriondo and Antonio Models, namely: (1) planning instrumentsconsisting ofdetermination of instrument objectives, Determination of instrument objectives, determination of competencies tested, determination of the material being tested, grid arrangement, writing items based on the principles of developing Aggression Questionnaire [9], [60]- [64], compilation of scoring guidelines, Item validation and repair items; (2) trying out the instrumentconsisting to expert validation, the instrument which consisted of 94 items became 88 items that had been repaired for further testing; (3) establishing instrument validity and reliability with activity trying out the instrument and (4) interpreting the assessment scores [65].

Validity
The concept of validity is very important in a measurement. An instrument can be said to be valid when measuring what should be measured. The development of the Aggressive Behavior Inventory (ABI) instrument is evaluated whether it is able to measure what should be measured. In this case the extent to which the instrument measures the aggressive behavior of students. Validity analysis uses Principal Component Analysis (PCA) of residuals, which measures the extent to which the diversity of ABI instruments measures what should be measured. PCA analysis uses 2 parameters, first the value of total raw variance in observation (minimum 20%) and second value of total raw unexplained variance (minimum 15%) [

Validity of respondents
The instrument validity of respondents uses variable maps that can show the distribution of students' abilities on the left and the level of difficulty items on the right [57]. Further it is conveyed in Figure 1 below.  Based on Figure 1, the first left wright map shows that there is one student (295L) whose level of ability is higher (+0.32) than other students. There are also ten students (94P, 194P, 96P, 284P, 290P, 167P, 208P, 286P, 171P, and 349P) with a low level of ability (-1.86 to -2.91 logit) which shows the ability to answer low questions P52 (-0.80 logit) is not able to answer correctly. From the results of the analysis of map variables for female students (94) it can be stated that the lowest aggressive level with the mean value is (-2.91 logit). While male students (295) can be stated that the highest aggressive level with the mean value is (+0.32 logit).
Second, the right map explains the distribution of logit points in items. Item P15 is a problem with the highest difficulty level (+0.69 logit), which means that the probability of all students working correctly on this problem is small. can do it right, low logit value (-0.80 logit). With the statement P52 is "When talking about bad friends, I better go".
Third, comparing the distance between M-S-T (mean, 1SD and 2SD) on the above variable maps shows that the left side of the maps of the distribution of students' ability is greater than the distribution of the level of items on the right. In this context, the items show diversity, but the distribution of 360 students' ability is wider and wider. This means that the ability of 360 students is unable to reach the items with high ability.
Fourth, comparing the mean value of the person and the mean value of the item. The mean person value of 360 respondents is -0.85, while the mean value of the item is +0.00. This shows that the ability of the person is too low from the problem difficulty level.

Reliability
The reliability of an instrument refers to the stability of a measurement and consistency in measurement. To obtain information about the reliability of the person and the reliability of the items can be displayed in a statistical summary. The results of the statistical summary are explained further in Table 3 below.
Furthermore, person and item grouping can be known from the separation value by using the strata person formula H, so the value of H = [(4 * separation) + 1] / 3 [57].

Differential item functioning (DIF)
Measuring instruments and items can be biased because of differences in which certain items will favor one particular type (eg gender, family background, etc.). In Table 4 the following DIF analysis results are displayed, which can be determined by the probability value below (0.05) showing items that are biased [57].   Table 4 above, it appears that 22 items that are not biased are P2, P4, P9, P31, P34, P41, P42, P45, P49, P50, P56, P58, P61, P62, P64, P67, P72, P75, P78 , P82, P84, and P86. The number of items that are biased shows that differences in students' assessment of aggressive behavior are influenced by a variety of things, namely gender, parental educational background, culture, and economic level of parents.

Rating Scale Validation
The validity of the rating scale is very important in the measurement, because the rating scale is used to test the verification of the rating of the choice used. In the ABI instrument, it uses answer choices in the form of likert rating for each item. Respondents gave answers in accordance with their situation on each item given. Respondents' answers are seen based on whether the choice of answers given by respondents moves to the leftmost column 1 with the choice Always or the rightmost column 5 with the option Never. This choice contrasts the level of students' aggressive behavior in each item. More is presented in Figure 2 below. In figure 2 above shows the number 1 = always, 2 = often, 3 = rarely, 4 = sometimes, and 5 = never. Further to know the size of the ranking validity called Andrich Threshold, which shows the transition that occurs in decision making by respondents from one rank to the next [66]. Andrich Threshold value that moves from option 1 (none), then to option 2 (-0.50 logit), choice 3 (-0.30 logit), choice 4 (-0.19), and choice 5 (+1.00 logit). It can be seen that the value of Andrich Threshold moves from none then negative and leads to positive sequentially indicating that the five choices given are valid for the respondent.

Discussion
Based on research results. Measurement of aggressive behavior of junior high school students, classified valid and reliable helps further researchers to uncover aggressive behavior. Rasch analysis conducted can see the suitability of person and item, and can conduct analysis up to the item and item level level, analysis at the instrument level can also be done Empirical research found many researchers who created instruments of aggressive behavior in the fields of health, sports, social and education. Development of the Agregession Questionnaire (AQ) given to high school students in Egypt with a sample of 510 free bias between men and women [61]. Limitations in the Aggession Questionnaire (AQ) is the small number of items (29 items). So the researchers developed the Aggressive Behavior Inventory (ABI) by making physical, verbal, angry, and hostile forms of aggressive behavior into 12 indicators with a total of 47 items so that they were more representative of aggressive behavior that often occurs in junior high schools in Indonesia.
Furthermore, when compared to other instruments such as SDAS (social dysfunction and aggression scale) instruments, it is not only to measure aggressive behavior but can also predict aggressive events as recorded by SOAS-R (staff observation revised scale aggression). Therefore, it is very useful to apply both instruments, SDAS and SOAS-R, as well as in recording aggressive behavior [68]. Furthermore, aggression in sports, the development of a scale to measure aggressiveness and anger in competition. The Competitive Aggressiveness and Anger Scale (CAAS) looks at how aggressiveness in sports, individuals who behave aggressively due to regulations that allow individuals to engage in aggressive behavior but anger that causes athletes ignores the rules [70]. The limitation of the scale is only the understanding of aggression in sports and incomplete measurement of anger. This instrument does not precisely measure aggressive behavior in junior high school because this CAAS instrument is designed to look at aggressive behavior in sports. So the need for the ABI instrument by looking at aggressive physical behavior, anger and animosity in junior high school.

Conclusion
Aggressive behavior is one of the behaviors of students that requires attention from various parties. Schools as a place of formal education have a responsibility in dealing with aggressive behavior of students. The conclusion of this study is that the ABI instrument is valid and reliable for measuring the aggressive behavior of junior high school students with a total of 47 items. The advantages of the ABI instrument are being able to measure in the form of verbal, emotional anger, and physical displayed hostility that is related to behavior that is often done in junior high school.
This instrument is an alternative that can be used by field study teachers and counselors to reveal the level of aggressive behavior of junior high school students. appropriate for aggressive children. Just as subject teachers use instruments that aim to improve learning strategies. All parties in the school, namely subject teachers, counselors and administrators have responsibilities and have important roles. One who has a role important is the guidance and counseling teacher or counselor. One of the functions of guidance and counseling is the prevention function, namely efforts to intervene in the need for assistance.
The results of the instrument can also assist the school in designing programs to prevent aggressive behavior of students by completing the necessary facilities and infrastructure