Health & Fitness

A revised teacher rating scale for Reactive and Proactive Aggression

A revised teacher rating scale for Reactive and Proactive Aggression
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Journal of Abnormal ChiM Psychology, Vol. 24, No. 4, 1996 A Revised Teacher Rating Scale for Reactive and Proactive Aggression Kim Brown, 1 Marc S. Atldns, 2,5 Mary L. Osborne, 3 and Mary Milnamow 4 A teacher rating scale of reactive aggression, proactive aggression, and covert antisocial behavior was evaluated in a normative sample of third- to fifth-grade predominantly white lower middle class boys N = 186). Factor analysis revealed independent and internally consistent Reactive Aggression six reactive items), and Proactive Aggression five proactive items, five covert items) factors. Although the factors were intercorrelated r = .67), and each factor was significantly correlated with negative peer social status r = .26 for each, controlling for grade), the independence of the factors was supported by the unique relation of Reactive Aggression with in-school detentions r = .31), controlling for Proactive Aggression and grade. These results supported the reliability and validity of Reactive and Proactive Aggression as rated by teachers, which shouM facilitate further research of these constructs. A recent model for subtyping childhood aggression distinguishes between reactive and proactive forms of aggression (Dodge & Coie, 1987). Reactive Manuscript received in final form July 25, 1995. This research was supported in part by an NIMH First Award to Marc Atkins (MH4682), and support from an NICHD Mental Retardation Research Center Core Center Grant (DH26979). Portions of these data were presented at the annual meeting of the Society for Research in Child and Adolescent Psychopathology, Sarasota, Florida, in February 1992. The authors are grateful to Paul McDermott and Tom Power for editorial comments, and to Gail Apfel, principal, and the teachers, staff, and students of Stonehurst Elementary School for their enthusiastic participation in this research. 1Department of Counseling Psychology, West Chester University, West Chester, Pennsylvania 19383. 2Institute for Juvenile Research, University of Illinois at Chicago, Chicago, Illinois 60612. 3Department of Pediatrics, Robert Wood Johnson Medical School, New Brunswick, New Jersey 08903-0019. 4Department of Psychology, Temple University, Philadelphia, Pennsylvania 19122. 5Address all correspondence to Marc S. Atkins, University of Illinois at Chicago, Institute for Juvenile Research, 907 South Wolcott Ave., Chicago, Illinois 60612-7347. 473 0091-0627/96/0800-0473509.50/0 © 1996 Plenum Publishing Corporation  474 Brown, Atkins, Osborne, and Milnamow aggression, considered to be a defensive response to a perceived threat, fear, or provocation, has theoretical roots in the frustration-aggression model posited by Dollard, Doob, Miller, Mowrer, and Sears (1939) and later revised by Berkowitz (1990). According to this model, aggressive be- havior is always preceded by frustration (e.g., an unexpected barrier to goal attainment). The key proposition in Berkowitz's analysis is that negative affect of any type will activate anger-related feelings and thoughts and, ul- timately, an aggressive response (Berkowitz, 1990). Dodge and Coie pro- posed that reactive aggression is evidenced by a hostile attributional bias--that is, a tendency to attribute hostility to ambiguous provocations. Proactive aggression is defined as an unprovoked, aversive behavior intended to harm, dominate, or coerce another person. The theoretical ba- sis for this construct is Bandura's social (1973) learning theory, which de- fines aggression as a learned behavior controlled by external rewards. Two types of proactive aggression were proposed by Dodge and Coie (1987): Instrumental-proactive aggression, which is governed by reinforcement principles such as aggression for object acquisition; and hostile-proactive aggression, which involves domination or intimidation of another (i.e., bul- lying). Hostile-proactive aggression is presumed to relate to tendencies to overvalue the outcome of aggression and to underestimate its impact on victims. Instrumental-proactive aggression appears similar to covert antiso- cial behavior such as lying and stealing (Loeber & Schmaling, 1985), al- though this has not been tested empirically. In an attempt to define these constructs, teacher ratings of 12 items describing proactive aggression, reactive aggression, and nonspecific aggres- sion were completed on 259 elementary school boys (Dodge & Coie, 1987). The results were entered into a principal components analysis from which two factors were retained, each comprised of the three items loading most strongly on the factor. The first factor was labeled Reactive Aggression, accounting for approximately 69% of the total variance, and the second factor was labeled Proactive Aggression, accounting for approximately 6% of the variance. The factors had high internal consistency (alpha = .90 and .91, respectively), although the correlation between the factors was .76, in- dicating approximately 50% shared variance. As predicted, Reactive Ag- gression but not Proactive Aggression was associated with hostile attributional biases (Dodge & Coie, 1987). Price and Dodge (1989) found that teacher ratings of reactive and proactive aggression in a normative sample of kindergarten and first-grade boys each were related significantly to negative peer social status. Although these results are promising, there are several limitations of this rating scale which indicate the need for a revision. First, proactive ag- gression was defined with hostile aggression items only, despite, as noted  Reactive and Proactive Aggression 475 above, the theoretical relation of proactive aggression to both instrumental and hostile aggression. Second, items were retained which loaded highly (i.e., greater than .40) on both factors. In fact, when the factor loadings reported in Dodge & Coie (1987) were tested for significance at p < .01, 11 of the 12 items had significant loadings on both scales. 6 Third, the de- cision to retain the Proactive Aggression factor is questionable given an eigenvalue for this factor of less than 1.00 (Gorsuch, 1983). The goal of the present study was to examine the reliability and validity of a newly developed teacher rating scale assessing reactive and proactive aggression. The content validity of items included in the scale was evalu- ated, and the factor structure and internal consistency of retained factors was assessed. The concurrent validity of the retained factors was assessed by comparison to peer social status and to in-school detentions, the later being a measure of highly disruptive school behavior. METHOD Teacher Rating Scale A 28-item rating scale was developed with equal item representation assessing proactive aggression, reactive aggression, covert antisocial, and prosocial behavior derived from a review of existing scales (Dodge & Coie, 1987; Loeber & Schmaling, 1985; Stott, McDermott, & Marston, 1987). Covert antisocial items (e.g., lying, stealing) were included to sample in- strumental-proactive behaviors. Prosocial behaviors were included to mini- mize response set bias. Items were presented on a 3-point scale (0 = never, 1 = sometimes, 2 = very often). The content validity of the 21 antisocial items as related to the theo- retical constructs of proactive aggression, reactive aggression, and covert antisocial was assessed by an evaluation of the agreement of six inde- pendent raters (one undergraduate psychology major, two master's level psychologists, and three doctoral level psychologists). Raters were provided with definitions of each theoretical construct and asked to identify the con- struct associated with each item. Perfect agreement was obtained on 17 of 21 items. On three items, one rater differed from the others: One proactive item was scored as covert ( Changes the rule to help him win ), a second proactive item was coded as reactive ( Hurts others to win a game ), and one reactive item was coded as proactive ( Won't admit anything is his fault ). Another reactive item was coded by three of six raters as proactive 6Complete results of these analyses are available from the first author.  476 Brown, Atkins, Osborne, and Milnamow ( Fights with other children for no reason ). The kappa coefficient (Fleiss, 1971) for the total antisocial scale was .93. Subjects and Procedure The scale was administered to the teachers of third-through fifth-grade boys (N = 186) from 15 regular classrooms of a public school in an urban suburb of Philadelphia. The school served a predominantly white lower middle class community. Ratings were obtained only for boys to allow com- parison to earlier studies with male-only samples (e.g., Dodge & Coie, 1987; Price & Dodge, 1989), and due to constraints on teachers' time which pre- cluded the collection of ratings on the entire third to fifth grades. Negative Peer Status. Positive ( Name three students you like ) and negative ( Name three students you don't like ) nominations were col- lected from male peers in each class, and converted into z-scores within each class (Coie, Dodge, & Coppotelli, 1982). Scores for social preference (positive nominations minus negative nominations) were reverse coded to provide a measure of negative peer social status (Coie et al., 1982). In-School Detentions. Disciplinary records were reviewed for boys in grades 3 to 5 for the schoolyear in which data were collected. Detentions (N = 309) were coded as one of the following mutually exclusive categories: Inappropriate behavior (n = 155, 50.2%), physical aggression (n = 86, 27.8%), disrespectful behavior (n = 43, 13.9%), verbal aggression (n = 14, 4.5%), and destroying property (n = 11, 3.6%). For purposes of analysis, categories were collapsed to form a total score for detentions. RESULTS The means for the 21 antisocial items ranged from .12 to .71 SDs ranged from .38 to .75). The 21 antisocial items were entered into a prin- cipal-components analysis using an oblique (Promax) rotation. An oblique rotation was used rather than an orthogonal rotation as used by Dodge and Coie (1987) due to the expected high intercorrelation expected be- tween proactive and reactive factors (Frick, O'Brien, Wootton, & McBurnett, 1995; Harpur, Hare, & Hakstian, 1989). Criteria for the accep- tance of a factor solution required the factor to (a) obtain an eigenvalue greater than 1.00, (b) meet the requirements of Cattell's (1966) scree test, (c) contain a minimum of three items loading .40 or greater, and (d) be theoretically sound and meaningful. Results of the factor analysis indicated two factors which satisfied the requirements for factor independence. These results are presented in Table I.  Reactive and Proactive Aggression Table I. Rotated Factor Loadings for the Teacher Ratings of Aggression a 477 Factor Item I II Proactive aggression 3. Plays mean tricks (P) .84 .07 24. Tells things that aren't true (C) .82 -.02 5. Takes things from other students (C) .81 .06 12. Says mean things about others (C) .80 .09 19. Has hurt others to win a game (P) .78 .13 27. Threatens others (P) .78 .06 7. Picks on smaller kids (P) .78 .02 21. Gets others to gang up (P) .76 -.02 18. Does sneaky things (C) .69 .19 4. Acts out behind teacher's back (C) .57 .22 Reactive Aggression 11. Mad when doesn't get his way (R) -.02 .91 2. Gets mad when corrected (R) .00 .87 10. Blames others (R) .12 .76 20. Poor loser (R) .24 .72 17. Gets mad for no good reason (R) .19 .65 26. Won't admit his fault (R) .16 .57 Unclassified items 6. Needs to be the leader (P) -.04 .53 9. Causes trouble but not caught (C) .32 .26 14. Fights for no good reason (R) .57 .62 15. Changes rules to win (P) .43 .67 25. Writes things on walls (C) .42 -.10 Eigenvalue 11.32 1.93 Percent total variance 53.9 9.2 aEach item's theoretical factor in parentheses: C = covert antisocial, P = proactive, R = reactive. N = 186. The first factor contained five proactive aggression items and five cov- ert antisocial items. This factor, labeled Proactive Aggression, had an eigen- value of 11.32 and accounted for 53.9% of variance. The second factor contained six reactive aggression items. This factor, labeled Reactive Ag- gression, had an eigenvalue of 1.93 and accounted for 9.2% of variance. Items were unclassified due to high loadings on both factors (item 14), low loadings on both factors (items 9), and relatively low correlations with the factor on which they loaded (item 6, r = .51 with Reactive Aggression; item 25, r = .48 with Proactive Aggression). Item 15 was dropped from the Reactive Aggression factor to maintain the congruency of the factor. The retained factors had high internal consistency (alpha coefficients = .94 and .92 for Proactive Aggression and Reactive Aggression, respectively) and were correlated moderately highly with each other (r = .70, p < .001).
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks