|
Abstract
|
OBJECTIVE: To describe the development and initial psychometric features of a brief self-report instrument for assessing exposure to community violence, the Children's Report of Exposure to Violence (CREV). METHOD: The CREV was administered to 228 schoolchildren between the ages of 9 and 15 years. A subset (n = 42) was retested 2 weeks later. RESULTS: The CREV has good test-retest reliability, internal consistency, and construct validity. Analyses performed by race and gender revealed no differences in frequency of exposure to community violence as assessed by the CREV. CONCLUSIONS: This initial report suggests that the CREV has promise as a self-report questionnaire, and its construction represents a significant advance in the development of psychometrically sound instruments for use in community violence research.
VioLit summary:
OBJECTIVE:
This article by Cooley et al. described the development of the Children's Report of Exposure to Violence (CREV), a self-report instrument that assesses the effects of community violence exposure on children. Principal psychometric features of this measure were presented.
METHODOLOGY:
The authors conducted a primary analysis of quasi-experimental, cross sectional, data that were collected from 228 children for the development of the Children's Report of Exposure to Violence (CREV); a self-report measure designed as a screening instrument to assess the impact of community violence exposure on children. An operational definition of community violence, "deliberate acts intended to cause physical harm against a person or persons in the community," was utilized to generate the instrument items. The CREV was designed for children aged 9-15 years. 32 items were included in the instrument. 29 items asked respondents to rate the frequency of their experiences of exposure to community violence on a five point likert scale. The likert scale categories included no/never, one time, a few times, many times, or every day. The last three items on the instrument were open-ended questions that asked the children to note other exposures to violence not assessed in the previous items. The first 29 items were summed to obtain an overall score.
The authors reported that the instrument measures four aspects of violence exposure. These included: 1) the media content area (television or film exposure), 2) reported content area (persons' reports of violence occurrence), 3) witnessed content area (directly witnessed violence), and 4) the victim content area (violence that has been directly experienced). Three victim categories were included in this last content area: self, familiar people, and strangers. The number of items allocated to each content area varied according to the violent act and victim category. For example, the victim content area did not include items associated with being killed.
The authors divided the CREV into distinctive sections according to the victim type. For example, "These questions ask about violence against a stranger. A stranger is somebody you do not know." For each content area the victim of violence was specified. For example, media content area items asked "Have you ever watched.....on TV or in the movies?;" the reported content area items asked "Has anyone ever told you that.....?;" the witnessed content area asked "Have you ever seen.....?;" and the victim area items asked "Have you ever been.....?" The verbs threatened physically, chased, beaten, robbed, shot, stabbed or killed were used by the authors to describe the types of situations surveyed in the instrument.
Once the instrument was designed, children who attended public elementary and middle schools (grades 4-7) in urban and rural communities in South Carolina were asked to have parents sign consent forms so they may participate in the study. These children made up the sample. Children were in the same school district (mean grade was 5.63, sd= 0.96). Age mean for the sample was 11.4 years (sd=1.37; range= 9-15 years). 116 respondents were female (50.9%) and 112 were male (49.1%). African-americans (n=169) made up 74.1% of the sample. 45 caucasians (19.7%), 4 hispanics (1.8%), 3 native americans (1.3%), 3 asians (1.3%) and 4 bi-racial (1.8%) children comprised the rest of the sample.
The CREV was administered together with a background questionnaire (specific information on this questionnaire was not provided by the authors). The first author and two other researchers administered the test to small groups during school hours. The authors reported that verbal definitions of the term "violence" were provided by the children which revealed that the children understood its meaning. To familiarize the children with the likert scale format a practice question was administered; "Have you ever eaten ice cream?" Administration of the instrument took approximately 20 minutes.
The authors stated that children from the first three schools were provided the opportunity to participate in the test a second time. This second testing was completed by the students at home. These data were used to test reliability.
FINDINGS/DISCUSSION:
The authors reported that score range potential for the CREV instrument was 0-116. For the media content area it was 0-20, for the reported and witnessed areas it was 0-40, and for the victim content area it was 0-16. Score distribution was examined by gender, age and the overall sample. An exploratory principal component, oblique rotation factor analysis was conducted to test construct validity. Items with factor loadings greater than .45 and an eigen value greater than one, comprised the factors. Each factor consisted of more than two items and the authors reported that 42.9% of the total variance could be accounted for in a two factor model based on the first 29 items of the CREV. Of these items, 24 loaded onto one of two factors. The remaining items were not identified as separate factors. The first factor, Direct Exposure, consisted of 21 items associated with "real-life" violence displays by direct exposure or vicarious exposure. Media exposure, the second factor, consisted of three items and was assessed using TV or movie observation, or community violence. The authors reported that the various aspects of exposure (witnessed, victim, and reported) did not load onto separate factors.
The authors found no significant statistical difference between respondents who were tested a second time and those who were not. They suggested that the second 42 respondents were therefore representative of the total sample. Test-retest reliability measured at .75 for the total score, .78 for the Direct Exposure factor and .52 for the Media Exposure factor (p<.001). The test-retest reliability for gender was .70. The reliability score for ages 9 to 12 years was .70, and for ages 13 to 15 years it was .80. For the African-American children test-retest reliability was .76.
Scale intercorrelation scores among the CREV factors were: Direct Exposure factor and Total score, .98; Media Exposure and Total score, .56; and the Media Exposure and Direct Exposure factor, .44. The authors stated that each score, while related, measured a specific aspect of violence exposure.
Internal consistency for the two CREV factors and the overall total score was estimated using Cronbach's alpha. An overall alpha value of .93 was obtained for the Direct Exposure factor with total factor correlations. For the Media Exposure factor, an alpha value of .75 was found. This factor with total factor score correlations had an overall alpha of .78.
Males' and females' scores were compared using Hotelling's T squared. No significant differences between the two groups were found for race, grade or age. Although an overall significant difference in CREV scores for males and females was found using a multivariate analysis (p<.001), the authors found no significant univariate differences for gender on the CREV total or individual factor scores (p>.05).
When 45 caucasian respondents were matched by age, gender and grade to 45 african-american respondents, no significant demographic differences were found. The authors stated no significant differences were found between the african-american and caucasian reports of exposure to community violence on either the Direct Exposure factor, the Media Exposure factor or the Total score (p>.05).
The authors reported that even though the CREV was developed as an assessment instrument of various types of exposure to community violence (e.g., through direct contact, rumor, and/or varying forms of media) and different victim types (e.g., familiar people, strangers, self), the content areas did not form distinctive factors in the factor analysis. As a result, the authors suggested that the content areas (e.g., those that assess the experience of violence by different people in specific situations) be utilized to supply qualitative, rather than quantitative, information for use in the clinical setting. It was suggested that, in clinical practice, the CREV could be used to assess how the trauma of community violence affects children, otherwise not reported. CREV responses may also be used to obtain further information on reported occurrences. The authors contended that it is particularly important to assess the impact of community violence on both children and adolescents since community violence exposure has been found to have negative emotional effects in both these groups.
AUTHORS' RECOMMENDATIONS:
The authors reported that since how to organize criteria that demonstrate the severity of community violence exposure, was not understood, the CREV cannot be used, at this stage, as a screening instrument. They suggested that further studies are required to ascertain how the CREV may be utilized in this manner. The authors also suggested that additional studies should be conducted to further test concurrent and predictive validity of the CREV. The authors concluded that the CREV has great potential as a measuring instrument used to assess the frequency of children's exposure to community violence.
(CSPV Abstract - Copyright © 1992-2007 by the Center for the Study and Prevention of Violence, Institute of Behavioral Science, Regents of the University of Colorado)
Witnessing Community Violence
Community Violence Effects
Witnessing Violence Effects
Child Witness
Juvenile Witness
Survey Instrument
Instrument Development
Exposure to Violence
Late Childhood
Early Adolescence
Elementary School Student
Senior High School Student
Self Report Studies
South Carolina
Urban Environment
Urban Youth
Rural Environment
Rural Youth
Language: en |