This section provides JVQ users with guidelines for clinical interpretation and information on the reliability and validity of the JVQ-R2. It is also appropriate for individual users to develop their own strategies for interpretation, just as it is appropriate to select the most relevant modules or items for administration.
Can diagnoses be based entirely on JVQ-R2 scores?
We discourage using answers to structured self-report questionnaires as the primary basis for clinical diagnoses. This includes determinations of whether mandatory reporting is triggered. Rather, we suggest that for the purposes of clinical assessment and diagnosis, that these responses are better used as starting points for further inquiry.
It is the case that one major goal of the JVQ is to enhance the correspondence to official categories, especially for items in the Child Maltreatment and Sexual Victimization modules, which are designed to represent the types of offenses typically investigated by child protection agencies. Nonetheless, positive responses to these and other items that may represent reportable offenses are not diagnostic in and of themselves. Children may misunderstand the question or otherwise mistakenly give an incorrect positive response to a screener question. Additional one-to-one interviewing or other investigation would be required in order to determine whether a report to the JVQ constitutes a reportable offense.
Comparison to National Rates and Patterns
One unique feature of the JVQ-R2 is the availability of nationally representative data. These can be treated as norms to determine when a child’s reports are substantially higher than those of a typical child. These are also available by gender and for selected age groups. The rates are available in the following publication:
Finkelhor, D., Turner, H., Ormrod, R., & Hamby, S. (2009). Violence, abuse and crime exposure in a national sample of children and youth. Pediatrics, 124(5), 1411-1423.
Reliability and Validity
The JVQ-R2 was developed in an intensive process that is one of the most rigorous scale development projects ever conducted. The completed draft instrument has undergone extensive review to ensure that each item has conceptual integrity and is phrased in ways that are developmentally appropriate for children. For both research and clinical purposes, it is important to establish the reliability and validity of scores.
Review by professional colleagues. The first phase of review involved comments from researchers at the Family Research Laboratory and Crimes Against Children Research Center, including critiques by Murray Straus, Ph.D. and Glenda Kaufman Kantor, Ph.D.
The second phase of professional review sought feedback from additional experts in the fields of victimization and measurement. These included critiques by: Victoria Banyard, Ph.D., University of New Hampshire; Lucy Berliner, Ph.D., University of Washington; Kathy Kendall-Tackett, Ph.D., University of New Hampshire; Mary Koss, Ph.D., University of Arizona; James Lynch, Ph.D., American University; Harriet MacMillan, M.D., M.Sc., F.R.C.P. (C), McMaster University; and Joy Osofsky, Ph.D., University of New Orleans.
Focus groups with parents and teens. Review of the draft instrument by parents and youths was considered an essential step in developing an instrument that produces the most accurate data on victimization rates. The first stage of focus group review was parent meetings. Due to the length of the questionnaire, there were six groups, each focusing on one of the six original modules (a module on Extraordinary and Catastrophic victimization was later incorporated into other modules). The groups were recruited by: the staff of the University of New Hampshire Cooperative Extension office, an advertisement in the Concord (NH) Monitor newspaper, and personal contacts. The following groups participated: College of Lifelong Learning Behavioral Science students; the Association for American Mothers; three groups affiliated with the UNH Cooperative Extension; and a group of teachers.
The second stage of focus group review was adolescent meetings. These were conducted after the instrument draft was revised in response to the parent focus group suggestions. The groups were composed of adolescents recruited from a program for teen parents and a non-profit organization that provides support services to parents and teens.
The focus groups produced feedback about ways to make the language of the questionnaire simpler, ways to make the items relevant for youth living in both rural and urban settings, and word choices that are more likely to be understood and/or used by children in their everyday lives. For example, we received considerable feedback that “private parts” was the word for genitalia that was taught in most school and family settings, including in many child abuse prevention programs. Also, whereas some professionals during our peer review questioned whether today’s youth referred to each other as “boyfriend” or “girlfriend,” parents and teens made it clear that those are still the most familiar terms and that other alternatives such as “date” or “intimate partner” would be less well understood.
Cognitive interviews with youth. A semi-structured interview version of the survey was developed in order to assess young children’s comprehension of survey items on victimization. This is a key validity question that has been seriously under-studied to date. The cognitive interview obtains detailed narrative descriptions of all victimization reports as well as probes to assess comprehension, over-reporting, spontaneous organization of categories such as frequency, and possible under-reporting due to literal interpretations of items or discomfort related to disclosure. Many of these probes were developed from focus group comments as well as the measurement and clinical experiences of the research team.
Larry Ricci, M.D., and Kerry Drach, Psy.D. collaborated with the authors on collecting cognitive interview data. A clinic serving a high-risk sample, largely comprised of youth who had been substantiated on at least one form of child maltreatment, was selected as the site so that a sufficient number of narratives could be obtained with a relatively small sample size. Twenty-four children aged 6 to 15 participated. The results of the cognitive interviews led to additional simplification and shortening of items to maximize comprehension by the youngest children.
Psychometrics: Test-retest reliability, construct validity and other test characteristics. The test-retest reliability and construct validity of the JVQ were established in the Developmental Victimization Survey, the first national sample. The full results of our psychometric evaluation are available in Finkelhor, Hamby, Ormrod, & Turner, 2005 (see publications page). A few key results from this evaluation include:
- Although all respondents were informed that they could refuse to answer any questions they chose, there were only 16 refusals (unweighted) to answer a screener or a rate of .02% out of 69,020 asked screeners.
- There are no dramatic developmental discontinuities that appear across the transition between proxy respondents and self-reports.
- We found moderate, significant correlations with trauma symptoms for all the aggregates and for most of the individual screener items as well. The correlations are in the same range as those found in most assessments of community samples of victimized children.
- The instrument showed adequate test-retest reliability in a 3 to 4 week re-administration with a subsample of 100 youth and 100 caregivers.
- Consult the full paper for additional details. The growing empirical database on the associations of JVQ-R2 responses with trauma symptoms, suicidal ideation, and other factors also contributes to the establishment of construct validity.