Meta-analysis of studies assessing the efficacy of projective techniques in discriminating child sexual abuse

doi:10.1016/S0145-2134(98)00086-6

Child Abuse & Neglect

Volume 22, Issue 11, November 1998, Pages 1151-1166

https://doi.org/10.1016/S0145-2134(98)00086-6 Get rights and content

Abstract

Objective: This meta-analysis of 12 studies assesses the efficacy of projective techniques to discriminate between sexually abused children and nonsexually abused children.

Method: A literature search was conducted to identify published studies that used projective instruments with sexually abused children. Those studies that reported statistics that allowed for an effect size to be calculated, were then included in the meta-analysis. There were 12 studies that fit the criteria. The projectives reviewed include The Rorschach, The Hand Test, The Thematic Apperception Test (TAT), the Kinetic Family Drawings, Human Figure Drawings, Draw Your Favorite Kind of Day, The Rosebush: A Visualization Strategy, and The House-Tree-Person.

Results: The results of this analysis gave an over-all effect size of d = .81, which is a large effect. Six studies included only a norm group of nondistressed, nonabused children with the sexual abuse group. The average effect size was d = .87, which is impressive. Six studies did include a clinical group of distressed nonsexually abused subjects and the effect size lowered to d = .76, which is a medium to large effect.

Conclusion: This indicates that projective instruments can discriminate distressed children from nondistressed subjects, quite well. In the studies that included a clinical group of distressed children who were not sexually abused, the lower effect size indicates that the instruments were less able to discriminate the type of distress. This meta-analysis gives evidence that projective techniques have the ability to discriminate between children who have been sexually abused and those who were not abused sexually. However, further research that is designed to include clinical groups of distressed children is needed in order to determine how well the projectives can discriminate the type of distress.

Résumé

Objectif: Cette méta-analyse de 12 études évalue l’efficacité des techniques projectives à distinguer les enfants abusés sexuellement de ceux qui ne le sont pas.

Méthode: Une revue de la littérature a permis d’identifier les études publiées ayant utilisé des instruments projectifs avec des enfants abusés sexuellement. Seules les études mentionnant des statistiques permettant l’évaluation de l’efficacité en fonction de l’effet de taille ont été retenues pour la méta-analyse. Douze de ces études répondaient aux critères. Les tests projectifs concernaient le Rorschach, le Test de la Main (Hand Test), le Thematic Apperception Test (TAT), le dessin de la famille (Kinetic Family Drawings), le Human Figure Drawings, Draw your Favorite Kind of Day, le Rosebush, Visualization Strategy et le House-Free-Person.

Résultats: Les résultats de cette analyse ont donné un effet de taille global de d = .81, ce qui est un effet important. Six études ont comparé leurs résultats à un groupe clinique d’enfants non-perturbés mais qui n’étaient pas abusés sexuellement. L’effet de taille moyen était de d = .87, ce qui est impressionnant. Six études ont comparé leurs résultats à un groupe d’enfants perturbés, mais qui n’étaient pas abusés sexuellement et l’effet de taille est descendu à d = .76, ce qui correspond à un effet moyen à important.

Conclusion: Ceci indique que les instruments projectifs peuvent assez bien distinguer les enfants perturbés des sujets non-perturbés. Dans les études, incluant un groupe clinique d’enfants perturbés, qui n’étaient pas abusés sexuellement, l’effet de taille plus bas indique que les instruments n’arrivaient pas à distinguer le type de perturbation. Cette méta-analyse met en évidence que les techniques projectives peuvent distinguer les enfants abusés sexuellement des enfants non-abusés sexuellement. Mais de futures recherches sont nécessaires avec des groupes cliniques d’enfants perturbés pour déterminer comment les tests projectifs pourraient faire la distinction entre les différents types de troubles psychologiques.

Resumen

Objetivo: Este meta-análisis de 12 estudios evalúa la eficiencia de las técnicas proyectivas para discriminar entre niños sexualmente abusados y ninos no sexualmente abusados.

Método: Se realizó una revisión de la literatura para identificar los estudios publicados que utilizaban instrumentos proyectivos con niños sexualmente abusados. Los estudios que reportaron estadı́sticas que permitieran calcular el efecto del tamaño fueron incluı́dos en el meta-análisis. Se encontraron 12 estudios que se ajustaban a esos criterios. Los proyectivos revisados incluı́an el Rorschach, el Test de la Mano, El Test de Apercepción Temática (TAT), los Dibjujos de la Familia Kinética, Los Dibujos de la Figura Humana, Dibuje su Dı́a Favorito, El Rosebush; Una Estrategia de Visualización; y el Test del Arbol.

Resultados: Los resultados de éste análisis dieron un efecto de tamaño de d = .81, que es un efecto grande. Seis estudios incluı́an sólo un grupo normativo de niños no-alterados, no-abusados con el grupo de abuso sexual. El efecto de tamaño promedio era d = .87, que es un efecto de medio a grande.

Conclusión: Esto indica que los instrumentos proyectivos pueden discriminar bastante bien los niños alterados y no alterados. En los estudios que incluı́an un grupo clı́nico de niños alterados que no habı́an sido sexualmente abusados, el efecto menor del tamaño indica que los instrumentos fueron menos capaces de discriminar el tipo de alteración. Este meta-análisis comprueba que las técnicas proyectivas tienen la habilidad de discriminar entre niños que han sido sexualmente abusados y aquellos que no han sido sexualmente abusados. Sin embargo, se necesitan más estudios diseñados para incluir grupos clı́nicos de niños alterados para determinar qué tan bién pueden los proyectivos discriminar el tipo de malestar.

Introduction

HOW WELL PROJECTIVE instruments discriminate between sexually abused and nonabused children is evaluated by this author through the application of meta-analytic strategies Cohen 1988, Howell 1992, Hunter and Schmidt 1990, Rosenthal 1987, Rosenthal and Rosnow 1991, Wolfe 1986. Meta-analysis is concerned with drawing inferences about an area of research from the results of many different studies. Meta-analysis has become associated with one set of diverse techniques, largely due to a seminal paper by Glass (1976), who categorized three levels of analysis:

1.
Primary analysis: the original analysis of data from a research study.
2.
Secondary analysis: the re-analysis of the data from a research study.
3.
Meta-analysis: the statistical analysis of results from a large number of individual research studies so as to integrate their findings (pp. 3–8).

The statistical techniques incorporated by meta-analysis vary in complexity. In the most commonly used meta-analytic techniques, the common statistic selected is an effect-size measure, an index of how large or strong the relationship is between the variables. There are six stages (as cited in Brakewell, Hammond, & Fife-Shaw, 1995) involved in the Glass meta-analytic procedure (item 3 above):

1.
Select the independent and dependent variables of interest.
2.
Identify, locate and obtain all relevant and usable studies containing information of interest.
3.
Code each study for characteristics which might be a predictor of study outcomes, that is, might relate to the size of effects obtained.
4.
Calculate estimates of the effect-size (ES) for the variable pairs (independent-dependent) of interest.
5.
Calculate estimates of the effect-size across studies.
6.
Examine those study characteristics identified in (3) above which correlate with study effects (p. 391).

The analysis that follows applies the steps of the Glass (as cited in Brakewell et al., 1995) procedure with modification and the addition of recommendations by Hunter and Schmidt (1990) in order to assess the relationship between various projective instruments and identification of childhood sexual abuse (CSA). The modification is the restriction of sources from selected data bases (PsychLit, Eric, Psych Abstracts, and Medline). This decision extends the complication of the “File drawer problem” (Hunter & Schmidt, 1990, p. 506), which refers to the fact that journals tend to include studies that contain Type I errors and exclude the studies that contain Type II errors. Using Orwin’s (1983) formula for calculating the number of studies that would be needed to lower the ES to a small effect, it was determined that 38 studies with small or insignificant findings would be needed to over-turn the present ES. In addition, studies included in this analysis have been selected from a specific ten-year period, which further biases the sample. The restriction of databases and time period prohibits this analysis from being identified as a comprehensive meta-analytic review. It is, rather, a demonstration of applying selected statistical strategies to generate a minimum effect size which can then be used to examine the efficacy of certain projectives in assessing child sexual abuse. It can also be useful in a prospective power analysis of particular projectives in discriminating sexually abused children from nonsexually abused children.

Section snippets

Procedure

The data base sample of 16 studies employed 9 projective instruments. Twelve of these studies used projective instruments to assess sexually abused children, and four studies used projectives with physically abused children. All included two or more groups and were published between 1986 and 1996. The data reported in all of these studies are appropriate for calculation of effect size. The Rorschach Inkblot Test (Rorschach, 1921) was used in three of the studies and the TAT (Murray, 1943) in

Review of studies

Six of the studies on sexual abuse were designed to include a clinic group of psychiatric or distressed subjects who were not sexually abused. Most assessment instruments seem to be able to discriminate between child sexual abuse subjects and a normal group of subjects. However, all that can be inferred in this case is that the instrument is able to detect some type of distress from nondistress. The inclusion of a clinical group with no history of sexual abuse is important because it creates

Discussion

All of the 16 studies reached a common conclusion: projective techniques can identify children who have been sexually or physically abused from nondistressed children. The 12 studies of sexual abuse assessment in children that have been reviewed here all showed medium to large effect sizes with an average d = .81, a large ES (Table 4). This offers data about the ability of the various projective techniques to significantly discriminate between children with a known history of sexual abuse and

References (53)

∗Allan, J., & Crandall, J. (1986). The rosebush: A visualization strategy. Elementary School Guidance & Counseling, 21,...
B. Aron
A manual for analysis of the Thematic Apperception TestA method and technique for personality research
(1949)
∗Blaine, G. H., Bergner, R. M., Lewis, M. L., & Goldstein, M. A. (1981). The use of the objectively scorable...
G.M. Brakewell et al.
Research methods in psychology
(1995)
Buck, J. N., & Jolles, I. (1946). H-T-P. Western Psychological...
R.C. Burns et al.
Actions, styles, and symbols in Kinetic Family Drawings (KFD)
(1972)
∗Chantler, L., Pelco, L., & Mertin, P. (1993). The psychological evaluation of child sexual abuse using the Louisville...
H. Cohen et al.
Tasks of Emotional Development Test
(1971)
J.E. Cohen
Statistical power analysis for the behavioral sciences
(1988)
P. Elkisch
Free art expression

J.E. Exner

The Rorschach

(1978)

Exner, J. E. (1986). The Rorschach: A comprehensive system (2nd ed., vol. 1). New York:...

J.E. Exner

The RorschachA comprehensive system

(1990)

Exner, J. E. (1991). The Rorschach: A comprehensive system (2nd ed., vol. 2). New York:...

R. Fine

Manual for a scoring scheme for verbal projective techniques (TAT, MAPS, stories, and the like)

Journal of Projective Techniques

(1955)

G.V. Glass

Primary, secondary and meta-analysis of research

Educational Research

(1976)

F.L. Goodenough et al.

Goodenough-Harris Drawing Test

(1963)

∗Hackbarth, S. G., Murphy, H. D., & McQuary, J. P. (1991). Identifying sexually abused children by using Kinetic Family...

∗Henderson, O. (1990). The object relations of sexually abused girls. Melanie Klein and Object Relations, 8,...

∗Hibbard, R. A., & Hartman, G. L. (1990). Emotional indicators in human figure drawings of sexually victimized and...

D.C. Howell

Statistical methods for psychology

(1992)

J.E. Hunter et al.

Methods of meta-analysisCorrecting error and bias in research findings

(1990)

K. Kendall-Tackett et al.

Impact of sexual abuse on childrenA review and synthesis of recent empirical studies

Psychological Bulletin

(1993)

∗Kinard, E. M. (1980). Emotional development in physically abused children. American Journal of Orthopsychiatry, 50,...

E.M. Koppitz

Emotional indicators on human figure drawings of childrenA validation study

Journal of Clinical Psychology

(1966)

E.M. Koppitz

Psychological evaluation of children’s Human Figure Drawings

(1968)

Cited by (28)

The use of a favourite kind of weather drawing as a discriminatory tool for children who have experienced physical abuse
2015, Arts in Psychotherapy
Citation Excerpt :
Therefore, the FKW drawing proved to have discriminant validity as a tool to diagnose abuse. The results that the FKW drawing was correlated with the experience of being abused and could discriminate the degrees of abuse agree with Manning (1987) and Epperson (1990), as well as with the results of West's (1998) research in sexually abused children. On the contrary, however, Veltman and Browne (2001) found that abused children showed no difference in FKW drawing.
In this study, objective assessment criteria were developed to test the validity and reliability of using the Favorite Kind of Weather (FKW) drawing as a projective technique to discriminate between physically abused and non-abused children. Stratified sampling was used to select 577 third- to ninth-grade school children, of whom 295 were residing in shelters and 282 living with normal families. The study instruments were FKW drawing, the Child Abuse Experience Scale. The assessment criteria for the FKW drawings that were developed in three stages by four researchers: two graduate school students, one in master's and one in doctor's course, my supervisor, and me. The criteria comprised six formal and four content elements. A three-point scale was used for these ten items. In the study results, the inter-rater consistency in the FKW drawing among the three raters ranged from 94.0% to 95.8%, with an intra-rater consistency of 97.0%. The reliability was estimated to be Cronbach's α = .721 for the formal elements and Cronbach's α = .632 for the content elements. Based on these results, the FKW drawing was validated on the basis of construct validity, convergent validity, and discriminant validity, which confirmed its validity for use as a child abuse discriminator.
The prevalence of child sexual abuse in community and student samples: A meta-analysis
2009, Clinical Psychology Review
Studies conducted internationally confirm that child sexual abuse is a much more widespread problem than previously thought, with even the lowest prevalence rates including a large number of victims that need to be taken into account.
To carry out a meta-analysis of the prevalence of child sexual abuse in order to establish an overall international figure.
Studies were retrieved from various electronic databases. The measure of interest was the prevalence of abuse reported in each article, these values being combined via a random effects model. A detailed analysis was conducted of the effects of various moderator variables.
Sixty-five articles covering 22 countries were included. The analysis showed that 7.9% of men (7.4% without outliers) and 19.7% of women (19.2% without outliers) had suffered some form of sexual abuse prior to the age of eighteen.
The results of the present meta-analysis indicate that child sexual abuse is a serious problem in the countries analysed.
Dysharmonic children, sexual abuse and drawing: A comparative approach with the Van Hutton's diagnostic scale
2005, Annales Medico-Psychologiques
Les auteurs présentent la grille mise au point par Van Hutton (1994) à partir d'une revue de la littérature américaine et d'une expérimentation comparative, utilisant le test de dessin d'une personne et le test de dessin de l'arbre, de la maison et de la personne pour diagnostiquer les abus sexuels. Ils soulignent à la fois l'intérêt de cet outil sur le plan du dépistage et de la prévention, mais aussi ses imperfections actuelles. Aussi ont-ils entrepris d'en affiner la validation en montrant d'une part que l'échelle SRC de cette grille permet de différencier significativement les enfants présentant une dysharmonie d'évolution de structure psychotique des enfants abusés. Une étude plus analytique des items de l'ensemble de la grille suggère qu'elle doit permettre de démarquer également les caractéristiques projectives des enfants dysharmoniques de leurs homologues témoins, ce qui est confirmé dans ce travail au niveau d'une seule dimension psychologique prise en considération ici : les difficultés relationnelles. Des pistes de recherches futures sont suggérées.
The authors present Van Hutton's diagnostic scale “House-Tree-Person and Draw A Person as measures of abuse in children: a quantitative scoring system” (1994). This scale (HTP and DAP) was elaborated from a review of American research and from a comparative research conducted by Van Hutton using HTP and DAP test to diagnostic sexual abuse. A French translation of this scale has been realized and used to estimate its discriminative value in a French population of children. Increasing sexual abuses and complexity to establish diagnostic justify development of adapted instruments for evaluation in this field. Projective techniques like this scale, based on drawing production, bring a real contribution to perceive psychopathological signs expressed by children when confronted to a traumatism like sexual abuse. Methodological investigation has been realized, with the four scales of the Van Hutton scale: SRC (preoccupation with Sexually Relevant Concept), AH (Aggression-Hostility), WGA (Withdrawal and Guarded Accessibility) and ADST (Alertness for Danger Suspisciousness and lack of Trust). Comparison between three groups has been achieved: control group (39 French children without problems, 24 girls and 15 boys aged 5 to 12 years), abused children (Van Hutton population, 20 children, 15 girls and 5 boys aged 7 to 12 years), and psychopathological group (39 psychotic disharmonic French children, 29 boys and 10 girls, aged 5 to 12 years). Hypothesis were advanced in the way of significant differences between control and disharmonic children, compared to abused children, for each of the four scales. On an other way, no significant difference should be found between control group and disharmonic group concerning sexual abuse signs. In spite of reduced size of groups, results show interesting data, specially for SCR scale, which appears to discriminate very well abuse children from the others two groups. Significant differences are found for this scale, as waiting in the hypothesis, as well in total result than in sex-ratio comparison. But results obtained for the others scales (AH, WGA and ADST) show no significant differences between groups: psychological dimension explored by these scales are perhaps too general, not enough specific, to allow discrimination. At last, no significant difference has been found between control group and disharmonic group concerning sexual abuse signs, as waiting in our hypothesis. In conclusion, research presented here show pertinence of SRC scale of Van Hutton Diagnostic Scale, which differentiates productions of control and disharmonic children from those of sexual abused one. The authors point out the usefulness of the scale for a diagnostic and for preventive perspective, but also its actual limitations which implies further validation brought by study undertaken here. Further researches are suggested to set up validity on a largest way.
Build-a-Person Technique: An examination of the validity of human-figure features as evidence of childhood sexual abuse
2005, Child Abuse and Neglect
This exploratory study examined the discriminant validity of 10 human-figure features commonly used by many proponents of the draw-a-person (DAP) projective technique as evidence of childhood sexual abuse. Two exploratory features were also examined.
Rather than drawing human figures, 64 children (M = 8 years, 9 months), including 19 sexually abused children, 26 nonsexually abused but emotionally/behaviourally troubled children, and 19 nonabused, nonclinical children, were instructed to build male and female figures using groups of prefabricated pieces of human-figure body parts. Unbeknownst to the children, each group of parts contained a potential sexual abuse feature (or features).
None of the human-figure features under examination, either individually, or in combination, distinguished the sexually abused group from the other two groups of children.
In accordance with years worth of DAP projective technique research, no support for the validity of the human-figure features under examination as evidence of childhood sexual abuse, was found.
French-language abstract not available at time of publication.
Spanish-language abstract not available at time of publication.
Response to Garb et al.'s comments about meta-analysis of studies using projectives with child sexual abuse subjects [2]
2000, Child Abuse and Neglect
Projective techniques and the detection of child sexual abuse [1]
2000, Child Abuse and Neglect

View all citing articles on Scopus

View full text

Original ArticlesMeta-analysis of studies assessing the efficacy of projective techniques in discriminating child sexual abuse