Paper
Thursday, July 22, 2004
This presentation is part of : Instrument Development
Revalidation of the Barton Hospital Picture Test
Mary Megel, PhD, RN and Margaret E. Wilson, PhD, CPNP. College of Nursing, University of Nebraska Medical Center, Omaha, NE, USA
Learning Objective #1: Describe the revised Barton Hospital Picture Test
Learning Objective #2: Describe the psychometric properties of the revised Barton Hospital Picture Test

Objective: To assess the revised Barton Hospital Picture Test (BHPT), a projective instrument for measuring stress in hospitalized children, for test-retest reliability, inter-rater reliability, and discriminant validity.

Design: Descriptive, correlational approach to instrument development.

Population, Sample, Setting, Years: The population includes non-hospitalized and hospitalized children aged 5 through 9 years. The sample of 75 children included 41 from rural and urban communities and 34 from hospital settings.

Concept or Variables Studied: Stress associated with hospitalization in childhood is the primary variable studied.

Methods: The original BHPT was found to depict out-dated hospital scenes that did not reflect current reality in terms of equipment, nurses’ uniforms, and the experiences of the children. The eight pictures in the projective test were revised. Children were recruited from the hospitals and community, and after obtaining informed consent from a parent and assent from the child, the children were shown each of eight pictures of hospital experiences and asked to “Tell me a story about this picture.” Each story was audiotaped, transcribed verbatim, and divided into sense lines. Each sense line was coded by two investigators as “no stress” or one of four types of stress: anxiety-fear, anxiety-defense, dependency, or aggression. A small subsample of community children was tested within two weeks of the initial interview for test-retest reliability.

Findings: Using two-tailed t-tests, the stress scores for the hospitalized children were significantly higher than the community children. Interrater reliability was 0.84 (kappa coefficient). Gender and age were not significantly related to stress scores. There was no significant change over time in mean stress scores for children who were retested.

Conclusion: The results support the discriminant validity and test-restest reliability of the revised BHPT.

Inplications: The revised BHPT may be used in research involving stress in hospitalized children.

Back to Instrument Development
Back to 15th International Nursing Research Congress
Sigma Theta Tau International
July 22-24, 2004