Files and links (1)
Friedman_Stephanie_19987.84 MB
PDF Restricted Access, VIEWABLE UPON REQUEST: contact archives@drexel.edu
Dissertation
Behavioral-analytic model of test construction: a measure of caregivers' cancer knowledge
Doctor of Philosophy (Ph.D.), Allegheny University of the Health Sciences
May 1998
DOI:
https://doi.org/10.17918/00008797
Abstract
Family members and caregivers of cancer patients often experience psychological distress and emotional burden. One area that has been identified as potentially important to reducing caregiver distress is the provision of information about the disease, its course, treatment, and prognosis (Houts et al., 1991). However, psychoeducational interventions aimed at providing caregivers with increased knowledge have shown varied effects in reducing distress. One possible reason for these varied effects may concern the active measurement of caregiver knowledge. To best assess the potential efficacy of such psychoeducational programs, an empirically-derived measure would provide the best evaluation of actual changes in caregiver knowledge. Therefore, this manuscript will describe the development and initial psychometric assessment of a caregivers' cancer knowledge questionnaire (Cancer Knowledge Questionnaire; CKQ). The development of the CKQ is in accordance with the behavioral-analytic model of test construction (Goldfried & D'Zurilla, 1969). As such, it was constructed with an emphasis on content validity. A preliminary study consisting of semi-structured interviews with lay-caregivers and healthcare professionals revealed four specific topic areas identified as "useful" and helpful in reducing caregiver distress: (a) background and general cancer information, (b) emotional effects of cancer, (c) physical effects, medications, and treatments, and (d) social services, support, and acquisition of information. The alternate-choice CKQ was derived from the information gleaned during these interviews. The methodology for test construction, item validation, and psychometric evaluation followed empirically-supported recommendations specific to writing multiple-choice tests (Haladyna, 1994). Five studies are described. Three item analysis studies with a total of 272 participants resulted in 36 content-valid alternate-choice questions with established plausibility of distractor options. Questions reflect three content domains (facts, concepts, principles), two cognitive domains (defining behaviors, predicting behaviors), and two content dimensions (procedural and declarative knowledge). Study #4 with 47 participants evaluated test-retest reliability which was estimated to be high (Pearson r = .75, p = .000). The fifth study, contrasted groups comparison of 26 oncology experts and 26 matched-controls, demonstrated the ability of the CKQ to discriminate between persons with and without knowledge about cancer. Implications of the studies' results and plans for future studies are discussed.
Metrics
11 Record Views
Details
- Title
- Behavioral-analytic model of test construction
- Creators
- Stephanie H. Friedman
- Contributors
- Christine M. Nezu (Advisor) - Drexel University, Allegheny University of the Health Sciences (1996-1998)
- Awarding Institution
- Allegheny University of the Health Sciences
- Degree Awarded
- Doctor of Philosophy (Ph.D.)
- Publisher
- Allegheny University of the Health Sciences; Philadelphia, Pennsylvania
- Number of pages
- xi, 120, [93] pages
- Resource Type
- Dissertation
- Language
- English
- Academic Unit
- Allegheny University of the Health Sciences (1996-1998); Clinical and Health Psychology [Historical]; School of Health Professions (1996-1998)
- Other Identifier
- 991021888758604721