Improved Methodology: A Factorial Survey Study


Principal Investigator: RUTH E LUDWICK
Affiliation: Kent State University
Country: USA
Abstract: The purpose of this study is to evaluate the validity and reliability of the factorial survey method for implementing an experimental design. Second, we test a method to access nurses in nursing homes as subjects. This preliminary work will serve as foundation for an R01 grant submission. Validity will be assessed by use of the recent case method identifying the isomorphism of the vignettes with real life situations. Test-retest reliability of responses will evaluate the stability of both individual and collective judgments. Strengths of the study include expansion and evaluation of the factorial survey to nursing; and ease of implementation of an experimental design that does not compromise day-to-day nursing practice. The use of semi-structured interviews will be used to determine whether any resident data was missing in the vignette or whether additional interventions might be appropriate. Based on previous studies, three dependent (patient problems) and 10 independent variables (patient conditions) related to elder patient acute confusion are used in the vignettes. Additionally, the impact of two nurse characteristics and organizational characteristics are assessed. Two hundred nurses working in nursing homes will each be given three vignettes outlining a clinical problem for which a clinical decision response is asked. The sample of 600 vignettes will allow for a power of .85 likelihood of detecting a 3% increment to R-Squared. A test-retest of the vignettes will be done on a 10% subsample of respondents who score as either high or low outliers. This 10% subsample will also participate in the semi-structured interviews, which will immediately follow the test-retest of the vignettes. Descriptive statistics, regression, and a repeated measures within subjects ANOVA will be used to analyze the data.
Funding Period: 2003-06-01 - 2004-11-30
more information: NIH RePORT