Construct Modeling, Using Rasch Diagnostics and Analysis to Refine an Instrument that Measures Constructs of Diversity

Thursday, 15 July 2010

Mary Bourke, RN, MSN, PhD
April Mouser, BSN, MSN
School of Nursing, Indiana University, Kokomo, IN

Learning Objective 1: understand Construct Modeling which is a process used to develop survey instruments.

Learning Objective 2: understand how Rasch analysis allowed a more appropriate analysis of survey data and psychometric properties of the instrument.

Purpose: This study used Rasch Model diagnostics and analysis to refine an instrument that will be used to measure nursing program outcomes related to constructs of diversity and ethics. In order to facilitate measurement of these constructs an instrument was created based on construct modeling (Wilson, 2005). The process includes the following: construct map, the design plan, the outcome space, and the measurement model. Methods: Several iterations of the instrument were developed after consultation with peers and content experts in diversity and ethics. The pilot stage of instrument development began by surveying freshman and senior nursing students. Results: Results were analyzed using the Rasch Model of statistical analysis. Rasch analysis allowed a more appropriate analysis of the data and provided detailed information about the psychometric properties of the instrument. Conclusion: Tool diagnostics were performed as follows: category frequencies and average measures, as well as, threshold estimates, probability curves and category fit statistics. Validity was articulated within the context of the Rasch model. Person and item reliability indexes were clarified. Lastly, Dif analyses was performed to determine instrument sensitivity to the impact of instruction related to tolerance towards diversity and ethics during four years of baccalaureate education.