SAMPLE SIZE CALCULATION IN STRUCTURAL EQUATION MODELING OF EQUILIBRIUM

Abstract

The aim: To pave the way and exemplify sample size calculation to studies with complex data structures describing equilibrium. Materials and methods: Try is probated to apply ad-hoc power analysis to structural equation modeling (SEM) of equilibrium. As example we use theoretical structural equation system that describes equilibrium of price, quality and comfort of health services developed by Dranove D., Satterthwaite M. (1992). We show the way of transition from theoretical balance models to SEM that can be processed with common statistical tools. SEM is prerogative for such transition as demonstrated in the paper. We introduced some new ideas to support transition. We use Satorra & Sarris (1985) method of ad-hoc power analysis to SEM. Results: The sample size to support error types 1 and 2 at arbitrary accepted levels 0.05 and 0.2 is 400 at least to test the influence of equilibrium price (p*) on equilibrium quality (q*). 600 sample size is needed to check for the influence of equilibrium price (p*) on equilibrium comfort (c*). Sample size of 600 is required to test hypothesis on informational noise about quality influences equilibrium value of quality. Conclusions: It’s new ground that we are tentatively exploring in paper concerning SEM of equilibrium. The main challenge as we see it is the transition from theoretical balance models to SEM that can be processed with common statistical tools. SEM is prerogative for such transition as demonstrated in the paper.