BAYESIAN STRUCTURAL EQUATION MODELS: ANALYSIS OF TWO ALTERNATIVE MODELS ABOUT SAFETY PERCEPTION

Lately, the need to incorporate Bayesian estimations more strongly into the field of Structural Equation Models (SEM) has become evident. In fact, this approximation, added to those coming from robust estimations, is the newest approximation in structural estimation. The aim of this paper is to evaluate the effect of Bayesian estimation in obtaining structural parameters in both structural models dedicated to safety perception. The latter feature has become an especially prominent subject in big cities nowadays and it has been traditionally approached through SEM. This was done by adapting the results obtained by Carro et al. [3] into two structural models for safety prediction and perception and by treating them with structural parameter estimation with a noninformative prior function and under the Bayesian inference assumptions. The results we obtained showed a regular systematic effect as regards obtaining standard errors in structural estimations systematically smaller than those obtained by means of robust estimation techniques. These results insist on the conception of Bayesian estimation as a more appropriate option in order to obtain more efficient estimations.