A problem in Strategic Human Resource Management about evaluation of the impact of a training programme on performance has been faced in the case of areal spatially referred data. Generalized log-linear Poisson models with and without spatial component (with Besag specification) have been considered and compared. A Bayesian modelling approach has been adopted in the statistical data analysis. In order to avoid computational slowness of Markov Chain Monte Carlo simulations, the Integrated Nested Laplace Approximation method has been used in model fitting. It has been shown that the spatial structure, when considered, may improve the identification of the real drivers of performance. It follows that investigating hidden spatial structure in the data set should be a good practice when using data analytic tools in Strategic Human Resource problem solving. The interactions and interplay between areas and customers who live and act in geographically adjacent districts should be always considered in model fitting;
Evaluating a Program on HR with Spatially Structured Performance Data
Marcello De Giosa
2019-01-01
Abstract
A problem in Strategic Human Resource Management about evaluation of the impact of a training programme on performance has been faced in the case of areal spatially referred data. Generalized log-linear Poisson models with and without spatial component (with Besag specification) have been considered and compared. A Bayesian modelling approach has been adopted in the statistical data analysis. In order to avoid computational slowness of Markov Chain Monte Carlo simulations, the Integrated Nested Laplace Approximation method has been used in model fitting. It has been shown that the spatial structure, when considered, may improve the identification of the real drivers of performance. It follows that investigating hidden spatial structure in the data set should be a good practice when using data analytic tools in Strategic Human Resource problem solving. The interactions and interplay between areas and customers who live and act in geographically adjacent districts should be always considered in model fitting;I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.