The experimental plans in the psychometric field (but also in marketing, customer satisfaction and other areas) are commonly characterised by a variety of aspects of which we need to verify the dependence with a variable which identifies various populations. Secondly there is the problem of searching for the single factors which significantly contribute to the explanation of the phenomenon. This requires the use of procedures which control the multiplicity of the considered factors. The single aspects are often in turn generated by a combination of several sub-aspects (items, pathologies or behaviours). The presented instrument provides for a (strong) control of multiplicity (FWE) for the main factors and supplies the p−values for the sub-aspects of the factor itself. Should there be categorical variables, a suitable decomposition of the variable itself makes it possible to consider the single modalities of the random variable as sub-aspects of the factor and makes it possible to calculate the p−values for each one. An application to data from psychometric experiments is presented.
Non Parametric Multifocus Analysis
BONNINI, Stefano;
2003
Abstract
The experimental plans in the psychometric field (but also in marketing, customer satisfaction and other areas) are commonly characterised by a variety of aspects of which we need to verify the dependence with a variable which identifies various populations. Secondly there is the problem of searching for the single factors which significantly contribute to the explanation of the phenomenon. This requires the use of procedures which control the multiplicity of the considered factors. The single aspects are often in turn generated by a combination of several sub-aspects (items, pathologies or behaviours). The presented instrument provides for a (strong) control of multiplicity (FWE) for the main factors and supplies the p−values for the sub-aspects of the factor itself. Should there be categorical variables, a suitable decomposition of the variable itself makes it possible to consider the single modalities of the random variable as sub-aspects of the factor and makes it possible to calculate the p−values for each one. An application to data from psychometric experiments is presented.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
