Browsing by Author "Cunha, Liliana"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
- Development and validation of a health and work survey based on the rasch model among portuguese WorkersPublication . Barros, Carla; Cunha, Liliana; Baylina, Pilar; Oliveira, Alexandra; Rocha, ÁlvaroThe purpose of this study was to develop the Health and Work Survey (INSAT) and examine the validity of the discomfort rating scale. Data were collected from 706 Portuguese workers from six economic sectors with the support of the Health and Work Survey (INSAT - Inquérito Saúde e Trabalho). The INSAT is a self-administered questionnaire to assessing working conditions, health and wellbeing, and to provide information to the occupational health systems in the organisations. For the survey instrument validation, the Rasch Partial Credit Model (PCM) was used to analyse item fit statistics. From the application of PCM, Person Separation Reliability was obtained (0.8761) and the value can be considered very good (>0.8). From the statistical analysis, the Overall Model fit information, given by Outfit Mean square/Infit Mean square, is between 0.5 and 1.5, meaning "Productive for measurement" and "acceptable fit overall". The INSAT items can generate predictable response patterns. We recommend that the INSAT discomfort rating scale and some other items should be reviewed in future works. In any event, this tool proves to be useful in assessing the relationship between work and health and in evaluating key main risk factors, helping to prevent problems and improving occupational health systems.
- Pointer dynamics as cognitive biometricsPublication . Cunha, Liliana; de Magalhães, Sérgio Tenreiro; Sá, Vítor J.; Santos, HenriqueBiometric authentication technologies aim to increase the precision of security systems and have given proof of their effectiveness. Cognitive biometrics, in particular, has shown the potential of raising security levels, although lacking somewhat in terms of precision and adequate research. Therefore, this study analyses the viability of the usage of the Pointer Dynamics biometric technique as cognitive biometrics, under certain conditions. To this end, an algorithm was developed, whose identity validation criterion is the execution of the individual pattern in the manner of interaction with the mouse, when selecting the grids of the graphic secret. Authentication and fraudulent authentication tests were conducted with two types of images, one with a supposed emotional meaning for the user, and one without. This was done so as to allow the comparison of results brought forth by the application of the technique with and without cognitive stimulus. Although results do not allow definite conclusions to be drawn, the fact is that the False Acceptance Rate of both the cognitive tests and graphs present marked relative variations, for thresholds between 4 and 7, which indicate that it is likely that the cognitive stimulus has increased the resistance to attempts of fraudulent authentication.