GNGTS 2015 - Atti del 34° Convegno Nazionale

Findings. A representation of the CATI survey data is proposed in Fig. 1. The graph shows the mean scores of the factors divided into two subgroups for PGA>0.15 (red line) and PGA<0.15 (blue line). The black dotted line indicates a hypothetical perception level in case of PGA>0.15. Means and standard deviations are shown in Tab. 2 according to questionnaire factors and PGA. Tab. 2 - Means and standard deviations for factors and PGA. Questionnaire Factors     PGA<0.15     PGA>0.15      General Mean StDev Mean StDev Mean StDv Hazard 3.03 1.053837 3.53 1.225083 3.23 1.153096 Exposure 4.11 1.587015 4.18 0.814899 4.14 0.776802 Vulnerability 2.83 1.251397 2.77 1.359365 2.80 1.296184 People and Community 3.60 1.099334 3.76 1.190593 3.66 1.139365 Phenomenon 4.39 0.682505 4.57 0.818415 4.46 0.746004 Total 3.59 1.158801 3.76 1.261732 3.66 1.204460 Tab. 3 - Differences between means of the CATI and the web survey. Hazard Exposure Vulnerability People & Phenomenon Community CATI [PGA<0.15]- -0.58 -0.41 -1.55 -0.84 -0.58 Web data Zone 3-4 CATI [PGA>0.15]- -1.26 -0.19 -0.83 -0.54 -0.76 Web data Zone 1-2 Furthermore, we compare data collected on the web (N=8,572) with data of the CATI survey (N=4,012), as shown in Fig. 3. As a whole, the data collected on the web have higher perception scores than the CATI survey data. Tab. 3 reports the differences between CATI survey and Web data sets. Fig. 2 – Distribution in classes of Hazard perception values for the PGA intervals. 296 GNGTS 2015 S essione 2.3