GNGTS 2013 - Atti del 32° Convegno Nazionale

overlaps with the latter. However we did not have direct access to the INGV list and emails were sent by staff of the Institute. After one invitation and two reminders, we received 241 valid questionnaires from the GNGTS list (response rate = 16.4%) and 138 from the INGV list (response rate unavailable) 2 . The low response rate from GNGTS is quite common in online survey: our previous questionnaire on science communication, addressed to a large sample of Italian academic scientists (ISAAC wave 1), registered a response rate of 17.4%. Unlike this case, however, we don’t have a priori information about the target population of Italian seismologists, so we are not able to evaluate the representativeness of our sample. The latter can be characterized as follows: • mean age is 46.1 • 64.9% are men • 12.9% are university professors, 52.3% are research officers (at university and/or research institutes), 12.1% are technicians and 22.7% have other roles/positions (PhD students, post-docs, professionals, etc.) • 62.1% belongs to academic/scientific associations or did so in the past Moreover, based on reported main research interest, about 60% can be unambiguously classified as seismologists, while the rest are spread in other research fields (applied geophysics, volcanology, environmental geophysics, other). The questionnaire included 19 closed-ended questions (52 items) focused on 4 topics: communication activities, attitudes toward the public of science, politics and the role of scientists, the case of l’Aquila (opinions on the sentence, its consequences, media coverage and media treatment of the case, the role of Giampaolo Giuliani). Twelve final questions provide information on the respondent. Most opinion questions are Likert-type and require respondents to rate sentences on a 1-5 scale with labels only at extreme points (1=completely disagree; 5=completely agree). Principal component analysis and homogeneity analysis (Cronbach’s alpha) revealed that the answers to three sets of questions can be summarized with summative scales. After appropriate reverse scoring, item ratings were averaged and three synthetic variables were created: attitude toward the sentence (6 items, alpha=0.83), attitude toward the media (5 items, alpha=0.65) and attitudes toward the role of Giuliani (5 items, alpha=0.70). All the resulting variables were re-scaled 0-10 where 10 means the most negative and 0 the least negative attitude toward the object probed (sentence, media, Giuliani). Results. In this section, we comment upon the distributions of the most relevant items proposed in our survey. Particular attention will be devoted to any differences between INGV and GNGTS sub-samples. 3 Topic 1: communication practices and attitudes. The answers to the question about involvement in communication activities during the last 5 years show that INGV researchers are more involved in such activities than GNGTS researchers, both during seismic events (35.8% vs. 23.9%) and during non-seismic periods or in low-risk areas (60.9% vs. 52.7%). INGV researchers, indeed, feel more prepared for public speaking during seismic events than GNGTS researchers (43.7% vs. 39.1%) and are more likely to have attended at least one specific course of risk communication (24,6% vs. 14.6% *). Consistently, a higher percentage of INGV researchers thinks that is important to attend courses on risk communication, public speaking and media analysis. The necessity of a specific formation, as opposed to a generic communication experience, is acknowledged more by INGV than by GNGTS researchers (76% vs. 63%*). However, data on participation in specific communication courses show that this research community (and/or their research institutions) pay more attention to the 2 It should be noted that the GNGTS email list includes also a number of people not directly involved in seismology. Thus, the response rate could be higher than actually calculated. 3 Statistically significant differences ( p value < 0.05) are marked with an asterisk. 136 GNGTS 2013 S essione 1.1