Abstract
The lower degrees of the macroseismic scales are estimated mainly using the diagnostic related to the percentage of people who felt the earthquake (“felt” percentage). However, estimating this percentage by examining only a sub-sample of people may be biased, as those who complete the questionnaires are predominantly those who have felt the earthquake. One way to solve this problem is to send a specific request to them when an earthquake occurs. The 'Hai Sentito il Terremoto' (HSIT) site began implementing this strategy in December 2008 by allowing users to register on the website. Thanks to contributions from registered users, the HSIT database has been populated with "not felt" reports, which since 2009 have outnumbered the “felt” ones. The diagnostics of registered and unregistered users were analysed separately in a sample of municipalities, showing no difference between the two types of users with the exception of “not felt” reports number. Despite the availability of the "not felt" reports, the experimental “not felt” number is still underestimated when compared to those indicated in the Mercalli-Cancani-Sieberg scale. It was therefore necessary to multiply the number of "not felt" reports by a correction factor, estimated to be 6, to properly use the “felt” percentage as a diagnostic effect. The macroseismic intensity and event magnitude do not affect this correction factor. By using both on-request reports and a correction factor it is possible to overcome the problem of underreporting "not felt" reports, more accurately estimate low intensities, and better define the “felt”-“not felt” boundary.