The link between meteorology and social/behavioral science is an interesting field of study. I'm certainly no expert on that. However, one of my grad school professors was interested in that field. If I remember correctly, the correlation between false alarms and people not responding to the next severe threat was pretty weak. For many people, they will either take watches and warnings seriously or not based on their own personal trust in the forecast(or belief in science in general). That trust is probably independent of any previous false alarms. I'm sure we've all heard people say "they never get it right". That statement is obviously false and they probably know that at some level, but you would never get those people to take shelter during a warning. They would probably need to visually confirm a tornado before they would seek shelter.