That was probably a horrible title, but those length limits are killer :)
An example of what I'm talking about:
Statistics show that overall crime rates in the USA have been trending downwards for quite some time, yet many people in the general public think that, say, their kids are at huge risk for abduction by strangers. This can be chalked up to enhanced crime reporting and media coverage in the 21st century.
It's not happening more often, but it's being reported more often, so people think there's more of it. (Child abductions)
Statistics show that most uses of force by law enforcement is justified, and most police behave themselves and do their job honorably. However, since the media plays up the outrage when the statistical outliers happen (someone gets shot in murky circumstances), people begin to think that cops as a whole are all trigger-happy, un-empathetic morons.
It's not happening more often, but it's being reported more often, so people think there's more of it. (Police misconduct)
Is there a specific term that describes this phenomenon?
I'd guess that "confirmation bias" kind of applies, because when you start looking for bad things, you find evidence of more bad things, but it's really easy to forget the stats in question - however, I was wondering if there's something more specific?