If you talk on a cell phone, it has been demonstrably shown that you are less focused and while driving the chances of having an accident increase because of the distraction factor. I don't think that really needs to be cited.
sure, if one person does it, it's trivial, but we have a whole society who does it than the chances spike as well, even if it's random.
As I wrote at the start, the issue is one of altering a very tiny risk to a somewhat higher or very tiny risk. It's not that there is no increase in risk at all associated with cell phone use. It's that it is still a very small risk. The risk picture is not altered substantially.
What you get at in the second paragraph is a mindset I think is dangerous to liberty. There are over 300 million people in the United States and millions...probably tens of millions and maybe hundreds of millions of vehicle trips every day. And there are undoubtably (for example) millions and probably tens of millions of instances in which cell phones are used while driving every day. Very small risks translate into what appear to be large numbers when the numbers of times those very small risk are incurred is very large. So if you think in terms of managing us as though we are a population of animals (yes, we're animals but you know what I mean) instead of treating us as individuals you're going to find plenty of reason for controlling our behavior in order to reduce very small risks to somewhat smaller very small risks.
If you're looking at the individual you might think something like this:
"if that person makes choice B instead of choice A, the chance that someone will die will be one in 1 million instead of 1 in 2 million. That means the chance that nobody will die will be 99.9999 percent instead of 99.99995 percent. Either way the chances that nobody will die approach virtual certainty. So there's no reason to interfere with the choice."
But if you're looking at that choice being made 10 million times per day in a large population you start thinking, "We could "save" over 1800 lives per year by stopping people from making that choice."
The problem with letting the second way of looking at things govern society is, again, that we make all kinds of choices that alter risk like that. The scenario I gave about the choice between staying home to watch a movie or going to a theater to watch one is indeed an example of one; whether you choose to accept it or not. And if we are going to live anything like free lives that's the way it's going to be. Since we have accepted the premise, we have opened the door for people to use it when they decide that a particular choice should be denied.
When you're looking at something like telling someone who is driving a vehicle that they must wear a seat belt or can't use a cell phone you are interfering with their choices because of something that might but probably won't happen in their individual case. And the probability that it won't is overwhelming. It's not like there is even a good chance, by any reasonable standard of what a "good chance" is, that it will happen.