I think I can, somehow. I know it sounds like pseudoscience, and I know it can be confirmation bias at work, but it happens to me quite often that I somehow "know" someone will be calling me shortly. I need to construct experiment to test this. All I can think of right now is to write the time at which I thought someone will call, and to record the times someone did call.
Alternative explanation is telepathy, but I think it's less probable ;)
Good luck constructing an experiment that (a) can measure this effect, and (b) pass muster as science. It would need to be well-controlled -- it would have to collect a guess about who is calling for each and every phone call, and compare the guesses with the outcomes. It would also have to accommodate the nonrandom nature of calls -- times and days of the week.
It's often true that someone will selectively remember those times when he guessed right, and forget the other times. It's also true that, if a person reliably calls at a certain time or day of the week, that biases the experimental outcome.
If Joe usually calls on Wednesday, and if on Wednesday you guess it's Joe before answering, how much weight should we place on the guess?
I'm not being critical, just pointing out some of the factors that prevent this sort of thing from being very useful as scientific evidence.