What are the chances that a deadly asteroid will hit Earth in the next decade? What are the chances that an asteroid that will kill multiple people will hit Earth in the next decade?
 A: Dr. Phil Plait has written about this extensively.  He has a book (Death from the Skies) with a chapter that deals with this.
He has a blog entry about this very subject as well (in addition to a link to one just talking about getting hit by a meteorite).  Here is an excerpt:

what are the odds of getting killed by one?
Turns out, they’re a lot higher! Why?
Because you are small and the Earth is big, getting knocked on the noggin by a meteorite is a low odds event. But a big meteorite, say one 100 yards across, doesn’t have to directly fall on top of you to shuffle you off this mortal coil. It could land kilometers away and the blast wave (or the heat) could do you in. And a bigger one can land hundreds of kilometers away and still snuff you out, especially if it hits in the ocean and causes a big tsunami to march over the beaches and coastlines.
However, big asteroids coming in and whacking us are much rarer than small ones; if you go out on a clear night you might see a dozen meteors caused by rocks smaller than a grain of sand, but you could wait 100 million years for a dinosaur-buster. You have to account for that as well. This is a calculation worth doing, because a) a lot of people fret about it, and b) it could in fact mean the end of all life on Earth. That might be worth knowing.
Astronomer Alan Harris has made that calculation. Allowing for the number of Earth-crossing asteroids — the kind that can hit us because their orbits around the Sun intersect ours — as well as how much damage they can do (which depends on their size), he calculated that any person’s lifetime odds of being killed by an asteroid impact are about 1 in 700,000.
One out of seven hundred thousand! That’s still pretty low… and certainly not enough to lie awake at night worrying about it.
But there are a few important things to consider.
1) A big asteroid is rare, but one bigger than about 10 km across would kill everyone, all 6 billion of us. That skews the odds. If one of those hit every 100 million years, then your lifetime odds of dying in an impact is 100 million years divided by 70 years = 1 in 1.5 million.
A small impact might happen 1000x more often (every 100,000 years), but might only kill 1/1000th as many people, so the odds are roughly the same. Weird.
2) We are lousy at understanding low probability events. I know that 1 in 700,000 is a ridiculously low probability, but it’s hard to grasp. As a comparison, you’re more likely to die in a fireworks accident. But what’s funny is, this is a slightly higher chance than being killed by a terrorist! Despite propaganda to the contrary, the odds of any given person being killed by a terrorist attack are incredibly low. While terrorist attacks in the long run are a near certainty, the odds of you getting killed are very low.
It’s like the lottery: someone wins every time (eventually), but chances are it won’t be you.
Worrying about preventing a terrorist attack is a good idea, but (unless you work in a high-risk job) worrying specifically about dying in one is not*.
Incidentally, you have about the same odds as being killed on an amusement park ride. Wheee!

It really all depends on how you look at it.  In his blog entry, he does get into how as a species, we actually have the ability to prevent this sort of a disaster with a relatively modest amount of spending and effort.  So the odds may actually decrease as we get our act as a species together.
A: For statistical purposes, the difference between the odds of two people dying and the odds of one person dying are not signficant so, depending on who you reference...
1 in 200,000
1 in 700,000
1 in 40,000 << This seems pretty high, considering that this is about the rate of death by chicken pox.  However this also includes comets and other things besides just "meteorites"
1 in 765,000,000 << risk due to major impact
A: Hm. To answer that question, one would simply count such occurrences for a given time in the past, and divide it through the timespan. Then do some corrections for the increased number of cities in the last centuries. 
The only bigger impacts which where reported from people was the Tunguska-Event 1908 and 1490 in China, 10 000 people died, according to the German wikipedia on asteroid impact and the corresponding article in English, though the article says that at least some astronomers find the number of deaths implausible.
There I found the calculation of a big enough impact for climate changes every 500 000 to 10 million years (not very accurate, is it?). 
I can sell you an insurance policy. ;)
A: The chances are of the order of 1 in 10,000. You can derive this number by assuming the most probable impact as the chance order of magnitude (there's a 1 in 10,000 chance in 2019).
All that you wanted to know about near Earth objects, dangerousness and probability of impact is here: http://neo.jpl.nasa.gov/risk/
A: The odds of a very-high-casualty rate impact in the next decade are even lower than previous answers have stated, since there are ongoing surveys looking for just such dangerous "near Earth objects" and "Earth approachers" and there are no known near-term threats: JPL's Current Impact Risks
New survey instruments are coming online now. For instance, the Pan-STARRS system is designed to cover the entire sky several times per month. 
Of course, no survey is going to pick up every meteor capable of killing a person or causing people to die (pea-sized meteor hits airplane pilot, etc.) 
