Ok, let's have a look at some stats here.
The number of children killed or seriously injured in a road traffic accident in Great Britain in 2009 was 2,671, of which 1,660 were pedestrians.
"?Reported Child casualties fell by 6 per cent. The number of children killed or seriously injured in 2009 was 2,671 (down 5 per cent on 2008). Of those, 1,660 were pedestrians, 7 per cent down on 2008. 81 children died on the roads, 43 less than in the previous year, a reduction of over a third."
That means that only about 1,000 children were injured in cars out of maybe 10mio (my guesstimate). So, the odds of being injured would be about 1 in 10,000 (less than the chance of a woman dying in childbirth, for instance). As for dying, it is less than one in 100,000. Then I can ask the question: in how many of these injuries/fatalities would a rear/front facing child seat make a significant difference. If we assume 1/3 (I doubt it is that high), we are talking taking a 1/30,000 risk of any significant increase in injury or a 1 in 300,000 chance of an increase in mortality.
The numbers above are really tiny. Unless you are of the school that believes safety is paramount and wraps your child in cotton wool all the time, then you have to balance safety versus comfort and quality of life. For me, after about 1 year old, being able to face forwards and interact with one's environment and other members of the car trumps the miniscule extra risk taken.
So, in conclusion, YABU!