MarkCC from Good Math, Bad Math has discussed an article in the Manchester Evening News about a lottery scratch-off game in Britain. The premise is that the purchasers are given a target temperature, and they want to scratch off to find temperatures that are lower than the target.
But they can't do it. People don't understand negative numbers. In the particular case given, it seems fairly unforgivable, since temperature is something people experience every day. In fact, I remember in first grade (when I would have been 7 years old), being asked some question like "What is five minus six?", and understanding the significance of negative numbers.
I want to speak of something else, though. It's the usage of 'smaller' to describe order. This was brought up in the comments, and I really think that people should stop talking about -10 as being 'smaller' than -5. Yes, -10 is 'less than' -5 in the usual order on the reals, but, to me, 'smaller' seems to imply a statement about magnitude, not order. When I'm confronted with the term 'smaller', I often find that I need to consider carefully whether it is magnitude or order that is being discussed, because both can happen even in very similar contexts.
Consider limits, for example. What is the behavior of f(x) as x approaches minus infinity? How about as x approaches zero? Now, which of these could be spoken of as being the behavior of f(x) as x becomes very small? For my money, x becoming very small means x approaching zero, but it's not hard to imagine that someone might mean x approaching minus infinity. And, in particular, there is no problem speaking of the limit from the right as x approaches zero as being 'x getting smaller', but what about the limit from the left? Then x is increasing, but maybe also getting smaller.
The imprecision of language is something that is problematic in math as well as 'the real world', and is apparently extra-troublesome when they intersect. In math we often use many different terms to mean the same thing (or the same term to mean many different things), and for this reason I've always found textbooks that make a point of mentioning alternate terminology to be very helpful. It would be nice if we could all agree on things and choose nice, distinct terms, though. Well, perhaps one day.