Playability of a word indicates how useful it is to know a word in a Scrabble game because it is likely to be playable as the best move in a particular situation. It is measured by how much worse off you are if you don't know that word. The difference in equity point between that word and the next best play (i.e. equity loss) is tallied over many games to calculate its playability score; the higher the better. (For an explanation on equity loss, see my other answer here)
See here for further explanation http://crosstables.livejournal.com/24367.html
Taking your Q example, it's easier to understand the concept from the reverse angle: your Q will be unplayable if you don't know QIN and it will severely hamper you (i.e. big equity loss), so QIN has high playability score. On the other hand, not knowing ERE is not as bad if you know RE (for example), so even if that situation happens more often, the cumulative effect will not be as bad as that one time you had Q and don't know QIN, hence the lower playability score of ERE.
Another thing to note is that words with more anagrams tend to have lower playability score. Take the highly valued RETINAS rack, high in probability and (deservedly) must be learnt. However RETINAS itself has low playability score, because most likely most of the time if you don't know RETINAS but you know RETAINS (or any of the other anagrams) and the other anagram can be played, no big damage done and equity loss is small.