hmmm... just wondering what the statistics are representative of in that case if every year there is a chance (but unknown %) of different people who votes. A variance in percentages for any one category could very well be due to a different spread of voters as opposed to any 'growth' or 'decline' in the collection of an existing pool of voters?