Sir, - In these days of computer research, and in a land so attuned to the principles of proportional representation, is it not odd that serious voting analyses like your own of October 28th for Laois Offaly only reported the numbers of first preference votes? The answer, of course, is no, because unfortunately, in a PR-STV count, the voters unused preferences are never totalled. Why not, we may ask? Or perhaps we prefer not to, for such knowledge could well indicate a certain inadequacy in PR-STV.
Consider, if you would, the case when there are three candidates - G, K and R - running neck and neck for the last seat. Mr K, let us say, has the lowest total, so he is eliminated, and of his subsequent transfers, a few go to G but more to Ms R. So she wins. But neither G's nor R's other preferences were counted what if they had all been for K? Maybe K would have been a fairer winner.
In Professor Dummett's words, "the outcome [of a PR-STV vote] are often highly arbitrary". How odd, therefore, that the Constitutional Review Group chose not to examine the defects of what is nevertheless a pretty good system. Let us hope the All Party Committee will be a little more attentive, for PR-STV could certainly be improved, either by using a simultaneous PR List topping up process - which is what the Maltese chose to do with theirs in 1987 - or by combining it with the Borda preferendum in a system known as quota preference score.
One obvious advantage of such an improvement is that The Irish Times would then be able to do even more accurate analyses. It would also mean that the overall electoral system was fairer. - Yours, etc.
Ballysillan Road,
Belfast 14.