Preparing some slides on the ecological fallacy for my lectures next week. Struck by the fact that politicians' failure to understand the ecological fallacy (as illustrated in this Election Data thread) might end up playing a pivotal role in driving all of our futures...1/?
The Brexit vote may hang on defections to support govt from Labour MPs representing Brexit voting seats, particularly those like Caroline Flint who argue they must respect the will of their voters by supporting Brexit. But that arg rests (in part) on a logical mistake 2/?
Lab MPs representing Brexit voting seats infer that, as a result, the people who voted them in support Brexit. But it ain't necessarily so. It is perfectly possible for a Lab seat to vote heavily for Brexit yet for a majority of Lab voters *in that seat* to have voted Remain
Not just possible, but based on @election_data estimates, true in many cases. So if Lab MPs justify voting for Brexit by saying "this is what the people who elected me want", they are making a mistake. Often, in fact, most of the constituents *who voted for THEM* backed Remain.
@election_data Given the closeness of the Parliamentary arithmetic, it is quite possible that Lab MPs who erroneously infer that a majority for Brexit in their seat means a majority for Brexit among their voters may determine the form of Brexit we end up with.
@election_data PS - there is a different version of this arg which is less subject to the ecological fallacy. If an MP says "I must represent the views of *all* my constituents, and most of them voted to leave". However, if that's yr posn then you need to think about non voters' prefs too
@election_data So in the case of Caroline Flint's Don Valley, 68% *of those who voted* in 2016 backed Brexit (per Chris Hanretty's estimates). But that's actually only 41% of the constituents in her seat. Another 40% didn't vote at all. And that's *eligible* constituents, not everyone...
@election_data If an MP is bound to represent all those who live in their constituency, doesn't that also include those who aren't eligible to vote - such as people too young to vote in 2016, and settled migrants without voting rights?
@election_data So if the arg is "I must represent views of the people who voted me in", then the ecological fallacy is a big problem. But if arg is "I must represent views of all my constituents", then the views of those who didn't vote in 2016 need including too.
