Sunday 20 January 2013

Psephology 6 - Using Voter Turnout data

The snow is deep on the ground and it's way below zero, or so it seems, so I'm staying in to play around with some data. It would be interesting to do the statistics to see if deprived wards also have a poorer voter turnout - voter apathy is also a factor in such areas (according to Lewis, 2008, Open University). So I did exactly that. I worked out the average voter turnout percentage for each ward, to see which have persistent high or low turnouts. I also added the average LibDem % vote for each ward to the graph.


Yes, they're both percentages, but you can't assume that virtually the whole of Millfield turned out to vote LibDem. We're just looking at general trends here, and slight correlations; ecological fallacy, I think they call it if you push data too far. The point of the graph is to show trends:-  good vote but low turnout - we must be doing something right. Or high turnout and low vote. Oops, get leafleting.

Some wards have a very high turnout, Fulwell and St Chad's for example, which just happen to be the highest Conservative voting wards in the city. Can one infer from that that Conservatives are more conscientious? Wards like St. Anne's, Hendon and Southwick on the other hand, have very low turnouts in comparison and these are the poorest part of the city. I know it's stereotyping but I can't help but think of voter apathy and poverty being linked. On the other hand, in line with the theories,  it may not be pushing it too far to suggest voter turnout as a measure of how deprived an area is. See:-

The two sets of data do not always match. Millfield has both high turnouts and higher LibDem voting. However, St. Chad's and Copt Hill have a lower than expected LibDem vote by this criteria; are the LibDems staying home in these areas? St Chad's, my ward, is a swing ward, heavily contested between Labour and the Conservatives. Interestingly, Hendon seems to buck the trend, having a lower turnout, but with a higher than expected LibDem vote. Is there something about the LibDem message that appeals to lower income voters. I would say the tax threshold, but this data is from the last 8 years.

That data was from percentage turnout, but what about the actual numbers. Can they be useful? Let's look at the average numerical turnout by ward.

Ward
Average Turnout
Ward
Average Turnout
Barnes
4024
Shiney Row
3695
Castle
2859
Silksworth
3416
Copt Hill
3494
Southwick
2904
Doxford
3270
St Anne's
2692
Fulwell
3955
St Chad's
3628
Hendon
2760
St Michael's
3855
Hetton
2711
St Peter's
3630
Houghton
3463
Washington Central
3628
Millfield
2595
Washington East
3602
Pallion
2684
Washington North
3066
Redhill
2816
Washington South
3373
Ryhope
3208
Washington West
3457
Sandhill
3033





Comparing the actual and percentage data, in theory, it should closely match, but actually, there is a small variation in the graphs. Unfortunately, my graph doesn't translate from excel very well. Here's a close up though.



I wondered first whether it was an artefact of the scale I was using, but the trends of the lines should be broadly similar. It's just possible that I'm seeing an artefact of the difference in size of the wards. So in Redhill, it's easier to elect a candidate, because 33% is 1400 people, whilst in Pallion, 33% is roughly 1750 people. Data like this makes my brain whirr like a pachinko machine!
So off I went to calculate exactly how many people in each ward equal one percent of the vote?

Ward
Number of people equalling 1%
Ward
Number of people equalling 1%
Barnes
90
Shiney Row
102
Castle
87
Silksworth
85
Copt Hill
91
Southwick
84
Doxford
79
St Anne's
87
Fulwell
92
St Chad's
80
Hendon
90
St Michael's
92
Hetton
90
St Peter's
87
Houghton
91
Washington Central
90
Millfield
81
Washington East
88
Pallion
76
Washington North
89
Redhill
86
Washington South
82
Ryhope
82
Washington West
91
Sandhill
86







I used the 2012 election turnout figures to calculate this. Some  variation - 76 people equal a voting percent in Pallion, compared to 92 people in St Michael's. Not huge differences, except in Shiney Row where one percent is 102 voters!

Out of interest, there are some theories as to the percentage of the electorate that are politically active - from the 2009 Audit of Political Engagaement 6 at

This finds that in 2009, in the UK population:-

12% were very interested in politics
40% were fairly interested in politics
30% were not very interested in politics
17% were totally uninterested in politics

I suppose only the first two groups vote. Also in this survey, 5% claim to know a great deal about the subject!

Next time - National election results for Sunderland.