My long-term readers may recall something I wrote in November 2003 about an eccentric misrepresentation committed by then MP Anne Campbell in her constituency propaganda. To refresh your memory, she published this plot of her election majority over time,
unaccountably omitting the much larger majority she won in 1997. As I pointed out, a more honest version of the graph would have looked like this,
creating an altogether different impression of the trend she was trying to depict.
Some time ago one of her other constituents emailed her to ask why she thought this sort of thing acceptable. A copy of her answer came into my possession. I wouldn't usually publish such correspondence -- and I do so here without the permission of either party -- but in my view an MP's opinions of her constituents are a matter of legitimate public interest, as is dishonest propaganda. So here we go: (my comments are interspersed with the text)
The idea that the graphic is misleading is astonishing and could only have been insinuated in the way that it has by a LD supporter which I am sure Chris Witter is.
Note here (a) the slightly surprising confusion of the URL of my web log and my name; (b) the bald, certain and quite wrong claim about my political affiliation. Regarding the latter I am reminded somewhat of a story -- recounted, I think, in one of John Simpson's volumes of autobiography -- of Brian Redhead interviewing Nigel Lawson, then Chancellor of the Exchequer, on the Today Programme. In response to some difficult question other Lawson blurted out, ``You're only saying that because you vote Labour.''
Redhead calmly turned off Lawson's microphone and said, to the listening nation, ``We will now have thirty seconds' silence in which you can reflect upon the enormity of claiming to know how I vote in a secret ballot, and the nation can reflect upon the failure of your economic policies.''
We continue from culumny to irrelevance:
Is he implying that I am somehow ashamed of my 1997 majority and that a large majority in 1997 somehow diminishes a very healthy majority in 2001. If anyone should be trying to hide the 1997 result it is the LDs who did worse in 1997 than they have done for many years.
-- and from irrelevance to sophistry, and on to dishonesty:
I did promise to vote against top-up fees in 1997 and I did - immediately after the election when the government legislated to prevent universities from charging top-up or variable fees. I repeated the promise in 2001 and I am saying the same thing now, even though the government appear to have changed their minds. I have been wholly consistent.
I wrote a clarification on this point back in 2003. I had said `top-up fees' when I meant `tuition fees'. Anne, also, was apparently confused on this point. In 1997 she promised that ``Labour will not allow universities to introduce tuition fees''; yet in 1998 she did not vote against the Teaching and Higher Education Act which brought in tuition fees; nor did she even back this rebel amendment which would have satisfied her commitment to a ``new system of financial support [i.e., grants] [which would] ensure that students have enough to live on while studying.''
I don't understand why Parties do this kind of thing except of course, that occasionally people do get taken in by it. I do have a higher opinion of the Cambridge electorate.
I close with a picture of Anne as she chose to be depicted in a recent issue of Anne Campbell Reports, under the headline, ``Safer and Cleaner Streets'':
Comments
Posted by Matt Daws, Saturday, 30 April 2005 12:09 (link):
It's this sort of nonsense which really turns me off politics. We could be arguing about policy, but no, it keeps coming down to "we must beat the Lib Dems and the Tories" (foam at mouth). I do wonder if some Labour MPs even remember *why* they must beat the Lib Dems and the Tories: it's almost become a knee-jerk reaction, as if the election were nothing more than a popularity contest: "we must win, it doesn't matter how we win, or what we do to win, we must win". Surely there are some people left who think that maybe perverting all of one's policies (not to mention credability, personal morals and so forth) simply to win might not be the end goal of this exercise?
Posted by Roy Badami, Saturday, 30 April 2005 20:38 (link):
It wouldn't be appropriate to include the 1997 election in a constituency majority trend analysis of this type because the 1997 election was close to the the time when the Tory government was replaced by a Labour government, creating a discontinuity in the political timeline.
The 1992 and 2001 elections are both statistically significant because there was no such discontinuity surrounding either of those elections, and hence they are directly comparable. This is the prefered analysis, and the one that Anne Campbell chose to use.
The alternative analysis would have been to compare the 1979 and 1997 elections (these being the last two elections which occured close to discontinuities in the political timeline), and if you do that you will again see that Labour support in Cambridge demonstrates an upward trend. Neither of these elections are as statistically significant as the 1992 and 2001 elections, but their level of significance is nonetheless broadly comparable, permitting a direct comparision in a trend analysis.
It would be highly unlikely for too such comaparisons to give the same result solely by chance, to the extent that such a chance result would be statistically insignificant in the extreme. Therefore the fact that both comparisons (of different elections held in different circumstances) show statistically consistent trends indisputably proves that the trend for Labour support in Cambridge is an upward one.
-roy
Posted by Chris Lightfoot, Saturday, 30 April 2005 21:16 (link):
I disagree. If, as you say, there is a discontinuity between the two, a trend drawn across it are unlikely to be very useful. (Though I admit I'm not really sure what the implicit model you're using here is.)
More generally, we'd expect support for the incumbent to fall during the lifetime of a government. (Admittedly this pattern is by no means evident in Cambridge in particular, but then there's not all that much data in one lone constituency; in any case in five out of nine cases, support for the incumbent fell during each government, in one it remained about the same, and in three it rose.) On this basis it is the comparison between 1997 and 2001 which is most relevant to the question of support for Anne circa 2003.
And none of this really affects the question of whether Anne's plot is misleading. It clearly is: she has taken a series which is defined only at three points, and missed out the middle one, presumably because including it would have given a pessimistic impression of her chances. If she were actually intending to convey an analysis like yours she would have put in the 1997 result and explained why it was not relevant.
Posted by Roy Badami, Saturday, 30 April 2005 21:44 (link):
Er, my comment was a spoof :-) I evidently failed to make it non-sensical enough, though...
I'm impressed that you managed to read any meaning into it, given if was content free :-)
Posted by Helen Wright, Sunday, 1 May 2005 22:01 (link):
I'm impressed that you managed to read any meaning into it, given if was content free
Ah, but he's had a lot of practice recently, reading the various parties' manifestos.
Posted by Chris Lightfoot, Monday, 2 May 2005 01:25 (link):
A successful one, evidently!
I looked at it for some time trying to decide whether it was serious and eventually decided that it was, on the slightly unfair basis that I couldn't remember having seen you write such a spoof before. Of course this has the rather uncharitable interpretation that I thought it plausible that you would write such rubbish as the above. The best I can say in my defence is that it's always best to take criticism seriously in case you have misunderstood it....
No, equally tragically, it's completely genuine.
Posted by Roy Badami, Monday, 2 May 2005 02:24 (link):
Oh, the graph is fine, really -- it's just a bit overwhelming. When you first look at it, it looks very much like a random collection of red, yellow and blue squiggles... :-)
I was just idly looking for some irrelevent detail that could distinguish the 1997 election from the other two, and the obvious one was that it resulted in a change of government, so I wrote a spoof justification around that...
Sorry if I wasted your time on the graph -- I thought the phrase constituency majority trend analysis was sufficiently over-the-top to give it away from the first sentence.. :-)
-roy
Posted by Roy Badami, Saturday, 30 April 2005 22:07 (link):
Though actually, your graph does rather have the look of a spoof, too... :-)
Posted by Matthew Byng-Maddick, Monday, 2 May 2005 15:41 (link):
Hi Chris - pointed to this entry by Pete (Clay), after mentioning an interesting discrepancy in Oona King's campaign literature for Bethnal Green and Bow (the consituency where I live), and felt that it was possibly worth detailing here. To put it in context, the campaign is based on the fact that the tories are the only real contender in the constituency (I'm inclined to agree with this, for what it's worth), and that any vote not for her is a vote to let the Conservative candidate in (obviously Oona and I disagree somewhat on this front).
In two separate pieces of campaign literature she publishes various bar graphs about constituency returns, much as Anne did in Cambridge. The first contains the results for the 1997 and 2001 General Election returns:
1997: Lab: 20697, Tory(note that she uses Tory rather than the more traditional "Con"): 9412, Lib Dem: 5361, Green: 1666, BNP: 1267. The last three of these have a brace with a caption "Can't win here" above it.
2001: Lab: 19380, Tory: 9323, Lib Dem: 5946, BNP: 3350, Lib: 2963, Real Lab: 1117, Green: 812, and again a brace from Lib Dem downwards.
The second leaflet has the following set of results:
2001: Lab: 19380, Tory: 9323, Lib Dem: 5946, Green: 1666, BNP: 1267 (and that brace, again...)
It's also interesting that all of her campaign literature is for local policies and doesn't mention the scary voting record she maintains on national and international issues.
Of course, this year will be different anyway, as we have the rather high profile "Respect" campaign conducted by George Galloway, and his bus. I have no idea which is correct and whether this is just a typo or some such somewhere, but the general misuse of statistics is certainly interesting...
Posted by Peter Clay, Tuesday, 3 May 2005 23:20 (link):
It's amusing that she's campaigning on local issues; I just threw away a leaflet from the Conservative council candidate, who was seemingly campaigning on national issues like immigration. Bad candidate, no biscuit.
Posted by Matthew, Wednesday, 4 May 2005 18:55 (link):
Lord, Brian Redhead was a pompous arse, wasn't he? I remember another, David Dimbleby, saying much the same thing on April 10th 1992 when Dennis Skinner said, 'Everyone knows you are a Tory'.
Post a new comment.
Comments copyright (c) contributors and available under a Creative Commons License. See also the comments policy.