Fixed that for you: dodgy Liberal Democrat election graphs, 2017

Fixed that for you: dodgy Liberal Democrat election graphs, 2017

Elections are a time for many things. I’ve come to think of them as a time for sniggering at parties’ desperate attempts to sway voters by presenting a dubious grasp of mathematics in graph form and then smugly fixing them. I’m not that good at maths. But apparently I’m better than people who think they can run the country.

The Liberal Democrats are serial offenders. The three examples of dodgy graphs I’ve found in London so far are all from the yellow corner. But the Conservatives in Scotland seem fond of a graph this year, so if you have any capital-centric examples – of any hue – that I can pick apart instead of doing something constructive, hit me up on Twitter.

Hackney North and Stoke Newington

Here the Lib Dems are relying on by-election figures to persuade voters they really can win in a seat where Labour’s Diane Abbott is defending a majority of 24,016. Up 33%, they claim! From a base of what I have no idea, but I’m guessing it’s some kind of average vote spread from the 2015 general election.

(Here’s a good moment to point out that any rise looks wonderful if you start from a low base. An increase of 100% looks impressive, but not if your vote share actually rose from 1% to 2%. It’s only to be expected their share would go back up.)

They’re not including the Oldham West and Royton by-election of December 2015 in their “principal authority by-elections”, probably because the Lib Dem candidate in that one got exactly the same share as at the May 2015 general election – and a lost deposit both times. Yet in most of the by-elections since May 2016, the Lib Dems have either lost votes or got about the same number but a higher share of the total because of low turnout. Or they picked up votes, but fewer than 1,000.

I mean, in Tooting, this is what happened to parties’ vote shares in the June 2016 by-election compared with 2015.

The “33% rise” figure in the Hackney North and Stoke Newington leaflet has been derived entirely from the more recent by-elections in Witney and Richmond Park. Both were unusual. In Richmond Park, it’s thought voters punished Zac Goldsmith for his outrageous mayoral election campaign by turning to the candidate (Sarah Olney) most likely to defeat him. In Witney, the Lib Dems recorded a massive result to come in second in the contest to succeed David Cameron as MP, and we can assume this was on a wave of anti-Brexit sentiment in a Remain area – not forgetting that Cameron was the man who got us into this referendum mess in the first place.

Not so promising now, is it?

Let’s look at what would happen in Hackney North if the Lib Dems actually did manage a 33% increase in their vote share compared with two years ago and took all their extra votes from Labour.


Let’s be a little fairer, and do the same job on the 2010 general election result, as that’s probably more indicative of where Lib Dem sentiment is these days (i.e. there’s less visceral hatred of them).



Oh, Liberal Democrats, why do you do this? In Vauxhall, they’re using betting odds to claim that they are closing in on Labour Brexiter Kate Hoey. Now, apart from the fact that betting markets aren’t always a totally accurate predictor of the future, the odds offered are based entirely on who’s gambling, how much and on what. It’s not exactly hard data.

 And don’t get me started on that arrowhead at the top of the Lib Dem bar, making the race seem closer.

Thing is, that even without that pointy bit the Lib Dems and Labour were pretty damn close on 1 May, according to Paddy Power. I’ve turned the odds into probability percentages because odds make my head hurt.

But the problem with odds is that they change. Here’s what Paddy Power was offering for Vauxhall four weeks later, on 30 May (yesterday).

The Vauxhall leaflet shows everything that’s wrong with the Lib Dem obsession with graphs. There’s a grassroots swell of opposition to Hoey because of her prominent Leave stance and the constituency voting about 77% Remain. Why not press that point home instead of simply having to have a pretty graphic?

Greenwich and Woolwich

This is exhausting.

OK, point one: the figures in this leaflet appear to be based on Projected National Share (PNS), which is something the BBC comes up with after a set of local elections to estimate how the parties would have done if the whole country had voted, instead of just in some regions.

YouGov has a good explanation of why you can’t then translate that into a general election. The takeaway is that people vote differently in local elections to general elections. So it’s probably not a good idea to bang on about it on your leaflets. Especially when the PNS in question actually looks like this.

Point two: remember earlier, when we talked about the low base? It was always inevitable that the Lib Dems would have some kind of bounceback after their 2015 disaster. A 10% increase isn’t that exciting in context.

Point three: the PNS tells you bugger all about how any one particular race is going. As I’ve already written, the Lib Dems are potentially on course to do well in Greenwich and Woolwich, but let’s not pretend they’re 13-14 points ahead of Labour.

Point four: the Lib Dems are the only party that can stop a Hard Brexit are they? And how’s that going to work? Do they reckon they’ll be in power on 9 June? Or the official opposition? What about Labour’s policies on stopping an economically suicidal deal? Putting this caption on that graph is just… I mean, stop it, Lib Dems. Just stop it.

PS (1 June, 2017)As John Griffiths points out in the comments, I’ve got this one wrong: it’s not parliamentary by-elections the Hackney North graph is covering, it’s “principal authority by-elections”, i.e., council by-elections. This distinction got past both me and Dave Hill, erstwhile proprietor of this site, so I’m going to argue that if the language is arcane enough to mislead us, it’s going to mislead casual voters.

It also makes the graph even more bloody meaningless. At least by-elections to Westminster are comparing like with like (ish; bearing in mind that by-elections are usually an excuse to kick the party in power). Council by-elections? Are they freaking kidding? That election for a local councillor that hardly anyone turns out for, never mind realises is happening? Go away, Liberal Democrats, and take a good look at yourselves.

 Rachel Holdsworth is a former political correspondent with Londonist and is now a freelance writer and editor.

Categories: Analysis, Maps & Stats


  1. John Griffiths says:

    Interesting piece and as a LD I agree with a lot of your analysis…but check again Joe Rochard’s graph which is nothing sbout %s. The 33 is the net gain in Council by elections since May 2016…also nothing to do with Parl by elections. Still it’s a pretty fair cop. Especially like your disection of LD use of the latest betting odds in Vauxhall. Whatever next?!

    1. Rachel Holdsworth says:

      Oops, that’s my bad, mixing up council with parliamentary by-elections! That passed me and Dave (doing the editing) by, which kind of makes me even more annoyed – at myself, of course, but also at the Lib Dems for using language that’s designed to confuse. Jeez, how’s the casual voter meant to pick this apart?

Leave a Reply

Your email address will not be published. Required fields are marked *