Local Council by-elections 31st March 2022

[ricmk]

jamesdoyle the GWBWI is just a bit of fun but I think you’ve hit on the 2 weakest elements to me:

-treating not standing as getting 0 votes
-giving little credit for holding seats. Many by-elections are in tricky circs replacing a Cllr with a big personal vote so holds can be big deals.

On the first there’s a big difference between a party not being organised enough locally to stand and them genuinely having little support. Feels to me that a derisory vote share is true but can a history of not standing reduce the score (maybe by 1/2 either way if you haven’t stood recently)

On the second I would simply deduct a few points from the ‘par’ score for a defending party where there’s a long term incumbent to replace.

[johnloony]

Apr 2, 2022 12:14:29 GMT ricmk said:

jamesdoyle the GWBWI is just a bit of fun but I think you’ve hit on the 2 weakest elements to me:

-treating not standing as getting 0 votes
-giving little credit for holding seats. Many by-elections are in tricky circs replacing a Cllr with a big personal vote so holds can be big deals.

On the first there’s a big difference between a party not being organised enough locally to stand and them genuinely having little support. Feels to me that a derisory vote share is true but can a history of not standing reduce the score (maybe by 1/2 either way if you haven’t stood recently)

On the second I would simply deduct a few points from the ‘par’ score for a defending party where there’s a long term incumbent to replace.

GWBWI is just a load of gibberish, codswallop and nonsense anyway so it doesn’t matter in the slightest if he adds, takes away, or takes the cube root of any losses, gains, swings, incumbencies, percentages or hippopotamuses at all.

[The Bishop]

Apr 2, 2022 12:14:29 GMT ricmk said:

jamesdoyle the GWBWI is just a bit of fun but I think you’ve hit on the 2 weakest elements to me:

-treating not standing as getting 0 votes
-giving little credit for holding seats. Many by-elections are in tricky circs replacing a Cllr with a big personal vote so holds can be big deals.

On the first there’s a big difference between a party not being organised enough locally to stand and them genuinely having little support. Feels to me that a derisory vote share is true but can a history of not standing reduce the score (maybe by 1/2 either way if you haven’t stood recently)

On the second I would simply deduct a few points from the ‘par’ score for a defending party where there’s a long term incumbent to replace.

The reasons the byelection happened can also be relevant here - Labour's vacancy this past week being a case in point. It could easily have been a lot worse.

[grahammurray]

Apr 3, 2022 10:57:29 GMT The Bishop said:

Apr 2, 2022 12:14:29 GMT ricmk said:

jamesdoyle the GWBWI is just a bit of fun but I think you’ve hit on the 2 weakest elements to me:

-treating not standing as getting 0 votes
-giving little credit for holding seats. Many by-elections are in tricky circs replacing a Cllr with a big personal vote so holds can be big deals.

On the first there’s a big difference between a party not being organised enough locally to stand and them genuinely having little support. Feels to me that a derisory vote share is true but can a history of not standing reduce the score (maybe by 1/2 either way if you haven’t stood recently)

On the second I would simply deduct a few points from the ‘par’ score for a defending party where there’s a long term incumbent to replace.

The reasons the byelection happened can also be relevant here - Labour's vacancy this past week being a case in point. It could easily have been a lot worse.

Genuine question.

Is there any empirical evidence that the cause of a council by-election has a material influence on the result?

[minionofmidas]

Apr 3, 2022 11:05:24 GMT grahammurray said:

Apr 3, 2022 10:57:29 GMT The Bishop said:

The reasons the byelection happened can also be relevant here - Labour's vacancy this past week being a case in point. It could easily have been a lot worse.

Genuine question.

Is there any empirical evidence that the cause of a council by-election has a material influence on the result?

Quantifying that is a very reasonable future use of the GWBWI!

[Merseymike]

Apr 3, 2022 11:05:24 GMT grahammurray said:

Apr 3, 2022 10:57:29 GMT The Bishop said:

The reasons the byelection happened can also be relevant here - Labour's vacancy this past week being a case in point. It could easily have been a lot worse.

Genuine question.

Is there any empirical evidence that the cause of a council by-election has a material influence on the result?

There can be. Labour had a councillor in Sefton who was the Mayor even though he had only been on the council for a couple of years. He was caught as was his wife for major disability benefits fraud largely shown up because of his mayoral activities. The subsequent by-election in a Lab-Con marginal saw Labour's vote collapse and a third place. It was undoubtedly because of these circumstances.

[Khunanup]

Apr 2, 2022 11:10:22 GMT jamesdoyle said:

GWBWI

LDm +35
Con -1
Lab -57
Grn -92

This might cause some head scratching, which I think would be entirely reasonable: why do the LibDems get such a (relatively) good score, the Greens such a poor one, and why do two Con holds end up as a (marginally) negative score?

The first two are connected, so I'll take them together. Part of the GWBW score is calculated from the change in vote share. This is normally fine, but in Melton this week we had a good example of what can happen when a party fields candidates only sporadically. The Greens are particularly prone to this. In Melton, the Greens stood last time amd got a very decent vote share. So not standing this time meant a significant negative score - essentially the entirety of their negative score above. But looking at this history, last time was the _only_ time they've stood in this ward in getting on for 20 years. So a one-off good result has meant they are heavily penalised. Gut feeling tells me this is wrong.
The LibDems are the mirror image - they stood for the first time in ages, if not ever, and did pretty well - not least because, presumably, they picked up a big chunk of 2019's Green vote. So going from nothing to 24% of the vote gave them all their positive score for the week. Again, this is overstated (to me - I expect to you too?).

This issue - one-off results significantly skewing scores - is something I'm working on for the 2022-3 version! I expect that reworking this area might mean that parties that stand sporadically (Greens, and to an extent the LiBDems) should see a dampening effect (in both directions) on their score swings.

For the Cons, the issue is that I discount holds and vote share changes in seats that have been solidly held in the history given; obviously Melton Sysonby was a very solid seat for the Cons, and Mayfield was solid - but on a very short record.

Again, this is something I've been looking at. Firstly, my current algorithm basically looks at how solid a seat is over the record available, so Mayfield is counted as being as solid as Sysonby. Which my gut tells me is overstating it. And secondly, for seats as safe as Sysonby, essentially the algorithm discounts the defending party's performance to zero. Again, my gut tells me that there should still be some value, albeit a small one, in getting a decent result, or an increase in vote share, in safe seats. So this is something else for the 2022-3 GWBWI. This should work in favour of parties defending more seats, so Con and to an extent Lab should benefit from this.

ETA :
1) Although I don't expect a response, if anyone's got any thoughts on those comments, I'd welcome them, for or against!
2) My delay in posting was, as I said yesterday, because I was going to Portsmouth to see a play. Well, it was very much worth it - The Lion In Winter was one of the best things I've ever seen on a stage. It was a studio production, in the round, in a tiny space (Khunanup and cogload (I think?) will know the Square Tower), so there was nowhere to hide flaws in performances - but that was fine, bcause there weren't any). It was directed and acted superbly. I'm stilll buzzing about it.

Know the Square Tower very well! Not too many ex-gunpowder stores built in the 15th century where you can go to a play, get married or attend a market...

[jamesdoyle]

Apr 3, 2022 12:44:26 GMT Khunanup said:

Know the Square Tower very well! Not too many ex-gunpowder stores built in the 15th century where you can go to a play, get married or attend a market...

No more plays now, though - that was the last production there. They can make more money from a Saturday wedding than a week's play run.

[jamesdoyle]

Apr 2, 2022 12:14:29 GMT ricmk said:

jamesdoyle the GWBWI is just a bit of fun but I think you’ve hit on the 2 weakest elements to me:

-treating not standing as getting 0 votes
-giving little credit for holding seats. Many by-elections are in tricky circs replacing a Cllr with a big personal vote so holds can be big deals.

On the first there’s a big difference between a party not being organised enough locally to stand and them genuinely having little support. Feels to me that a derisory vote share is true but can a history of not standing reduce the score (maybe by 1/2 either way if you haven’t stood recently)

On the second I would simply deduct a few points from the ‘par’ score for a defending party where there’s a long term incumbent to replace.

Thanks for that ric, and sorry I didn't reply to you yesterday, things got rather busy and it slipped my mind. To take your points in order:

1. 'a bit of fun'.
Well, I hope it could be, or become, something more than that: the aim is to make it an objective measure to forestall spin, opinion and prejudice, of which,I think, there has been less on these threads since I started this, whether or not there's a connection! Happy to accept it's not there yet, but with input like yours I hope to get there. When I started posting this I knew that it couldn't yet be solid, but doing so encourage me to keep doing it so I could spot the flaws, get feedback, and also get some data as to what numbers would or would not be 'good' or 'bad'.

2. 'difference between a party not being organised enough locally to stand and them genuinely having little support'
Yes, I think this is fair. I will work on both sides of this: unreasonable bonuses for standing after not doing so, and unreasonable penalties for not standing after doing so. A 'reduce by half' rule is probably actually a bit crude, as there are degrees of 'not standing very often' when you have a reasonably long history to the ward, so I think it should be possible to weight this more gradually.

3. 'a long term incumbent'
This is very interesting, and one I hadn't thought about. I think at the moment this is going to be out of scope. When I started doing this, I wanted to restrict it to easily and consistently available information, so the index is built around the data that tony provides each week: number of seats for each party on the council, ward election history (including who is defending), and parties standing. Obviously this is outside that, but I think there might be merit in that - but then I think I'd also have to include some way of factoring in when a former councillor stands again. And commit time to finding that information. None of these are numeric, so I'd have to look at how to turn it into a numeric value. Undoubtedly worthwhile, but how much so I don't know.

Apr 3, 2022 10:57:29 GMT The Bishop said:

The reasons the byelection happened can also be relevant here - Labour's vacancy this past week being a case in point. It could easily have been a lot worse.

Thanks, bish. Like ric's incumbency point, I don't doubt this can be a factor - but it's very definitely subjective. What causes would I include, how would I rank the importance of different causes, and how do I weight them? Plus, how do I find out the necessary detail? Definitely WIBNI, but no more than that for the foreseeable.

Thanks to all for comments: I have been, frankly, baffled that people on here who spend huge amounts of time analysing elections, forecasts, polls and boundaries - and I enormously value the work that such people do - are so dismissive of an attempt to analyse another part of psephology: argue about it, yes, but why dismiss it?

[Crimson King]

Just wanted to say I do think there is value in this, and I wouldnt worry that you cant make it perfect. At (a not insignificant) least it gives a starting point on top of which people are free to argue that other factors mean it is better or worse than the raw numbers suggest.

A further thought - does to formula take account of movements in polling since the election in which the vote share changes are based?