What is the lesson from #HartlepoolByElection for forecasting future elections?

I have 2 hypotheses.

1. Hartlepool was behind the curve in #GE2019 & the by-election represents catch up.
2. #Brexit realignment of British politics is still continuing.

Let's look at them... /1
The 1st can be confirmed just by looking at CON vote share changes between 2010 & 2019 in Hartlepool's 4 neighbours.

Easington > from 14% to 26%
Sedgefield > from 23% to 47%
Stockton N > from 26% to 41%
Redcar > from 14% to 46%
Hartlepool > from 28% to 29%

So the jump ...

/2
... to 52% in this by-election can be put down to #Hartlepool copying its neighbours 2 years late.

But could this indicate the Brexit realignment of 2019 was incomplete & there could be more Hartlepool's?

The answer is yes. Note how nearly all BXP voters went to CON ...

/3
... in Hartlepool.

How many more LAB seats could CON win if all BXP+UKIP voters switched to CON?

Answer is 35 where CON is already in 2nd place to LAB.

The list is shown below with seats ordered by the combined BXP+UKIP 2019 vote.

I've not included the 2 Barnsley seats...

/4
... & Blaenau Gwent where BXP is 2nd to LAB.

To see if it's feasible for all BXP+UKIP voters to switch to CON, my table above has a LV DIFF column.

This is the difference between a Possible CON19 share & actual CON19 share.

Possible CON19 = 0.7 * %LV16 + 0.2 * %RM16

I ...
/5
... used this Possible CON19 as 1 of 7 models I used in my #GE2019 forecast where I was the most accurate forecaster (see bit.ly/2Q2qepM).

The 70% & 20% came from what the final polls said (bit.ly/3b9KbCu) and latest polls confirm this is still the case...

/6
So Hartlepool, since it voted 70% leave, 0.7*0.7 + 0.2*0.3 = 55% is the Possible CON19 share.

Since they got 29% in 2019, the LV DIFF is 55-29% = +26% as shown.

They actually got 52% in the byelection which means there was 3% spare.

As my table shows, 8 of the other 35 ...

/7
... have spare votes less than or equal to Hartlepool & are probably not at immediate risk.

The other 27 though must be regarded at risk for LAB if polls remain where they are today.

Top of list after Hartlepool is Ed Milliband in Doncaster.

Next is Hartlepool's neighbour.

/8
... Easington, 1 of 7 more NE seats that are vulnerable.

As bad as it is for LAB in NE, it's even worse in Yorkshire where 11 seats will be firmly in CON's cross-hairs.

So to repeat, #HartlepoolByElection was definitely a case of the seat catching up with the red wall...

/9
... collapse of #GE2019 and keeping up with the Joneses.

But before we get onto the question if the Brexit realignment is still continuing today, Labour has to shore up its defences in at least 27 seats that will share Hartlepool's fate unless action is taken.

/end

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Nigel Marriott

Nigel Marriott Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @MarriottNigel

8 May
Dan Norris has won #WestOfEnglandMayor for #Labour making it one of the few LAB gains in #LocalElections2021.

In doing so, he also wrecked my forecast (see bit.ly/3uqY0Uu) the Conservatives would just hold on!

So what went wrong?

My...

@adam_ldr @StephenSumner15

/1 Image
... blog makes it clear that my forecast was in fact the minimum CON needed to do to win & was based on a number of assumptions.

The final vote tally saw LAB with 40k votes ahead of CON & crunching the numbers breaks the reason for this into 4 parts.

First an aside.

/2
When analysing elections using Supplementary Vote, it's more correct to look at number of votes rather than vote shares.

That's because there are a number of steps to the final vote totals & errors can multiply along the way.

So the 1st reason for the 40k difference is ...

/3
Read 8 tweets
8 May
WECA Mayor count takes place today covered by @adam_ldr @StephenSumner15

I've updated my forecast based on the PCC counts yesterday (Orig FC here bit.ly/3uqY0Uu). PCC had an IND candidate but WECA doesn't.

My FC of 1st prefs is unchanged, CON lead by 5pts.

But ... /1 Image
... PCC saw 51% of eliminated candidates votes passed to LAB & CON as 2nd prefs.

This is high for an SV election.

I haven't any other SV election exceed 43%. Normally it's less than 40%.

Reason appears to be twofold.

Bristol T/O was higher than expected (54% of WECA) ...

/2 Image
... compared to 51% of WECA in 2017.

Additionally, the Green vote was very good in Bristol & it's very plausible that Green voters are giving Labour 2nd prefs in much higher numbers.

As charts below show, it seems CON 2nd prefs are coming from LD & IND.

For my revised ...

/3 ImageImage
Read 8 tweets
3 Nov 20
Parliament votes tomorrow to confirm that England will go into its 2nd #COVID19 lockdown from Thursday. Does the data show that it's necessary?

We know that deaths are rising and at present are on trend to overtake the April peak around the 3rd/4th week of November.

But ...1/n
... if cases are already in decline by now, then what will a 2nd lockdown do to help reduce deaths?

At best all it can do is accelerate an existing decline or at worse prevent cases from increasing again.

So are cases about to turn? Nationally, the picture is poised. /2
The 7 day centred moving average has turned but potential revisions to the last 7 days could reverse this (red triangles are a short CMA projection based on previous restatements).

The thing to note though is that the latest 7 day CMA is as of 26th October, 1 week ago.

... /3
Read 11 tweets
3 Nov 20
#USPresidentialElections2020 today so I thought I would list some facts and figures you might not be aware of.

This the 59th presidential election in USA and the 56th competitive one.

3 elections saw a candidate elected unopposed 1789, 1792, 1820.

1/n
Of the 55 competitive elections, 36 saw the winner win a majority of both the electoral college (EC) & popular vote (PV).

One, 1800, needed a contingent election in the House of Representatives as well, an oddity that led to the 12th Amendment that could play a part in 2020.

/2
14 of 55 competitive elections saw the winner take the majority of the electoral college with a plurality of the popular vote.

Lowest PV share to win an EC majority was 39.8% by Lincoln in 1860 (his name was not on the ballot in the South) followed by Nixon 68 & Clinton 92.

/3
Read 12 tweets
13 Feb 20
If this goes viral, I'm claiming the credit for the hashtag!

The #stats behind "Moving to a world beyond p<0.05" bit.ly/2we28Ox should be called #Pexit.

Yes there are parallels between #Pexit & #Brexit as I will explain! /1
First look at the timing.

March 2016 - ASA statement on misuse of p-values bit.ly/2ONg5sX

June 2016 - The UK votes to leave the EU bit.ly/2OMy2be

They were preceded by decades of debate on the use & misuse of P-values & the UK's relationship with the EU.

/2
In September 2016, I attended the @RoyalStatSoc conference in Manchester. #Pexit was a keynote session where @AmstatNews gave its reasons for its statement.

See youtube.com/watch?v=B7mvbO…. I am in the front row taking many notes!

9 mins into that clip is a slide explaining ...

/3
Read 24 tweets
1 Dec 19
Are we going to see a repeat of #GE2017 ? In my Friday post marriott-stats.com/nigels-blog/uk… summarising latest poll trends, I said Labour need to focus on C2DE, Leavers & Lib Dems.

With the polls we've had over the weekend, this thread summarises what I am seeing. /1
CON-LAB lead is unchanged from the start of the election since both LAB & CON are up 7 pts. If this remains the case, it will be a Conservative majority.

Note figures for the latest week includes some from end of last week. My week is defined to be Thurs-Weds. /2
LIB DEMS 1 - Labour needs to first recover its #GE2017 voters lost to the Lib Dems. Having lost 40% of LAB17 voters by the time Theresa May stood down, they've recovered a third. Of the remaining 2/3, half are still with LD&GRN and half with CON&BRX. /3
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!

:(