Thread by @fascinatorfun

Remain United - Methodology

⁦@thatginamiller⁩ ‘s The RemainUnited.org tactical voting site is now up and running.

remainunited.org/methodology/

Sometimes there are competing possibilities. This is how they sorted decisions on that basis.

B4B used a markedly larger cohort (associated with improved confidence intervals).

Let’s keep our eye on the results of new polling.

It has the advantage of being pretty up to date polling 1st week in Nov).

It uses MRP analysis, the Rolls Royce of constituency polling

But the polling sample, though quite large , is about 1/7th the size of @BestForBritain

@remainutd suggest concentrating on these

There are a load of Tory seats however, where a small extra push could result in a further swing.

So let us push them hard.

It is worth noting that these are NOT all the swing seats. There are plenty more

So Cheltenham @Chelt4Europe seems to be considered a safe LibDem Seat now even though it is currently held by Tory, Alex Chalk.

That seems to be true on both Best For Britain and Remain United.

It is also worth noting that Remain United has said that they expect to tweak the site in the coming days and week with new polling and also feedback.

So discrepancies worth notifying to both BestForBritain and RemainUnited (as moth use MRP methods) so they can #data as updated.

This is the link where you can enter the postcode. And get a constituency analysis.

remainunited.org

Correction. It uses RRP analysis which has sprung from MRP family so I am curious about any differences between BestForBritain (MRP) and Remain United, especially when polling dates overlap more and as the campaigns progress.

I admit the far larger cohort-46k that Best for Britain uses plus multiple pollsters, (unlike RU that uses ComRes that did that awful Telegraph poll a couple of months back) tends to make me rely on Best For Britain first and if in doubt.

Confidence intervals matter in marginals