[ad_1]
It is very rare for a polling company to admit to making a mistake that affected the polls it published. It is so rare that we can speculate that if polling companies themselves have discovered mistakes in the past, the companies concerned have remained silent and hoped that no one else would notice unless they assumed that there was an extraordinary rogue result due to a sampling error or sampling error. like. It is better to remain silent than to suffer reputational damage by admitting a mistake, so polling companies can give reasons.
However, in the case of the latest Redfield and Wilton Scottish polls for the UK general election, it may be more difficult to spot an apparent error.
At face value, the headline results seem quite plausible. SNP 39%, Lab 29%, Con 22%, LD 6%, with fieldwork from 2n.d for 5Th March. The SNP lead is 10%, quite unremarkable and consistent with the generally reduced lead that has been seen since February. The SNP should be relieved that its vote share hasn’t fallen amid a less-than-friendly leadership election, while Labor will be disappointed. The Conservatives will be encouraged by the recovery that appears to be only 3% below the level of GE 2019. But at face value there is nothing to ring alarm bells.
Dig deeper into the data table, and something is wrong. The weighted results, before the final voter filter is applied, have SNP and Labor levels at 30% each, along with 14% don’t know and non-voters. Stripping DKs and DVs out, the vote share before turnout weighting is 35% SNP, 35% Labour, 19% Con, 6% LD.
Indeed, the R&W turnout adjustment to the March 2023 poll has changed the pegging level to a 10% SNP lead over Labor in the headline result. All the more surprising given that turnout adjustments in the previous November 2022 R&W poll left the SNP’s lead over Labor unchanged at 10%.
So something doesn’t smell. And when looking at the details of R&W’s turnout adjustments, one thing stands out. 34% of those who voted Labor in the 2019 GE were recorded as “probably” or “definitely” not voting in the current GE, compared to just 1% of Labor 2019 GE voters in the November poll. In contrast, the proportion of SNP, Conservative and LD 2019 voters who are unlikely to vote is now 11%, 10% and 8% respectively.
Adjusted for R&W turnout, the large difference between Labor and the other parties is unbelievable, as is the scale of change since the previous R&W poll.
The data suggest that R&W must have made some mistakes when processing the polling data, like a transposed formula in the spreadsheet or an incorrect decimal point. If not, then R&W must come up with a plausible explanation for the seemingly implausible. And in the meantime, in the absence of these explanations, it is reasonable to conclude that Labor may have been level with the SNP in the general election polls in Scotland.,
Wulfrun Phil
[ad_2]
Source link