Polling Correction
The Scotsman wishes to clarify its reports of polling data produced by polling partners Savanta Comres.
Data on which our coverage of the December, January and February polls is based was incorrect due to a weighting error made by Savanta Comres.
Instead of weighting by likelihood to vote at the correct step, the step was missed, which meant weighting was turned off for independence and Holyrood voting intentions.
Savanta Comres have apologised for the error in a statement and thanked the British Polling Council for their co-operation and patience while it was fixed.
The error means the data overstated support for the SNP and Yes and understated support for No and Scottish Labour in the three polls.
The correct data also shows the first No lead in the Scottish independence voting intention in February and the SNP being predicted to fall short of a majority for the first time in the same poll.
The correct data for the polls is as follows:
December
Indyref (when don’t knows are excluded)
Yes 55%
No 45% Constituency voting intention (VI) SNP 52%
Labour 19% Conservative 19%
Liberal Democrats 7%
Other 3%
List VI
SNP 39% Conservative 20%
Labour 20%
Liberal Democrats 7% Green 11%
Other 3%
January
Indyref (when don’t knows are excluded)
Yes 52%
No 48% Constituency VI
SNP 49% Conservative 20%
Labour 20%
Liberal Democrats 8%
Other 4%
List VI
SNP 40%
Labour 20% Conservative 18%
Liberal Democrats 9% Green 10%
Other 3%
February
Indyref (when don’t knows are excluded)
Yes 49%
No 51% Constituency VI
SNP 50% Conservative 23%
Labour 19%
Liberal Democrats 6%
Other 2%
List VI
SNP 39%
Labour 21% Conservative 21%
Liberal Democrats 8% Green 9%
Other 2%