Updated Lib Dem constituency polling scorecard: and the flaw in Ashcroft’s evidence

Here’s the latest summary, after the new constituency polls from Lord Ashcroft.

As before, his methodology is key to whether you think his polling is a good guide to Liberal Democrat prospects. Here’s the case against, as I put it in Liberal Democrat Newswire #53:

The key voting intention question asked people to think about how they would vote in their constituency – which usually gets more accurate constituency results (and a higher Lib Dem share) than simply a question about the general election in abstract. However the polling questions didn’t name candidates. Where there is a high profile candidate, naming the candidate usually gives a further boost to their party. Those who have done such constituency polling before with and without candidate names expect naming a high profile Lib Dem to lift the party’s support further by between five and ten percentage points.

To his credit, Lord Ashcroft has addressed these sorts of criticisms, but he’s done so without evidence:

I have not gone so far as to name individual candidates, as the Lib Dems do in their own private polling. Doing so usually boosts the Lib Dem vote share (especially when, as in the Lib Dems’ research, the voting intention question is preceded by a warm-up question asking whether the respondent has a favourable opinion of the incumbent, of whose name they are reminded). Whether this produces a more accurate assessment of real voting intentions is a different question. Indeed I have coined the term “comfort polling” to describe the practice of parties conducting research in such a way as to maximise their own apparent vote share.

On balance I continue to think that when people are prompted to consider their own area and the local candidates, an MP’s personal reputation should be baked in to their voting decisions.

No actual evidence presented for his viewpoint yet followed by “I continue to think…”? Ashcroft’s right to criticise comfort polling but this is comfort punditry, not evidence-based punditry (on which point see Stephen Tall’s excellent post).

So with those heavy caveats applied, here’s the overall scorecard at the moment based on his polling, and note the repeated examples in the list of what a big difference polls that named candidates have produced.


Lib Dem ahead (by over 10%)

  • Carshalton & Wallington
  • Colchester
  • Eastbourne
  • Eastleigh
  • Southport
  • Sutton & Cheam
  • Thornbury & Yate

Lib Dem ahead (by 6-10%)

  • Cambridge (March 2015. Up from 1-5% behind)
  • Cheltenham
  • Hazel Grove
  • Kingston & Surbiton
  • Lewes

Lib Dem ahead (by 1-5%)

  • Bermondsey & Old Southwark
  • Birmingham Yardley
  • Brecon & Radnorshire
  • Cheadle
  • North Cornwall (Tied in June and September 2014. In 1-5% ahead category March and April 2015)
  • St Ives (March 2015. Unchanged)
  • Torbay (March 2015. Up from tied)


  • None

Lib Dems behind (by 1-5%)

  • Berwick upon Tweed
  • Berwickshire, Roxburgh & Selkirk (April 2015)
  • Portsmouth South
  • Sheffield Hallam (March 2015 and April 2015)
  • Taunton Deane

Lib Dems behind (by 6-10%)

  • Mid Dorset & North Poole
  • North Devon (March 2015. Down from 1-5% behind)
  • St Austell & Newquay (March 2015. Down from 1-5% behind)
  • Solihull
  • Wells

Lib Dems behind (by over 10%)

Possible gains (1-5% behind):

  • Watford

Possible gains (6-10% behind):

  • Oxford West & Abingdon

Possible gains (over 10% behind):

  • Camborne & Redruth (March 2015)
  • Hampstead & Kilburn
  • Harrogate & Knaresborough
  • Newton Abbot
  • Truro & Falmouth

Note: Maidstone and The Weald has not been polled by Lord Ashcroft, but a Liberal Democrat constituency poll put the party 5% behind.

Updated to add link to Stephen Tall’s subsequent post also about Ashcroft polling and to add in a couple of seat polls missed originally. Updated further 17 April, 25 April, 29 April and 1 May to include extra Ashcroft and Lib Dem constituency polling.

Leave a Reply

Your email address will not be published.

All comments and data you submit with them will be handled in line with the privacy and moderation policies.