Polls from NOP
NOP has been polling since 1959 and has published 600 polls so far; the latest was on 1 May 2005.
Parties
| Pollster | Share (%) | Date |
|---|---|---|
| Lab | 36.0% | 2005-05-01 |
| Con | 33.0% | 2005-05-01 |
| LD | 23.0% | 2005-05-01 |
Coalitions
No polls available yet.
Accuracy at past elections
Average error 2.4 percentage points across 13 elections.
Systematic over- and understatement, per party
Mean signed difference between this pollster’s final poll and the result. A positive figure means the party was stated higher than it finished.
| Party | Mean difference | Elections |
|---|---|---|
| Labour Party | +2.3 percentage points | 13 |
| Liberal Democrats | -0.9 percentage points | 13 |
| Conservative Party | -0.9 percentage points | 13 |
Per election
| Election | Final poll | Days before | Parties scored | Difference |
|---|---|---|---|---|
| 2005 general election | 2005-05-01 | 4 | 3 | 0.3 percentage points |
| 2001 general election | 2001-05-31 | 7 | 3 | 3.5 percentage points |
| 1997 general election | 1997-04-29 | 2 | 3 | 4.0 percentage points |
| 1992 general election | 1992-04-07 | 2 | 3 | 4.0 percentage points |
| 1987 general election | 1987-06-10 | 1 | 3 | 2.3 percentage points |
| 1983 general election | 1983-06-08 | 1 | 3 | 2.3 percentage points |
| 1979 general election | 1979-05-01 | 2 | 3 | 1.3 percentage points |
| October 1974 general election | 1974-10-02 | 8 | 3 | 3.9 percentage points |
| February 1974 general election | 1974-02-23 | 5 | 3 | 2.2 percentage points |
| 1970 general election | 1970-06-12 | 6 | 3 | 2.6 percentage points |
| 1966 general election | 1966-03-27 | 4 | 3 | 1.1 percentage points |
| 1964 general election | 1964-10-09 | 6 | 3 | 2.5 percentage points |
| 1959 general election | 1959-10-02 | 6 | 3 | 1.1 percentage points |
How this is measured: for each past election we take each pollster’s last poll in the 30 days before polling day and compare it, party by party, with the result. The figure is the mean absolute difference across the parties that appear in both. Only parties the pollster actually reported are counted. A late shift in opinion is not the same thing as a polling error.