UK Election Data Vault

Voting Intention for Next UK General Election by Pollster

Pollster ratings included in parenthesis

Pollsters Ratings

Our pollster ratings are meant as a guide to recent performance of pollsters in UK general election polls. They are based on three components: accuracy, longevity and membership of the British Polling Council.

Seventy percent of a pollster’s score comes from their accuracy in predicting the most recent three UK general elections. The measure used depends on the differences between the national share of the vote (for the three largest parties) and the final poll fielded before polling day. A perfect prediction of each of the last three elections would contribute seventy points to a pollster’s score.

Ten percent of a pollster’s score comes from their longevity in polling UK general elections. A pollster is awarded ten points if the published a poll conducted during each of the last three parliaments (i.e. between 2010-15, 2015-17, 2017-19) New pollsters have a longevity score of zero.

Twenty percent of a pollster’s score comes from membership of the British Polling Council (BPC). Members receive twenty points and non members zero.

Pollster ratings are used in the calculation of our polling average. They have been back calculated so contemporary ratings are used to weight the poling average. For example, the pollster rating used to weight polls for the 2010 general election are based on accuracy in predicting the 1997, 2001 and 2005 general elections.

Calculation of pollster ratings

The calculation of our pollster ratings takes place in three stages. Firstly, we calculate accuracy of each pollster. To do this we take the last poll published before each general election and find the sum of squares of the differences between the poll result and final election result. Only polls taken in the final 60 days before an election are used in pollster ratings

We use polling results for Conservatives, Labour, Liberal Democrats, the SNP and Plaid Cymru. Each party has equal weight in the ratings, however, polling for the bigger parties tends to influence the final ratings more as the polling misses tend to grow with the vote share of a party. Table 1 shows the calculation of accuracy ratings for the 2019 general elections for Great Britain. We do similar calculations for polls in Scotland and Wales.

Table 1: Pollster Performance in 2019 general election

Pollster	CON	LAB	LIB	SNP	PLC	Sum of Squares of Differences	Raw Rating
Opinium	45	33	12	4	0	0	100
Ipsos-MORI	44	33	12			1	99
Kantar	44	32	13	4	0	3	97
Number Cruncher Politics	43	33	12	4	1	3	97
Panelbase	43	34	11	4	0	5	95
YouGov	43	34	12	3	0	5	95
Deltapoll	45	35	10	4	0	8	92
Survation	45	34	9	4	1	10	90
Focaldata	42	34	14	3	1	14	86
BMG	41	32	14			19	81
Savanta Comres	41	36	12	4	0	23	77
Election Result	45	33	12	4	0

We do not base our pollster ratings purely on one election, we prefer to gauge accuracy in the longer term. Hence the second stage is to calculate an average trend rating over the past three elections. Table 2 shows the calculations for the latest trend ratings. All raw ratings have a floor of 25. In these cases either a pollster did not poll a particular election or they performed poorly in terms of accuracy.

Table 2: Pollster Performance in general elections 2015-2019

Pollster	Raw Rating 2015	Raw Rating 2017	Raw Rating 2019	Trend Rating (Average of 2015, 2017 & 2019)
Opinium	85	71	100	85
Ipsos-MORI	83	71	99	84
Kantar	25	90	96	70
Number Cruncher Politics	25	25	96	49
Panelbase	52	73	92	72
YouGov	52	73	91	72
Deltapoll	25	25	92	47
Survation	74	92	90	85
Focaldata	25	25	86	45
BMG	75	28	81	61
Savanta Comres	84	48	77	69

The final stage of the calcuation brings together the three elements of our ratings. In addition to accuracy, the other two elements are longevity and membership of the BPC. The Longevity weigtht is worth between 0 and 10 depending on how many recent election a pollster has polled. BPC members get 20 points and non-members get 0.

Final ratings are measured out of 100. Trend ratings are weighted down so they form 70% of the final rankings. They are then added to the longevity weight and BPC member weight to give the final raking score. These scores are given a letter grade to ease interpretation. Table 3 shows the calcuation of pollster ratings for all pollsters who polled the 2019 election and have published at least one poll since. Seven new pollsters have published a poll since 2019, six of them (Redfield & Wilton, Omnisis, PeoplePolling, More in Common, FindOutNow, Techne and JL Partners) are members of the BPC and have ranking score of 41 (D). The other new entrant (More in Common) is not a member of the BPC and has a score of 21 (F).

Our highest ranked pollster at the moment is Opinium who have an accuracy rating of 85, along with full marks for longevity and BPC membership. Amoung the pollsters who have polled at least one election before, the worst performing is Number Cruncher Politics with a score of 41 (D). This is mostly due to lack of longevity, they scored very well in the one election they have polled.

Table 3: Calculation of pollster ratings for next general election

Pollster	Trend Rating	Trend Rating Weight	Longevity Weight	BPC Member Weight	Final Ranking Score	Pollster Grade
Opinium	85	60	10	20	90	A-
Ipsos-MORI	84	59	10	20	89	A-
Kantar	70	49	10	20	79	B+
Number Cruncher Politics	49	34	7	0	41	D
Panelbase	72	51	10	20	81	A-
YouGov	74	52	10	20	82	A-
Deltapoll	47	33	7	20	60	D+
Survation	85	60	10	20	90	A-
Focaldata	45	32	7	20	58	D+
BMG	61	43	10	20	73	B
Savanta Comres	69	49	10	20	79	B+