Poll 1 (18/04) - 18 April 2017

About the survey

 

Fieldwork dates: April 18th 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 1,000 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org/

Poll 2 (21/04) - 19-21st April 2017

About the survey

 

Fieldwork dates: April 19-21st 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,027 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 3 (24/04) - 21-24th April 2017

About the survey

 

Fieldwork dates: April 21-24th 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,024 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 4 (28/04) - 26-28th April 2017

About the survey

 

Fieldwork dates: April 26-28th 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,012 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 5 (02/05) - 28th April - 1st May 2017

About the survey

 

Fieldwork dates: 28th April - 1st May 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 1,970 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 6 (05/05) - 3rd-5th May 2017

About the survey

 

Fieldwork dates: 3rd-5th May 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,020 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 7 (08/05) - 5th-7th May 2017

About the survey

 

Fieldwork dates: 5th-7th May 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,038 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 8 (14/05) - 12th-14th May 2017

About the survey

 

Fieldwork dates: 12th-14th May 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,030 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 9 (21/05) - 19th-21st May 2017

About the survey

 

Fieldwork dates: 19th-21st May 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,004 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 10 (26/05) - 24th-26th May 2017

About the survey

 

Fieldwork dates: 24th-26th May 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,044 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 11 (30/05) - 26th-29th May 2017

About the survey

 

Fieldwork dates: 26th-29th May 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,002 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 12 (02/06) - 31st May-2nd June 2017

About the survey

 

Fieldwork dates: 31st May-2nd June 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,051 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

Poll 13 (04/06) - 2nd-4th June 2017

About the survey

 

Fieldwork dates: 2nd-4th June 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 2,000 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org

  • Poll 14 (07/06) - 6th-7th June 2017

About the survey

 

Fieldwork dates: 6th-7th June 2017

 

Interview method: Online

 

Online Sampling Method: A nationally representative sample was selected at random from the NewVista panel of 200,000+ adults, with sample selected in proportion to population distribution.

 

Sample size: 1,532 GB adults aged 18+

 

Data weighting: In both cases, data were weighted to the profile of all adults aged 18+. Data were weighted by sex, age, social class, household tenure, work status and region. Targets for the weighted data were derived from the National Readership survey, a random probability survey comprising 36,000 random face-to-face interviews conducted annually.

The data were further weighted by declared votes in the 2015 general election. The weighting scheme is designed as follows: 

Weighting by past votes – online 

The sample were fully weighted to the 2015 General Election result. Recall on online tests has been superior to that achieved over the telephone, and given the resultant minimal impact of past weight voting online, making provision for faulty recall seems unnecessary. 

Weighting for turnout  

ICM has constructed a turnout probability matrix based on an interlocking combination of age and social grade. Turnout estimates from various sources, most notably but not limited to the British Election studies were factored into the matrix in order provide a best estimate for each demographic subgroup, modelled to population incidence and real turnout levels from the 2015 General Election.

Weighting for political interest 

The data is weighted to the standard question on level of interest in politics, from BES 2015. 

Voting intentions: 

ICM derives vote intentions from 2 questions. First of all respondents are asked if they are registered to vote in UK elections. Those not registered are excluded from vote intention calculations. Then, respondents are asked how likely it is that they would be to go and vote in a new election, although this question is no longer used to create turnout factors. Those who say they will vote are asked to say which party they would support in a new election. Respondents are then asked whether they voted in 2015 and which party they voted for in that election. The vote figures shown in the tables are calculated after ICM has excluded those who are not registered, say they will not vote, refuse to answer the question or don’t know who they would vote for (but see right)

Adjustment process 1: ‘Partial Refuser’ Reallocation 

A.) 75% of 2015 Conservative and Labour voters who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015.

B.) 50% of 2015 voters for all other parties who refuse to answer the vote intention question or say they don’t know, are added back to the party they voted for in 2015. 

Adjustment process 2: ‘Total Refuser’ Reallocation 

‘Total refusers’ are people who refuse/DK their future vote intention AND also refuse/DK who they voted for in the previous General Election (2015). Given the lack of any political information about such respondents to date, ICM has excluded them from the vote intention figures. However, our post-2015 Recall Survey revealed that Total Refusers (who were subsequently willing to tell us what they did in the 2015 General Election) split disproportionately across different parties. Indeed, one important observation was that more than half of all Total Refusers actually voted Conservative, with more than twice as many voting Conservative than Labour.

Our new adjustment thus reallocates some Total Refusers back into the poll sample. This is achieved in the following way:

 

1. The number of Total Refusers on any poll is multiplied by the proportion of Partial Refusers who were (already) re-allocated in Adjustment Process 1. (For example, if 60% of Partial Refusers were added back, then 60% of Total Refusals will be added back).

 

2. Total Refusers are then multiplied by each party’s share of reallocated Partial Refusers. (For example, if 40% of already allocated Partial Refusals were 2015 Conservative voters, then 40% of remaining Total Refusals will be reallocated to the Conservatives).

 

3. ICM’s default position is that Total Refusers at least look like Partial Refusers in terms of political make-up. However, given the findings of our Recall Poll, we believe that Total Refusals are probably even more proConservative than pro-Labour. In order to allow for this, the share of Total Refusals added back to the Conservatives is increased by 20% (for example, from the 40% mentioned in (2, above) to 60%), with a corresponding reduction of 20% in the share of Total Refuser reallocation to Labour.

 

Questions: 

The tables on this site show each question, in full, in the order they were put to respondents, all response codes and the weighted and un-weighted bases for all demographics and other data including but not limited that published .

 

Further enquiries: martin.boon@icmunlimited.com

   

British Polling Council: ICM is a member of the British Polling Council and abides by its rules. http://www.britishpollingcouncil.org