The public, campaign professionals, news media organizations and interest groups are being inundated with polling results now that we are in the final stages of the 2008 election cycle. We are already witnessing reports about who is winning, losing, and why a candidate is stronger or weaker than the other as campaigns position for earned media and fund raising dollars.
Party Identification (ID) is one of those demographic sub-data points receiving a lot of attention this cycle. Party identification may be reported based on a poll respondent’s declared or partisan registration status in states where this data is available and applicable or, more commonly, based on a “self-identified partisanship” demographic question in polls. This data is very useful for both public review and internal strategic analysis.
We see many reports that Democrats have a certain percentage advantage compared to four years ago, etc. This polling data is catching legitimate changes in voter attitudes. However, self-identified partisanship is a moving target. An individual who identifies as a Democrat today may have said Independent a year ago and Republican four years ago. This is a natural adjustment as people’s attitudes and perceptions of the political parties change over time.
The current partisan shift and trends are fueled by negative perceptions of the GOP brand, which causes fewer people to admit they are affiliated with / support the GOP, even if their underlying beliefs are more in line with Republicans. Obama’s appeal to elements of the electorate who previously didn’t participate (younger voters) and those who are shifting their affiliations based on their attraction to Obama, also fuel the changing partisan identification.
Self-identified partisanship shouldn’t be relied upon as an indicator of voting behavior in isolation. Due to its fluidity, we strongly recommend all polling analysis and public released polls include an evaluation based also on self-described ideology. This sub-group is frequently arrived at by a question such as the following (or similar iterations):
Ideology tends to be a more stable, underlying factor that drives an individual’s voting behavior. It is an excellent variable to evaluate in conjunction with and context of changing partisan affiliations. It can verify strong partisan shifts in favor or against a candidate. It allows for a more realistic, accurate assessment of the political environment if the underlying ideological make-up of the surveyed electorate is out of alignment with a shifting partisan composition.
For example: If a Congressional District shows a 10 point increase in Democratic identification from two years ago, but the District remains a moderate to conservative leaning area, then the surface movement among partisans may be tempered by the voters’ underlying ideological leanings. This situation would show favorable trends for a Democrat, but not a fundamental change in the voters’ likely behaviors, although the partisan shift will have an impact on the immediate election.
Alternatively, in an area that has a 10 point Democratic gain from the last cycle and shows a moderating, more liberal electorate than previous elections, is catching a fundamental change in the electoral make-up and an indication of a long term, more stable movement in the Democrats favor.
The two scenarios where Democrats receive a similar gain in self-identified partisanship will require a different tone of message because of the underlying ideological composition of the District. A campaign that fails to realize the ideological difference between a liberal Democrat and a moderate to conservative Democrat or a center-right, Independent leaning area where self identified Democratic affiliation is increasing can be disastrously off message.
In conclusion, opinions of partisan sub-groups should not be reported or analyzed in isolation of the voters’ ideological position. Publicly released polls should include sub-group analysis by partisan and ideological breakouts and all reporting on polls should include similar evaluations. We also recommend campaigns utilize ideological data in their internal strategic analysis, which will lead to better, more accurate strategic recommendations and decisions.
Party Identification (ID) is one of those demographic sub-data points receiving a lot of attention this cycle. Party identification may be reported based on a poll respondent’s declared or partisan registration status in states where this data is available and applicable or, more commonly, based on a “self-identified partisanship” demographic question in polls. This data is very useful for both public review and internal strategic analysis.
We see many reports that Democrats have a certain percentage advantage compared to four years ago, etc. This polling data is catching legitimate changes in voter attitudes. However, self-identified partisanship is a moving target. An individual who identifies as a Democrat today may have said Independent a year ago and Republican four years ago. This is a natural adjustment as people’s attitudes and perceptions of the political parties change over time.
The current partisan shift and trends are fueled by negative perceptions of the GOP brand, which causes fewer people to admit they are affiliated with / support the GOP, even if their underlying beliefs are more in line with Republicans. Obama’s appeal to elements of the electorate who previously didn’t participate (younger voters) and those who are shifting their affiliations based on their attraction to Obama, also fuel the changing partisan identification.
Self-identified partisanship shouldn’t be relied upon as an indicator of voting behavior in isolation. Due to its fluidity, we strongly recommend all polling analysis and public released polls include an evaluation based also on self-described ideology. This sub-group is frequently arrived at by a question such as the following (or similar iterations):
- How would you describe your own political beliefs -- very liberal, somewhat liberal, moderate, somewhat conservative or very conservative?
Ideology tends to be a more stable, underlying factor that drives an individual’s voting behavior. It is an excellent variable to evaluate in conjunction with and context of changing partisan affiliations. It can verify strong partisan shifts in favor or against a candidate. It allows for a more realistic, accurate assessment of the political environment if the underlying ideological make-up of the surveyed electorate is out of alignment with a shifting partisan composition.
For example: If a Congressional District shows a 10 point increase in Democratic identification from two years ago, but the District remains a moderate to conservative leaning area, then the surface movement among partisans may be tempered by the voters’ underlying ideological leanings. This situation would show favorable trends for a Democrat, but not a fundamental change in the voters’ likely behaviors, although the partisan shift will have an impact on the immediate election.
Alternatively, in an area that has a 10 point Democratic gain from the last cycle and shows a moderating, more liberal electorate than previous elections, is catching a fundamental change in the electoral make-up and an indication of a long term, more stable movement in the Democrats favor.
The two scenarios where Democrats receive a similar gain in self-identified partisanship will require a different tone of message because of the underlying ideological composition of the District. A campaign that fails to realize the ideological difference between a liberal Democrat and a moderate to conservative Democrat or a center-right, Independent leaning area where self identified Democratic affiliation is increasing can be disastrously off message.
In conclusion, opinions of partisan sub-groups should not be reported or analyzed in isolation of the voters’ ideological position. Publicly released polls should include sub-group analysis by partisan and ideological breakouts and all reporting on polls should include similar evaluations. We also recommend campaigns utilize ideological data in their internal strategic analysis, which will lead to better, more accurate strategic recommendations and decisions.
