Why do pollsters collect party affiliation information?

[SuperDemocrat]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

 

[Al Azar]

Just to remind the mouthbreathers that most of the country does not lean Republican.

 

[Unkotare]

So they know how much beer to buy?

 

[Toro]

Pollster's ask but don't segregate based on party affilition.

 

[Dekster]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

It is so they can try to align the sample pool with known demographics related to past elections for similarly situated people. For instance, if black republicans make up 3% of the population, then they want as close as possible to 3% black republicans in the poll, and so forth.

 

[Al Azar]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

It is so they can try to align the sample pool with known demographics related to past elections for similarly situated people. For instance, if black republicans make up 3% of the population, then they want as close as possible to 3% black republicans in the poll, and so forth.

If they alter the sample to achieve quotas, it's not a random sample. It's an engineered sample to produce an artificial result.

 

[Dekster]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

It is so they can try to align the sample pool with known demographics related to past elections for similarly situated people. For instance, if black republicans make up 3% of the population, then they want as close as possible to 3% black republicans in the poll, and so forth.

If they alter the sample to achieve quotas, it's not a random sample. It's an engineered sample to produce an artificial result.

Some do, some don't. I think the polls associated with universities tend to try to hit the demographics as close as possible because the students aren't that adept at being able to properly weight polls to bring them into alignment. A couple times I have been called by university pollsters who thanked me but said that they were looking for people in different demographic groups in our area to poll. Some of them never asked much other than who I was voting for which made me think they weren't real polls. I have been polled enough I tend to get a feel for them, especially after taking some of those "few minutes" half hour polls on candidates, parties, and issues

 

[Al Azar]

Sounds to me like rolling a die and not recording a repeated number because it's only supposed to come up 1/6th of the time.

Either the sample reflects actual results or it doesn't, and if it doesn't, it has no value.

 

[konradv]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

If for example a district is 1/3 R, 1/3 D and 1/3 I and half the people polled say they're R, it isn't going to be accurate.

 

[SuperDemocrat]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

If for example a district is 1/3 R, 1/3 D and 1/3 I and half the people polled say they're R, it isn't going to be accurate.

That makes a little sense but if you wanted to get a sample that reflects the area accurately why not use some other bit of information such as male/female ratio or height, age, etc, etc, etc. Wouldn't that just do just as much good in ensuring that your sample is accurate?

 

[konradv]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

If for example a district is 1/3 R, 1/3 D and 1/3 I and half the people polled say they're R, it isn't going to be accurate.

That makes a little sense but if you wanted to get a sample that reflects the area accurately why not use some other bit of information such as male/female ratio or height, age, etc, etc, etc. Wouldn't that just do just as much good in ensuring that your sample is accurate?

How do height and weight effect political leanings? I'm sure they try to get close to 50/50 Ms vs Fs. Age would also be a factor that's controlled for in a good poll.

 

[SuperDemocrat]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

If for example a district is 1/3 R, 1/3 D and 1/3 I and half the people polled say they're R, it isn't going to be accurate.

That makes a little sense but if you wanted to get a sample that reflects the area accurately why not use some other bit of information such as male/female ratio or height, age, etc, etc, etc. Wouldn't that just do just as much good in ensuring that your sample is accurate?

How do height and weight effect political leanings? I'm sure they try to get close to 50/50 Ms vs Fs. Age would also be a factor that's controlled for in a good poll.

It doesn't but if you want to make sure your sample is 'spread out' so you don't get to many people of one segment then you may want to make sure the individuals reflect the true population. It would be unusual to get a sample where everyone is above 6'11'' tall since that is way above the population average but if you sample's average height is closer to what is known to the population's average height then it maybe a better sample. I just don't see why party affiliation information is the only way to accomplish this.

 

[konradv]

I'm wondering what is the purpose of knowing a person's party affiliation when they take a sample to see how people are going to vote in the election? Lets assume that they say they are democrat but are going to vote Trump. Why does that change the percent of people who are voting for Trump vs Hillary? Why is that information even relevant for the poll in the first place? I just kind of think that it shouldn't matter unless you are taking a poll polling how many people are planning to vote for the other party's candidate. A straight up poll of how people are going to vote shouldn't need party affiliation information. A democrat who is going to vote for Hillary is still a vote for Hillary. Her/his party affiliation didn't change that person's answer.

If for example a district is 1/3 R, 1/3 D and 1/3 I and half the people polled say they're R, it isn't going to be accurate.

That makes a little sense but if you wanted to get a sample that reflects the area accurately why not use some other bit of information such as male/female ratio or height, age, etc, etc, etc. Wouldn't that just do just as much good in ensuring that your sample is accurate?

How do height and weight effect political leanings? I'm sure they try to get close to 50/50 Ms vs Fs. Age would also be a factor that's controlled for in a good poll.

It doesn't but if you want to make sure your sample is 'spread out' so you don't get to many people of one segment then you may want to make sure the individuals reflect the true population. It would be unusual to get a sample where everyone is above 6'11'' tall since that is way above the population average but if you sample's average height is closer to what is known to the population's average height then it maybe a better sample. I just don't see why party affiliation information is the only way to accomplish this.

Who says its the only way? It's the only way you mentioned in the OP, but sex and age are also considered. Party info is important, but only as one of several qualifiers in a well done poll.

 

[Redfish]

OK, time for some facts

1. 40% of americans self identify as conservatie, 20% as liberal
2. pollsters are paid to do polls
3. pollsters want to give their customers the result they want
4. polling more dems than pubs will make the result favor dems
5. 1000 out of 330,000,000 is not a statistically meaningful sample
6. the media lies
7. the media has a political agenda
8. social media is 90% bullshit
9. Hillary Clinton is a terrible candidate and a terrible human being
10. Trump could win

 

[konradv]

OK, time for some facts
1. 40% of americans self identify as conservatie, 20% as liberal
2. pollsters are paid to do polls
3. pollsters want to give their customers the result they want
4. polling more dems than pubs will make the result favor dems
5. 1000 out of 330,000,000 is not a statistically meaningful sample
6. the media lies
7. the media has a political agenda
8. social media is 90% bullshit
9. Hillary Clinton is a terrible candidate and a terrible human being
10. Trump could win

Facts or excuses? Why does the most successful pollster out there have Hillary at a 4-1 advantage?

http://projects.fivethirtyeight.com/2016-election-forecast/

 

[Redfish]

OK, time for some facts
1. 40% of americans self identify as conservatie, 20% as liberal
2. pollsters are paid to do polls
3. pollsters want to give their customers the result they want
4. polling more dems than pubs will make the result favor dems
5. 1000 out of 330,000,000 is not a statistically meaningful sample
6. the media lies
7. the media has a political agenda
8. social media is 90% bullshit
9. Hillary Clinton is a terrible candidate and a terrible human being
10. Trump could win

Facts or excuses? Why does the most successful pollster out there have Hillary at a 4-1 advantage?

http://projects.fivethirtyeight.com/2016-election-forecast/

because that's the result they are being PAID to produce. try to pay attention.

 

[konradv]

OK, time for some facts
1. 40% of americans self identify as conservatie, 20% as liberal
2. pollsters are paid to do polls
3. pollsters want to give their customers the result they want
4. polling more dems than pubs will make the result favor dems
5. 1000 out of 330,000,000 is not a statistically meaningful sample
6. the media lies
7. the media has a political agenda
8. social media is 90% bullshit
9. Hillary Clinton is a terrible candidate and a terrible human being
10. Trump could win

Facts or excuses? Why does the most successful pollster out there have Hillary at a 4-1 advantage?

http://projects.fivethirtyeight.com/2016-election-forecast/

because that's the result they are being PAID to produce. try to pay attention.

Easy thing to say. Who's paying you? Your regurgitation of alt-right talking points is highly suspect.

 

[Redfish]

OK, time for some facts
1. 40% of americans self identify as conservatie, 20% as liberal
2. pollsters are paid to do polls
3. pollsters want to give their customers the result they want
4. polling more dems than pubs will make the result favor dems
5. 1000 out of 330,000,000 is not a statistically meaningful sample
6. the media lies
7. the media has a political agenda
8. social media is 90% bullshit
9. Hillary Clinton is a terrible candidate and a terrible human being
10. Trump could win

Facts or excuses? Why does the most successful pollster out there have Hillary at a 4-1 advantage?

http://projects.fivethirtyeight.com/2016-election-forecast/

because that's the result they are being PAID to produce. try to pay attention.

Easy thing to say. Who's paying you? Your regurgitation of alt-right talking points is highly suspect.

I have posted facts. You are free to believe the far left propaganda if you like. I really don't care what you believe.