Standard Deviation

TCY Junkie · 11-08-2008, 12:26 PM

Normal distribution is 68.27 percent. I was eating out and they had election news on and it stated 1 out of 20 black people didn't vote for Obama in New York City. I'm not going to look up the vote breakdown and just assume they are right like most people do when they hear things on TV.

I hate the freaking news and figures like these. I really can't stand politicians so I can't understand the situation. Someone please explain. I didn't post in that one thread because the one time I opened it up everyone was attacking each other and said I wasn't going back in there. Maybe I should have posted this in the punch in the face thread, but I'm truly curious. I have things to do, maybe I will have an answer when I get back this evening.

CraigSca · 11-08-2008, 03:50 PM

Is there more to your question in the other thread? I'm not sure I'm understanding what you're asking here.

lordscarlet · 11-08-2008, 09:46 PM

Yeah, what's the question?

TCY Junkie · 11-08-2008, 11:21 PM

Quote:

Originally Posted by CraigSca

Is there more to your question in the other thread? I'm not sure I'm understanding what you're asking here.

No, I opened the other thread like a month ago and it irritated me.

I just figured with all the factors I could think of the vote should have been closer to 10 to 1 with a sample size that big. I don't really watch the news, does skin color really matter that much? I'm not upset at all. If I was a person that got depressed, I would be depressed by these stats. I glanced for some details of the breakdown, but I got tired of thinking about it.

I just find it hard to believe they did that good of a job at getting black people that would vote for him out to vote and other black people not to vote. If they really did that good of a job than hats off to them. Here I will say it: I don't think this is normal breakdown if you took race out of the issue and that is depressing to me. Sure a few would vote that way, but not enough to scew the data. Am I wrong in thinking the percentage shouldn't be 95 percent with race out of it, if so tell me things. I really don't expect a response that will satisfy me and that also is slightly depressing in itself.

JonInMiddleGA · 11-08-2008, 11:23 PM

I think I'm more confused about what you're trying to figure out now than I was from just the first post.

edit to add: Yes, 95% of blacks who voted casting ballots for Obama sounds on par with most of the other data I've seen.

If anything, it's actually lower than the New York state figures from the CNN exit polls, which showed Obama with 100% of the votes by black males in the Presidential race & 99% of the black female voters, rounding up to a 100% result for all black voters. Hell, Kerry had a 90% to 9% edge over Bush with black voters in NY in 2004 (1% went to Nader), nothing really surprising in either of those numbers to anyone who has been paying much attention.

thealmighty · 11-08-2008, 11:55 PM

BTW, saying normal distribution is 68.27 percent doesn't make any sense in and of itself. 68.27% is the percent of whatever population that would fall within 2 sd (one to negative + one to positive, if I recall) of a norm. Normal distribution is an entire curve, not just the percent you gave.

TCY Junkie · 11-09-2008, 02:40 AM

Quote:

Originally Posted by thealmighty

BTW, saying normal distribution is 68.27 percent doesn't make any sense in and of itself. 68.27% is the percent of whatever population that would fall within 2 sd (one to negative + one to positive, if I recall) of a norm. Normal distribution is an entire curve, not just the percent you gave.

Yeah, I didn't have the numbers. I basically made it impossible to understand because I didn't put everything down. I'll just assume 90 percent is the average over the years. I wonder what the percentage really is. I also wonder what percent one standard deviation is. And the reason for this thread how many standard deviations is it above this year? How much of an outlier is Obama? Obama might not be an outlier but I think he is.

Has the percent ever been lower than 1 in 15?

Quote:

Originally Posted by JonInMiddleGA

I think I'm more confused about what you're trying to figure out now than I was from just the first post.

edit to add: Yes, 95% of blacks who voted casting ballots for Obama sounds on par with most of the other data I've seen.

If anything, it's actually lower than the New York state figures from the CNN exit polls, which showed Obama with 100% of the votes by black males in the Presidential race & 99% of the black female voters, rounding up to a 100% result for all black voters. Hell, Kerry had a 90% to 9% edge over Bush with black voters in NY in 2004 (1% went to Nader), nothing really surprising in either of those numbers to anyone who has been paying much attention.

Thanks for the last election data. I would have guessed it been around 1 in 8 at most, instead of 1 in 10. Still a jump to 1 in 20 doesn't make me feel good. What kind of increase did Obama have with other groups and as a whole? I guess you could tell where to get the info, but I don't think I have anymore questions on this. I was just curious after seeing 1 in 20 the other day if race scew the data. I think I have my answer.

lordscarlet · 11-09-2008, 06:43 AM

I think there is either some major thought you're leaving out here, or some misunderstanding of how statistics work. When it comes to standard deviation, I don't know that you can apply it here. When you're just talking about the probability of someone voting for obama, the outliers are people that voted for mccain or an independent. You're not going to end up with 68% voting for one specific candidate just about anywhere.

Having said that, you seem to be "disturbed" that 95% of black voters would vote for Obama. I'm not sure why that is. I live in Washington, DC and 93% of all voters voted for Obama. Does it bother you that almost 1 in 20 voters, regardless of race, age, sex, etc voted for Obama in DC?

And, for reference, I don't know what they're calling "New York City", but manhattan was 85% for Obama as a whole.

rowech · 11-09-2008, 06:50 AM

Within one standard deviation is considered to be 68% and within two standard deviations is considered to be 95%. Anything outside of two standard deviations is considered unusual. However, these percentages compare how closely something is compared to a mean. It has nothing to do with straight percentages.

Your better use of this data would be to take 50-100 different polls and find the average percentage of blacks for the question you care to know about. Calculate the standard deviation and then add/subtract that number twice from the average you calculated. That will give you the range of values you would expect for blacks voting based upon your data. (example: 92-99%) Then anything outside that range would be considered unusual.

Klinglerware · 11-10-2008, 08:34 AM

As Jon states, that 95% of African-Americans in NYC voted for Obama is both plausible and unsurprising, since the African-American demographic typically votes overwhelmingly Democratic election after election. (Even though the finding is certainly plausible, TCY Junkie raises a good point about the validity of the source--the exit-poll breakdown just for sampled NYC precincts exists, but it would not be available to the media due to sample size issues. So, unless that source referred to some other post-election polling or if they got their hands on data they weren't supposed to have, I'm wondering if they are misrepresenting the exit poll or just making the data up.)

As to TCY Junkie's initial question, race is not the primary driver of the demographic splits you are seeing in the poll. As Jon states, the black vote was almost as highly democratic in 2004. Also, I was a little astounded that the percentage of white males who voted for McCain in some deep south states was almost as high. In Mississippi and Alabama, the white male vote for McCain was 85%-90%. But I looked at the 2004 exit polls, and the splits were almost as high for Bush.

So, now that we've established that these are not outliers but longer-term trends, we can safely conclude that Obama's race wasn't a big driver of the vote in 2008. Issues and values are more significant. At the root, the vast majority of the African-American electorate perceives the Republican party as a party that does not promote policies that positively address the interests of large segments of the African-American demographic. If the perception is in fact true, then why should a black voter vote Republican if it is not in his or her interest? If the perception is false, then the Republican party should do a better job marketing itself to that demographic. Of course, it is also possible that since African-Americans are not central to Republican electoral strategy, Republicans actually have ignored them--which does promote the perception (whether deserved or not) that Republicans and African-Americans don't share the same interests.

Likewise, for White Males in the Deep South and the Democratic Party. A large percentage of this demographic perceives the mainstream Democratic Party as almost hostile to their values, beliefs, and policy interests. So, why should they vote for a party that doesn't share their values? If seen in this light, voting behavior does make a little more sense.

Finally, I'm not saying that race didn't matter--that is almost naive. We've seen small bumps in some of the poll results that does suggest that race might have played a very small role in changes in voting behavior. But like I said, interests and values, and which party better shares these with each individual voter seems to matter much more.

11-08-2008, 03:50 PM	#2
CraigSca Pro Starter Join Date: Jul 2001 Location: Not Delaware - hurray!	Is there more to your question in the other thread? I'm not sure I'm understanding what you're asking here. __________________ She loves you, yeah, yeah, yeah, yeah! She loves you, yeah! how do you know? how do you know?

11-08-2008, 09:46 PM	#3
lordscarlet Pro Starter Join Date: Oct 2005 Location: Washington, DC	Yeah, what's the question? __________________ Sixteen Colors ANSI/ASCII Art Archive "...the better half of the Moores..." -cthomer5000

11-08-2008, 11:23 PM	#5
JonInMiddleGA Hall Of Famer Join Date: Nov 2000 Location: Behind Enemy Lines in Athens, GA	I think I'm more confused about what you're trying to figure out now than I was from just the first post. edit to add: Yes, 95% of blacks who voted casting ballots for Obama sounds on par with most of the other data I've seen. If anything, it's actually lower than the New York state figures from the CNN exit polls, which showed Obama with 100% of the votes by black males in the Presidential race & 99% of the black female voters, rounding up to a 100% result for all black voters. Hell, Kerry had a 90% to 9% edge over Bush with black voters in NY in 2004 (1% went to Nader), nothing really surprising in either of those numbers to anyone who has been paying much attention. __________________ "I lit another cigarette. Unless I specifically inform you to the contrary, I am always lighting another cigarette." - from a novel by Martin Amis Last edited by JonInMiddleGA : 11-08-2008 at 11:32 PM.

11-08-2008, 11:55 PM	#6
thealmighty Pro Starter Join Date: May 2003 Location: heaven	BTW, saying normal distribution is 68.27 percent doesn't make any sense in and of itself. 68.27% is the percent of whatever population that would fall within 2 sd (one to negative + one to positive, if I recall) of a norm. Normal distribution is an entire curve, not just the percent you gave. __________________ Check out The Unofficial FOFC Movie Guide Here

11-09-2008, 06:43 AM	#8
lordscarlet Pro Starter Join Date: Oct 2005 Location: Washington, DC	I think there is either some major thought you're leaving out here, or some misunderstanding of how statistics work. When it comes to standard deviation, I don't know that you can apply it here. When you're just talking about the probability of someone voting for obama, the outliers are people that voted for mccain or an independent. You're not going to end up with 68% voting for one specific candidate just about anywhere. Having said that, you seem to be "disturbed" that 95% of black voters would vote for Obama. I'm not sure why that is. I live in Washington, DC and 93% of all voters voted for Obama. Does it bother you that almost 1 in 20 voters, regardless of race, age, sex, etc voted for Obama in DC? And, for reference, I don't know what they're calling "New York City", but manhattan was 85% for Obama as a whole. __________________ Sixteen Colors ANSI/ASCII Art Archive "...the better half of the Moores..." -cthomer5000

11-09-2008, 06:50 AM	#9
rowech Pro Starter Join Date: Feb 2006	Within one standard deviation is considered to be 68% and within two standard deviations is considered to be 95%. Anything outside of two standard deviations is considered unusual. However, these percentages compare how closely something is compared to a mean. It has nothing to do with straight percentages. Your better use of this data would be to take 50-100 different polls and find the average percentage of blacks for the question you care to know about. Calculate the standard deviation and then add/subtract that number twice from the average you calculated. That will give you the range of values you would expect for blacks voting based upon your data. (example: 92-99%) Then anything outside that range would be considered unusual. Last edited by rowech : 11-09-2008 at 06:51 AM.

11-10-2008, 08:34 AM	#10
Klinglerware College Starter Join Date: Dec 2003 Location: The DMV	As Jon states, that 95% of African-Americans in NYC voted for Obama is both plausible and unsurprising, since the African-American demographic typically votes overwhelmingly Democratic election after election. (Even though the finding is certainly plausible, TCY Junkie raises a good point about the validity of the source--the exit-poll breakdown just for sampled NYC precincts exists, but it would not be available to the media due to sample size issues. So, unless that source referred to some other post-election polling or if they got their hands on data they weren't supposed to have, I'm wondering if they are misrepresenting the exit poll or just making the data up.) As to TCY Junkie's initial question, race is not the primary driver of the demographic splits you are seeing in the poll. As Jon states, the black vote was almost as highly democratic in 2004. Also, I was a little astounded that the percentage of white males who voted for McCain in some deep south states was almost as high. In Mississippi and Alabama, the white male vote for McCain was 85%-90%. But I looked at the 2004 exit polls, and the splits were almost as high for Bush. So, now that we've established that these are not outliers but longer-term trends, we can safely conclude that Obama's race wasn't a big driver of the vote in 2008. Issues and values are more significant. At the root, the vast majority of the African-American electorate perceives the Republican party as a party that does not promote policies that positively address the interests of large segments of the African-American demographic. If the perception is in fact true, then why should a black voter vote Republican if it is not in his or her interest? If the perception is false, then the Republican party should do a better job marketing itself to that demographic. Of course, it is also possible that since African-Americans are not central to Republican electoral strategy, Republicans actually have ignored them--which does promote the perception (whether deserved or not) that Republicans and African-Americans don't share the same interests. Likewise, for White Males in the Deep South and the Democratic Party. A large percentage of this demographic perceives the mainstream Democratic Party as almost hostile to their values, beliefs, and policy interests. So, why should they vote for a party that doesn't share their values? If seen in this light, voting behavior does make a little more sense. Finally, I'm not saying that race didn't matter--that is almost naive. We've seen small bumps in some of the poll results that does suggest that race might have played a very small role in changes in voting behavior. But like I said, interests and values, and which party better shares these with each individual voter seems to matter much more.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)