<p>Now that is a fun topic…why do we have so little about who people voted for, but all kinds of stuff about why? as of this poiint we have little if no data from 2004, and rumors I have heard and a trying to verify, is that this year, in most cases the polls lined up pretty close to actual #s, EXCEPT in Virgnia and Montana</p>
<p>anyone have any information</p>
<p>as well, seems like much information is under lock and key, why who knows…but it is interesting that will the faulty machines that this information is not as readily available</p>
<p>just makes one wonder</p>
<p>as for reliability, well, statsitically for many many years, the exit polls have stood up to what actual results were, except in 2000, 2002 and 2004…did the masses suddenly change that year? statistically, unlikely, but just throwing it out there</p>
<p>Actually, I don’t think the results in 2004 were really that bad. What WAS bad was the early release of incomplete, unanalyzed poll data too early. What do I mean here? That the raw numbers are not what good pollers use. Depending on their sampling scheme, response rates and so forth, they will adjust the raw numbers to gain greater accuracy. But this takes time and often additional data collected separately from the actual poll. For instance, the raw numbers may be adjusted for turnout. </p>
<p>I won’t bore you with the process. If you’re curious, there are some good websites that discuss this in detail.</p>
<p>At any rate, the difference this year is that the pollsters kept their data unler lock and key so to speak (actually in a room with no phones and no cell phones) until analysis was finished and the release time arrived. In the past, there were early leaks. Like 04.</p>
<p>Actually, I thought the polls were abysmally off this year (but not in a result-changing way). The Pew Center came out with a big poll on November 6 saying that the conservative base was energized and the gap between likely voters planning to vote Dem or Rep had narrowed to 4-7%. In retrospect, that was a howler.</p>
<p>In Pennsylvania, the polls had Bobby Casey up over Santorum by about 10%, but the lead had come down a bit and people were talking about Santorum’s momentum. Casey won by 20%. He won even if you exclude Southeastern PA and Allegheny County. In fact, Casey’s margin over Santorum was almost exactly the same as Rendell’s over Swann for governor, when Rendell had polled much better than Casey. Dems retained Senate seats in New Jersey and Maryland with margins of 8-10%, despite those races being called “toss-ups”.</p>
<p>I don’t think the real pollsters believed that. Santorum pulled his network advertising about 10 days before the election, going solely with cable. That was likely because the national party pulled their money. They knew it was over.</p>
<p>The exit poll data will emerge in the post mortems over the next few weeks. Keep an eye on C-SPAN. For example, the DLC will surely hold a press briefing and release their analysis.</p>
<p>Some tidbits I’ve seen:</p>
<p>The Democratic candidates got nearly a third of the white evangelical vote.</p>
<p>The Democratic candidates got their usual majority of women voters, but also won a majority of male voters for the first time in recent memory.</p>
<p>The Democratic candidates totally wiped out the inroads the Republicans had made with Latino/a voters, getting 75% of those votes (I think I remember the percentage correctly).</p>
<p>Because they are based on statistical processes, no ONE poll is all that accurate. So it is easy, but trivial, to make statements like JHS did about ONE poll. </p>
<p>It is also interesting to note that polls have their greatest predictive power for presidential races, less for governorships and even a bit less for senatorial races. Few house races get serious polling numbers even released.</p>
<p>The pollster.com blog has some nice technical analyses of polls, but you’d better like the technical side to appreciate and understand the work.</p>
<p>Interesteddad, </p>
<p>you are right. The interpretation has already begun. One of the big less obvious and less publicized parts of polls are the demographic questions, and the kind of analysis they enable. To me, that’s the fascinating part. The election returns tell us who won. The exit poll data tells us more about why, and by whom.</p>
<p>Glad that the republican precinct people call me to remind me to vote. If it wasn’t a young lady at the other end of the line, I’d use words that are not polite. But she got the message-and she ‘understood.’</p>
<p>CGM was their election fraud this year. You ranted alot about how bad these new machines are and how easy they are to cheat. So was there election fraud. I am interested to hear your excellent analysis because judging by your previous comments I assumed you would come out strong against the results of this election.</p>
<p>I personally do not feel there was any election fraud, but I also felt that way before the election when there was the whole talk about machines…</p>
<p>Election fraud has been going on for years. Some years its more obvious than others. And it has been practiced by both sides and is well documented for past years. </p>
<p>This year? Just wait a few months. </p>
<p>Get used to the accusations and concerns. It is part of the process. Some play hard. Some play too hard. Some get in trouble. And life goes on.</p>