Pollster.com

Articles and Analysis

 

POLL: TeleResearch Indiana (4/25-29)


TeleResearch Corp/RTV6

Indiana
943 likely voters
Clinton 48, Obama 38

 

Comments
djneedle83:

Trash Poll..

Just read the internals

67% Female
33% Male

African AMericans

69% Female
31% Male

We know the constant trend has been 55-58% Female.

And it also mentions that Obama supports from Males went from +20 to -4 over the span of a week. The poll is too inconsistent and it's polling method is not rigorious enough for solid results.

____________________

djneedle83:

Trash Poll..

Just read the internals

67% Female
33% Male

African AMericans

69% Female
31% Male

We know the constant trend has been 55-58% Female.

And it also mentions that Obama supports from Males went from +20 to -4 over the span of a week. The poll is too inconsistent and it's polling method is not rigorious enough for solid results.

____________________

PHGrl:

djneedle83,
can you share the internals? curious minds want to know..

____________________

J_from_Germany:

PHGrl, here is the link:
http://www.theindychannel.com/download/2008/0501/16107250.pdf
To me it seems, that they weighted the results (50%female/50%male), but I'm not sure. Maybe Eric can clarify this?

____________________

PHGrl:

Danke! J from Germany!
the results dont appear weighted at all.. and Wow. 72% of the democrats polled were female?!

____________________

John S:

This poll must be paid for by Hillary

____________________

J_from_Germany:

I still think, that ths poll is weighted. When you look at page 7 and page 8 the "total" is always the exact average between the female and the male vote, which leads to the assumption, that the total is weighted (50% female/50% male). This would actually oversample the male vote. But still the results look almost to good for Hillary.

____________________

cinnamonape:

I also have my qualms. They don't explain their methodology very well. And a five day poll, COMMISSIONED, to follow the controversy from BEFORE Wright made his comments through the week?

They did daily tracking?

Was it the same number every day? Was the sex ratio the same? Or did the weigh the results somehow each day so that all demographics were constant. If you do that sample sizes become very critical. It works out to less than 200/day.

So the question arises how much is the variability is simply random walk?

And why do they only report the results on the final day. Shouldn't they roll all the results into one, which would make the results far closer? Then the sampling would be comparable to other rolling polls.

Something's very odd about the poll and its presentation.


____________________

cinnamonape:

Surveys a crock. If you average out the 5-day results you get a 46%-41% average which is statistical insignificant from the first three days. I think that it may be showing an effect but not one as great as they think. The error rates per day will go up a lot if they are sampling and manipulating samples under 200 individuals.

____________________

cinnamonape:

Actually the five-day average would be more like 46%-42%. If they are going to report only the final day they should note the actual sample size for that measure (likely under N-200) and revise the error statistic upward, considerably.

Otherwise the larger sample size and error is ONLY valid for their rolling poll over 5-days.

Eric...you should either remove the dot from your sample or use only the collective data in your analysis. This one isn't what it seems.

____________________

Eric:

cinnamonape,

The final page of the pdf is a bit confusing, though I spoke with TeleResearch and they did confirm that the 48-38 is the cumulative based on five nights of polling, the 47-40 a cumulative of four nights of polling, etc. Aside from the first night (showing 45-43), individual nightly results were not released.

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR