The Gallup Daily (National)
Obama 47, Clinton 45
Obama 45, McCain 45... Clinton 46, McCain 45
Rasmussen Reports (National)
Obama 46, Clinton 42
McCain 48, Obama 42... McCain 48, Clinton 44Article
Should be O 46, C 42 for Ras.
Posted on April 20, 2008 1:43 PM
Is it just me, or is anyone else troubled by the over-interpreting that Gallup is doing in their comments to their poll results, such as this:
Obama's support began to erode slightly even before the highly publicized April 16 Democratic debate in Philadelphia, and fell more significantly in the two days immediately after it. His advantage whittled away to a 1-point deficit in April 16-18 tracking, the first time since mid-March that Clinton's share of the vote exceeded Obama's, albeit by a statistically insignificant margin.
Mark and others have repeatedly made the point that deviavions like these could simply be random sampling noise. Has Obama's support in the population moved ahead again after falling behind yesterday? Who really knows? Yesterday's and today's three day averages are not statistically different from one another. Indeed, if you're an Obama supporter, you might look at the current trend line since March 28 (the graphic they are currently presenting) and conclude it is just as equally possible that Clinton's yesterday one-day move ahead is a statistical fluke out of 21 days of plotted polling? Or, if you're a Clinton supporter, maybe he is still behind and yesterday's polling is the outlier. I'd like Gallup's write up more if they sticked more to providing the numbers and long-term trends without providing so much commentary on the short-term trends that often seems to be there simply because they feel a need to provide it.
If Gallup really feels the need to provide commentary, how about more of the long-term trends in support on education levels, etc. or the aggregation of polls across battleground states? That analysis is interesting. This short-term commentary makes me think Gallup is unsophisticated when I know they are not.
Posted on April 20, 2008 3:07 PM
I read the Gallop Obama/Clinton tracking poll almost every day. I think they do a very good job with these "color" comments they add to their poll results. They repeatedly remind the reader of what is a statistically meaningful difference and what is not.
Their commentary is aimed at people who don't understand polls well and aren't familiar with the MOE. They also remind the reader that they use a three day running average.
I think their comments are well balanced. Their headline said " Obama Holds Slight 47% to 45% Advantage" and the first sentence of their commentary (which you omitted) said:
Obama's largest lead to date in the Democratic nomination race came less than a week ago when he led Clinton by 11 points, 51% to 40%. However ...
Posted on April 20, 2008 3:53 PM
I've got to agree with Michael McDonald on this, esp. with respect to short-term trends. I too feel better when the pollsters are referees, keep score and make measurements - and leave the comments to others (like us).
Of course some consensus perspective is helpful for orientation of the non-expert (like myself), but if it is intrusive it makes me uncomfortable about the biases that may have shaped the poll itself.
Heisenberg is not dead.
Posted on April 20, 2008 4:12 PM
I concur as well. Gallup needs to report the data and leave the commentary out of it, as it could very well be sample noise. Note that the poll is now 47-45. Why a 3 point swing in one day? The moe is never understood by most - rarely factored in either by the pundits.
In PA especially, with moe and large undecideds, the span is rather large. I tend to believe susa there.
Posted on April 20, 2008 4:22 PM
Is that a typo on the Rasmussen Clinton-Obama numbers? It's 42% Clinton, 46% Obama on their website.
Posted on April 20, 2008 4:39 PM
I think Gallup usually does a pretty good job of restraining themselves from reading too deeply into or ascribing too much meaning to the tracking poll tea leaves, though today's does seem more colorful than usual. It varies in part because Gallup doesn't use the same writer every day. Compare Frank Newport's more stoic description yesterday of the first time Obama didn't lead in a month, for example.
Posted on April 20, 2008 5:10 PM
Comments: (you may use HTML tags for style)
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.
Please email us to report offensive comments.
See our comment policy here. Note that we require commenters to share their email address via Typekey. We will never share your email address with anyone without your explicit permission.
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