Misreporting Polls: Republican’s “Historic” Lead Over Democrats
Misreporting Polls: Republican’s “Historic” Lead Over Democrats
Written by Rev. Robert A. Vinciguerra Saturday, 11 September 2010 19:50
A Gallup poll released on August 30th showed that the Republican Party had a lead over Democrats in the Generic Congressional Ballot poll by ten points. A percentage so wide hasn’t occurred since 1942… At least not in Gallup’s tracking.
Observers from both the blogosphere and respected media outlets immediately lauded the results as “historic” and made speculations about the Democrats impending electoral doom in the upcoming November midterms.
Was the poll really historic? Was the news handled appropriately by the national media? The short answer to both questions is a resounding no.
Before determining what level and type of news coverage would have been fitting for the situation, it is important to first understand what the poll really said and how to read it.
The poll in question was conducted by Gallup, a reputable polling firm, from August 23rd to August 29th 2010 with a sample of 1,540 people with a four percent margin of error. The question asked is essentially whether or not the respondent would rather vote for a democratic or republican candidate. The results were 51% to 41% in favor of the GOP. Indeed, it does show a ten point preference for republicans.
However, a single snapshot of a tracking poll id no more statistically useful than asking a blind man to identify a model of car by only feeling its tires.
Tracking polls take a series of snapshots of the mood of the electorate during a specific window. The polls are repeated so that the disposition of the voters can be effectively tracked. Each poll could be completely accurate within the margin of error. But before we can know with any degree of certainty, the poll must be compared to the previous tracking poll, the following tracking poll, and other similar tracking polls taken by other firms in the same timeframe. Only then can an accurate image emerge.
In this case, Gallup’s previous tracking poll, which concluded on August 23rd, showed only a three point lead for Republicans (47% to 44%). And what of Gallup’s following tracking poll? It shows that Republicans and Democrats are exactly tied at 46% on the Generic Ballot.
How could that be? Did the GOP really gain seven points in just six days? Did Democrats really make up a ten point deficit? Of course not. By plotting out the Gallup poll data for the past we can see from the exponential trend line (green) that in fact the average results show a gradual decline in Democratic support. It quickly becomes apparent that the poll shows a 10 point lead for the GOP is an outlier.
Did Gallop do anything wrong? Probably not. Nate Silver’s FiveThirtyEight give Gallup their 11th highest rating. Sometimes the samples are just bad and anomalous results occur. However, that’s not to say that Gallup behaved in a responsible manner. It’s likely that they also had suspicions that the poll was an outlier, yet they released a statement calling the lead “unprecedented,” a good way to get their name in the news. And it’s not unprecedented, not really. Perhaps in their history, but Rasmussen has been releasing polls showing a nine to ten point lead for republicans since January.
Gallup is a business, though, so their actions are understandable and nearly forgivable. The media, on the other hand, doesn’t get off so easy.
The “historic nature” wasn’t merely discussed by the usual suspects. It’s expected that partisans, politicians, conservative media, and right-leaning blogs would jump all over the poll results to promote their own self-interests. However, the “historic” angle of the story wasn’t just propagated by the RNC, Drudge, or Fox News, it was prominently promoted by legitimate news agencies such as NPR, The Washington Post, and CBS News just to name a few.
The media knew then, or at least they should have, and there’s no excuse as to why they didn’t, the same thing that everyone knows now. The “historic polls” wasn’t historic at all, and it was an outlier.
Though they may have forgotten so, it is the media’s job to tell the public who’s right and who’s wrong, what’s true and what’s not, what’s real and what isn’t, and to do so with authority on issues that affect the people who they write for. It’s a responsibility at which they have been failing spectacularly. Stories that aren’t important (see: Brittany Spears) should be left to the gossip rags where they belong. When a leader promotes an out and out lie (see: death panels), then the media must come out and say so.
By elevating this particular poll and telling an incomplete and inaccurate story, the media has done damage. They’ve altered people’s perceptions of reality which could have real world consequences. Some voters have been emboldened by the narrative of the story, while others have been discouraged, and not all of them will be aware of the Gallup poll released on September 6th which shows a dead heat, or the August 31 National Journal poll which shows Democrats leading by two.
The powered wielded by the media is necessary for any healthy democracy. For the populus to be well informed the media must be religiously responsible in their reporting. Their erosion of responsibility in exchange for ratings has upon this country a great illness which has yet to be cured.