Congress vs. Nickelback: The real action is in the cross tabs

by Andrew Gelman on January 8, 2013 · 5 comments

in Methodology,Public opinion

I can’t tell if this is a spoof:

When asked if they have a higher opinion of either Congress or a series of unpleasant or disliked things, voters said they had a higher opinion of root canals (32 for Congress and 56 for the dental procedure), NFL replacement refs (29-56), head lice (19-67), the rock band Nickelback (32-39), colonoscopies (31-58), Washington DC political pundits (34- 37), carnies (31-39), traffic jams (34-56), cockroaches (43-45), Donald Trump (42-44), France (37-46), Genghis Khan (37-41), used-car salesmen (32-57), and Brussels sprouts (23-69) than Congress.

Congress did manage to beat out telemarketers (45-35), John Edwards (45-29), the Kardashians (49-36), lobbyists (48-30), North Korea (61-26), the ebola virus (53-25), Lindsay Lohan (45-41), Fidel Castro (54-32), playground bullies (43-38), meth labs (60- 21), communism (57-23), and gonorrhea (53-28).

Here’s the basic pattern:

It’s interesting that Congress is so unpopular among conservatives, given they control half of it. Or, conversely, that it’s so popular among people who are very liberal, given that Congress mostly appears in the news as an opponent of Obama. With n=830 and 13% of respondents characterizing themselves as very liberal, that’s 108 very liberal respondents so the standard error on their estimates is .5/sqrt(108) = .05. So these numbers don’t look like pure noise.

But the real action is in the crosstabs.

Surprisingly, liberals and conservatives have roughly the same opinion of Fidel Castro (relative to Congress):

True to type, conservatives sympathize with playground bullies, and liberals are a bunch of big babies:

Amazingly enough, gonorrhea nearly holds its own among moderates. If supporting an STD is what it takes to be a political moderate, I don’t want to be one:

Finally, I noticed that Brussels sprouts are unpopular among extreme liberals. Maybe this fits in with the “big babies” stereotype again:

This looks like it’s straight from the Onion:

“We all know Congress is unpopular,” said Dean Debnam, President of Public Policy Polling. “But the fact that voters like it even less than cockroaches, lice, and Genghis Khan really shows how far its esteem has fallen with the American public over the last few weeks.”

But all those crosstabs . . . they look real. So I don’t know what to think.

P.S. I hate to link to a robopoll. I think robopollsters are worse than cockroaches, the Kardashians, and carnies (but better than Ebola, root canal, and John Edwards).

{ 5 comments }

Nadia Hassan January 8, 2013 at 11:24 pm

It would be decidedly snorefest in comparison to some of those colorful categories, but nevertheless interesting to see how Congress stacks up against institutions like the banks, oil companies, and so forth…

Dale January 8, 2013 at 11:33 pm

Not a spoof. PPP just has a sense of humor/knows how to make good linkbait. (A previous poll found that 43% of voters would, if they saw mommy kissing Santa Claus, tell daddy.)

The Political Omnivore January 9, 2013 at 4:33 pm

As noted above, quite real. They are known for it–and remember, as “bad” as robo-polling is, PPP was one of the more accurate pollsters for 2012 (so, shame, Monkey Cage–I expect big-data style analysis from you–not raw robo-phobic emotion!). In any event, when asking for unpopular things for their poll, I suggest Polling Organizations (they regularly re-tweet their hate-tweets).

I tried to interview them during the 2012 cycle–but they wouldn’t return my tweets or email–so I interviewed an empty chair version of them instead: http://politicalomnivore.blogspot.com/2012/09/the-omnivore-interviews-public-policy.html

John Dickey January 9, 2013 at 10:36 pm

I would think that some of the advances in poststratification would fix many of the data problems we would expect out of a robopoll. How am I wrong? We would prefer door-to-door polls but we changed because of economics. Why is this type of data so much worse?

Sebastian January 11, 2013 at 2:21 pm

Andrew doesn’t think robocalls are scientifically bad, he thinks they’re ethically bad: You’re harrassing people at home – preferrably at dinner time – and then don’t even have them talk to a real person. Not sure I agree 100%, but I do agree with the general sentiment.

Comments on this entry are closed.

Previous post:

Next post: