Paul Krugman wondered what a poll would look like asking people what has happened to the federal budget deficit over the past three years or so, while writing about Rand Paul either lying or being unforgivably ignorant about this issue.

He linked to a poll from 1996 showing the public wrongly "knowing" that the budget deficit had increased during the Clinton years when it had plummeted and then turned into a surplus.

Google jumped in and ran a survey which showed similar results: a plurality of Americans believe that the budget deficit has "increased a lot" since 2010, when in fact it has fallen from $1.3 trillion in late 2009 to an estimated $700 billion by the end of this year. Only 11.8% of Americans correctly know that the deficit is plummeting.

Before I break out the results by age and region, I'm going to preemptively address criticism that a Google survey is a worthless poll. To the contrary, Google became a premier national pollster in 2012 when its brand new commercial survey service -- Internet-only -- was ranked by Nate Silver as the second most accurate pollster of the cycle, beating CNN, Ipsos, SurveyUSA, PPP, Washington Post, ABC News, Rasmussen, Gallup, and everyone else except IBD/TIPP.

Many Internet-only pollsters (serious attempts) were amongst the most accurate, with Internet pollsters taking up five of the top eight spots, and four of the top two. Google Consumer Surveys was #2.

I've calculated the margin of error on this survey (at the time of writing) to be 3.59% for weighted responses, an acceptable margin for a national poll.

Although Google as a pollster has significant credibility now, the question it asked was poorly constructed. Based on intuition, I find it likely that too many respondents confused the national debt with the budget deficit. The debt has increased significantly since 2010 and if that's what people meant when they responded, then many Americans actually are pretty well informed.

Overall

But since that's not what was asked, we can't just make that assumption and throw away the results. 42.1% of respondents wrongly said they believe that the deficit has "increased a lot", while 22.6% wrongly believe that it "stayed about the same". Roughly the same number of people make up the rest: 12% wrongly believe the deficit has "increased a little", 11.5% wrongly believe it has "decreased a little", and just 11.8% correctly know that it has "decreased a lot".

If all respondents believed this was about the debt, then 41.2% know the truth while a collective 57.9% are misinformed to varying degrees. If respondents correctly understood this question to be about the budget deficit, a painful 88.2% of Americans are misinformed and just 11.8% know the truth.

There is no meaningful difference between men and women. Only 9.4% of women and 13.3% of men correctly believe that the deficit has fallen a great deal since 2010.

Age

The best informed group are 18-24, where 21.2% know the truth about the deficit and 34.6% "know" the opposite (big increase). The most uninformed group, are 65+, where only 8.7% know the facts about the deficit and 51.5% believe the exact opposite of what's happened.

There are no ideology or party identification breakouts and these respondents aren't all voters, but we can draw some conclusions from that data set. Young voters, especially 18-24, tend to be strongly liberal. These people support same-sex marriage in the 70% range and would like to see most drugs decriminalized. Older voters, especially 65+, tend to be quite conservative, approving of same-sex marriage (if memory serves) in the low 30s, if it's even that much that strongly oppose abortion and decriminalization.

Although older people are the biggest consumers of products like national evening news broadcasts and cable TV news, it's safe to say that younger voters are better informed generally because of their comfort (aka obsession) with technology and information.

Still, it's likely -- for those that find this useful -- that liberals are probably better informed on the deficit than conservatives are. But given these numbers it's pathetic all around. More 18-24 year olds "know" the wrong thing and the older you go, the worse that gets.

Population density

The differences between urban, suburban, and rural respondents isn't substantial, but it is interesting. The further you get away from cities, the less informed people are about the deficit. 38.8% of urban respondents wrongly "know" that the deficit has "increased a lot" while only 14.2% know the truth. That gap widens with suburban respondents 41.5% to 11.7%, and it's really bad with rural respondents where 51.2% know the wrong thing, and barely 6% know the truth.

There are similar trends by income. The $100,000 to $150,000 income bracket (no data available above that) is the most uninformed, with 47.1% wrongly thinking the deficit has increased by a lot. The best informed bracket is $0-25,000 per year (a substantial number of these people are technically and literally living in poverty) where 15.9% correctly know the deficit has plummeted.

But again, it's still pretty ugly. Urban respondents may be the best informed of all, but that's not saying much when 85.8% know the wrong thing. And the least informed of them all, rural respondents, get it wrong at a 94% clip.

Region

The response varies mostly by severity of being wrong when it comes to region. The South and Midwest look the worst for "increased a lot", at 48.7% and 46.1%. But all four regions are roughly the same, between 13.4% and 11.2% answering correctly.

Politics

I tried to come up with the best informed combined demographic, with my first guess being men between 18 and 24, urban, Southern, making between $0 and $25,000 per year. But there were no results for that demo at all. So draw your own conclusions. I think the overall message here is that first, people are probably confusing debt with deficit and this survey is of limited use, and second, if that's not the case then this country is epically ignorant about the state of affairs in government spending.