As you can see, just over a third of respondents overall (36 percent) said they believe that all or most of the government statistics are reliable and accurate. Among Democrats, a narrow majority (53 percent) trusts at least most of the numbers; among Republicans, it’s merely 24 percent.
Note that these responses are not terribly different from those YouGov got when it asked this question in January 2015, which suggests we can’t pin the skepticism on our current president. In 2015, 45 percent of Democrats and 19 percent of Republicans saw all or most government statistics as reliable and accurate.
(Another survey I wrote about in the fall, from Marketplace-Edison Research, also found a substantial data trutherism gap between Donald Trump and Hillary Clinton voters, but the question was worded differently.)
On the unemployment rate specifically, just a quarter of respondents overall in the most recent YouGov poll think the numbers are accurate. Democrats are more likely to buy them than Republicans are, but in neither group does a majority believe they’re accurate. Pluralities of Democrats, Republicans and independents believe the figures undercount the number of unemployed.
To be fair, this skepticism may reflect concerns about the long-standing methodology (e.g., not including people who want jobs but have given up looking), rather than belief in conspiracy theories or fears of malevolent data manipulation.
On some other questions in this survey, it’s easy to understand the connection between partisan affiliation or ideology and views of the data — and also, the direction in which respondents believe the numbers to be wrong.
On crime, for example, Republicans are much more likely to believe rates have gone up, despite official statistics showing they’ve fallen — though in no group does the majority believe crime has gone down.
Likewise, there’s a great degree of skepticism among both Republicans and Democrats about the Congressional Budget Office’s forecast that the Republican health-care bill would lead to 24 million people losing insurance. But a plurality of Democrats (46 percent) believes the CBO score understates the rise in uninsured, while exactly the same share of Republicans believe the CBO overstates this figure.
On at least one question, on the accuracy of the decennial census data, it’s harder to draw a direct connection between partisan affiliation and trust in the numbers. Half of Americans think the census undercounts the number of people in the country, but the share is higher for Republicans than the other groups. Theories as to why?