IN HER defence, Hillary Clinton did warn that she would be “grossly generalistic” before she began. “You can put half of Trump supporters into what I call the basket of deplorables”, the Democratic nominee for president of the United States said at a fundraiser on September 9th, before classifying her opponent’s voters as “racist, sexist, homophobic, xenophobic, Islamophobic, you name it”. Conservative politicians and pundits pounced on the comment, comparing it with Mitt Romney’s ill-advised denigration in 2012 of 47% of American voters as “dependent” and “entitled”. Seeking to defuse the firestorm, Mrs Clinton apologised the next day—though only for having assigned a number, 50%, to the share of Mr Trump’s voters she believes are beyond salvage. She held firm on the assertion that such unsavoury characters lurk among Trumpistas in unspecified quantities.
Mrs Clinton’s guess about the magnitude of the “deplorable” fraction of Mr Trump’s enthusiasts can still be subjected to a rough fact-check. On July 30th and August 6th, YouGov included in its weekly poll four questions about “racial resentment”, which seek to measure attitudes regarding race relations. At first glance, Mrs Clinton’s 50% estimate looks impressively accurate: 58% of respondents who said they backed Mr Trump resided in the poll’s highest quartile for combined racial-resentment scores. And at a lower threshold of offensiveness—merely distasteful rather than outright deplorable, say—91% of Mr Trump’s voters scored above the national average.
Nonetheless, caution is required before declaring Mrs Clinton’s claim vindicated. First, our racial-resentment index, constructed from a standard battery of questions used in political-science studies for decades, is only an indirect proxy for racism itself. For all the pride in political incorrectness that Mr Trump has brought into vogue, people remain hesitant to admit their prejudices to pollsters. Instead, researchers measure racial resentment using questions on preferential treatment for blacks, such as strong disagreement with the statement that “generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class”.
Moreover, Mrs Clinton accused Trumpistas of far more prejudices than racism alone. Regarding her charge of homophobia, 51.8% of Mr Trump’s partisans—again, just above her suggested figure of half—do support a hypothetical constitutional amendment that would allow states to ban gay marriage. But it is of course possible to support this policy for reasons other than bias against homosexuals, just as it is possible to oppose affirmative action for reasons other than bias against racial minorities.
And regardless of how close her estimate was to the statistical truth, the backlash to Mr Romney’s 47% speech should have been enough for a cautious politician like Mrs Clinton to realise that insulting precise fractions of Americans is bad politics. There is no shortage of images of indisputably deplorable behaviour among Mr Trump’s backers. Even if the wonky, data-hungry Mrs Clinton is familiar with the polling figures that could conceivably support her position, sometimes it’s wiser to let a picture take the place of a thousand statistics.