There they go again.

Conservative media figures, apparently disheartened by recent poll results showing presumptive Republican presidential nominee Donald Trump heading in the wrong direction, are once again claiming biased or unreliable pollsters are unfairly weighting results against their party. If this sounds familiar, it’s because they did the same thing in 2012, spending months attempting to “unskew” polls showing Mitt Romney losing, only to watch him be soundly defeated on election day.

In the run-up to the 2012 election, conservatives consistently complained that polls showing President Obama in the lead were inaccurately counting the gap between self-identified Democrats and Republicans. According to this school of thought, the polls were being “skewed” to show Romney losing. One blogger, Dean Chambers, took the data in the polls and reweighted them with a partisan split friendlier to Republicans resulting in “unskewed” polls showing Romney easily winning. Chambers’ work -- which was more akin to wishful thinking than academic analysis -- was nonetheless widely cited by conservative media as evidence of a concerted effort to influence the results of the presidential election in Obama’s favor.

The polls were not skewed. An average of 2012 election polling predicted that Obama would win by 0.7%. In reality, the victory was by a margin of 3.86%. If anything the polls undercounted Obama’s support.

Polls can of course go up and down, and the occasional outlier is inevitable. But the argument that the partisan split that pollsters report as they survey voters is somehow skewed to help Democrats is a conspiracy, not actual analysis.

Despite this, conservative media are once again pushing the “unskewed” theme as recent polls show Clinton leading Trump.

This time, the charge against the polls is being led in part by the candidate himself. Trump recently responded to a poll showing him losing with tweets that complained “The @ABC poll sample is heavy on Democrats.  Very dishonest - why would they do that?” and “The ‘dirty’ poll done by @ABC @washingtonpost is a disgrace. Even they admit that many more Democrats were polled.”

The ABC News/Washington Post poll in question shows Clinton ahead of Trump 51%-39%.

FiveThirtyEight’s Nate Silver noticed the reboot of the “unskewed” theme and asked, “Has anyone seen Donald Trump and Dean Chambers in the same room together?”

Conservative media figures have also zeroed in on the ABC/Wash. Post poll for criticism. On Fox News’ Fox and Friends, co-host Steve Doocy said that in the methodology for the ABC/Washington Post poll “they actually talked to 12 percent more Democrats than Republicans,” adding, “According to the Gallup poll, there are 3 percent more Democrats in the country than Republicans, so it looks like they've got a favorite in it.” During the same segment, co-host Brian Kilmeade explained to viewers, “So far Donald Trump leads in most independent polls.” This is true, if by “most,” Kilmeade meant none of the last 21 polls included in Real Clear Politics’ general election polling data.  

The methodology for the poll, conducted by Langer Research for ABC/Washington Post, addresses the partisan breakdown":

Partisanship can follow political preferences, and in this poll Democrats account for 36 percent of all adults and 37 percent of registered voters – a non-significant (+3) difference from last month. (The former is numerically its highest since 2009, the latter, since 2012.) Republicans account for 24 percent of all adults and 27 percent of registered voters, about their average in recent years, with the rest independents.

This accounts for little of the shift in voter preferences, however. Even using the same party divisions from last month’s ABC/Post survey, in which Trump was +2, he’d now be -8. The reason, mentioned above, is his comparatively weak performance among Republicans – 77 percent support – compared with Clinton’s support among Democrats, 90 percent. 

A Reuters/Ipsos poll showing Clinton with a 13% lead over Trump prompted an outburst as well.

On Fox News host Sean Hannity’s official website, a blog post complained the poll “is heavily skewed.” On his June 27 radio show, Hannity cited the partisan breakdown and described it as a “misleading poll” because the media is “in the tank for Hillary.”

Hannity apparently didn’t learn his lesson about attempting to unskew polls in 2012, when he was saying things like, “These polls are so skewed, so phony, that we need to start paying attention to what’s going on so that you won’t be deflated.”

In a post purporting to highlight “More Polling Tricks” from an “EXTREMELY SKEWED” poll, conservative blogger Jim Hoft of Gateway Pundit complained this week that “Reuters freighted their poll with 20 percent more Democrats than Republicans” and concluded that “we can safely say that Trump appears to be in much better shape than the poll suggests and could likely be headed to a landslide victory in November.”

Hoft made a similar argument in September of 2012, complaining that a CNN poll showing Obama leading Romney “drastically oversampled Democrats to get this stunning result.” He then went on to cite Dean Chambers, who said that when “unskewed” the CNN poll showed Romney leading by eight percent.

Perhaps remembering how much egg the conservative media had on its face after the 2012 debacle, Fox News contributor Brit Hume tried to steer his fellow conservatives away from repeating their mistakes.

In an appearance on America’s Newsroom, Hume noted that Trump “couldn’t stop talking” about polls showing him in the lead during the primaries, but now “his supporters, the ones I hear from anyway say that the poll is rigged, and all the rest of it.” Then he told host Martha MacCallum, “I don't think your viewers should pay too much attention to that. Look at the polling averages. Look at all the polls put together, to see what you get. And I think the picture's pretty clear. He's trailing, but not insurmountably.”