RJ poll has some of the screwiest internals I've seen since....2010

The Las Vegas Review-Journal poll suggests state Sen. Mark Hutchison is crushing Assemblywoman Lucy Flores among Hispanics, that Republican candidates are destroying Democrats in Las Vegas, where Democrats have a 14 percentage point edge and that GOP state Sen. Barbara Cegavske is getting 42 percent of the African-American vote.

In other words, it makes no sense.

That doesn't mean Hutchison is behind -- he almost surely isn't. Or that Cegavske isn't running close to Treasurer Kate Marshall -- she may be. Or  that Adam Laxalt isn't losing to SOS Ross Miller in the race for AG -- he almost certainly is.

But there is no reason to have any faith in these numbers based on the internals of the poll, which are posted on SurveyUSA's web site. (The RJ used to post demographics when it released polls but not with this one.)

This is reminsicent of four years ago when the RJ did polling on the cheap and had Sharron Angle defeating Senate Majority Leader Harry Reid, who won by almost 6 percentage points. Survey USA actually is highly rated, but if this poll is right, it's dumb luck.

Here's a list of questionable internal numbers:

►The overall sample is 53 percent male and 47 percent female. The reverse is likely to be closer, as it was in 2012 and most elections. Immediate red flag.

The party breakdown shows 41-35 in the Democrats' favor. While the rest of the poll seems skewed GOP, those numbers might slant it the other way. Even if the margin is close -- it's really about 5 percentage points -- GOP turnout is likely to be heavier so the electorate almost certainly won't be that favorable to the Democrats. Anyone ever hear of weighting?

Hispanics are at 15 percent of the electorate, which may be close. But the numbers are very odd, to say the least: Hutchison beating Flores by 17, which seems insane; Cegavske beating Kate Marshall by 3, which seems unlikely; and Laxalt beating Miller by 6 in that demographic, which is bizarre, too.

The regional numbers seem way off, too. Hutchison is handily beating Flores in Clark County, including by 23 in heavily Democratic Las Vegas proper -- I don't think so. On the other hand, Flores is in a dead heat with Hutchison in Washoe, which seems far-fetched. Cegavske is beating Marshall by 21 in Las Vegas, which also seems nuts. And Laxalt is winning by 6 in Las Vegas and losing by -- wait for it -- 21 in Washoe? What?

There are other strange results, too -- Flores losing by 20 among younger voters, for example. But you get the point -- there is so much that doesn't seem right here, so much contradictory information that it can't be taken seriously.

My question: Did anyone at the RJ look at these numbers and say, "Maybe we shouldn't release this."


SurveyUSA says the margin of error is 4.2 percent. But with those anomalies, I'd say it's infinite.