Internal polling by congressional candidate Clint Didier says he’s leading the race to replace Doc Hastings, but details in the documents suggest it’s anyone’s race to win.

In answer to the question, “If the election for Congress was held today, of the following candidates, which would you vote for?”

Didier was chosen by 17.5 percent of respondents, followed by 9.75 percent for former state Department of Agriculture Director Dan Newhouse and 7 percent for state Sen. Janéa Holmquist, R-Moses Lake.

Conducted by Lansing, Mich.-based Team Telcom from May 26 to 28, the poll surveyed 400 “likely voters” in the district. Didier has touted a lead in paid advertisements since mid-June without citing the specific numbers, but his campaign released the polling documents to the Yakima Herald-Republic last week.

More relevant for the shape of things to come is that 233 respondents — 58.25 percent of the sample — said they either don’t know who they’ll vote for, refused to answer or said they wouldn’t vote for any of the 12 candidates in the race. Just 70 respondents said they’d vote for Didier, 39 for Newhouse and 28 for Holmquist.

Furthermore, Didier’s ads say the poll had a margin of error of 2.5 percent, which Jay Leve, editor of SurveyUSA polling, based in New Jersey, said is “impossible” with a sample size of 400.

Leve said it would take a sample of 1,500 likely voters in the district to have a margin of error of 2.5 percent. The margin of error on Didier’s poll is likely 5 percent or greater, Leve said.

A margin of error that size creates scenarios where Newhouse could be leading Didier based on the polling data and Holmquist not much further behind. But Leve said the most important statistic is the majority who didn’t pick a candidate.

“It suggests no one is paying attention yet to this contest,” he said. “It would be premature to draw any conclusions.”

Stuart Elway, president of Elway Research in Seattle, agreed.

“It means a lot for Didier because he gets to say he’s ahead, but in terms of indicating how this race is going, it’s not very good at that,” Elway said.

In an interview Friday, Didier’s lead political consultant, Doug Simpson, said he’s not concerned about a majority of the sample not choosing a candidate. He said he interprets the percentages to show a majority of Republicans have made up their minds and it’s mostly Democrats and independents who are undecided.

Simpson did not return numerous follow-up calls to explain how the campaign determined the poll’s 2.5 percent margin of error.

The 30 remaining votes by respondents were split among seven of the nine other candidates: nine for Democrat Tony Sandoval, five for Republican George Cicotte, five for Democrat Estakio Beltran, four for Independent Josh Ramirez, three for Independent Richard Wright, two for Republican Gavin Seim and two for Republican Glen Stockwell.

Republicans Gordon Allen Pross and Kevin Midbust didn’t receive any votes in the poll.

Arguments about the validity of the poll aside, Didier’s advertisements touting its results aren’t entirely accurate. The ads say Didier leads by a “more than a 2 to 1 margin.” It’s true that respondents favored Didier by a ratio of more than 2 to 1 over Holmquist. But the poll shows he falls shy of a 2-1 ratio against Newhouse.

Simpson, who wrote the ad, said he was trying to convey that Didier got more votes than Newhouse and Holmquist combined.

“I’m not going to argue about” the discrepancy, Simpson said. “It’s just a technical slip of the tongue.”

Simpson said he will change the language describing Didier’s lead according to the poll in future ads.

Almost 74 percent of respondents were age 60 or older. A little more than 42 percent either tended to vote Republican or were “mostly Republican,” with 25 percent saying they are independents and almost 24 percent saying they tend to vote or mostly vote Democratic.

A slight majority of the respondents, 52.5 percent, were women, and 47.5 percent were men.

The respondents were contacted on landline telephones and cellphones.

The Top Two primary is Aug. 5. Ballots will be mailed out no later than July 18.