Public Policy Polling (PPP) sparked controversy Wednesday after the left-leaning firm declined to release a survey it conducted last weekend that accurately forecasted the successful recall of a Democratic state senator from Colorado.

{mosads}The survey PPP conducted, but did not release, showed Colorado District 3 Sen. Angela Giron (D) would be recalled by a 54 percent to 42 percent margin.

“In a district that Barack Obama won by almost 20 points I figured there was no way that could be right and made a rare decision not to release the poll,” Director Tom Jensen wrote in a post on the firm’s website. “It turns out we should have had more faith in our numbers because she was indeed recalled by 12 points.”

Nate Silver, whose FiveThirtyEight blog at TheNew York Times accurately predicted every state in the 2012 presidential election, criticized the firm over Twitter.

VERY bad and unscientific practice for @ppppolls to suppress a polling result they didn’t believe/didn’t like. http://t.co/tOp5PhUbnf

— Nate Silver (@fivethirtyeight) September 11, 2013

PPP fired back, saying Silver has in the past admitted to withholding data that he wasn’t comfortable with.

@fivethirtyeight @ppppolls Nate U said on BS report (5:42) U didn’t post new model on elec day b/c of same fears.  http://t.co/994WaJ8yzc

— Josh (@bosdcla14) September 11, 2013

The polling outlet followed that up with a string of tweets saying it was just trying to add some context to a local race where no other polling existed.

@ZacMcCrary @DCCyclone In a race where there was no other polling out there we were just trying to give some context into the results

— PublicPolicyPolling (@ppppolls) September 11, 2013

The National Rifle Association backed the recall effort in Colorado, which targeted two Democrats for supporting gun control measures the group said weakened Second Amendment rights. 

PPP says the NRA successfully made the recall election about gun rights in general, rather than the specific laws Giron helped pass.

“If voters made their decision based on the actual pretty unobtrusive laws that Giron helped get passed, she likely would have survived,” the firm wrote. “But the NRA won the messaging game and turned it into something bigger than it was- even if that wasn’t true- and Giron paid the price.”

The firm argued that one of the reasons it released the data after the election was to show the disconnect between the election results and Coloradoans’ general support for Giron’s legislation.

@MysteryPollster didn’t release to show we had race right, released to show disconnect btwn recall result, positions on specific legislation

— PublicPolicyPolling (@ppppolls) September 11, 2013

PPP was one the most accurate polling outlets for the 2012 presidential election, forecasting Obama to win a 2-point victory while hitting the critical 50 percent mark. Obama won by 3.9 percentage points, 51.1 to 47.2.