EDMONTON - When the smoke cleared Monday night, the Alberta provincial election results stunned pundits, pollsters and party faithful alike.

The underdog, scandal-beleaguered PC Premier Alison Redford -- came from way behind to win a majority over the upstart Wildrose Party, who had already been all but sworn in by a bevy of polls from different sources. 

How could so many polls be so wrong?

According to Chris Baker, president of Continuum Research, a Canadian company that has done polling for all major Canadian political parties, modern technology doesn't do some political polls any favours. 

Polls that are slammed together with robo-calling and interactive voice recognition, 1,000-calls-today-and-results-tomorrow for $999.99 -- aren't always that accurate. 

"It's bad research, but it's dirt cheap," he said. 

The dirty little secret of the public opinion research industry? 

"Good, fast, cheap -- pick two," Baker said. 

For parties or businesses with good, accurate public opinion research, they can hold a definite advantage over competitors. 

"Bad information is worse than no information," Baker said. 

Who answers the phone? Where do you start making calls? When do you stop? How random is it? Do you test the honesty of the respondent? Do you determine the likelihood that they will actually go to the polls and cast a real ballot? 

"If you assume that everyone who answers the poll is going out to vote, you're making a huge mistake," he said. 

And it's as easy as poor question design, sampling bias, and lack of data verification, Baker said. The difference could be the difference between a political scientist and a pollster. 

"Frankly, the public opinion industry has got to come to grips with this because it's killing our industry." 

Intense riding surveys, laser-focused, that can give a party accurate, on-the-ground information that can inform their campaign decisions, can run $10,000 -- a figure out of reach in many ridings. 

Top leadership candidates who are trailing are often heard to say "the only poll that counts is election day." But in the case of parties with full campaign coffers, don't bet on it. 

In Alberta 2012, the Progressive Conservatives likely had laser-sharp, top-notch internal polling more finely tuned than anything rolled out by news organizations, Baker said, citing tools such as focus groups and surveys, that can give feedback on everything from what a candidate wears on the campaign trail to whether it's time to switch from talking about former premier Ralph Klein to former premier Peter Lougheed. 

"They were probably fully invested in getting a victory for their client," Baker said. 

And while candidates may talk about "listening to Albertans" -- a phrase that can sometimes be a euphemism for watching the polls for direction -- polling cards are typically played close to the vest. 

"Any candidate that says they're driven by polling, doesn't get it," Baker said. "You don't say, 'My polling advisor told me this.'" 

Going forward, newly elected Redford is in the political catbird seat, Baker said. 

"She's a stronger premier than she was two days ago -- the election did her some real favours," he said. "It allowed her to clean house, it demonstrated she can win and pull the Tories from the jaws of defeat ... it revitalized the reputation of her staff as well as her own reputation." 

Redford is getting a lot of credit from the Tory rank and file for focus and discipline and getting the job done, Baker said.  

"She has renewed her star ... the morale is going to be tremendous," Baker said.