The Los Angeles Times roared out of the gate this morning by taking on the politics of fast food. At 5 a.m., it posted a provocative headline: “Nearly 90% of fast-food workers allege wage theft, survey finds.”

It may not get much more attention, at least from big media firms. The survey grows from the curiosities of “Low Pay Is Not OK,” a group that’s fighting for a fast-food minimum wage of $15 an hour. And the numbers themselves come from Hart Research Associates, a public opinion firm that has done polling with NBC News and the Wall Street Journal.

To compile the results, Hart conducted an online survey of 1,088 respondents via Facebook “who work in fast food restaurants in the top 10 metro areas across the country.” Recruitment got a little help from inducements, as the Facebook ad above notes: “FAST FOOD EMPLOYEE? 2 MINUTES, AND YOU COULD SCORE $100 BUCKS.” Guy Molyneux, a partner at Hart Research Associates, told the Erik Wemple Blog, “They had to also confirm to us that they work in a fast-food restaurant.”

Helpful, but not good enough for outlets that have promulgated exacting standards as to what survey results are worth passing along to the public. The Associated Press (AP) has placed its clearance bar in its famous stylebook. Here’s the wire service’s filter for opinion surveys: “Only a poll based on a scientific, random sample of a population — in which every member of the population has a known probability of inclusion — can be considered a valid and reliable measure of that population’s opinions.” Examples of research that doesn’t meet this criterion include various online polls in which respondents are “self-selected,” often “including ‘professional respondents’ who sign up for numerous surveys to earn money or win prizes.”

Hart acknowledges that an online survey isn’t the “gold standard” of opinion research, to use the term cited by Molyneux in an interview with the Erik Wemple Blog. The best approach, says Molyneux, is a random sample of adults who would be contacted and asked if they worked in the fast-food industry. Fewer than 1 percent of the U.S. population, Molyneux estimates, would answer “yes” to the fast-food-employment question, and so it’d be “prohibitively expensive” to snare fast-food workers in a random sample polling methodology. The Erik Wemple Blog asked Low Pay Is Not OK how much it paid for the survey; it didn’t share.

Hart found that the respondents that it secured through online recruitment checked out: That is, the numbers of people who reported working at McDonalds, Burger King, etc., roughly reflected the proportion of franchisees for those chains, and reported information about wages and the like also tracked with existing data. “We feel very good that we got a broad sample of fast-food workers that would be a random sample of adults,” Molyneux said. Of the survey’s key figure — that 89 percent of respondents reported some kind of wage theft — Molyneux says, “Could the real number be 83 or 84 percent? Yes. Do we think it could be 53 percent? No.”

In its initial writeup of the poll, the Los Angeles Times cited the sponsorship of Low Pay Is Not OK along with the survey’s methodology and its geographic base (Los Angeles, New York, Chicago, Dallas, Houston, Philadelphia, Miami, Atlanta, Boston and Washington, D.C.). What it didn’t do was acknowledge the difficulties of generalizing the results to a national population. Have a look at the story’s first sentence:

The vast majority of fast-food workers in the U.S. say they’ve been the victims of wage theft, according to a survey released Tuesday.

A more accurate formulation:

The vast majority of fast-food workers in a non-random-sample web survey that included gift-card inducements in ten major metro areas say they’ve been the victims of wage theft.

Nor did the story, as originally posted, contain any comment from the fast-food industry. Nancy Sullivan, a spokesperson for the Los Angeles Times, rebuts as follows:

The survey in question was embargoed until this a.m., and our story published online once that restriction was lifted (and long, long after our print edition had been produced for the day). The Times followed typical journalism practices and ethics, including clearly stating who conducted the survey (Hart Research for the Low Pay Is Not OK campaign) and that it was part of an advocacy campaign (“The campaign is part of an effort to raise minimum wage for fast-food workers to $15 an hour and to push companies to let their workers unionize without retaliation.”). As you are well aware, endeavoring to get comment before the embargo would have violated its terms. We will be updating the post will additional comments and have yet to determine if the story will be part of tomorrow’s print edition.

Bold text highlighted to provide a disclaimer: Whether the Erik Wemple Blog is aware of such restrictions on an embargo or not, that’s not good journalism.