The regular release of a potentially market-moving economic survey two seconds early to deep-pocketed traders has triggered an investigation — and a flurry of questions.

This week, New York Attorney General Eric T. Schneiderman announced that the privately generated survey will not be made available early to a select group of traders while his office conducts a broader investigation into whether such practices are fair.

By contrast, the Securities and Exchange Commission has not publicly voiced concern about these types of early releases, which do not violate insider trading laws or the many disclosure rules that govern the market.

The stark difference between the reactions reflects fundamental confusion about how to cope with the ultra-high-speed trading that now dominates the U.S. market, with participants raking in billions of dollars in the past few years by using sophisticated technology to execute transactions at blink-of-an-eye speeds.

The deal Schneiderman unveiled involves a consumer confidence survey compiled by the University of Michigan and distributed exclusively by Thomson Reuters. Since late 2008, Reuters has charged a premium — an extra $6,000 — to clients who want the market-moving survey results two seconds before other clients.

“What you’re seeing unfold is a backlash against high-frequency traders,” said Charles Jones, a professor at Columbia University’s business school. “Regulators are trying to figure out how to police a two-tiered system between the very fastest traders and everyone else.”

The information at issue is not government data, which is owned by taxpayers and therefore is not fed early to a select group. But it is one of the most significant privately generated pieces of information for the market.

The survey is released to the public at 10 a.m. every other Friday. But some clients pay to get it five minutes ahead of the public, and others pay the extra $6,000 to get it two seconds before that. Under the Schneiderman agreement, Reuters can still distribute the results to clients five minutes early, but no earlier than that.

Some experts say that the data is intellectual property that is costly to collect and that its owners should be able to sell it how­ever they wish, as long as it is available under the same terms and conditions to anybody willing to pay for it.

It’s no different than selling train tickets, said James Angel, a Georgetown University professor. Riders can pay a premium for the high-speed train or less for the slower one.

Reuters “charges through the nose for the people who want it first, then they charge less to the people who want it less fast, and finally they give it away to people who are willing to wait and read about it in the paper,” Angel said.

The advantage does not come at the expense of mom-and-pop investors, who are not in the game of responding to news within milliseconds, he said. The fiercest competition is among the high-frequency traders, to whom fractions of a second matter.

But when Reuters suspended its earliest feed, some of those traders were relieved that they no longer had to pay the extra cash to keep up with their high-speed competitors, said a trading expert who did not want to be named because Schneiderman’s investigation is ongoing.

Reuters says that it has done nothing wrong. In a statement, the company said it believes that news-gathering companies can legally distribute nongovernmental data and exclusive news to subscribers. And that’s how some ethics experts see it, too.

“I agree with the objective of an even playing field,” said Richard Painter, who served as chief White House ethics lawyer in the George W. Bush administration. “But I don’t think that the New York attorney general ought to be trying to regulate the press and telling them how to distribute information to their subscribers. . . . It’s a slippery slope.”

The SEC declined to comment on the matter. Paul Rowady, a senior analyst at TABB Group, said the agency is probably wary of the slippery slope.

“You can see that several Pandora’s boxes can open up,” Rowady said. “Do we apply rules to all private data? You can pick one or two or a subset of data items known to be catalytic in the market. But where do you stop and draw the line? ”