God help the poor news consumers of America, especially the would-be voters.

President Obama’s decision to reject the Keystone XL pipeline last week incited a new wave of coverage and speculation about how many jobs the line would create. Unfortunately, many outlets are still citing inflated and unreliable industry figures in the tens to hundreds of thousands while ignoring more modest and trustworthy approximations from academia and government, which place the total anywhere from 2,500 to 6,000.

The media were obsessed with the jobs number in 2011 and developed a preference for 20,000. A wide variety of conservative politicians and industry groups, from House Speaker John Boehner to the American Petroleum institute, have cited that figure, and some reporters have mistakenly attributed it to a variety of research firms that have issued reports on Keystone XL. But make no mistake, the number comes directly from TransCanada, the company that wants to build the 1,700-mile pipeline, which would transport crude oil from Canadian tar sands to Gulf Coast refineries in the United States. TransCanada first mentioned it, with no explanation, in a February 2 press release, but didn’t explain its math for another eight months.

Here’s how the company arrived at the figure:

Construction of the 1,600 mile pipeline is broken down into 17 U.S. pipeline spreads or segments, with 500 workers per spread—that’s 8,500 jobs Keystone XL also needs 30 pump stations worth tens of millions of dollars. Each station requires 100 workers—that’s 3,000 jobs. Add another 600 jobs that would be needed for the six construction camps and tank construction at Cushing, Oklahoma. A project of such magnitude needs construction, management and inspection oversight—that would create 1,000 jobs, bringing the overall Keystone XL total to 13,000 direct, on-site jobs.

Now, compare the math in the company’s press release to the information that it provided to the US State Department for its final Environmental Impact Statement (EIS), delivered in August. With regard to the seventeen “spreads,” it read:

Approximately 500 to 600 construction and inspection personnel would work on each spread, except for the proposed Houston Lateral which would require approximately 250 workers. Each spread would require 6 to 9 months to complete. Construction of new pump stations would require 20 to 30 additional workers at each site. Construction of all pump stations would be completed in 18 to 24 months. Tank farm construction would require approximately 30 to 40 construction personnel over a period of 15 to 18 months.

Notice that in its press release, TransCanada omitted the durations of employment and inflated the number of pump-station and tank-farm jobs in order to arrive at 13,000 construction jobs. To that, it arbitrarily added 7,000 manufacturing supply jobs in order to get to 20,000 jobs. Most reporters published only that number despite the fact that, based on the information provided by TransCanada, the State Department’s EIS said Keystone XL “would result in hiring approximately 5,000 to 6,000 workers over the three-year construction period.”

In September, researchers at Cornell University’s Global Labor Institute used the information in the EIS to come up with an estimate that was even more modest. Factoring in the various durations of employment, it calculated that “on-site construction and inspection creates only 5,060-9,250 person-years of employment (1 person-year = 1 person working full time for 1 year). This is equivalent to 2,500-4,650 jobs per year over two years.”

One of the researchers told InsideClimateNews that the difference between the Cornell and State Department estimates is attributable to the fact that the State Department includes a number of workers that TransCanada has already hired, while the Cornell study addressed only new jobs from pipeline construction. Whatever the case, the State Department and Cornell figures are clearly more reliable than those from TransCanada, which has a history of toying with numbers.

After the 20,000 “direct jobs,” the second most popular tally cited by the media has been 118,000 “spin-off jobs,” and it, too, has a long, convoluted history. It comes from a June 2010 report from The Perryman Group, a financial analysis firm based in Texas that was hired by TransCanada to evaluate Keystone XL. That report estimated, in a very non-transparent way, that the pipeline would create 118,935 person-years of employment over the hundred-year life of the project.

TransCanada first cited the figure in a September 2010 press release (which failed to mention the century-long time frame). But two months later, it issued another press release that dispensed with person-years and mistakenly referred to “118,000 spin-off jobs” instead. The error was repeated in dozens of media reports.

Curtis Brainard writes on science and environment reporting. Follow him on Twitter @cbrainard.