When the state Department of Labor recently released adjusted data that showed Connecticut added 12,200 jobs in 2015 — less than half of the 26,900 initially reported — some residents were quick to question and criticize state officials for being misleading.

But there is a reason for the seemingly large discrepancy, according to the Labor Department: monthly job figures and the year-end totals are tallied using two different sets of data.

The Labor Department reported nonfarm employment data each month last year, as it does every year. To do so, it relies on information culled from a partnership between the state Department of Labor and the U.S. Bureau of Labor Statistics, according to Patrick Flaherty, assistant director of research and information at the state DOL’s Office of Research and Information.

Every month, BLS surveys a certain number of Connecticut employers — about 4,000 of the 116,000 worksites in the state — and asks them to estimate how many people they employ, along with what those workers earn for wages, Flaherty said.

The Labor Department then uses those estimates to gauge the change in statewide employment from the previous month, he said. As with any survey, he added, the data can be skewed depending on what companies are surveyed.

When polling a sample of the larger workforce, the employers chosen could “miss,” meaning they don’t represent broader statewide employment trends, Flaherty said.

“We think that may be what has happened,” he said. “The survey just missed this year. Those 4,000 employers were not representative of the state as a whole.”

Those surveyed, for instance, may have had particularly — and atypically — strong job growth that bolstered the monthly figures.

Year-end job totals, on the other hand, are devised from a different data set pulled from the Quarterly Census of Employees and Wages. In that, nearly all employers in the state report the number of employees they have, and what they earn, to the state and federal labor departments, according to Flaherty.

That data, he said, is much more comprehensive than the monthly survey data. The census includes figures from all employers who participate in the unemployment insurance system, or about 95 percent of payroll employers in Connecticut, he said. It does not include self-employed individuals or independent contractors.

Employers are required to report the data quarterly but, by the time they submit the data and Labor Department officials tabulate it, there is lag time in reporting it to the public, Flaherty said. That’s why the 2015 total was not released until this month.

“There’s no mystery to that; it takes time,” he said. “We don’t know the accurate count of exactly how many people were working in 2015 until into the following year. We don’t get an accurate, hard count, until some months later.”

State labor officials previously had more control over collecting that data, which is now collected by the U.S. Department of Labor, Flaherty said. The state was more accurate when it was in charge of the data, he said, since it is more familiar with Connecticut employers and job trends here than the federal department.

The state Labor Department strives to be as accurate as possible, he said. He added that, while the 2015 job growth discrepancy may seem large, “in the overall scheme of the 1.7 million people employed (statewide), we’re not off by all that much.”

Gov. Dannel P. Malloy downplayed the job growth reduction at a media briefing following last week’s State Bond Commission meeting, blaming the discrepancy on the federal labor department.

“Every time those numbers go down, (the news media) make a big deal out of it,” Malloy said, noting the state can only report figures based on the federally culled data. “The best measurement (of employment data) is by year, the second-best measurement is by quarter, the least best estimate is by month,” he said.