Here’s a quick quiz: Do more people moonlight today than in the 1990s? Are workers without a high-school degree more likely to work a second job than college graduates? And do immigrants moonlight more often than native- born Americans?

I’d bet most people would answer yes to all three questions. If so, they would be wrong on all three — at least according to the official data. The surprising reality of second jobs in America underscores the importance of using real data to examine the labor market. Casual guesses can be misleading.

As part of its monthly survey of households, the Bureau of Labor Statistics at the Department of Labor collects information about workers who hold more than one job. In January 2012, about 5 percent of workers indicated they worked at least two jobs. That share is lower than it was during the 1990s. In January 1997, for example, more than 6 percent of workers had multiple jobs.

Why the decline? No one seems to know for sure. One possibility, noted by Larry Katz of Harvard University, is that it is a statistical quirk: Workers are now more likely to work for temp agencies than during the 1990s, and may report only one “job” even though they work at multiple locations through the agency. But the rise in temp work is unlikely to be large enough to explain the change in the moonlighting rate.

Another possibility is that, in the aftermath of the financial crisis, workers lack the opportunity to work a second job. Some empirical studies have suggested that moonlighting is pro-cyclical: It rises in a strong economy and declines in a weak one. Recently, as Catherine Rampell, an economics reporter for the New York Times, has observed, the states with the highest rate of multiple-job holding have been the ones with the lowest unemployment rates. If that inverse relationship between moonlighting and joblessness is universal, then the overall rate may be depressed by the high number of people out of work nationally.

The problem with the Great Recession explanation, though, is that the rate of multiple-job holding started to decline in the late 1990s, and over the past five years it has been relatively flat.

The percentage of people holding multiple jobs has held steady for the past few months. That means it can’t answer another riddle: Why, in recent months, has the increase in the number of employed people differed so much according to whom you ask? The figure reported in the Bureau of Labor Statistics monthly household survey has been growing more rapidly than the number reported to the bureau by companies. One reason the two approaches can differ is that some workers have more than one job (so businesses count them more than once).

Over the past year, the increase in employment in the household survey has been 8 percent higher than that in the survey of companies. In January, the difference was larger still: Households reported 631,000 added jobs, while businesses reported only 243,000. (The lower number was better publicized.) Yet, in that month, the share of people holding multiple jobs did not change, so it can’ t explain the difference.

Even accounting for all the known differences between the surveys — for example, the household one counts self-employed people but the business one doesn’t — still leaves job growth in the household survey at 491,000, more than double that in the other one. This remaining difference suggests that the poll of businesses may be undercounting job growth, which would be good news for the lab or market. One reason is that new companies are likely to be under-represented in the business survey as the economy begins to recover.

The data on moonlighting contain some other surprises, as well. In 2011, for example, less than 3 percent of workers without a high-school degree worked multiple jobs, according to an analysis by Steven Hipple of the Bureau of Labor Statistics. That compared with almost 4 percent of workers who did have a high-school degree, almost 6 percent of those who had a bachelor’s degree and more than 7 percent of those with a doctorate.

Perhaps less surprising are the reasons people give for working a second or third job: Among those who hadn’t finished high school, 80 percent said “to meet expenses or pay off debt” or “to earn extra money.” Among those with a doctorate, by contrast, less than 25 percent offered those reasons. Almost a third said “enjoys the second job.”

Hipple also shows that only 3.1 percent of immigrants worked a second job, compared with 5.3 percent of native-born Americans. Again, that is not the pattern that might have been expected (although part of the difference may reflect second jobs in the gray or informal economy that may not be reported in surveys).

At least some of the data conform to common perceptions, however. About a quarter of male firefighters work a second job, along with a fifth of male paramedics and more than a tenth of female dental hygienists.

But the biggest take-away from the data on multiple-job holding is that anecdotes and perceptions can be misleading. So it is important that we have evidence to check them. The Obama administration’s budget for fiscal year 2013 includes $618 million for the Bureau of Labor Statistics, a modest increase of $9 million over the current level. Part of the extra money would finance a supplement to the household survey, which would, among other things, ask questions about why people hold more than one job. This would be money very well spent.

Peter Orszag is vice chairman of global banking at Citigroup and a former director of the Office of Management and Budget in the Obama administration.