The Census overcounted residents in Boston by 14.3 percent, the highest error rate in the nation. But the overcount was nearly matched by the number of people the Census missed in its 2010 count, according to a new report released by the US Census Bureau.

Taken together, the figures mean that the Census over-counted by 1.27 percent — a statistically insignificant figure given the Census survey’s margin of error of 3.8 percentage points.

Advocates said the errors highlighted the potential for problems in Census data gathering — errors that could reduce the city’s funding sources, which are often based on Census information.

“I’d be interested in the demographics of where there was undercounting and where there was overcounting,” said Cheryl Clyburn Crawford, codirector of MassVOTE. “Did the undercount take place among new citizens who are typically hard to count? Or did we overcount where college students live?”

The Census made the analysis based on surveys conducted in randomly selected neighborhoods several months after the Census collection had wrapped up. The survey results were then compared against Census data.

Mark Melnik, deputy director for research at the Boston Redevelopment Authority, said the survey sample size was so small — approximately 350 Boston households — that the figures could not be considered significant.

In fact, he said, the survey wasn’t meant to be interpreted at the local level. The local counts were merely taken to help spot check the accuracy of the national level data, which was more accurate this year than in decades past.

He noted that the local survey won’t influence population estimates issued before the 2020 Census.

“At the end of the day, if it’s not statistically significant, then it doesn’t call anything into question,” he said.

Patrick Cantwell, an assistant division chief at the US Census Bureau, said the bureau had not pinpointed a reason for Boston’s unusual number of overcounts, in part because the survey is an estimate.

But he said such errors often occur in places with large numbers of college students, who are counted once at their college residence and again in their hometown. Undercounts tend to occur among transient populations, such as renters and young people.

Cantwell did not know whether Boston historically has had a higher rate of Census errors than other cities.

The errors found in the survey will not be used to correct 2010 Census results, but rather to improve 2020 Census gathering methods, he said.

“We are looking at their error structure to see what we can do to improve,” Cantwell said.