The financial stress on American farms, already documented by government economists, is being reflected in new census data showing a decline last year in the nation's farm population.

Although there is disagreement over their accuracy, figures just released by the Census Bureau show that the country's farm population dropped last year by about 7 percent, from 5.7 million rural residents in 1984 to 5.3 million. The 1985 numbers were a footnote to a report on the 1984 farm population issued jointly last week by the Census Bureau and the Agriculture Department.

By comparison, the national farm population during the 1970s showed an average loss of about 3 percent.

Diane DeAre, a Census Bureau analyst, said that she stood behind the 1985 numbers as a reflection of the growing economic problems in agriculture.

"We expected to see some impact of the farm crisis on population . . . but not necessarily a change of this size," DeAre said.

But Calvin L. Beale, a demographer who oversees population studies at the USDA, said he was skeptical of the sharp changes reported in 1985 by the Census Bureau. "There is a decline that reflects the farm crisis, but I don't think the actual change was 7 percent," Beale said.

Beale said he thought the differences could be attributed to changes in the bureau's sampling methods. He said that the bureau's 1984 figures may have been statistical aberrations, and that most samples are not accurate enough to permit a fair year-by-year comparison.

"The basic question is: What's going on in the farm economy?" Beale said. "The [population] numbers reflect it . . . .With these economic conditions, we will have smaller numbers than these. And there is gradually an increase in the number of people who are hired workers rather than owners."

Where do displaced farm residents go?

"Not all of them go anywhere," DeAre said. "Some may remain on their farms but are not counted if annual sales drop below $1,000 [the bureau's standard for defining a farm]. Others die; a widow may sell the farm property. In the 1970s, there were fairly large declines in farm population, but the amount of land in farms did not decline . . . . There was a lot of 'people loss' not related to farms going out of production."

Other key points in the report for 1984:

*Median farm-family income was $18,925 in 1983, one-fourth less than that of nonfarm families. Farm residents had a lower unemployment rate than nonfarm workers (3 percent, compared to 7.7), but farm residents often hold more than one job. Farm work counts as employment even if an off-farm job is lost.

*Approximately 3 percent of farm residents were black, with 99 percent of them living in the South. Three of five farm residents in the West were described as Hispanics. Even though the Census Bureau reported that black farm population had dropped from 242,000 to 148,000 between 1980 and 1984, both Beale and DeAre said the survey samples were possibly misleading, and thus the figures were not reliable.

*And for the singles crowd, n.b.: More men than women can be found on farms -- 107 males for every 100 women, compared with 93 men for every 100 women off the farm. Eligible men -- the unmarried or widowed -- are concentrated in the 20- to 24-year-old class and the 65-or-older group.