WHEN PRESIDENT Reagan told a group of school children last week that unemployment fell in March, he introduced a new quirk into the already complicated business of measuring unemployment. Apparently he has decided to seasonally adjust his choice of statistics so as to put the best possible face on them.

Up until now the administration has been content to cite the "seasonally adjusted" unemployment measures normally used in gauging the health of the economy. The reason that the adjusted figures are usually cited is that they are a much better indicator of how the economy is doing. Even in good times, unemployment tends to go up and down over the year because of seasonal factors. For example, Christmas shopping gives a boost to retail employment, cold weather hurts construction, agriculture and other out-of-doors work, and summer sends hordes of students into the job market.

To give a clearer picture of how the economy is doing, the Bureau of Labor Statistics adjusts each month's figures to take account of these normal fluctuations--the "funny way of counting" to which the president referred. BLS publishes both adjusted and unadjusted data so the user can make comparisons. In recent months, the adjusted rates have been lower than the unadjusted rates because unemployment tends to rise when the weather is bad even in good times. For example, in February the actual number of people unemployed was 10.4 million and the unadjusted rate was 9.6 percent. The seasonally adjusted figures showed only 9.6 million unemployed and a rate of 8.8 percent.

Now that spring is more or less here, employment is beginning to pick up a bit and the unadjusted numbers are beginning to look better. The unadjusted unemployment rate is still a half point higher than the adjusted rate, but the raw data showed a drop in the number of unemployed people of 88,000 --the number that the president singled out to make his point to the children. The reason that the seasonally adjusted rate went up instead of down is that joblessness usually drops much more than that in March--and that's a good sign that the economic situation has deteriorated.

We expect that the president will soon want to make another seasonal adjustment in his choice of statistics. In May and June the schools will close, and the flood of students seeking jobs should knock the unadjusted unemployment rate up by more than half a percentage point. Will he love the unadjusted data in June as he did in March?