YOU WOULD BE well advised to look with a jaundiced eye on any projections for the future that assume a certain number of babies will be born. Nothing is more mysterious or harder to predict than a society's birthrate. Almost no one anticipated the high birthrates of the 1947-62 baby boom --rates that were, in fact, much higher than those experienced by any other industrialized country in this century. But by the early 1960s, almost everyone had come to assume that the baby boom was the norm.

People were surprised again when birthrates fell sharply in the years after the early 1960s. The failure to anticipate these changes caused all manner of problems, from a shortage of school classrooms in the 1960s to the Social Security revenue shortfalls that now seem likely to occur in the next century.

Now the birth statistics may be about to confound the experts again. The fertility rate--the number of live births per 1,000 women age 15 to 44--has remained close to the same low figure for the last 10 years. But the birthrates for women of different ages have changed in different directions over the past three years.

Women under 25 are having fewer babies. That's part of a long-term trend; it was a sharp decline in births among women under 25 that ended the postwar baby boom in the early 1960s. But the birthrate among women over 30 increased significantly between 1980 and 1982. Women over 30 who gave birth in the early 1980s tended to have above-average levels of education and income. The widely publicized increase in births among the trendy set may be the harbinger of a much larger trend.

The higher fertility rate among women over 30 has been enough to hold the overall birthrate steady, and may do so for most of the 1980s. If this happens, those who foresaw fewer births, on the theory that baby boom women would have passed peak childbirth years in the middle 1980s, would turn out to be wrong: there won't be a decline in the number of women in their thirties until some time in the 1990s. The overall result would be somewhat more Americans than most demographers have been expecting--and, again, a set of new unanticipated problems and benefits.