IT has been a banner year for scientists trying to date the age of the universe. They still don't know exactly where it falls within a generally accepted range of 10 billion to 20 billion years. Yet new methods and better data have given them what they consider an encouraging sense of progress."I suppose that's true" that there's been progress this year, says cosmologist Mark Birkinshaw of the Harvard-Smithsonian Center for Astrophysics at Harvard University. He adds, "The numbers [used in the computations] are considerably more solid now." Massachusetts Institute of Technology astrophysicist Bernard Burke agrees. He says: "It [now] seems to me our general picture of the universe really isn't far out of kilter." He has joined MIT colleagues Joseph Lehir and Jacqueline Hewitt, and David Roberts of Brandeis University, in research that they described in the magazine Nature last July as giving them "some confidence that the scale [age and size] of the universe is indeed known to about a factor of two." That's saying a lot for a science in which untested assumptions about the nature and behavior of cosmic objects underlie all methods of age estimation and contribute as much to scientists' conclusions as do observational data. And it's in those assumptions that the main challenge to pinning down the age of the universe now lies, according to Dr. Birkinshaw. He sees a warning in the fact that different age-determination methods are not converging toward a common estimate, even though they now use more precise data. He suspects "this may be telling us there are things we don't understand." The trick in estimating the universe's age is first to gauge the rate at which the universe is expanding after the "big bang" primordial explosion from which cosmologists believe the cosmos emerged. Because of this expansion, galaxies and other objects move away from Earth with speeds proportional to their distance. The farther away they are, the faster they go. And the faster they go, the redder their light appears. Astronomers measure an object's recessional speed by observing this "red shift." The rel ation between an object's velocity and its distance is determined by a constant named for the late Edwin Hubble, who discovered the cosmic expansion. It, in turn, is related to the age of the universe. So once astronomers have a numerical value for this constant, they can use it to estimate that age. Estimates of the Hubble constant have ranged roughly between 50 and 100. Values near 50 imply ages around 20 billion years. A value of 100 gives a younger age of around 10 billion years. Astronomers are uncomfortable with such a young estimate because it's younger than some star formations with ages around 15 billion years. Hubble constant estimates have had large uncertainties. This year, several values have been published that individually are more precise but still do not agree with each other. George Jacoby at Kitt Peak National Observatory near Tucson, Ariz., and John Tonry of MIT came up with a value range of 76 to 86, implying a relatively young universe. They used modern techniques to identify objects of known intrinsic brightness in the Virgo cluster of galaxies and made a new, more precise, estimate of the cluster's distance by observing how dim the objects appear as seen from Earth. Using a different method, Birkinshaw and his Harvard colleague Jack Hughes came up with an estimate of around 57, favoring an older universe. Burke and his colleagues, using still another method, came up with Hubble constant estimates from as low as 46 to a high of 69. The point, Burke says, is not that different methods give different estimates but that these estimates all fall within a range of roughly 50 to 100. Nevertheless, Birkinshaw notes, "we will only really begin to understand" what is going on when the various methods of estimating the Hubble constant - and hence the universe's age - are refined to the point that their conclusions start to agree. Only then will astrophysicists be able to have full confidence in the assumptions that underlie those methods.