Pinpoint predictions are hard to come by. You can not say with any reasonable level of accuracy: “Two weeks after this Thursday, the airport temperature will be 65 degrees, the humidity 78%, the barometer 26.5 and rising.” But you can often say “In the latter part of the month, barring any unexpected developments, we can expect cooler but relatively humid days with rising pressure.”
I think sociology can eventually aspire to that level of predictive power.
At the moment, the field may not be too much further along than an analogous weatherman who could only tell you “This winter will be cooler and wetter than this past summer”, but I think that will change over time.