Hard sciences are those sciences in which precise accurate and repeatable measurements are possible. Such high quality measurements in turn allow the creation of testable hypothesis which in turn create highly predictive models of the measured phenomena. Physics is the "hardest" science because it has the easiest to measure phenomena. As you move away from physics, the measurements get harder, less precise, less accurate, they hypothesis less and less testable and the resulting models less and less predictive.
It's probably better to think of mature predictive science verses immature non-predictive science. Each scientific field starts out unable to measure, test or predict anything. It spends a lot of time simply observing and labeling things kind of making a basic inventory of phenomena. Then it starts developing measuring methods, then methods of testing hypothesis and eventually producing useful predictive models.
Astronomy was the first science because the heavens where the easist phenomena to accurately measure. Long before people understood what celestial objects were they could measure their motions and predict their paths. Physics followed soon after, and chemistry shifted from the taxonomy stage or naming stage to the measuring stage in circa 1800 and then generated testable hypothesis by they 1820s and predictive models by the 1850s. That's when chemist started really being able to synthesize things.
Biology started out in the early 1800s naming things, started measuring populations in the 1830s and by 1860 Darwin had created testable predictions. A lot of subfields like biochemistry and cellular biology, zoology etc split off. Some being more mature than others. By 1947, synthetic Darwinism could make testable predictions about populations and genetic relationships subsequently proven by DNA assay.
Psychology and Social Science are even today only at the basic observation and labeling stage. The phenomena they seek study really can't be measured at all. How does none reliably measure love, hate, anger, pride etc? Even with FMRIs, such measurements are dubious. Psychologist make few testable hypothesis and have no predictive models of human behavior.
Social Scientist are not even really though the labeling stage yet. How do you measure "Social Justice" even assuming anyone has bother to actually define the phrase. Sociologist have in the past century or more produced no testable hypothesis and have produced no predictive models.
Frankly, as a student of science history, the pretense of the Social scientist to be considered a hard science is simply revolting, considering all the damage they've done in the arrogance, Anyone remember eugenics? Say, how did those inter city housing project turn out? Or my favorite, explain to us again how you spent 50 years calling anyone who said homosexuality was innate and based in biology a Nazi fascist and then suddenly, in the early 90s just did a 180 turn about based on no new evidence whatsoever? Social scientist have never produced a single testable prediction nor a single predictive model, yet they still con us into basing social policy on the fatuous pronouncements usually using the most poor and vulnerable members of our society as their guinea pigs.
If a group of claimants cannot offer up testable hypothesis it is not a hard science. Their statements have no predictive value, and their assertion are merely personal opinion with no more weight than anyone else's.
I would thing that when someone ask a world building question under the label of "hard science" they expect an answer based on solid measurement, testable hypothesis and proven predictive models, not idil fad driven guesses of the secular version of a self-appointed priest class.
Try this out, can you think of several classic science fiction stories based on physics that still stand the test of time? I can, stories from 1920s, 30s, 40's, 50s etc years still retain the validity of their physics.
Now, how about classic science fiction stories based on the social sciences of the 1920s and 30s, you know when scientific racism and eugenics where the big rage in the social sciences? Brave New World is still a good read, but it was supposed to be a satire of many of the idea of the time, particularly the idea of social scientist that they could engineer society. Or how about stories influenced by the utterly fictional, completely made up Freudian psychology? back in the 50s, 60s and 70s various forms of Marxism, behavioralism and social engineering were all the rage. How do those stories stand the the test time? Any classics you can think off? Any stories you can read and say, wow this based social sciences fads back in the sixties still rings true even nearly 50 years later?
No you can't because their ideas are gibberish and get turned over and tossed down the memory hole every generation. Social science can't offer a writer any help at all. A writer has just as good an idea about society as as any "social scientist" because the writer has just done just as much science on the matter i.e. none.
Stories based on the social science fads du jour will likely ring hollow in their own day and certainly won't be classics remembered decades from now.
If you want the hard science tag to mean anything other than "some guy who managed to get tenure said so," then restrict it to the predictive sciences.