It is a matter of logic. The logic is that the agency of any decision cannot be identified other than by choosing what it is. That is simply 1+1=2, it is incontrovertible fact that the conclusion can only be reached by choosing it. That the agency of a decision can only be identified by choosing what it is, by expression of emotion with free will, is the proper root meaning for opinion, as categorically distinct from fact.
Only opinion in identifying agency can leave the freedom in the concept of choosing in tact. Facts use a logic of being forced, cause and effect. Evidence forces to a conclusion, resulting in a 1 to 1 model of what is evidenced. Obviously the concept of choosing requires freedom, so to propose agency as fact, is to impose the logic of being forced on the concept of choosing, which breaks down the concept. The force inherent in facts, is a logical contradiction with the freedom in choosing.
And agency is where the spirit is at, the soul, so … what are you doing circumventing it? Sociology inundated with atheists / materialists much who don’t like to hear of the soul or spirit? Tough luck for them, then they can only focus on cause and effect mechanisms in human behaviour, not on choosing.
All subjectivity has this logic, including taste and whatnot. When we look at the physics of tasting, then we will find that it operates by choosing. That is what common discourse says, it is how common discourse functions.
You just use a wrong concept of choosing, which is based on sorting out the best result. The correct definition of choosing is in terms of spontaneity. And that means there is all sorts of freedom at various levels, down to freedom in the way cells do stuff.
If the weather can turn out several different ways, then it means it is decided. This is only weird to you because you conceive of choosing as sorting out the best result. With a sorting algorithm, the result is forced by the sorting criteria, and the data to sort.
What you conceive of as choosing, is a complicated way of combining choosing with sorting. First it is chosen what is good (spontaneously), and then these chosen goods serve as sorting criteria.
So as a sociologist you might be interested in these sorting criteria that are used, values, which are highly manipulative in forming opinions. Do people have sorting criteria in which black is bad, and white is good, or something. Yes probably, children generally are scared of the dark, so that provides a sorting criteria for black as bad, which can then be used for black skin color as bad, or something like that.