The primary gist of what you wrote is important for people to grasp. Allow me to expand on it a bit, because thoughtless attitudes about "data" are pervasive, perhaps especially among the SV crackpots.

"Data" comes from datum [0], that is, what is given. What are the data or givens of measurement?

Whenever we measure something, we do so from the standpoint of some prior conceptualization. It makes no sense to speak of measurement apart from some conceptual context, as the measurement is of something as it is understood. It is through this conceptual background that we can situate some thing as a measurement, as data, and understand the meaning of this measurement, infer implications, and so on. Some call this the theory-ladenness of observation.

So you cannot say "Data! QED.", first, the meaning of the given is inaccessible without knowledge of its nature and the prior knowledge that allows us to locate the data in the appropriate context, and second, because data are not arguments. Data are used in arguments.

So if your conceptual context is flawed, your measurements are vulnerable, both in their motivating rationale and in their interpretation. A little error in the beginning leads to a great one in the end. And there's a lot of crap people carry around in their conceptual baggage.

So, we have at least three attack surfaces: the conceptual presuppositions of a theory, the theory, and the data sought to corroborate the theory.

Of course, theory-ladenness does not necessarily entail relativism [1]. So, the point isn't that we can't know anything, so anything goes, or that we don't know anything, so burn it all down. The argument is that we should be more cognizant of the bases of our justifications.

[0] https://www.etymonline.com/word/data

[1] https://edwardfeser.blogspot.com/2025/08/hanson-on-observati...