Is there a logical fallacy for differing abstractive levels. As in, you start talking on the level of demographics until someone counters with verifiable data on the same level of abstraction, at which point the original voice drops down to anecdote.
There are multiple options depending on how the anecdote is presented within the context of the argument.
To take an example I recently came across, many autistic people like to be referred to like that instead of as "person with autism." If I point out a survey of 1k autistic adults showed a strong preference, and you reply "well, my sister has autism and she would hate to be called that" then we can either take this to be a false generalization or moving the goal posts depending on how the words are meant. There's a related case where it's a straw man.
If the intention is "my sister has preference x and therefore it's correct for everyone" then it's a false generalization, because you're portraying your sister's preference as generalizable without justification.
If it's changing the topic of the argument from "what do autistic people prefer" to "what does this particular person prefer" it's moving the goal posts.
In both these cases it's also a non sequitur. Any argument where the conclusion doesn't follow from the premises is a non sequitur
The strawman occurs if the point is that your survey doesn't apply to your sister, and therefore the data is wrong, because you're misrepresenting my position as "all autistic people..."