I do object to the broad brush that human studies are ALL garbage science. Anything that relies on humans responding to questionnaires is inherently not "science" but there are ways to conduct those studies using considerable more subjects and individuals without obvious agendas. I've seen several. I've also seen the garbage kind.

With enough participants, well written questions that aren't subject to preconceived beliefs, and critical reviews of the studies' conclusions, it is entirely possible to end up with reasonable, decently falsifiable conclusions.

Not as solid as firm data without subjectivity, but reasonable conclusions that anyone else using the same standards could replicate.