A survey with 70 recipients has a 95% confidence level with a 12% margin-of-error and can be generalized to effectively any population size. (using this calculator). - disclaimer: I'm not sure if the above is only relevant for 2-option surveys. If it is, then we consider only the 53 yes-or-no results, giving us a 14% margin-of-error.
.
No.
Quote
Sampling ratio (sample size to population size): Generally speaking, the smaller the population, the larger the sampling ratio needed. For populations under 1,000, a minimum ratio of 30 percent (300 individuals) is advisable to ensure representativeness of the sample. For larger populations, such as a population of 10,000, a comparatively small minimum ratio of 10 percent (1,000) of individuals is required to ensure representativeness of the sample.
Even within those restrictions the the sample must be randomly selected from the population. The make up of opinions on this board are markedly different than those on Reddit which has a larger population.
These survey results are limited in scope, from a biased subset, and can't be extrapolated to the general population.
Quote
If we *don't* assume that the forum sample roughly accurately reflects the full population of BG3 players, then we can't use this poll to say anything and it's basically useless
