I assume you read at least the phrase you quoted. In that very phrase, two things are compared: a) the full game, b) the DLCs. Assuming that indeed data is biased, it should be biased in both elements of the comparison. Ergo, DLCs overall ratings are still significantly lower than the full-game's.
Now, if we get technical, I am happy to get technical too.
Lower DLC ratings does not necessarily mean DLC are viewed worse than the base game. Fewer people buy DLCs, leading to fewer reviews, so each review has a higher weight than the reviews for the base game. This problem with unequal sample sizes can lead to wrong conclusions unless accounted for.
It's ironic that you question "biased" data using an opinion. Is there any fact that proves that indeed steam reviews are biased towards negativism?
I wasn't saying this explicitly applies to the steam reviews (it probably does, this would be just my opinion), just more as a general concept, which is established in marketing science. So not an opinion, but science based information.
Regarding your specific example: "$10 for 4 animals is too much". I don't see how a person writing that review is ignoring other content in the pack. They are most likely fine with the rest of the content, but very disapointed with the amount of animals. Not ignoring anything. (This is just an assumption, so is yours).
This was just a random example, though I have seen reviews stating something along the lines that "$10 for 4 animals means $2.50 per animal and thats too expensive", in which case they are explicitly not assigning value to the other content, this was more what I was meaning to provide as an example. Also, you are also making an assumption by saying they are most likely fine with the rest of the content, this is an excellent example how you and I both have different biases in interpreting the same data (with no real way to tell who is correct).
Why negative critics are extreme and positive are not extreme?
Sorry I wasn't clear here, I was including extreme positive in this statement
But it is the only meaningul data that is avaliable (as far as I am aware).
Objectively, what data is used to know what the PZ players want, then?
You are assuming this is the only data they have. They also could have independent market research, maybe use focus groups, possibly use player usage data (I'm not sure which diagnostic data, if any, are sent to Frontier). There could be other sources of information that you and I likely will never see. Even if they don't, using biased data isn't inherently bad IF you correct for it. If the data is biased but has been corrected for such bias, it is still usable to draw conclusions.