All the current answers are very good, but I want to propose a different point of view.
Even though I would say the vast majority of games don't get their stars right is out of straight laziness and/or ignorance. The same goes for firearm sounds, physics (ever played a racing game? any racing game?), history (for historical games) and many other fields.
However, if I were directing a game, I may not aim for realism in certain elements such as the star field, and I would do so intentionally, on aspects that are not crucial to gameplay.
The reason behind this is that the more you aim for realism, the larger the player expectations will be on that realism. And while player expectations have no limit, budget and time do.
So suppose I got a star chart for my game, and I use that as my skybox. Players who notice that, will then criticize that the stars are 3 degrees off, or that based on the vegetation you see in the game, you can estimate the latitude where the game is taking place, and that the star field doesn't correspond to such a latitude.
So even if I fix that, more players would then criticize that you can't see the ISS, or that some star has an apparent magnitude of 4.5, but in the game it looks like 4.8.
Rinse and repeat.
The point here is that, the more you aim for realism, the bigger your players' expectations will be. If your game requires realism in some aspect, then by all means dedicate your time and money to maximizing this realism. For all the rest, I would probably intentionally make it unrealistic, so I can keep my focus and that of my players in the parts of the game that actually matter.