Reviews will typically mention which version they were, but in general, there are very few differences between them these days, unlike back in the 6th gen or early 7th gen. Games like Cyberpunk are outliers.

Starfield is not a bad game. In a lot of ways, it's a very good one. My biggest complaints with it, personally, are all the ways that it should have been modernized but refused to, falling back on what worked over a decade before it came out without turning an eye toward its contemporaries and the improvements they've made to the same formula. I find Steam reviews to be a valuable data point among plenty of other data points, but user reviews being that much lower than the critic average doesn't mean the critic score is a problem.

For an example of a game where critics reviewed it less favorably than the user score, see Mad Max or Days Gone, which might be explained as games where the initial sales weren't strong, and people who found it later, often at a discounted price, were pleasantly surprised compared to its reputation. There's also the likes of SkillUp's review of Ghost of Yotei. That game has largely reviewed very well by other outlets, but he found his review to be out of sync with his audience. If you're a reviewer who plays dozens of games per year, your opinion of a formulaic open world game might be very different from someone who plays 3 games per year and hasn't gotten sick of it. Both are valid points of view.