Research experience is definitely important, but it's also important to not over-simplify. My friend went to top ranked school without any research experience at all.
But one data point showing that someone with a specific set of GPA/GREs/research experience got into School X doesn't mean that if you have the same then you will also get in. It doesn't mean that's the minimum standard to get admitted, or even that it's the typical acceptance.
I think graduate school admissions is much more complicated than 2 or 3 dimensional space. Also, even though big programs may admit 50 ish people per year, the amount of available data is pretty sparse. For most applicants, we might only know a few people at a certain school well enough to really know their profile, and the aggregate data published by the grad schools does not show the whole picture. And the aggregate data can only be published for some of the stats, typically the ones that are easily quantifiable such as GPAs and test scores. However, it would be wise to remember that just because there are mean X, Y, Z scores published, it does not mean that X, Y, Z scores are more/less important than the unpublished factors.
I think a significant chunk of the evaluation of the applicant is qualitative rather than quantitative. The goal of the admissions process is to find the candidates that will be most likely to succeed in the program. To me, this means the admissions committee is looking for a candidate that possesses a certain set of abilities and character traits. Few programs are going to say "We want someone with a 3.95 GPA". Instead, I think the real evaluation criteria is something like "We want someone who is able to learn independently", or "We want someone who can excel in an academic environment". One of the ways that an applicant shows this is through excellence in previous coursework (i.e high GPA).
So, for things like GPA, or research experience, or GREs---my advice is to think of them as ways you can demonstrate your excellence/fit for the program. They are ways that you can show that you meet the criteria. But, these scores themselves are not the criteria!
Note: The above is a slight simplification. For programs that get huge numbers of applicants, they might use past experience to filter/narrow down the applicant pools using cutoffs in these scores. Usually these cutoffs are well below the published mean/median values because schools understand that these metrics are just one potential way to evaluate. But when you are preparing your application, you would be better off thinking about how you can demonstrate you are a good candidate rather than worrying about the quantitative values of your scores.