The problem with applying "survey" or "study" data is similar to using poll data, in that detailed results are not, and cannot be truly representative in detail, only in overall trends.
there may be 8 million new gun owners out there this past year (2?) but they aren't a homogenous group. In fact, the only thing we know with certainty is that they legally purchased a firearm. Aside from that small fact, we're ALL different and if the batch isn't properly mixed, you don't get a representative sample.
Ask a few dozen or even a few hundred people anything, and then ASSUME their answer is valid for millions of people? Sure, the news does it all the time, people make a living doing it. Politicians do it. Doesn't mean its accurate, only that its popular and SOMETIMES represents the majority of opinions.
You got 8 new gun owners in your family and 3 of them became shooters, that's great. I got zero (0) new gun owners in my family though several guns were purchased. No NEW enthusiasts, either.
Also don't forget that those 8 million "new" guns include ALL guns, a sizable percentage of which ae going to be rifles and shotguns that don't "go to the range" twice a month, or even once (particualry during the winter) and a good chunk of the handguns don't either. A lot of people, first time buyers are buying "just in case" and may only shoot one box of ammo OR LESS, ever.
Also, if you're trying to use those numbers to figure future ammo demand, do remember that there are also a lot of guns that are still "under the radar" and don't show up in official counts, but their owners buy ammo, too.
My "crystal ball" has turned black, with a white circle and a large number 8 on it. Lately, when I ask it a question the usual result is "answer uncertain ask again later..."