In my opinion the federal level government should mete out rules that are unquestionable. There's nobody of sound mind that would genuinely and rationally argue that stealing could be anything less than harmful to a society, and so it makes sense to lay out federal restrictions on such. But I am certain there are a good deal of people that would genuinely argue, with some evidence to support their cases, that smart phones are something less than harmful towards the development of children during school time.
To understand why I'd make this distinction, imagine this is something you would personally disagree with. Once something is federally banned it is quite difficult to unban it, and more importantly it also precludes any further studies of the effects of it on a large scale. By contrast imagine one region in France locally banned smartphones and another did not. And then after 20 years we see measurably different outcomes outside of these two regions previous trajectories. That would be quite informative and evidence that could help push more local regions to push for bans.
Local level bans allow citizens to choose to either opt-in or opt-out of certain rules and thus live in conditions that are closer to their ideal. Federal level bans, by contrast, are ultimate unless you choose to move to another country which is a much more substantial change than moving to another district/city/state/etc.
Citizens are equal everywhere, therefore it makes little sense not to enact public health measures nationally.
It seems even more questionable, to me, to enable a measure for 20 years in one region using their children as guinea pigs, but that would most likely be unconstitutional anyway.
That's how it works. A law is debated, voted on and if accepted then enacted. Another way to call it is "progress". It works the same for... well mostly everything that becomes law.
But since you're not French, you can easily wait 20 years and then decide if it was a good thing or not. For now, France has debated the subject and taken a decision, and I don't think it needs advice on whether the political process should have been different or not.
---
And how would you propose measuring whether or not this experiment will be a success, lacking any control to measure it against? If it turns out French students are performing worse 20 years from now there will certainly have been countless other 'experiments' carried out since then, meaning any causality is going to be extremely difficult to trace. And similarly if they are performing better. By contrast when things are enacted locally we can often get samples of groups where things are very similar except for [x], which can help determine causality.
Instead it all turns into, as usual, retrospective confirmation bias. People take what they want to be true, and attribute positive results to that while taking what they don't want to be true and attribute negative results to that. It's hardly a logical way of trying to achieve genuine "progress."
As said elsewhere in that discussion, most school (all of them?) are already banning smartphone, so it has already been tested before.
This seems like a typical French statist approach...
But the real question is why are you questioning the way the French are running their own country? I don't think they are really asking for guidance on how to use the French political structures.
Boy, we are on Reddit...
I cannot find a single reason to do it at any other level than national.