So a couple of things here: with my interests coming from working in health and cognitive psych, and more recently working as a UX researcher in games. I've recently done some work that was focused on kids gaming activities during lockdown.
If I had to comment on the general body of research on screen time, I'd say two main things: the negative effects they are showing are small (and I mean that formally - as in, the effect sizes are small. Way too much focus on showing significance at the moment without considering the practical impact, but that's early days research for you), and the quality of the body of research itself is low (I don't mean that as a slight: again, it's partly a result of an immature research field. Not enough studies, poor replicability, little ability to generalise, few good systematic reviews and meta-analyses)
Here's the other thing with the research, from talking to researchers and attending conferences: too many of them have low digital literacy themselves. They barely differentiate between using devices to do different activities - watching videos vs playing games for example - and certainly are not at the level of differentiating between the nuances of various games. If I asked you 'are games good for you' - you'd ask "which games?" The same questions should be applied to screen time. When Times Table Rockstars and PUBG are both falling under the same metric, something is going very wrong.
But again, all understandable. Researchers - psychologists especially - need simple metrics to start out with, and a screen time is easy to obtain on a mass scale. As time goes on, I think the direction of travel is that we will begin to tease it out: we will break down and classify device activities better. And we will begin to address covariates, moderators and mediators, which are just barely addressed in the current research.
Eg: household income - this impacts access to devices but also then impacts access to quality of apps and games available. With less money available for purchases children are shuttled towards FTP games, which are often a mess of dark patterns, microtransactions, pushy notifications, and terrible ethical practices overall.
There are issues around underlying conditions - some studies do not even account for increased device use in children with limited mobility, which could be enough to create the very modest effects shown. Attention also: be wary of blaming attention issues on prior use of screen time, when even the most negative of studies have not been willing to establish causality/direction of cause here. The evidence is extremely weak on attention, cognition , etc.
And lots of other similar issues.
The other big consideration that I think changes the screen time debate is relative risk. How much more risky is greater screen time than other activities? One comparison I saw was that it's about as increased a risk as having a diet that included potatoes. Although any nuanced discussion of relative risk needs to look at the combined risk of eating potatoes AND screen time, of course.. but the message is that screen time (and games) are best managed as part of an overall pattern of risk and behaviour, with no good evidence that it has a worse impact on children's lives than many of the other behaviours we encourage without even thinking about it. We have to look at how things like Fortnite and Roblox have been used in the last few months: for some children they've been their main channel for socialising with their friends.
But anyway, to get off the academic stuff, many of the parents I spoke to were coming around to the view that if they could find apps and software that they felt confident about, they would probably lift screen time restrictions. This is a good direction to move in I think: things like Osmo and the Namco-Bandai one (have forgotten the name) are learning tools that children view as primarily enjoyable games, and that's great. More digital literacy in parents, in educators, in researchers will be what helps them to make better choices and guidance about what devices and apps children use.