The acceleration point for both age groups studied is 2012. What happened that year? The article doesn't try to answer this. Might be mentioned in the study I suppose.
The real culprit is probably more in line with far more alternatives to reading for entertainment.
Up until the 2010s I think it was still a lot less socially normal to play a lot of games. We reached a tipping point somewhere that went from gaming being a sometimes activity for kids to basically every kid plays games
Most of the people in my high school in the 2000s didn't play games as a primary hobby. Only a few of my friends had a PC for games or a console. It wasn't nearly as ubiquitous as it is now.
Halo 2 was when many people started gaming
What you see is that before the anomaly year that was 2020, the scores were slightly lower as compared to 2012, but they were still not as low as early 2000s or before.
Another thing to notice is that the higher percentile performance largely remains flat but lower percentiles seem to be suffering from a drop.
While social media and phones could be a part of it, if they were the only factors, we wouldn’t see such a disparity between different percentiles. I suspect educational policy. No Child Left Behind (NCLB), no matter how controversial, coincided with the improvements in the scores. In 2012, a bunch of states were allowed to have more flexibility from NCLB. In 2015, it was replaced with Every Student Succeeds Act which allowed states to set their own standards. I think this was the single biggest contributor to the declining scores.
Prior to that, I stopped reading because video games were easy to get lost in endlessly. At the time, I recall I was probably playing a lot of League of Legends, TF2, Minecraft, and probably some others -- all of which I felt I could pretty much sink an infinite amount of time into, at the time.
Did it work? :)