Dr Jared Cooney Horvath

Now, a lot of people think maybe that has something to say about tech, and it does, but I also think that has a lot to say about learning and actually what is required for learning.

If they would have stopped 20 years ago and said, okay, what is learning before shoving these tools in front of kids, maybe we would have saved 20 years of downslope to get to this point where we're like, you know what?

Humans matter.

Teachers are important.

Depth matters.

None of it was driven by data or learning.

There wasn't a single decision made, I would say, across EdTech ever for learning.

So we've got data out the wazoo that shows the more time kids spend on tech at school for learning purposes, the more learning goes down.

And it's basically linear.

It's not...

Some is okay, a lot is bad.

It's just kids who don't use it at all, great.

Kids who use it a little bit, a little bit worse.

A little bit more, a little bit worse.

To the point where kids who are using it six hours a day, basically all online, are two-thirds of a standard deviation worse on average than kids who don't touch tech ever at school.

So it's the data across standardized tests, international tests, national tests, local tests, all shows that same thing.

to which a lot of people say, well, cool, that's just correlation.

Correlation isn't causation.

Very true.

How do you make correlation causation?