I think the main reason would be that they can compare their scans directly against the original models.

If they were to use real world objects, any comparison would have to use another scanner, which would introduce its own bias.

That seems backwards. Keys, cups, and pens are pretty trivial shapes. (Assuming you can get good angles on the cup so that it doesn't have air pockets.) More complex shapes like the weird statues they're using seem like a more robust test.

Keys and pens have fine details that would be pretty impressive to scan accurately.

This method would only work with objects that are rigid in all orientations, too - so bunches of keys are out.