Thanks. I did consider that there was probably a capability to make a master/linked shape, after I decided I need them a little larger, but I didn't anticipate that need in the beginning. (In this particular use, the dots can be fine when the whole image is at a certain scale, but when shrinking the image, the dots are the only part that need to be relatively larger to compensate.)
Still, I think the app should default to scaling objects individually. It might make sense to spread the objects when scaling up if the objects are grouped, but as individual objects, it makes more sense to scale the objects without changing their origins. In other words, selecting a bunch of objects and scaling up should have the meaning, "make these objects bigger", only.
Another way to describe it is: If you have several 28 pt text objects and select all of them, then set font size to 14 pt, they all change their font size but they remain in their original positions. However, if you use the transform tab to scale 50% (*=.5), it shrinks the collection (altering locations)—the same as if you scaled by corner-dragging the selection rectangle.
I don't know, this is so obvious that it seems like I'm missing an alternate way of doing it—though I don't see any indication of that in the docs.