If you install the tread first and then put in the riser, the riser will add stability to the tread by holding down the back of the tread; then it isn't just all on the nails and glue. This could theoretically matter if you had two adults step on the front of the same stair at the same time, especially if they are running down the stairs. Having the riser go first would make it so there is nothing at the back of the tread helping hold it down, and with enough force, physics states that the stair tread could basically "see-saw" on the top of the edge below it and the entire stair could come up.
Now, this shouldn't probably happen if enough nails and glue are used, but...
I'd rather make the weakest point be the actual nose of the tread (the nose would break off) rather than risk the entire stair coming up. At some point, no matter how well you install the stair, if you put enough weight on the front edge of the nose, something is going to give out. What you have to decide is what should give? Putting the riser on top of the tread at the back of the tread would probably make it so that rather than the whole tread flipping up, some part of the tread (presumably the nose) would simple break off instead. I guess it's kind of 6 of one, half dozen of the other at that point though, because either way you have a broken stair. The real question, to which I have no answer, is which design can hold more weight before it gives out. My gut says that putting the riser over the stair would hold more weight.
I've also seen people cut the back of the tread and the bottom of the riser at an angle so that the angles are flush with each other (the riser is still on top of the tread, but, the tread extends to the back of the riser). They made the statement that this gives a nicer looking joint/seam. That seems like the best of all worlds to me.