SNM 2016-03-02: Single Closest School Assignment

2016-03-02: Single Closest School Assignment

If you ditch the PPS hard boundaries and simply assign each kindergartner to her closest school, this is what would have happened to the kindergarten class of 2014:

(Click to open PDF.)

As with yesterday's post, this assignment disallows transfers and uses "optimal section counts" based on PPS's assessment of actual building capacities. Distances are driving (by-car) distances. Since the historical boundaries in our data do not cross the river, and the district has tended to treat the schools on opposite sides of the river independently from one another, cross-river assignments (even shorter-distance ones) are not allowed.

Compared to the current hard boundaries, this "Single Closest School" model is:

a little better (optimal!) at reducing commute distance;
just as bad at balancing enrollments;
just as bad at diversifying school populations.

The one big difference is that with "Single Closest School", no one has to argue about where to put boundaries.

So, if you are ok with the lousy enrollment balancing performance of Hard Boundaries, then you should strongly consider Single Closest School. Unless, however, you would miss the arguing, and the gerrymandering, and the arguing. Then you should stick to the Hard Boundaries.

(No, we don't actually recommend Single Closest School, or Hard Boundaries, since we are not ok with the arguing, the gerrymandering, or the lousy performance.)