My favorite is changing the tests every few years. You're still testing, so people can't call you on that, but the test results aren't like for like, so you can't compare them and see how things are going on a longer scale.

That and the focus on point in time test results to evaluate a school rather than analyzing results taken from multiple periods. X% of Y grade students at grade level in year Z doesn't tell you much about the quality of the school in year Z, unless you know where those students were at in year Z - 1.