In the NCLB era, it has become a rule of thumb that Test Scores Go Up. That is, they are required to go up, and in aggregate, they do. This is partly because schools, teachers and kids adjust to the demands of a particular test over time, including but not limited to actual improvements in teaching and learning.
It is to at least to a small extent due to an increase in various kinds of cheating at the school level as pressures rise over time.
Finally, there is pressure to quietly make the tests easier in content or scoring. As Todd Farley recounts in Making the Grades, this may require no more than a visit to the scoring center from a mid-level state official and a quiet conversation behind closed doors about the "proper application of the approved rubric" to generate a different score distribution.
Or perhaps the cut scores are quietly moved in a technical meeting.
Or there's a presumption that making the tests a graduation requirement will lead to a large bump in scores, as was the case with the MCAS in Massachusetts.
From 2007 to 2012, the percentage of students in Rhode Island scoring Substantially Below Proficient has moved from 51% to 40%. The entire premise of using the NECAP for a graduation requirement is dependent on that number going down to no more than about 10% (at most!). That has to happen, or it will be politically untenable.
If you're used to looking at changes in state test scores over time in the abstract, it was reasonable to assume that this would "just work," because Test Scores Go Up but perhaps that doesn't factor in the quiet shenanigans by states and testing companies.
Because this is a multi-state testing consortium, RIDE can't wink-wink nudge-nudge the testing company and to get the scores up. Vermont and New Hampshire would have to agree as well, and since they aren't using the NECAP for graduation, they have no incentive to play along.
Again, this may be the shape of things to come in the Common Core era.