After an initial report of findings on the effect of testing interruptions in Indiana, legislators asked whether the overall finding of “no impact” could be a function of two factors cancelling each other out: some students were adversely impacted by the interruptions, while others were able to take advantage of the interruptions by changing their answers from wrong to right when they resumed the test after the interruption. The analyses provided in this paper show that there likely was such an effect, but its impact on the overall results was so small that it did not affect their interpretation.
