Is Going Beyond Rasch Analysis Necessary to Assess the Construct Validity of a Motor Function Scale?