Evidence That Selecting an Appropriate Item Response Theory–Based Approach to Scoring Surveys Can Help Avoid Biased Treatment Effect Estimates