If research participants are assessed with a particular questionnaire early in a study, but the researchers change the questions for later assessment so participants can't simply repeat answers, what potential threat to internal validity would be likely in the study?