Assessing conceptual learning within undergraduate biology programs

True False Or Multiple Choice, What's The Better Question Type?
Dr Brian Couch from the University of Nebraska-Lincoln presented his work studying the most effective method of framing questions to test student understanding. Couch explained that there is an assessment triangle connecting observation, interpretation and cognition, and how the assessments that test these elements are often limited by the resources devoted to their creation. Instructors create assessments to evaluate preparation, training and justify their ascension to tenure but may often miss the mark in adequately measure understanding. Couch's study looked at students in a molecular biology capstone course that was taken at the end of 4 years of training. Knowledge was assessed to improve the quality of instruction and shift thinking to program-level thinking since the assessment theoretically reflected the outcome of a completed education. It is important to remember that there are many intermediate stages that exist along the progression from novice to expert and along the way, ideas may be internalized that are incoherent and/or only partially correct. To test knowledge, the most common question type is multiple choice, where a right answer is chosen but only true/false statements are detected (one answer is right and three answers are wrong). Couch believes that true/false questions are a stronger question type because it can detect misconceptions as statements that may be incoherent or only partially correct. Couch devised an experiment where the two question types were invertible for use to assess the same knowledge. After administration, he compared the outcomes. One note on grading: in a multiple choice question, credit for a right answer is all-or-nothing, but in a multiple true/false question, it is possible to award partial  credit. Students scored ~67% on the multiple choice and ~72% on the multiple true/false. When looking at the types of questions there were answered incorrectly, Couch found that incorrect answers are not uniform or clustered but rather randomly distributed. This means that students' misconceptions are not fixed to one topic, but rather there are aspects of multiple topics that result in misconceptions that occur globally. Another way to say this is if you are making a sandwich, misconceptions of how to make the sandwich are not limited to what the components of the sandwich are. Instead, misconceptions can occur in what instruments are needed to make a sandwich (i.e., knife, spoon, scissors), whether the sandwich  can be made on a plate or paper towel, and whether the sandwich can be stored in a bag or in a sock. Couch believes that multiple choice questions systematically underestimate students who correctly endorse correct/true statements or incorrectly endorse incorrect/false statements, conveying the false notion that students understand a topic and providing a poor estimation of topic mastery. It seems that a better way of distinguishing between picking the correct answer to a question and actually knowing something is using multiple true/false questions.