These are then categorized, answered, rated, and commented on by other users, who can pass time by going through them one after another.
We categorized answers 1-3 as "dentulous" and answer 4 as "edentulous".
For each question, a set of five categorized answers were provided, and respondents were expected to choose the single best fitting answer to each question.
I also reviewed responses to questions about additional credentials to eliminate redundant or incorrectly categorized answers (e.g., answering other--massage therapist for question 4).
We categorized answers into a wide range of possibly "correct" values of 1000 to 3000 calories, <1000 calories as "incorrectly underestimated," and >3000 calories as "incorrectly overestimated".
Categorizing answers to open-ended questions is challenging.
For the analysis of current practice, we categorized the answers 'always' and 'frequently (in more than 50% of cases)' as "positive", and the answers 'never' and 'sometimes (in less than 50% of cases)' were categorized as "negative".
We reviewed students' initial responses to the Human Evolution Question and categorized all answers as either: "definitely containing a misconception," "probably containing a misconception," and "not containing a misconception".
For the Revision Question, we measured whether students rejected misconceptions by categorizing their answer as showing signs of "no improvement," "partial improvement," or "complete improvement".
We categorized the answers to these questions, as the underlying variables are naturally categorical (like symptoms from lower airways) [ 23].
Because the data were skewed toward the higher values, we re-categorized the answer choices into three groups: "poor or fair," "good," and "very good or excellent".
