Statistics Concept Inventory
A concept inventory is a battery of questions that test a student’s understanding rather than knowledge of procedures from a topic.
To construct these, people often leverage past questions that are known to have mixed understandings to highlight misconceptions.
I believe we can also deconstruct applied problems into combinations of statistical concepts that can counter students from “pattern matching” problems, given pattern matching tends to occur in the context where there’s a one to one mapping between concept to problems. In other words, the combinatorial nature of these problems should deter students from pattern matching problems and actually understanding the concepts.
These are organized mostly following the scientific method:
- Observe by summarizing data with graphs and models
- Hypothesize by [modeling][model/README.md]
- Collect the right kind of data to rule the question
- Analyze by quantifying the uncertainty in data we collect
- Decide and report with uncertainty
How to make concept change?
- dissatisfaction with the old concept
- the new concept is intelligible
- the new concept is plausible
the new concept is fruitful
- beliefs are relational entities, concepts are bigger?