M6L2: The Mechanics of Item Analysis

Item Analysis

Checking how well individual test questions performed using numbers so teachers know which ones were good and which were confusing.

Item Difficulty Index (p-value)

The percentage of students who got a test question right ($p = \text{Correct} \div \text{Total}$).

Optimal Difficulty Range

The sweet spot for a question's difficulty, which is between $0.30$ and $0.70$ so it isn't too easy or too hard.

Item Discrimination Index (D)

A score showing whether a question successfully separates top-performing students from struggling students.

Upper and Lower Groups

Splitting the class into the top $27\%$ and bottom $27\%$ of scorers to see how each group answered a question.

Very Good Discrimination Item

An item with a discrimination score of $0.40$ or higher that clearly tells the smart students apart from the rest.

Negative Discrimination Value

A massive red flag where low-scoring students got a question right more often than high-scoring students, usually meaning the answer key is wrong.

Distractor

An incorrect multiple-choice option designed to look believable by using common student mistakes or misconceptions.

Non-Functional Distractor (Dead Wood)

A wrong choice that absolutely nobody picked, which makes guessing easier and needs to be replaced.

Miskeyed Answer

When a question's official answer key is wrong, causing the smartest students in the class to accidentally pick a distractor instead.