- Open the CSV (comma-separted variables) file created by FormScanner.
- Download ProcessBubbles from the Downloads tab.
- Enter the number of questions (not including ID #) in cell C1, and the answer to each of the questions in the same row.
- Delete cells A2 and A3. Copy the results from Formscanner in row 2 and below. The cell after the last row must be empty.
- Type ctrl-p to process the information.
- Correct scores are yellow; incorrect scores are green; and invalid answers (either multiple response when not warranted or no answer at all) are red. The following information is provided:
- Item facility: the fraction of students who chose the correct answer on that question (# correct/total # of students). The item facility is also known as the question difficulty. Low values indicate difficult questions. A good exam should have questions with a variety of difficulties.
- No. of students: the total number of students who took the exam.
- Q1: the lower quartile
- Q2: the middle quartile or median of the class' scores
- Q3: the upper quartile
- Avg top --: the average of the students in the top quartile (>= 75
^{th}percentile). In this example, the set included 2 students. - Avg bottom --: the average of the students in the bottom quartile (<= 25
^{th}percentile). Here, it included 2 students. - Discrimination: The averages of the top and bottom quartiles are given also for each question (correct=1, incorrect=0). The discrimination is the difference between these two values. A large discrimination (the max is 1) means the well-performing students scored well on that question; a low discrimination means poor-performing students scored well. The discrimination should be >0.20; if the discrimination is less (or negative) check the key or consider revising the question.
See this site for more information about item analysis: http://sites.psu.edu/itemanalysis/ |