WeBWork Quality Assurance and Improvement 3 — Questions and Answers
Question 1: When conducting QA on a WeBWorK problem that uses MathJax rendering, which of the following should be verified?
- That all LaTeX expressions render correctly across major browsers (Correct answer)
- That students cannot view the page source
- That MathJax is disabled for mobile users
- That equations are stored as images rather than LaTeX
Correct answer: That all LaTeX expressions render correctly across major browsers
MathJax rendering must be verified across major browsers because rendering inconsistencies can cause students to see malformed equations.
Question 2: A WeBWorK problem generates randomized parameters but two different seeds produce identical problems. What does this indicate?
- The randomization is working correctly
- The random parameter space is too small or constrained (Correct answer)
- Seeds are being ignored by the server
- The problem requires a restart to fix
Correct answer: The random parameter space is too small or constrained
Identical problems from different seeds indicate the parameter space is too narrow, meaning the randomization provides insufficient variety.
Question 3: Which QA practice helps ensure WeBWorK problems remain accessible to students using screen readers?
- Using only graphical answer boxes
- Providing alt-text for images and using semantic HTML where possible (Correct answer)
- Requiring Flash-based rendering
- Disabling keyboard navigation
Correct answer: Providing alt-text for images and using semantic HTML where possible
Alt-text for images and semantic HTML are essential accessibility practices that allow screen readers to convey problem content to visually impaired students.
Question 4: During QA review of a WeBWorK problem pool, an instructor finds that 30% of problems have a randomized parameter that can produce a zero denominator. The best fix is to:
- Delete all affected problems
- Add a condition to exclude zero from the denominator's possible values (Correct answer)
- Warn students not to enter zero
- Change all fractions to whole numbers
Correct answer: Add a condition to exclude zero from the denominator's possible values
Adding a conditional exclusion for invalid parameter values (like zero denominators) is the correct fix and preserves problem randomization integrity.
Question 5: What is the recommended QA step after importing a new OPL (Open Problem Library) problem set into a WeBWorK course?
- Immediately assign it to students
- Preview each problem and test answer submission before assigning (Correct answer)
- Archive the original OPL files
- Disable randomization for the first week
Correct answer: Preview each problem and test answer submission before assigning
Previewing and testing each problem confirms it renders correctly and the answer checker works as expected before students are exposed to it.
Question 6: In a WeBWorK improvement cycle, which data source provides the most actionable feedback about specific poorly-performing problems?
- Student end-of-semester surveys
- Per-problem statistics showing attempts, scores, and completion rates (Correct answer)
- Server uptime logs
- Course enrollment numbers
Correct answer: Per-problem statistics showing attempts, scores, and completion rates
Per-problem statistics directly identify which problems have low completion or high attempt rates, pinpointing items that need revision.
Question 7: Which WeBWorK setting controls how many attempts students have before an answer is marked incorrect permanently, and is critical to QA for fairness?
- Max attempts per problem (Correct answer)
- Reduced scoring period
- Open date
- Showwork mode
Correct answer: Max attempts per problem
The max attempts setting determines fairness by controlling how many tries students receive, making it a key QA parameter to review before publishing a set.
WeBWork Quality Assurance and Improvement 3
When conducting QA on a WeBWorK problem that uses MathJax rendering, which of the following should be verified?