Placement Test Demo Peter Liu AITS Yale University

  • Slides: 8
Download presentation
Placement Test Demo Peter Liu AITS, Yale University Email: peter. liu@yale. edu 1

Placement Test Demo Peter Liu AITS, Yale University Email: peter. liu@yale. edu 1

Placement Test Project Some Backgrounds for Placement Test Project • Every year, there about

Placement Test Project Some Backgrounds for Placement Test Project • Every year, there about 1000 new students, who will need to take the Math Placement Test from July to mid of August time-frame. After taken the test, students will print out the Placement Test results (automatically graded at Tests & Quizzes tool) to meet with their advisors at 3 rd week of August and discuss which Math course level is suitable to them. • Last Fall, Yale Math department has approached us and want to see whether we could make it a more automatic way to alleviate the burden of advisor and students. • Goal: Hope that 70% of the students will know immediately what course level to register without seeing an advisor. 2

Placement Test Project The Requirements and Rubric Algorithm • Students take placement exam. One

Placement Test Project The Requirements and Rubric Algorithm • Students take placement exam. One of the questions will be a multiple-choice request for scores on calculus Advanced Placement exams (levels AB and BC, each score in the range 0 - 5). • Analyze exam output. Each student's (tentative) placement will be computed by an algorithm, based solely on that student's answers on the placement exam • Students and department are notified of the (tentative) placement Students should be notified immediately after completing the exam, either at a website or by email. The placements should also be available to the department, for example as a downloadable spreadsheet. • Placement exam results available for export to other Yale systems 3

Placement Test Project The requirements and Rubric algorithm • Sample placement algorithm (to illustrate

Placement Test Project The requirements and Rubric algorithm • Sample placement algorithm (to illustrate form of algorithm; details aren't realistic) ab = score on AP Calculus AB (ab 5) bc = score on AP Calculus BC (bc 5) s 1 = total score on problems 1 - 10 (s 1 10) s 2 = total score on problems 10 - 15 (s 2 5) s 3 = total score on problems 15 - 20 (s 2 5) if bc == 5 and s 1 > 7 and s 2 > 4 and s 3 > 4 place in Math 120 printout “You might want to consider taking Math 230 - 231" else if s 1 > 7 and s 2 > 4 and s 3 > 3 place in Math 120 else if s 1 > 7 and s 2 > 3 place in Math 115 else if s 1 > 5 place in Math 112 else place in Math 112 4

Placement Test Project Design of Placement Test • Design a Rule Engine which will

Placement Test Project Design of Placement Test • Design a Rule Engine which will fit to generic Placement Test. Generic rubric rules for one placement level as shown below: Where Sk is the score of a part or one question Wk is the weight factor Pi is the user defined minimum score for i condition • Easy for user to define the Rubric conditions via UI. • Allow user easy to debug their Rubric Algorithm on Fly • Integrated with current ‘Tests & Quizzes’ tool and use the existing API/Listener to retrieve testing scores for further rubric analysis. 5

Placement Test Project Sample for Placement Rubric Level Set Up 6

Placement Test Project Sample for Placement Rubric Level Set Up 6

Placement Test Project Sample for Result for student: 7

Placement Test Project Sample for Result for student: 7

Placement Test Project Current Status • The Math Department has successfully completed the Math

Placement Test Project Current Status • The Math Department has successfully completed the Math Placement Test with around 1000 students, ended at August 20, 2013. • The Rubric Engine has hold up very well without any issue so far. • Future development? ? ? • More friendly UI • Allow to export data via Entity-broker API 8