Skip to content

Philosophy of STACK

Online assessment can effectively assess and give feedback to a large number of students. This frees up time that teachers would normally spend, and ensures consistent assessment.

Traditional multiple choice questions have a number of limitations, including:

  1. Only finite answers are possible. The teacher can put in "dummy responses" for common mistakes and give feedback based on those mistakes, but it is impossible to give feedback on student errors outside the given examples.
  2. You can only assess lower-order skills. "Give an example of..." type questions are impossible.
  3. It encourages strategic learning, i.e. instead of solving the problem, students will think about how to "trick the system" and find the right answer.
  4. Question distortion is especially problematic in mathematics. You may ask students to integrate a complicated function, and then give a list of potential answers. Differentiating the answers to get the original expression is a much easier process, and sensible students are likely to take the easier route, which is not what we wanted the student to practice.

The alternative is to use student-provided answer questions where the student's answer contains the content such as an algebraic expression.

STACK was designed to use computer algebra to assess students' answers which are algebraic expressions.

Design choices

Key design choices in STACK make it stand out.

Teachers should be able to write their own questions, with minimal coding skill

Teachers should take responsibility for their assessments. When assessing students' answers, STACK asks teachers to focus on the properties of students' answers, such as "algebraically equivalent to the teacher's answer", "factorised", etc.

  • Focusing on mathematical properties, such as equivalence, is a unique design feature of STACK.

Teachers can then pick suitable answer tests in a potential response tree and give marks and feedback accordingly. STACK does not require teachers to learn a coding language and assess students' answers by writing code such as :

If
 simplify(student_answer - teacher_answer) = 0
then
  mark = 1,
else
  mark = 0.

Students should not be penalised for poor computer skills

Online assessment should assess mathematics skills, not how well students know the specific syntax. For example, penalising a student for answering sinx instead of sin(x) is not fair.

  • Separating validity from assessment is a key design feature pioneered by STACK.

To ensure that students are marked for mathematical skills instead of computer skills, STACK separates "validity" and "correctness". When a student types an answer, it is interpreted by the CAS and a "validation box" is shown displaying how the student's answer is interpreted. This gives the student a chance to fix any syntax errors before their answer is marked.

Multipart questions should be possible

Multipart questions can be very helpful for students, for example to help guide a student through a new topic.

  • STACK completely separates input and assessment with a unique and flexible design.

A question can have unlimited input boxes, and unlimited potential response trees to handle the assessment. Each tree is not limited to a particular input, but instead has access to all the student's inputs. Hence, a tree assessing the correctness of part (b) of a question can use the student's answer to part (a) in its algorithm. This allows for follow-through marking, where a student be penalised for a wrong expression in part (a) but given credit for correctly substituting in values in part (b).

STACK is rich in features

STACK is designed to cover the needs of a large variety of users across mathematics and science.

  • Questions can be randomised to ensure different students see different variants of a question.
  • There are many different kinds of inputs. These are, for example, where the student enters a mathematical expression, or makes a true/false selection.
  • Partial credit is possible when an expression only satisfies some of the required properties.
  • Plots can be dynamically generated and included within any part of the question, including feedback in the form of a plot of the student's expression.
  • Students can work line by line reasoning by equivalence until they have a final answer in the correct form.
  • STACK supports working with significant figures and scientific units.

STACK is open source

The problems faced by teachers, particularly in University, are the same the world over. We want a system that encourages other developers to collaborate on improving STACK, but remaining free to use it in open or commercial settings.

  • STACK is the leading open source online assessment system for mathematics and STEM.

The business model for STACK relies on institutions collaborating on the infrastructure, remaining free to use the resulting system as the please, including in commercially as needed.

The documentation for STACK is available under the Creative Commons Attribution-ShareAlike 4.0 International License Creative Commons License.