MATLAB: MATLAB Grader marks the submission as incorrect if it contains statements like “clc” and “clear all”

"referencevariables"clearerrorgraderMATLABundefinedvariable

In MATLAB Grader, some assessment tests are being marked incorrect even when I have confirmed that the submitted solution is correct.

I noticed that my submission contains commands like "clc" and "clear". Do they have to do anything with submission.

Best Answer

This is expected behavior. The reason for that is because "clear" removes all the variables created in the reference solution. These variables are often used to compare against variables created in the learner solution. This results in the Variable Equals Reference Solution test type, and any assessVariableEqual tests comparing to a referenceVariable field, to be marked incorrect.

As a possible workaround, the instructors can include the following command when creating questions.

assessFunctionAbsence('clear', 'Feedback', 'Do not use clear in your solution');

Please refer to following documentation page for details about "assessFunctionAbsence" function.

https://www.mathworks.com/help/matlabgrader/ug/assessfunctionabsence.html

Related Solutions

MATLAB: My student’s submission in MATLAB Grader is correct. Why is it being marked incorrect?

There are several common reasons why an assessment may be marked incorrect even if the relevant code in the student solution is correct.

The student solution contains a syntax error. All assessments are marked incorrect and the syntax error message is shown as feedback in each assessment. Students should first correct any syntax errors and then resubmit their solution.
For numeric values, the tolerances may be inappropriate for the desired correct solution. Default tolerances are ±0.001 relative or ±0.0001 absolute. When the correct solution is very small, the answer may never be marked incorrect. When it is very large, small differences in how the calculation is done could result in a correct answer being marked incorrect.
Inconsistencies between the problem statement, the reference solution, and/or the learner template. Often, variables in the student solution are compared to variables in the reference solution. Verify that the reference solution is consistent with the problem statement, and that the learner template, if provided, is consistent with the reference solution. Sometimes in the editing process the problem statement may be updated but the reference solution overlooked, or the reference solution may have been updated but not the learner template.
Comparing a variable in the learner solution to itself instead of the corresponding variable in the reference solution. If using custom assessment code and comparing variables using the assessVariableEqual function, be sure to refer to reference variables using the structure referenceVariables.varName
'clear' or 'clear all' commands in submission can remove reference variables for the test cases. Make sure that the student is not using 'clear' command in the submission. As a workaround, include "assessFunctionAbsence('clear','Feedback','Do not use clear in your solution')" while creating questions.

If you have gone through the list of common issues mentioned above and are still having issues, please contact MathWorks Support.

MATLAB: Is it possible to suppress the default error message of the function assessVariableEqual

There is currently no way to suppress the default feedback when using the built-in assessment function assessVariableEqual. The alternative is to instead perform the grading using assert statements. Assert will only show the message you provide (see tips below for more).

To duplicate the behavior of assesVariableEqual, you must

First determine if the student variable exists.
Next, determine if the student variable is the same data type as the reference variable.
Then determine if the student variable is the same size as the reference variable.
Finally, determine if all the values in the student variable are the same as the values in the reference variable. For numeric values, apply a tolerance.

The first 3 tests are necessary to ensure the values can be compared without causing an error.

Here's an example (untested):

% does variable f exist?
assert(exist('f')==true,'The submission must contain a variable named f')
% is f a double?
assert(isa(f,'double'),'Variable f must be of data type double.')
% check size
assert(all(size(f)==size(referenceVariables.f)),'Variable size is incorrect');
% Check values using assert
assert(all(abs(f-referenceVariables.f)<0.0001),'The variable f contains an incorrect value');

Best Answer

Related Solutions

MATLAB: My student’s submission in MATLAB Grader is correct. Why is it being marked incorrect?​

MATLAB: Is it possible to suppress the default error message of the function assessVariableEqual

Related Question

MATLAB: My student’s submission in MATLAB Grader is correct. Why is it being marked incorrect?