why use automated checking?

In the discussion of [Schinke2014] (a review of assessment) the following sentence stands out:

“However, underlying the less encouraging news about grades are numerous opportunities for faculty members to make assessment and evaluation more productive, better aligned with student learning, and less burdensome for faculty and students.”

It is with idea in mind that automated checking of assessment should be implemented.

Indeed in [Wilcox2015] (a reflection on an implementation of automated checking) there are two initial questions:

  1. Does the automated check detract from student learning?

  2. Do the benefits of implementing automated checking outway the cost?

It is hoped that software similar to and like nbchkr answer the second question.

The first question however is positively answered in a number of pieces of work such as [Wilcox2015] itself where students reported a positive experience of using automated checking but also benefited academically which indicates a better overall learning process. This student satisfaction with the process is also reported in [Saikkonen2001]

Some of the downsides of human checking are listen in [Cheang2003]:

  1. Difficulty of judging efficiency and correctness;

  2. The fact that there can be multiple approaches to a problem that would be missed by a human checker.

  3. Emphasis on aesthetics. Note that given the modern emphasis on the importance of code readability I am not convinced by this particular downside.

  4. Inconsistency of human checkers

  5. Time: the workload of checking works is huge.

This last point is often mentioned in the literature and specifically [Schinke2014] highlight the importance of creating time and space for meaningful feedback through self and peer evaluation.

Self evaluation as a general pedagogic strategy relates well to automated checking as described by [Losada2010] where they prescribe giving a number of tests to students as part of the assessment. As part of the [Losada2010] a discussion as part of Bloom’s Taxonomy is given however this will not be discussed here given numerous downsides to the taxonomy (see for example: [Case2013]). [Losada2010] lists numerous advantages to automated checking:

  1. Fast feedback: in their case and similar to [Cheang2003], [Saikkonen2001] and others the particular framework being described was an “online” one that students could use to gain immediate feedback.

  2. Fairer grading;

  3. Permanent access;

  4. Efficiency;

  5. Fostering a positive attitude towards test driven development (TDD).

This last point relates to the testing practice in software engineering of writing a test before writing the software.

Specific strategies for writing checks are described in [Saikkonen2001] and [Wilcox2016] which give insight and guidance on writing using tests that allow for feedback that helps students identify errors. Contrary to [Cheang2003]’s suggestion that aesthetics having a major role in human checking being a negative, [Wilcox2016] points out that static tools can be used to check the code quality itself (all within the testing framework). One such example of this is in the Tutorial where a test is included to make sure that the code written is documented.

There are some negative aspects to automated testing another good quote from [Schinke2014] is:

“In fact, we have presented evidence that accuracy-based grading may, in fact demotivate students and impede learning.”

It was noted also in [Wilcox2015] that some students do feel that it the automated checks “unnecessarily strict”. Finally, [Wilcox2016] discusses some aspects of security and that automated testing can be done inside of a virtual machine to avoid running of malicious code.

These are all aspects to be considered when writing the specific checks for the assignments and not losing sight of the end goal which is to create a positive environment for student learning. Automated checking should not be thought of as a solution to a problem of assessment but hopefully a tool that enables better learning through:

  1. Timely and actionable feedback;

  2. The creation of space for productive learning activities.