15 December 2014

Struggling with assessment.

I'm in the final throes of course design for my bioinformatics class next semester. I've already written a bit about planning the course, and a little about my problems convincing students to take it. I've spent a lot of time getting the computer lab up and running, and a lot of time preparing course materials. Although I still need a few more students to enroll, I'm fairly certain I'll actually get to teach the course (and hey, if it doesn't make this semester, there's always two years from now? *eye roll*).

Here's where I am in planning. I've got a lecture that meets for three hours a week on Tuesday and Thursday mornings, and a lab that meets for three hours Thursday evening. Lecture is about general theories and concepts, while lab is about implementation of that content in coding and data analysis. The course content is split into two sections: the first six weeks is what I call the Bioinformatics Framework, where we talk about bioinformatics as a field of research/applications, managing data, developing pipelines, and hypothesis testing. The second part of the class is Applied Bioinformatics, where we'll cover several "vignettes" of bioinformatics applications, like sequence alignment, clustering/phylogenetics, and genome assembly. I'm pretty comfortable with this plan, including how it relates to my objectives for student learning.

My last big hurdle in course preparation is finalizing how I will assess student performance (i.e., giving grades). Because the class is based on skills development but also incorporates interdisciplinary thinking (biology + computer science), I'll need to implement a variety of assignment formats. I'm planning at least one formative assessment for students to turn in each week to make sure everyone's on the same page. I'm also going to have each student do a class project: researching a type of bioinformatic analysis not covered in class (like protein structure/folding, network analysis, metabolomics, etc). They will present their findings on the major challenges, methods, and applications in that topic to the class, so we'll get a broader feeling for research topics than what I'll have time to cover.

My problem is that I need to be able to explain exactly what students learned during the semester (summative assessment). This is partly for my own ability to track student performance, but also for reporting to departmental and university groups. However, I appear to have developed an allergy to things called "exams" (the most common form of summative assessment). I get anxious just thinking about having to write, administer, and grade an exam.

Azuki beans. They are pretty,
but I am no bean (or point) counter.
(thanks Wikimedia Commons)
I met with a fantastic instructional designer (Leslie Lindsey) from the aptly-named Office of Instructional Design last week, and we talked about different approaches to evaluating student performance. She validated me in my belief that I can give a course that does not include exam-based assessment. She helped me realize that my aversion to exams seems to be a fear of the reductionism of simply counting points to assess student learning, which seems to be required to give a final letter grade in the class. What she said to me blew my mind: "Think of assessments as a way of collecting data about student performance."

Oh, the irony! I'm teaching a class on using computers to analyze biological data and I failed to realize that assessing student performance and assigning grades is just another data analysis problem. I was getting bogged down in my imagined obligations as a professor, and not thinking about this enough as a data analysis problem. My problem is largely semantic, and perhaps I just need to think about offering a different kind of exam, designed to emphasize the things I value as a professor. I value steady, consistent effort by students throughout the semester, even if it means I need to keep up with grading on a weekly basis. I value student comprehension that allows conceptual synthesis and connection between topics, but understand this may take more time than is allowed in a class period. I don't want my need for data collection to adversely affect student grades when there are other, better means of assessing their understanding.

Ultimately, I've decided to use an evaluation strategy based on "units of assessment" that are graded on a similar (but adjustable and specifiable) rubric. Weekly assignments in both lecture and lab will count as 1 unit each. Research projects for each class will have multiple parts, each of which counts as 1 unit. For lecture, I'll have a day each for both the first and second part of the course for students to perform summative assessments. These assessments will include two parts (one in-class, one out-of-class) which count as 1 unit each. That means the summative assessments are weighted as a bit more important than weekly assessments. I'll average the rubric scores throughout the semester and convert to a letter grade. This seems much more palatable to me than assigning absolute point values or weighted percentages to every type of assessment. Also, I'm hoping it will capture student performance much more authentically than grading based on exams that occur on a few days throughout the semester.

I don't know if this makes sense to anyone else, but it's starting to make sense in my head?

No comments: