Using Project EDDIE modules in Freshwater Ecology
About this Course
EDDIE Module(s) Adopted and/or Adapted
This module introduces students to a unique ecosystem (karstic wetlands) that they are probably not familiar with. The module teaches them how to run and interpret diagnostic tests for linear models in R without requiring much coding experience. Sharing their results with their peers allows them to practice comparing models to determine which variables are the best predictors of periphyton biomass and food quality.
Relationship of EDDIE Module(s) to my Course
The module was taught about halfway through the semester during our lab period. We had previously completed the EDDIE ice phenology module in Excel so the students were familiar with linear regression. This periphyton module built on the quantitative skills they learned in the ice phenology module in two ways: 1) it was in R, rather than Excel and 2) it included testing the assumptions of linear regression and transformation of variables. The biological content was also a nice fit because the module includes some information on using diatoms as indicators of environmental conditions. This supports one of our field labs in which students sample and ID benthic invertebrates to use as indicators of stream water quality. Although students were not asked to prepare specifically for the module, we had been working in R for about a month and I did make sure that they were familiar with the basic ggplot syntax.
What key suggestions would you give to a colleague before they used the activity in their teaching?
I think it is important that students have some familiarity with R before using this module. I'd also recommend using Rstudio Cloud to avoid some of the challenges of getting R and Rstudio to run on different operating systems. Rstudio Cloud allows students to run R without having to download it to their computers. I set up the module as a class assignment which worked really well.
How did you address challenges in teaching with the module?
Although Rstudio Cloud solved some issues, most of the challenges were still related to R. The code was more complex than they were used to and we had numerous cases of accidental insertion and deletion of code when they were trying to run it. I addressed this by constantly moving from group to group to help fix code. In the cases that couldn't be fixed quickly, I had students skip that section of code and use their neighbors output.
The module reminded students that statistical tests have assumptions that should be tested before applying the tests. Testing assumptions and data transformation was not something they had done before. The module also gave them more practice interpreting which models were useful in helping us understand the relationship between periphyton and water quality.
I liked that the module included the code for behind the scenes pre-analysis steps such as replacing -99999 with NA and averaging subsamples. This gave students some experience with data management. The code was very well commented so students were able to understand what each step was accomplishing without having to know all of the syntax. They were able to modify the code to complete Activity C which gave them confidence that they can use R for data analysis.