The following is from Gabriel Weiderpass after TAing the lab for the first time in Spring quarter 2024.

“Hello Mark,

I TAyed the lab yesterday and you mentioned that you wanted some feedback on how it went. In general the experiment was great! It was interesting, getting the picture was non-trivial, the students took a reasonable time to finish, etc. Now more detailed comments about specific points

1) First the main problem that we faced is that the wiki and the phyton code were a bit too vague on what they had to do. So the students struggled a long time to even figure out what they were supposed to do.

For example after they “Split Channels to create three separate grey scale images from the R, G and B values.” they get three pictures in scales of R, G, and B. First, is the splitting in R, G, B the same as the one we defined on the phyton code, that is in Phyton we used 660nm for red, 530 for green, and 450 for blue. Is this what Image J is using? How do we even compare these R, G, and B splits of Image J with the phyton code? After that what are they supposed to do with these images? Can I get a table of intensity as a function of pixel distance to plot that on phyton? But then there is no part in the phyton code to import any data.

They were also very confused on what they should be doing on the last part. I told them that they were suppose to compare the peaks of their RGB split with the theoretical value of the plot above and therefore get the thickness of the film as a function of where you are in the image. This is what I understood and what I remembered from the lab training but it would be good if there was a bit more information on the wiki about their objectives.

2) The format of having the students read the paper beforehand worked. There were only two students (out of 11) that I noticed that hadn’t read the paper. One because he was destroyed in the quiz and the other because he was reading it when I arrived at lab.

3) Another small issue is that the students struggled to understand the importance of having good pictures to improve the data analysis. Even though I explained to them many times that having a good picture would produce a better reading in FIJI J and then would lead to more well defined peaks and make the whole analysis better, a lot of them didn’t got it. In the end I actually open the computer and analyzed a bad and a good picture and showed them the difference, in the bad picture the peaks are supper jagged and ill defined. It might be worth including in the wiki a small discussion on the importance of getting a good picture for the data analysis.

Those are all the points that I remember from yesterday.

Best,

Gabriel Weiderpass”