Monday, September 22, 2014

Digital Collaboration using Google Sheets

One of the tenets of active learning is to help students construct their own knowledge. To accomplish this, one approach is to have students collaborate during learning. My experience is that fostering an environment in which students are comfortable asking questions (and bringing questions to class) can be difficult, but it can be easier to have students perform small group work both in and out of class.

I like in-class group work for four reasons:
1) doing group work in class forces the instructor to be very clear about goals and the process that might be taken to reach the goal in a discrete amount of time
2) it is perhaps easier to convince students to be more involved in, and thus benefit more from, group work when you're face-to-face with them (they've already decided to attend class that day, so they're often committed to making the most of that investment)
3) the instructor can bring the groups together by the end of the class period to hear groups report out, to help students synthesize, and to help summarize the learning progress that occurs. This process should strengthen understanding of the material
4) having face-to-face discussions, which can be more efficient that electronic communication, can easily occur in class - there is no need to require groups to meet physically after class

Exercise Overview

When I wrote my first group exercise, which I delivered last week, I had one main goal: that it leverage the tablet. I picked a topic, probability and statistics, that I thought would be ideal for a tablet group exercise. We had been talking, in class, about Mendelian genetics, in which alleles of genes inherited from two parents combine randomly to produce variation in the subsequent generation. I wanted students to generate their own genotype data set that would digitally mimic the processes of meiosis (gamete production) and fertilization and then to compare their digital observation to the expected values that we can calculate from Mendel's two laws. This comparison takes the form of performing a chi-square analysis of the expected and observed data.

Set-up

Each group of five students was (digitally) given a set of instructions for the exercise. In advance, I had asked all of the students to download a free coin-flipping app (there are many available for all platforms).

Day 1

I walked the students through the analogy of flipping a coin (50% heads, 50% tails expectation) to what happens during meiosis in an A/a heterozygote (Mendel's First Law - segregation: we have a 50% chance of a gamete carrying one allele; 50% chance of the other). Thus, two students represent the two alleles of one gene. In each round of the exercise, both flip their digital coins: heads and heads = A/A, heads and tails = A/a, and tails/tails = a/a. Mendelian genetics suggests that these genotypes should occur in a 1:2:1 ratio.

To extend the analogy to Mendel's Second Law (independent assortment), two other students represent the alleles of a second gene (B) also brought together from two different gametes at fertilization. heads/heads = B/B, heads/tails = B/b, and tails/tails = b/b. Considering all possible combinations of alleles at both genes (1:2:1 ratio independently derived of another 1:2:1 ratio), one expects to see two-locus genotype results of the product of the two ratios (the 1:2:1:2:4:2:1:2:1 genotype ratio of a dihybrid cross).

What is the fifth group member doing while the other four are coin-flipping? The fifth member is the scribe: s/he is recording the results in a Google Sheets spreadsheet that will later be used by the group to analyze their data (the real reason for doing this exercise on tablets instead of on paper).

Tangent: Pre-set-up and Google Sheets

What did I have to do before class to set up this exercise? In addition to writing out the instructions for the coin-flipping process (although next time I will probably have students devise their own method for a way to obtain random one- and two-locus genotypes; I expect many of them will still arrive at coin-flipping as a good proxy), I also did some basic formatting of a Google Sheets template. The biggest issue was how to distribute the template to all of the students, while a) ensuring that only group members had access to each one, and b) not requiring me to manually "Share" each Google Sheet with each group member by adding their e-mail address via what might be called the "normal" process of sharing a Google Sheet:

1) click the blue "Share" button

2) enter e-mail addresses (preferably gmail.com addresses) of those you want to share with


















Instead, I received some great advice from a colleague, who suggested that I deploy the Google Sheet template via our learning management system (LMS; in this case, Blackboard). Here's the way that process worked:

A) I created the master Google Sheets template

B) I made a copy of that Google Sheet (and renamed each) for each group


C) In Blackboard, I created the number of student groups (group set) I wanted to have, and then enabled student self-enrollment (I didn't want to pick who was in each group).


D) After setting up the Groups, I returned to each Group's Google Sheet and selected the blue "Share" button. I opted to change the security settings to those seen below, where the only people who can access this file have to 1) be logged in to Google Apps via a fresnostate.edu gmail address (because Fresno State has a relationship with Google) and 2) have the link to the file. Note that I've also set the option that those who satisfy the above two requirements can edit (not just view, in which case they won't be able to modify the template - which is the point of doing this!)
E) Then I copied the URL for each group's file and pasted it into an initial thread that I created in each Group Discussion Board. Now, only the members of each Group can see that Group's file.

Day 1 (still) - face-to-face collaboration

Again, the scribe has the Group's spreadsheet open, and is recording coin flip results from four other students. After recording 16 rounds of four flips, the data generation phase is complete.

Post-class collaboration

This concluding the first class session, I asked each group to, after class, collaboratively count the number of "heads" (H) and "tails" (T) flips for each round and tally the number of occurrences of each of the sixteen possible combinations: HHHH, HHHT, HHTH, HHTT, etc. Then, each group converted the heads/tails nomenclature to genotype nomenclature (e.g. A/a; b/b). The great part about using Google Docs (or Sheets) is that these files are truly collaborative: all group members, if accessing the Google Sheet at the same time, can edit the document simultaneously, regardless of where the students and their tablets are located.

Day 2

In the next class meeting, I first walked the class through an example chi-square analysis: how to use formulae entered into the Google Sheet to calculate the chi-square test statistic value for each comparison of observation with expectation. We then discussed how to identify the number of degrees of freedom and the p value associated with the chi-square test statistic value. Then, I once again asked the students, in their Groups, to perform the chi-square test on their data.

We then walked through an analysis of the data, interpreting p values and determining whether any group's data were significantly different than the Mendelian expectation. Because I am the owner of all of the group Sheets, I opened each one and could display for the entire class the formula that each group member had entered to calculate chi-square values. Then, by copy-pasting, I created a new spreadsheet on the fly, in which I combined all of the individual group data into one total data set. Performing the chi-square analysis on these data allowed us to compare how an increased number of observations converged more exactly on the expected ratio. Doing all of this in class, with the instructor and students together, is a safe environment for students to explore and ask questions about p values (what they indicate, how to interpret p < 0.05, what "reject the null hypothesis" means, etc.)

Conclusion

While the set-up of this class was a bit tedious (creating eight different groups and the associated template files), I think that the effect on the students was strong, especially because it represented an authentic experience in genetics: tabulating data in a spreadsheet and using the spreadsheet to perform a statistical analysis. This is definitely an exercise we would not have done (or at least would not have as intimately involved each student in active learning) had each student not had a tablet computer in class.

Up next on Tablet Pedagogy

My approach for tackling common student mistakes made during chi-square analysis!

No comments:

Post a Comment

Have an insightful comment, best practice, or concern to share? Please do!