More diversity and more progress with CS teachers vs just on-line: Code.org
February 25, 2015 at 8:03 am 5 comments
Hadi Partovi of Code.org has a blog post (see here) with data from their on-line classes. He’s making the argument that classroom teachers are super important for diversity and for student success.
Learning #1: Classrooms progress farther than students studying alone
In the graph below, the X axis is student age, the Y axis is their average progress in our courses. The blue line is students in classrooms with teachers. The red line is students studying without a classroom/teacher.
Learning #3: The ethnic backgrounds of students with teachers are impressively diverse
The data below doesn’t come from all students, because (for privacy reasons) we do not allow students to tell us their ethnic background. This chart was collected via an opt-in survey of teachers in the U.S. offering our courses, and as such is susceptible to inaccuracy. The picture it paints helps confirm our thesis that by integrating computer science into younger-aged classrooms in public schools, we can increase the diversity of students learning computer science.
Entry filed under: Uncategorized. Tags: Code.org, computing education research, high school CS, K12, teachers.
5 Comments Add your own
Leave a reply to astrachano Cancel reply
Trackback this post | Subscribe to the comments via RSS Feed
Computing Ed Research - Guzdial's Take 
1.
shriramkrishnamurthi | February 25, 2015 at 8:13 am
I honestly do not understand the first graph (the one w/ Learning #1). Four year olds do so well, they don’t recapture that attainment level until about age 8 (in a classroom) or ever (outside)? Somehow turning 5 causes their attainment level to dramatically plummet? No group _ever_ gets beyond a 55% attainment level? (Is this an average across all members of that age?) Also, when they get to college age something is going on (maybe they’re too busy with college, whereas they had time to spare in high school?).
Do you really understand this? Do you trust this graph? (Could it be that there was only one super-class of four-year-olds, for instance?)
2.
Mark Guzdial | February 25, 2015 at 12:17 pm
I don’t think that the first graph is showing us learning. It’s showing us “average progress in our courses.” It’s not the case that every course/level/activity that is experienced leads to learning, but it is the case that you don’t learn from a course/level/activity that you don’t experience. I’ll bet that that the amount of learning per segment is much less before age 12 than after, so I’m pretty much ignoring that end of the graph. (The evidence that I’ve seen so far suggests that Piaget was right — kids figure out abstraction in the pre-teen years, and any “programming” before then tends to be without abstractions like variables and functions.) I can believe that the average user at a given age doesn’t cover more than 55% of the levels.
I interpret the results the way Hadi describes (caveat “learning” vs “progress”). Classrooms with teachers go further than those without teachers. That’s an interesting and useful result, especially for those who don’t see a value of teachers.
The second result in the post meshes with Yasmin and Debbie’s results (see here), so I buy that.
3.
astrachano | February 25, 2015 at 8:21 am
Shriram — data shouldn’t be questioned, but accepted with the interpretations provided by those with the data (tongue in cheek, in case that requires saying, couldn’t find an appropriate emoticon) There are too many questions about the data, and there’s no validation of it, to be able to understand, yet alone trust, the graph.
4.
shriramkrishnamurthi | February 25, 2015 at 10:53 am
[Clicks the figurative Like button]
5.
astrachano | February 25, 2015 at 8:22 am
And of course I meant “no validation of *them*” since data is plural