Saw this article in one of the ACM blogs that cited a recent study in an educational psychology journal on how teaching introductory Cs students to program by having them write code actually has a detrimental effect. More details here.
I absolutely disagree with the author that this study has implications in CS! First of all it cites a study done with teaching algebra so it is a different domain. I think that the problem is that there is too little programming going on rather than too much. I also wonder of anyone has considered that an introductory programming class probably has a mix of two very different types of students - one just exploring CS as one of many career options, the other the kind that took their dad's computer apart in the garage and started programming at age 8. The latter kind still have to take the introductory class because of degree requirements but now it gives you a class sample that's skewed and hard to make any conclusions about.
If everyone that took a introductory CS class were also required to have done some programming at middle school or high school level, that would make a huge difference. I have looked at some of the so called "introductory" CS classes in my school district's website - it has stuff like using word processing/spreadsheet programs. That is like saying learning to drive is "Introduction to automobile engineering". So no wonder these students get to their first CS class and feel lost..
I absolutely disagree with the author that this study has implications in CS! First of all it cites a study done with teaching algebra so it is a different domain. I think that the problem is that there is too little programming going on rather than too much. I also wonder of anyone has considered that an introductory programming class probably has a mix of two very different types of students - one just exploring CS as one of many career options, the other the kind that took their dad's computer apart in the garage and started programming at age 8. The latter kind still have to take the introductory class because of degree requirements but now it gives you a class sample that's skewed and hard to make any conclusions about.
If everyone that took a introductory CS class were also required to have done some programming at middle school or high school level, that would make a huge difference. I have looked at some of the so called "introductory" CS classes in my school district's website - it has stuff like using word processing/spreadsheet programs. That is like saying learning to drive is "Introduction to automobile engineering". So no wonder these students get to their first CS class and feel lost..
In school in India, in my first ever CS Programming subject, the teacher used to walk through a super simple example program and then give us a much more complex problem to program ourselves. Most of us couldn't figure it all out ourselves and so I thought this teaching process was really stupid. So if what the author is trying to say here is that a "go do all this coding and figure it out all yourself" won't really help - I agree with him.
ReplyDeleteBut that is not to say it should all be theory. Just that it should be a balanced mix, unlike how it was in my school. Also, like you point out, the effectiveness of the process very much depends on whether the introductory class is really introductory to all students in the class. I was a TA for a Java 101 course at ASU and my class had the whole range from CS geeks to students who took Java because they thought it was yet another spoken language like French or Spanish! So I totally relate to how much the teaching process should cater to the kinds of students in the class :).
I had the opposite experience in my school back in India. The teacher never deviated from whatever was already in the book (think it was Pascal) so I used to think - wow this stuff is so easy, till I got to a different class and it was like starting from scratch all over. :)
ReplyDelete