I've been reading Garvin-Doxas and Barker (2004) on communication patterns within computer science classses and it is fascinating to me, as a CS lecturer and a former CS student. For readers not in these geeky categories, it's still interesting because it gets to the heart of why sometimes even the best laid lesson plans for active learning encounter fail horribly. It's all to do with the unspoken prevailing culture for communication in classes within a department. If the normal form of interaction in classes is not supportive, classes which are based on collaboration and discussion will not work because people still behave in defensive ways. In fact, I think this kind of stuff should be included in lecturer training courses.
There are some technical definitions here which a worth getting straight to start with (if you want to check them out in detail see p4 of Garvin-Doxas and barker 2004). It's based on Gibb's 1960 work on a communication climate continuum: where there is a spectrum of communication patterns going from defensive to supportive. In education, both teachers and learners behaviours contribute to the climate of communication, but teachers are the ones who can change it because they typically have most influence. Typically higher education classrooms are stuck in a defensive pattern according to Gibb. (Although this may be changing. I hope.)
In a defensive climate, people are evaluative, controlling, strategic, neutral, superior and certain. And we talking learners as well as teachers, by the way. Two of these behaviours are of particular interest here; evaluative and neutral. Evaluative behaviour is when the person passes judgements on an idea or person, assigns blame, and questions other people's standards. Neutral behaviour is when the person fails to express concern for the people they are talking too: they are clinical rather than engaged.
A supportive climate occurs where people are descriptive, problem orientated, spontaneous, empathic, equal and provisional (ie. focuses on exploring a problem, and hearing other's points of view instead of enforcing their own). Descriptive means that the person avoids passing judgement. Problem orientated styles are when the person does't imply blame or lack of worth and expresses that they are working with others to solve a problem.
Students pick up what styles of communcation are acceptable in class very early by watching how the lecturer runs a class and how the other students behave. As I have discovered, once learners are set to expect one style, it is hard to retrain their expectations. A particular problem with defensive climates is that it reduces learners' sense of belonging, and female students in a male dominated environment are likely to be put off, not least because they are typically less confident of their skills. So we need to try to avoid this style in first year computer science classes where women are likely to be a minority (9% in my dept) and may choose to drop out if they don't feel they fit in. To be clear: I'm arguing that it is the culture in computer science classes which needs to change, not women. They will vote with their feet if they don't like it, to the loss of the field as a whole.
The researchers did ethnographic research in first year computer science class for a 3 year period in a research intensive university. They did classroom observation and interviews with staff and students. What they found was uncomfortably familiar to me. What I didn't realise was that this behaviour is perhaps unusual in other disciplines. Here are a couple of examples:
- The TA told the students "This assigment is going to be really easy. it took me fifteen minutes". Students found it hard and it took them 4 hours. When they mentioned this to the lecturer, he said "it's not that hard. You should be able to handle it since we've worked on this topic since the start of the course". The staff show lack of concern for students' difficulty and implies if students can't manage it they haven't been paying attention or working hard. When I was in first year programming classes I ended up in tears because it was so hard, and I thought everyone else found it easy. If you've ever learned to program, perhaps you can remember back to how emotional it can be: frustrating, demoralising and sometimes elating. (There is a term "prograsm" which refers to this elation. But that is just creepy. :-))
- The lecturer mentioned at the start of the class that some students are "rocket scientists" and celebrates examples of their success. This seems fine at first glance, but the problem was that the lecturers confused experience with intelligence. They didn't explain to the class that these rcoket scientists could do the advanced programming examples because they had prior experience. So the other students with no experience worried that they were stupid because they couldn't manage to write such programs themselves yet even though no one had taught them how. Often there is a range of prior experience in first year programming classes, and it is very hard to manage this as a teacher. As a student, I felt stupid. No lecturer ever told me that other students could do the assignments because they had been programming in their bedrooms since they were 10. No lecturer ever told me that I could learn to program too, but it would take practice and not to worry about it.
- Some students (mostly men) had a "strutting" behaviour, which means that they show off their knowledge/experience of computing as much as they can. It has the effect of reinforcing less experienced students' worries that everyone knows more than they do. An example might be of a student interrupting the lecturer with a fake question of the sort "But wouldn't it be better to use a function to do X?" I say fake question, because the strutter knows the answer already but is merely trying to show that they know this advanced concept which has not yet been covered. Or if another student gives an answer, a strutter might say "But wouldn't it be more elegant to do X?" A strutter might also point out syntax errors on the lecturer's slides. It's quite difficult to deal with this successfully because you don't want to exclude other students, but you don't want to crush the strutter who may have difficulties of his own. I have taught a few strutters, particularly in MSc classes.
Learning about strutting behaviour combined with the issue of different levels of background experience made me think about how best to support group learning for first year CS. It can be supportive for less confident learners to program in pairs, for example, and this is recommended in various articles relating to women in computing. But what if you end up in a group with a strutter? What if your group has a defensive communication climate, rather than supportive? From my experience, there is nothing more irritating than working with an arrogant nerd with poor social skills (and Big Bang Theory viewers will see what I mean). I can still remember two particular villains from my ugrad degree, who shall remain nameless. One was a patronising git and told my friend "Judy's a very nice girl but she just can't program C in her head." Another accused me of "over playing the girly card" because I happened to have a pink sports bag with me. Grrr.
Two practical steps might reduce potential strutting problems 1) Arrange for single gender groups (Vekiri 2008)and 2) Don't have group work outside the class until you have trained the groups in class ( modified from Wilson 2008). This can help women who might feel uncomfortable meeting men outside class hours, and you can make sure that people are behaving supportively in groups.
On a positive note, in my view my current university has fewer strutters than other hardcore mega-research oriented CS departments. My own ugrad university was filled with strutters and staff who were grown up strutters who colluded with them. (And also some loveable cuddly geeks like my husband in case he is reading this). It is an odd form of nerd machismo. If you work in another discipline, let me know: do you encounter alpha geeks of this sort?
Garvin-Doxas, K. and Barker, L. J. 2004. Communication in computer science classrooms: understanding defensive climates as a means of creating supportive behaviors. J. Educ. Resour. Comput. 4, 1 (Mar. 2004), 2. DOI= http://doi.acm.org/10.1145/1060071.1060073
Wilson, B. 2008. Improving comfort level of females in the first computer programming course: suggestions for CS faculty. J. Comput. Small Coll. 23, 4 (Apr. 2008), 28-34.
Verkir, I. (2008). Paper prepared for the OECD Expert Meeting on Gender, ICT and Education, Oslo 2-3 June 2008. Retrieved 21.4.09 from http://www.oecd.org/dataoecd/43/28/40832756.pdf.