Negative Social Stereotypes and Deep Learning

The next chapter to cover in my series of reflections on Ken Bain's What the Best College Teachers Do (started here and here) is on what teachers expect of their students. The opening story for this chapter should resonate with many readers here: it's about how negative social stereotypes affect student performance.

One of the interesting things that Stanford social psychologist Claude Steele found was "that the negative stereotypes sometimes had the strongest influence on the students who had all the confidence in the world, had not internalized any sense of inferiority, often had excellent preparations, and really cared about doing well academically." This, of course, would apply to the whole 'women in computer science' issue we talk a lot about.

It seems that a big factor in helping tackle these stereotypes is simply expecting the right level of performance from students. Don't go too easy on them and forgo learning just to get good reviews, but also don't pile on the homework and leave students feeling overloaded and alienated for the sake of increased difficulty.

More specifically, there are some tendencies underlying the best teachers' practices:

1. "Look for and appreciate the individual value of each student."
2. Have "great faith in students' ability to achieve."
   
3. Focus on the outcome and represent authentic goals with "trust, rejection of power, and setting standards." Replace the notion of power with the creation of opportunities.
   

In other words (or at least in my understanding), whether or not any stereotypes have affected students in the past, you start from scratch, assuming that each is capable of the very best. Yet you don't give up on them if they can't achieve it right away; rather, you mentor them, and help them get there.

All of this is said to "rest on an even more fundamental bedrock of ideas about the nature and meaning of learning." They key point from this section was that learning is not filling your head with information but rather constructing models of reality.

How can computer science be taught in this way?

I suppose I've touched upon my answer before, particularly in my teaching reflection. There are so many opportunities for activities ranging from the interactive CS Unplugged style to a construction of reality through carefully planned participatory code creation in class. Rather than be given information about how something works, let students slowly piece it together until the light bulb (hopefully) lights up. (And if it doesn't, more mentoring and helping is key, as mentioned above.)

I wonder if a shift in classroom practices for computer science would help retain more minorities? The research presented in this chapter suggests that it would, and fits with many of the other things I've read on the topic.