SSTiC 2014: Mark Guzdial on CS Education 2

There was no way I was going to blog David Johnson’s lecture on the travelling salesman problem (loads of cool empirical work): Subject is hard and David talks really fast šŸ˜‰

But here’s Mark part 2!

(We had a nice chat during lunch in which we poked at various things such as teaching how to teach CS esp to PhD students; various worries about peer instruction (some results show that showing the answer histograms between clicking and discussion doesn’t do a thing!; etc.)

This session is about education psych (in general).

Models of teaching: surprise, cs teachers generally don’t see themselves as nurturing or engaged in social reform.

Three common models: developmental, transmission, and apprenticeship.

Brown and Cocking 2000: How People Learn

Cognitivist Theories: learning = sense making

Not transmission.

Assimilation or accommodation. Accommodation requires new definition.

Models of memory


Make sense by creating connections between things we know and creating predictive mental models

Accommodation is hard

Easier to treat conflicting info as special case or ignore it as book knowledge

Whitehead: brittle knowledge

Lecture can work if students are trying to make sense

Test driven learning

Talent is overrated (motivation is most everything)

Duuuude! Gladwell? 10,000 hours?

Type decorations are hard to learn!

Transfer is hard (ie from situation to situation)

Marsha Lin.

Reflection: “Where would I use this next?”

Could either remember meta cognitive strategies or could recognize when a metacog strat (in general) would be helpful, but couldn’t remember what they were.

John Anderson, Rules of the Mind.

Students bind programming learning closely to syntax.

Schniderman: second programming language is harder than first (learning computation!)

Novices vs experts

Experts see patterns while novices see surface features

Experts know big ideas while novices can’t distinguish details from the main proposition

Teaching people how to learn (in abstracto) has failed

Sociocognitive theories of learning

Students want to become part of a community of practice

Legitimate peripheral participation (start at edges and move to the center)

(“Situated learning: legitimate peripheral participation”)

Expectancy-value theory

Eccles (1983) model of achievement-related choices

Dweck’s work on mindset

Fixed mindset vs growth mindset


Stereotype threat

Learning CS is hard

Rainfall problem (see Venable, Tan, and Lister, 2009)

Design of language makes a different!

Jim Spoor(?) built cognitive model for it

Proust (semantic debugging of rainfall problem)

Mike McCracken “Build a calculator”

Allison Tew (language independent validated test of intro CS knowledge)

MediaComp results are AMAZING.

Teaching computer arch on a gameboy: learning the same, but higher student liking of comp arch.