SSTiC 2014: Mark Guzdial on CS Education 3

July 14, 2014

CS Education research

Bias: US learning scientist with mostly CS degrees

1960-1970s: LOGO
1990s?: empirical worked killed LOGO
xxxx?: MIMS? Multi-institution something studies

The two cultures and the scientific revolution (CP Snow)

Computers and the world of the future
Snow: scientists and decision making
Perlis: The computer in the university
Study of process and should be ubiquitous

Elias responds: do we really need programming

2007 Economist Business by the numbers

We’re controlled by algorithms we don’t understand

Earliest attempt, BASIC in 1964

1968-1980: Logo

1984: Pea & Kurland killed Logo (perhaps not appropriately)

1990: Palumbo meta-review; no correlation between problem solving and programming

1987: Sharon Carver Dissertation

Cog model

Some evidence that programming helps student learn science and math (Idit Harel Caperton)

Yasmin Kafai replicated

1972: Dynabooks; Personal Dynamic Media; meta-medium

Squeak 1995 to Etoys to Scratch

1986: Andrea diSessa and Hal Abelson aim to support computational literacy (blend of Smalltalk and Logo ideas; Boxer)

1984: Lewis Johnson “Proust” (semantic error checker based on Rainfall Problem)

1988: Studying the Novice Programmer

1990: Jim Spohrer’s “Marcel”

Cognitive model of a student programmer

ToonTalk

John Anderson and Cognitive Tutors (Lisp and Pascal tutors)

Geometry tutor as well

Kafai tested in schools and it was a disaster (totally alien and conflict with other media)

Cog tutors work (for writing code; but nothing about debugging or design)

Are graphical languages better than textual

Petre and Green’s bottom line: every syntax highlights and obscures different pools of knowledge

1991 Petre and Green: Textual languages were easier to read

1993 Moher

2004 Moskal, Lurie, & Cooper starting wig Alice improved retention and performance

2009 Hundhuasen

Does algo viz help

2002 Hundhausen, Douglas, Stasko. Meta analysis

Not really

Building helps; not watching

Multiple symbol system hard on beginners

1993: Hundhausen: Observation-based is the way to go

2001: Kehoe, Stasko, and Taylor: viz students work longer

Not much CSED in 1990s

Sally Fincher Bootstrapping model

Weird problem: all papers rejected from biggest CSEd conference (more about teachers talking together)

John Pane 2002

Commonsense computing

ACM SIGCSE starts ICER

THEMES IN CS ED RESEARCH

discipline based education domains

No univocal evidence for success predictors

The computer boys take over — Nathan Ensmenger

What skills do CS teachers need?

Read and comment on code (but not write much)

Measure of PCK (Pedagogical Content Knowledge): Sadler et al

Knowing content

Knowing what students will get wrong

Few women and members of minority groups in CS: Western world phenomenon

Stuck in the Shallow End, Jane Margolis

Peak in 1983 (40%) now 12%

How do we teach everyone computing

Code.org

Raymond Lister on development path for programming

Is teaching computing more like mathematics or more like science?

Inquiry based learning?

%d bloggers like this: