# Mark Guzdial and Barbara Ericson

200302 @ [CCSE](https://www.mn.uio.no/ccse/english/about/news-and-events/events/2020/guzdial-ericson-seminar.html), University of Oslo

# Mark Guzdial

Talk is about "Teaching computing to reach a broader audience."

C.P. Snow the two cultures and the Scientific Revolution - the humanities/STEM split.

Computer Science was invented to teach _everyone_ about _everything_.

Terminology
* Computer Science: study of computers and all phenomoenn associated with them
* Computing: CS + CE + SE + IS + IT + ..., CS facing outward
* Programming: reading and writing in a notation for computers [missed some here]

Giving a historical overview from first mention of 'Computer Science' in the 60's and 'computing for everyone'
* Forsythe
* Alan Perlis
* Seymour Papert and Logo
* Andrea diSessa and Boxer, Computational Literacy
* Bruce Sherin and Boxer

## Three stories

### 1. Media computation at Georgia tech

Making computing (CS1) relevant for students, focussing on liberal arts, architecture, bussiness. Making material and assignments personally meaningful for students. Pass rate imporved a lot with the new course.

### 2. Changing pedagogy - teaching more than software development

Subject in Media Computation course: How does digital sound work? Demonstrating program written in block-based language GP to visualise sound waves, using his own voice, a harmonica and a ukelele.

Demonstrating JES, the IDE from the media computation course allowing for changing aspects of a sound(file). Asking us to _predict_ the outcome of a change. based on work from Eric Mazur, demo asking for prediction leads to mosty learning relative to demo without prediction or no demo at all.

Computer programs are also a useful way to communicate about process

### 3. What do students really need to know to use computing effectively?

Rich/Franklink study on K-8: https://dl.acm.org/doi/10.1145/3105726.3106166.

Working now with 'Task-specific programming': how can we use programming to enhance learing.
Tool: Data Visualization for Literacy. For example, 'discover' Rwanda genocide from life expectancy data (compare LifeExp data from Rwanda wirth another country).

# Barbara Ericson

Earlier work with teacher professional development, CS Unplugged, other efforts to help increase CS teaching to kids.

## Current work
* interactive ebooks for teachers and high school students (OER!)
* parson's problems
* helping underrepresented students succeed

Runestone _interactive_ Ebooks: 17 free ebooks. Running code 'in' the book, including Python Tutor.

Worked on adding other things:
* worked examples with by-line explanations
* 'worked examples plus practice'
* low-cognitive load practice
* multiple choice questions
* clickable code (click on these or these aspects of the code)
* Parson's problems (PP) (randomized code, ask to order correctly)

Introduces Cognitive Load Theory. Completion tasks are better than whole tasks. Influenced by prior knowledge.

Did research on Parson's problem. How do people interact with interactive elements in Runestone's 'How to think like a computer sciencist' book.. Parson's problem were used a lot!

Compared Parson's problems (PP) versus Fix the Code versus Write the Code. PP was fixed easier than Fix than Write. Learning gains (longer term) across all conditions.

Developed "Adaptive Parson's problems"
* help button after 3 incorrect solutions, make problem easier to solve
* choice of next question affected by how easy solved previous ones

Tested on teachers. Adaptive PP preferred over non-adaptive.

Ebooks have logs files of what readers do, studied those. Adaptive PP 2x more likely to get correct. Can ask what are the most common errors in solving PP.

Studied how teachers use the Ebooks and their opinions on it. Ebooks have audio tours on how to talk about code (!).

Now adding Open-Ended questions. 'Describe an array'; 'What questions do you ahve about arrays?' and then did analysis on responses. Got a lot more than expected. Called 'Learner sourcing' - using learners as a source for improving materials.

Power of Prediction - part of the Ebook.

## Future work
* interactive version of Python for Everybody
* add peer Instruction to Runestone
* personalize eBook content
* study cognitive laod using eye trackers

# Q&A

* writing PP? written in Restructured text (rst)
* BE: one thing to start with? Peer Instruction and PP
* MG: one thing to help achieve "reaching a broader audience"? Worried about computing (CS education) in schools. Catch the hype, teacher developemnts, integrating CS in other subjects.
* MG: do we need to havem or avoid, a task-specific programming language for each topic? Working on LiveCode with Elise Lockwood, very simple 'programming language', throw-away afterwards. Idea is to learn quickly, but use only short-term