Georgia Institute of TechnologyCS Education Research Group
View this PageEdit this Page (locked)Uploads to this Page (locked)Versions of this Page over TimePrintable Version of this PageHome PageRecent ChangesSearchSign In

Begin Study Design: Research Questions, Research Methods

Notes from Jan 26

Here's what came up in our brainstorm last semester.
  • CS 2110: Explains how a computer works using a simplified imaginary computer
  • CS 2261: Explains how a computer works using a GameBoy Advance architecture, which is more complex
Which is better and why? Issue of motivation vs. complexity Why does the GBA emulator "work"?

I'm sorry I wasn't able to attend the meeting but I should be able to be there in the future.
Just for the record I'm not 100% sure that I agree with this characterization:

CS 2110: Explains how a computer works using a simplified imaginary computer
CS 2261: Explains how a computer works using a GameBoy Advance architecture, which is more complex
CS2110 does actually attempt to explain how a computer works and uses an imaginary computer for the
first half of the course: Digital logic (which isn't really imaginary), datapath, instruction set, machine language, assembler. Then for the second half they program in C on a typical multiuser Linux system. A lot of time is spent on the machine model. Very detailed. There is a sense that students are having a thorough exposure to the C language.

CS2261 doesn't really talk about how a computer works except at the simplest level. Sort of a basic block diagram of the Von Neumann architecture. No discussion of digital electronics except occasionally very peripherally. No datapath, no instruction set, no machine language, no assembly. Almost the entire course will be C programming. By necessity a number of features of computers are brought into play that don't appear in CS2110: Sprites, Direct manupulation of video memory and bit-mapped graphics, Sound, interrupts, DMA, etc.

I guess the point I want make clear is that I don't perceive that this is two courses with the same objectives and just a different presentation methodology.

Bill Leahy


High level concepts

(or other things we think would be in both courses)
  • How does information go from user input to memory?
  • How can a make choices based on input?
  • How does memory structure influence program structure?
    • DMA, cache, I/O
  • High level concepts about computer org–those things that we remember after we've forgotten all the little details

Other topics to possibly explore

  • Cultural split in 2261 between the advanced and normal lab sections.
  • Gender issues 2110 v 2261.
  • Effect of the "why" factor in 2261. Here the content is tied to a purpose, like MediaComp, do we see similar reported satisfaction about the course being relevant to their lives?

Little details to remember to ask

  • Where are they in the program? How many credit hours do they have coming in? (Hyp: 2261 is less scary so they take it earlier in their career)

Outstanding Issues?

  • Lack of specific content overlap. - Perhaps not as big of an issue as we thought, if we go from a high level discussion.
  • Simply more innovative teaching practices in 2261, like the separated labs.
  • Motivations for being in the class. For example, 2110 is fodder for the uber-nerds and they eat this up. How can we compare them? - Perhaps it's not about comparing anything. Discovering this exact cultural issue is enough.
  • Tendencies due to cheating. Cheating could be a cause for flunking that we'll never know. This could be problematic for IRB if we want to ask about it.
    • What are the collaboration policies in the two classes?

Summary of where we are now

  • Intended Themes to Study: Content, Culture & Gender
  • Survey at the end with some content (general)
  • Focus Groups (with pizza)

Notes from February 2

CS 2110
  • Combination of computer organization & C
  • Some students don't see relevance of machine/low level/computer org, just want C programming
  • Startling DFW rates for women
  • One design goal: make systems more fun - don't drive people away from pursuing courses in that area
  • Content: datapath, instruction set, state machine, memory, C programming (dynamic memory allocation, pointer, linked lists, function pointers)

CS 2261
  • Original design - architecture for games, and small C programs
  • Intro to low lwevel programming, with goal of awareness, recognizing that they may not engage in this for their life/career
  • Use GBA for relevance
  • Includes a lot about game programming - drives syllabus, harder because its real; no abstraction, hand waving
  • Content: game programming - focus on GUI; more open ended assignments, especially for Project #2

High Level Research Question
Real vs simulated architecture?

Preliminary hypothesis: Real architecture is harder (more difficult to understand, master, requires more time on task) but is more motivating

Notes from February 16

General Notes

  • Main question: What is the effect of using a real architecture? - need to operationalize
  • Do students learn similar things?
    • Hypothesis: Students will learn the same
    • Alternate hypothesis: Students in 2261 will learn fewer things more deeply
    • Should we ask questions that we only expect one of the courses to know?
  • What about engagement, motivation?
  • Perceived difficulty?
    • Hours per homework/project/week?
      • Meeting criteria vs. playing?
    • Hypothesis: 2261 > 2110
  • 2261
    • Have to deal with annoying stuff, that is taken out of "virtual" architectures
    • More visual, no assembly language
    • Less structured, more choice on project
    • Split recitation model
  • PRS is used in both classes

Research Questions/Items to Explore

  1. Investigate similarity/differences in learning computer organization concepts, such as
    1. Given three different instructions, which are realistic?
    2. Given assembly language, predict what will happen
    3. Explain a given architecture model
    4. Byte structures (short, long, int)
    5. Given C data, draw memory
    6. Given C code manipulating pointers, draw memory
  2. Investigate engagement/motivation with course content
    1. How much fun did you perceive the class to be?
    2. Would you recommend the class to others?
    3. Did you like the class?
    4. How did you feel about the real vs. virtual architecture?
  3. Investigate the (perceived?) difficulty of the course
    1. How many hours did you spend on the last 3 homeworks?
    2. How many hourse did you spend on the last N projects?
    3. How hard was it?
    4. Homework diaries?
  4. Investigate the recitation model, particularly the split recitation model in 2261
    1. Did you like?
    2. In 2261, which one did you attend? Did you like the split model? Why?
    3. Records of attendance should be kept by TAs - how many people? which people? Is there a change in pattern of attendance at some point or do the same people attend the advanced recitation throughout the semester?

Content Comparison (compliments of Bill)

(Capital X indicates intensive treatment)

Topic21102261
Course Overviewxx
Binary, Hex, Octalxx
Floating Pointx
Fixed Point x
Datatypesxx
Bitwise operationsxx
Transistorsx
Gatesx
Functional Unitsx
Von Neumann Modelx
Instructionsx
Assemblyx
Low level I/OxX
Bitmap Graphics x
DMA x
Timers x
Trapsx
Tiles x
Sprites x
Subroutinesx
Sound x
Activation stacksx
InterruptsxX
Recursion (Assembly)x
Basic C Ideasxx
Variables (Memory Map)Xx
Control Structuresx
C FunctionsXx
C Recursionx
Pointersxx
Arraysxx
File I/O (Basic)x
StringsXx
Dynamic AllocationXx
Structuresxx
Linked Listsx
Void & Function PointersXx
Unions & Enumsx
Game Stuff x

HOMEWORK:

For next week, everyone write 1 to 2 general research questions related to one or both of the themes above. Post them here on the swiki page using the box below with your name next to them.

Question 1 (Gender):
Is there a gender imbalance between the normal and advanced lecture section? (Are women more likely to come to a specific one or both?) If there is an imbalance, what do the women report as a reason for coming/not coming to one of the sections?

Question 2 (Content):
How "useful" do students in 2261 tend to view the material? What about 2110 students? (Self-reported usefulness based on both how useful they think it will be in their academic and professional lives) How does this interact with self-reported satisfaction with the class overall?

Lana

Question1 (Culture&Gender)
I agree with Lana about the gender imbalance question, and I think maybe we can touch it more around the “Cultural split in 2261 between the advanced and normal lab sections”. For example, besides the possible gender imbalance, is there any other division between the attenders of normal labs and advanced labs, such as if they feel about 2261 relevant to their lives or not? If they think this course is important or interesting for them? the potential difference about the retention in advanced labs? How it is related to their exam results?...
Lijun
How does the ability to choose your own project influence engagement?
Are people actually aware of the alternative to 2110?
Testing add.
From what I gather, Bill also does imropv comedy. He is super funny and always comes up with the craziest stories to describe the concepts that we are learning. If you cannot get the class with Bill, I have also heard good things about Tom Conte who taught the other section. Conte has no sense of humor at all!

Last modified 1 August 2014 at 1:58 pm by Mark Guzdial