An Overview and Some Concept Maps

Jan 20, 2014 • Greg Wilson

Meeting of the Software Carpentry Instructors Study Group
Round 8.1 (2014-01-16)

“Never mind flying cars: I’ll believe the future is here when we get web conferencing working.”

19:00 call

  • James Harmon (Chicago/OTG)
  • Greg Wilson (Toronto / Mozilla)
  • Matthias Bussonnier (IPython & Institut Curie France)
  • Matt Hall (Agile Geoscience, Nova Scotia)
  • Likit Preeyanon (Michigan State University, East Lansing, MI)
  • Martin Paulo (Melbourne/Australia, V3)
  • Martin Callaghan (University of Leeds, UK)
  • Fernando Mayer (University of São Paulo, Brazil)
  • Simon Belluzzo (Australia, University of Melbourne)
  • Melanie Segado (Montreal, McGill University)
  • Neem Serra (St. Louis, Asynchrony)
  • Dana Miller (Singapore, NTU)
  • Jess Hamrick (UC Berkeley, CA)
  • Sheldon McKay (New York/Reactome)
  • Evan Bianco (Agile Geoscience, Nova Scotia)
  • Helen Yezerets(Indiana University)

14:00 call

  • Greg Wilson (Toronto / Mozilla)
  • Daniel Chen (New York, NY / Mailman School of Public Health
  • Samantha Ahern (Cambridge(England) / University College London)
  • Atul Varma (Brooklyn, NY / Mozilla Foundation)
  • Mark Stacy (Norman, Oklahoma) / University of Oklahoma
  • Nikolay Koldunov (Hamburg / Institute of oceanography)
  • Andrea Zonca (San Diego, Supercomputer Center)
  • Rob Beagrie (London, Imperial College)
  • Brian York (Baltimore, STScI)
  • Victor Ng (Toronto / Mozilla)
  • Brad Taber-Thomas (State College, PA / Penn State)
  • Patrick Marsh (NOAA Storm Prediction Center [Norman, OK)
  • David Schryer (University of Tartu)

10:00 call

  • Greg Wilson (Toronto / Mozilla)
  • Gabriel Devenyi (Hamilton / McMaster)
  • JC Leyder (Spain / European Space Agency)
  • Jason Orendorff (Nashville / Mozilla)
  • Abigail Cabunoc (Toronto / Ontario Institute for Cancer Research)
  • Brian Miles (UNC-CH)
  • Chris Friedline (Richmond / Virginia Commonwealth University)
  • Benjamin Bradshaw (Austin, TX / GTECH Corporation)
  • Jeramia Ory (Wilkes-Barre, PA / King’s College)
  • Michael Gatto (Tucson, U of Arizona)
  • Brenna O’Brein (Toronto / HackerYou)
  • Stephen Turner (Charlottesville, VA / University of Virginia)
  • Isabel Fenton ( London / Imperial College)
  • Joon Ro (Austin, UT Austin)
  • Alexis Pyrkosz (Michigan State University)
  • Kai Zhuang (Denmark Technical University — I’m late)

Readings

Practice and Reflection

  • Concept map by Wednesday of next week (Wed Jan 22) on the WordPress blog (http://teaching.software-carpentry.org/category/concept-map/ )
    • Something related to Software Carpentry’s core material that you could teach in 10 minutes or less
    • Keep it rough! (spend time iterating content not aesthetics)
    • tell us who your intended audience is
    • about 6 bubbles…ish
  • Comment on at least two other people’s concept maps (by Wed Jan 29)
  • We will meet again in the same time slots on Jan 30 (you can attend any slot you want)

Notes

Overall theme: educational psychology drives our actual two-day content

The preface repeated the phrase, “evidence-based” teaching quite a lot.

Theme 1: We can use web tools ourselves during our learning how to teach

  • E.g. we use Etherpad in class for note-taking
  • Advanced learners who might otherwise be bored can do this instead of drifting off to facebook during the lesson
  • Do it during our meetings to get some practice
      1. better notes
      1. teacher can confirm we heard correctly
      1. get to know each other
        • sense of community aids learning
      1. one potential challenge here, as someone who uses etherpad backchannels a lot — i sometimes get distracted by the backchannel and fail to focus on the speaker
        • some meetings have intentional periods of “silent etherpadding” to free ppl of having to switch btwn these contexts for brief periods, which is interesting.

ties in with digital literacies agenda that a lot of universities are pushing

Theme 2: Novice vs. competent practitioner vs. expert

(progression of stages, compressed from 5 to 3, here, based on work by Patricia Brenner et al) http://www.amazon.com/From-Novice-Expert-Excellence-Commemorative/dp/0130325228/

Aside here, the myth of different learning “styles”

http://www.changemag.org/Archives/Back%20Issues/September-October%202010/the-myth-of-learning-full.html

  • Novices don’t know what they don’t know
    • They ask questions that don’t make any sense because they have taken the knowledge and put it into the mental boxes that they already have
    • Our goal is to create those mental boxes: a conceptual map — must be patient, because this is simply their mental model, not blameworthy.
    • Unknown unknowns
    • Novices often have incorrect existing knowledge, incomplete existing knowledge.
    • primary task is not to give information, just to give the correct model (for example Bash software carpentry 15 commands/3 hours)
  • Competent practitioner: more or less correct map of the problem
    • Can solve routine problems in a reasonable amount of time with reasonable amount of success
    • doing routine things “routinely”
    • Understands the rules, can apply them correctly
    • Competent people can teach themselves
    • All you need on a day to day basis
  • If SWC gets people here, we can consider it a success (want to get people from novice to here, not from competent -> expert)
  • Experts: know when to break the rules
    • able to recognize: “This is a special case.”
    • Knows when to apply rules in which cases.
    • Can diagnose and reverse engineer problems much better than competent practitioners (better at reverse reasoning [effect—>cause rather than cause—>effect], which is harder).
      • Mozilla interviews: diagnose problems. Better at identifying experts
      • Does Test Driven Development help competent practitioners debug better without having to reach expert level?
  • You don’t have to be an expert in a domain to teach
    • Actually better if you’re not: expert blind spot
      • you learn 80% of what you teach / explain to someone else, have found teaching things have made my knowedge much better
    • You can’t see the world in the eyes of a novice
    • And experts make things that are actually difficult without hours of practice sound easy, which is demoralizing, because the expert can’t see the problem from the point of view of the novice without conscious effort.
    • “What does it look like to a novice”
  • Timescales
    • Tens to hundreds of hours to go from novice to competent (critically dependent on quality of instruction)
    • Thousands to go from competent to expert (critically dependent on practice and reflection)
      • Reflection is important here, not repitition of the task — the metacognition is important here, thinking about what you’re doing, learning, etc.
        • there are different forms for this — am currently going back and researching this as we are developing a new style of course and reflection journals are to form part of evaluation, am writing a guidance document
        • Examples of differences in expert pattern recognition — Michael Hansen @ Indiana U, eye tracking of programmers

Theme 3: teaching is a learnable craft based on empirical science

“I can’t make you a great dancer.  Maybe I can’t even make you a good dancer.  But if you do what I say, I can make you a better dancer.” from [All That Jazz](http://en.wikiquote.org/wiki/All_That_Jazz)

  • Research shows that retention improves if instruction presents and reinforces linkages as well as concepts
    • It’s not just what you know: it’s how quickly you can find it
    • Novices don’t have all the key nodes, few connections; Competent practitioners have lots of nodes, some connections; Experts don’t necessarily have more nodes, but have lots more connections between nodes.
    • Chess masters don’t evaluate future board positions faster; they don’t see bad moves. Expertise = *recognition *of correct solution.
    • give the brain an anchor -> graph search problem. More linkages = faster associations & more likely to be able to jump to next step in the problem
    • connections between ideas just as important as ideas themselves
  • Human short-term memory is very small: seven plus or minus two items
    • Probably actually smaller (5), thus limit introducing new steps or concepts to no more than 6.
    • Dealing with this, reduce the options (forget), or chunking
    • Chunking allows us to store several items as one (e.g., the spelling of your name) [using pattern recognition]
    • People don’t notice typos in things that are chunks in memory (e.g. why it’s hard to copy edit your own writing)
    • (For those curious, the original paper: http://cogprints.org/730/1/miller.html )
  • Practice reinforces concepts and linkages
    • strategy: teach for a few minutes then allow time to practice — might be harder for the teacher but much better for the learner. — standard practice in UK schools, short sharp explanation/demo followed by practical task
  • A concept map shows ideas and their linkages
  • Many ways to use them:
    • Lesson design: what am I teaching in the next 10-15 minutes? How are the ideas related? In what order will I present them? The map helps not locking you into a presentation order determined by how you started writing down bullet points. You can design the lesson by navigating the map in a way it makes sense. ~6 bubbles, ~10 arcs is max students can absorb in one sitting.
      • try tracing different paths to plan your lesson order and find a path that minimizes forward references
    • Forward references are extra cognitive load: every time you say “later”, you lose people
      • avoid forward references as much as possible
    • Preparation: hand out the concept map to students before the lesson
      • they can doodle, take notes on the map, add linkages (use map as scaffolding)
      • useful if they can give you back the maps
    • Presentation: write on board as you go through the lesson (see Cherubini and Venolia paper in the readings)
    • Note-taking: have students do their own concept map while you talk (and then discuss them with neighbors) (but this can be distracting during class when they should be practicing programming exercising, but useful for traditional note-taking classes. But have to be taught how to draw concept maps, most people probably don’t know)
    • Feedback: have them do concept maps at the end of the lesson to show you what they think they learned (less so for programming, more useful for theoretical classes)
    • Concept map can work as practice after a short period of learning (5-10 min) in a more “theoretical” course
    • value of concept maps come from its creation

Theme 4: feedback

  • The more polished something looks, the less honest the feedback you receive will be
    • unless you post it to Reddit
    • So sketch by hand, take a picture, and post that
      • reasons:
        • 1) put effort into content of concept map
        • 2) get more honest feedback (they’ll hold their punches if they know you put lots of effort into it)
  • More generally, people have to learn to give and receive feedback
    • Musicians, athletes, architects, and debaters all expect this as a regular and significant part of their training, yet this is lacking for those in the sciences
      • used to use 2 stars and a wish with younger students to get them to be more critical and meaningful
    • Giving feedback essential to learning how to accept it
    • Eventually, able to give yourself feedback (reflective practice)
  • One of the great changes in programming in the last decade is the rise of code review
    • Talked about since the mid-1970s
    • But only really practiced since distributed open source development left us with no other choice
    • Now a must-do for most programmers
    • Spreads the knowledge, but has to be learned and practiced
  • The same is true for teaching
  • We want you to learn how to reflect on teaching practice so that you can help everyone get better, including yourself
    • Commenting on other people’s concept maps is just as important as making one yourself