Preparing students for a research career: Gregory Abowd’s 30 PhD Graduates

Georgia Tech’s School of Interactive Computing did an article on my friend Gregory Abowd and his 30 PhD graduates, many of whom have continued in academia. You can find the article here.

The “Abowd family” is a real thing. The article ends talking about how Gregory and his students and their students get together at conferences. I’ve seen pictures of these events. There’s a strong sense of kinship and support in the group, inspired by Gregory.

Here at the University of Michigan, we have just hired two second-generation members of the Abowd family. Gabriela Marcu (see webpage here) and Nikola Banovic (see webpage here) both earned their PhD’s at CMU, working with former Gregory students Jen Mankoff and Anind Dey (who have now moved to U. Washington).  What’s striking to me about both Gabriela and Nikola is that they started down the path to academic research by doing undergraduate research with other Abowd graduates: Gillian Hayes at Irvine and Khai Troung at Toronto (respectively).

What does it take to support future academic researchers while they are still undergraduates?  Obviously, we don’t want all of our undergraduates to become researchers. But we need some. Academic researchers in computing perform a useful and important role. We particularly want more women getting into computing research, and kudos to Google for awarding fifteen grants to promote more women getting into computing research (see article here). We do not have enough CS academics today (as I described in this blog post), and that’s part of the struggle in dealing with the enrollment boom. So we want more — how do we get them?  What do we do at the undergraduate level to make it more likely that we get graduates like Gabriela and Nikola?

We need to expect that CS undergraduates will have careers other than software developers. We often build our undergraduate programs assuming that all of our graduates will become software developers, or will manage software developers. But you can do a lot with a CS degree. We have to build into our programs the features that will help students succeed in the career that they choose, including becoming academic researchers.

One of my colleagues in the Engineering Education Research program here, Joi Mondisa, researches mentoring. She just gave the first EER Seminar, and talked about the importance of being “treated/advised like family.”  Mentors give their mentees honest and valuable advice as if the mentee were a family member.

I suspect that that’s part of Gregory’s success — that the notion of being in the “Abowd family” is something that the members feel and actively participate in. That’s likely a lesson that we can use in the future. Personal mentoring relationships play a big role in encouraging future researchers.  I don’t know how to build personal “like family” research relationships into an undergraduate program, especially at the enrollment scales we see today. But it’s an important problem to think about, both because we should support a variety of outcomes for our CS undergraduates and because one way of managing the enrollment crisis is to grow more CS faculty.

 

The Backstory on Barbie the Robotics Engineer: What might that change?

Professor Casey Fiesler has a deep relationship with Barbie, that started with a feminist remix of a book.  I blogged about the remix and Casey’s comments on Barbie the Game Designer in this post. Now, Casey has helped develop a new book “Code Camp with Barbie and Friends” and she wrote the introduction. She tells the backstory in this Medium blog post.

In her essay, Casey considers her relationship with Barbie growing up:

I’ve also thought a lot about my own journey through computing, and how I might have been influenced by greater representation of women in tech. I had a lot of Barbies when I was a kid. For me, dolls were a storytelling vehicle, and I constructed elaborate soap operas in which their roles changed constantly. Most of my Barbies dated MC Hammer because my best friend was a boy who wasn’t allowed to have “girl” dolls, and MC was way more interesting than Ken. I also wasn’t too concerned about what the box told me a Barbie was supposed to be; otherwise I’d have had to create stories about models and ballerinas and the occasional zookeeper or nurse. My creativity was never particularly constrained, but I can’t help but think that even just a nudge — a reminder that Barbie could be a computer programmer instead of a ballerina — would have influenced my own storytelling.

I’ve been thinking about how Barbie coding might influence girls’ future interest in Tech careers.  I doubt that Barbie is a “role model” for many girls. Probably few girls want to grow up to be “like Barbie.” What a coding Barbie might do is to change the notion of “what’s acceptable” for girls.

In models of how students make choices in academia (e.g., Eccles’ expectancy-value theory) and how students get started in a field (e.g., Alexander’s Model of Domain Learning), the social context of the decision matters a lot. Students ask themselves “Do I want to do this activity and why?” and use social pressure and acceptance to decide what’s an appropriate class to take.  If there are no visible girls coding, then there is no social pressure. There are no messages that programming is an acceptable behavior.  A coding Barbie starts to change the answer to the question, “Can someone like me do this?”

Why Don’t Women Want to Code? Better question: Why don’t women choose CS more often?

Jen Mankoff (U. Washington faculty member, and Georgia Tech alumna) has written a thoughtful piece in response to the Stuart Reges blog post (which I talked about here), where she tells her own stories and reframes the question.

Foremost, I think this is the wrong question to be asking. As my colleague Anna Karlin argues, women and everyone else should code. In many careers that women choose, they will code. And very little of my time as an academic is spent actually coding, since I also write, mentor, teach, etc. In my opinion, a more relevant question is, “Why don’t women choose computer science more often?”

My answer is not to presume prejudice, by women (against computer science) or by computer scientists (against women). I would argue instead that the structural inequalities faced by women are dangerous to women’s choice precisely because they are subtle and pervasive, and that they exist throughout a woman’s entire computer science career. Their insidious nature makes them hard to detect and correct.

Source: Why Don’t Women Want to Code? Ask Them! – Jennifer Mankoff – Medium

International effort to improve data science in schools

I’ve been involved in this project over the last few months. (Where “involved” means, “a couple of phone conversations, and a set of emails about frameworks, standards, and curricula, and I missed every physical meeting.”) Nick Fisher has drawn together an impressive range of experts and professional societies to back the effort. It’s not clear where it’s going, but it is indicative of a growing worldwide interest in “data science” in schools.

The definition of “data science” is fuzzy for me, almost as fuzzy as the term “computational thinking.”  Does data science include computer science? statistics? probability? I think the answer is “yes” to all of those, but then it might be too big to easily teach in secondary schools. If we’re struggling to teach CS to teachers, how do we teach them CS and statistics and probability?

And if budgets and schedules are are a zero-sum game, what do we give up in order to teach data science?  For example, teacher preparation programs are packed full. What do we not teach in order to teach teachers about data science?

This group of experts knows a lot about what works in data science. Their opinion on what students need to know creates a useful measuring stick with which to look at the several data science classes that are being created (such as Unit 5 in Exploring CS). There’s some talk about this group of experts might develop their own course. I’m not sure that it’s possible to create a course to work internationally — school systems and expectations vary dramatically. But a framework is useful.

The aim of the International Data Science in Schools Project (IDSSP) is to transform the way data science is taught the last two years of secondary school. Its objectives are:

1. To ensure that school children develop a sufficient understanding and appreciation of how data can be acquired and used to make decisions so that they can make informed judgments in their daily lives, as children and then as adults

2. To inspire mathematically able school students to pursue tertiary studies in data science and its related fields, with a view to a career.

“In both cases, we want to teach people how to learn from data,” Dr Fisher said.

Two curriculum frameworks are being created to support development of a pre-calculus course in data science that is rigorous, engaging and accessible to all students, and a joy to teach.

• Framework 1 (Data Science for students). This framework is designed as the basis for developing a course with a total of some 240 hours of instruction.
• Framework 2 (Data Science for teachers). As a parallel development, this framework is designed as the basis for guiding the development of teachers from a wide variety of backgrounds (mathematics, computer science, science, economics, …) to teach a data science course well.

Dr Fisher said that the draft frameworks will be published for widespread public consultation in early 2019 before completion by August.

“We envisage the material will be used not just in schools, but also as a valuable source of information for data science courses in community colleges and universities and for private study.” For further information: idssp.info@gmail.com, or visit www.idssp.org

Applying diSessa’s Knowledge in Pieces Framework to Understanding the Notional Machine

In Lauren Margulieux’s blog where she summarizes papers from learning sciences and educational psychology, she takes on Andy diSessa’s 1993 paper “Toward an epistemology of physics” where diSessa applies his “knowledge in pieces” framework to how students develop an understanding of physics.  (See blog post here.)

The idea is that humans assemble their understanding of complex phenomenon out of knowledge of physical experiences, p-prims. Quoting Lauren:

Elements: P-prims are knowledge structures that are minimal abstractions of common phenomena and typically involve only a few simple parts, e.g., an observed phenomenon, like a person hitting a pen and that pen rolling across the table, and an explanation, like when people hit things, they move. P-prims are both phenomenological, meaning that they are interpretations of reality, and primitive, meaning that are (1) based on often rudimentary self-explanations and (2) an atomic-level mental structure that is only separated into parts by excessive force.

Cognitive Mechanism: P-prims are only activated when the learner recognizes similarities between a p-prim and the current phenomena. Recognition is impacted by many different features, such as cuing, frequency of activation, suppression, salience, and reinforcement. Because activation of p-prims depends on contextual features of phenomena, novices often fail to recognize relevant p-prims unless the contextual features align.

I find diSessa’s framework fascinating, and I’ve always wondered how we could apply it to students learning the notional machine (see blog post here on notional machine). My guess is that students use p-prims to develop their mental model of how the computer works, because — what else could they use? In the end, isn’t all our understanding grounded in physical experiences?  But using p-prims will likely lead to misconceptions since the notional machine is not based in the physical world.

Maybe this is a source of common misconceptions in learning computing.  The list of misconceptions that students have about variables, loops, scope, conditionals, and data structures is long and surprisingly consistent — across languages, over time.  What could possibly be the common source of all those misconceptions?  Maybe it’s physical reality.  Maybe students generally apply the same p-prims when trying to understand computing, and that’s why the same misconceptions arise. It’s sort of like using a metaphor to understand something in computing, but then realizing that the metaphor itself is leading to misconceptions.  And the metaphor that’s getting in our way is the use of physical world primitives for understanding the computational world.

Colleen Lewis, as a student of diSessa’s, uses the Knowledge in Pieces framework in her work.  In her terrific ICER 2012 paper, she does a detailed analysis of students’ debugging to identify misconceptions that they have about state. State is an interesting concept to study from a KiP perspective. It’s a common issue in CS, but less common in Physics. It’s not clear to me how students connect computational state to state in the real world.  Is it state like water being frozen or liquid, or state like being painted blue?  Do they get that state is malleable?

This is a rich space to explore in computing education. What are the p-prims for understanding the notional machine? How do students use the physical world to understand the computational one?

Read more of Lauren’s post here: Article Summary: diSessa (1993) Knowledge in Pieces Framework

South Carolina requires CS to fulfill high school requirement, and Keyboarding is no longer CS

Pat Yongpradit of Code.org shared some great news with me.  Well, it’s not really “new” — it happened back in March 2018. But it was something that both of us worked on, and it was great to finally see it happen.

South Carolina was one of the first ECEP (Expanding Computing Education Pathways) Alliance states. They had one of the first statewide summits on computing education (see blog post here). They were one of the first states to require computer science for all high school students.

The problem was that they didn’t actually require computer science. They allowed some 90 classes to count as CS, and only six actually contained CS content (like programming or algorithms). Even a course on “keyboarding” counted as “CS” under the South Carolina system. South Carolina resisted changing this requirement, as Tony Dillon of the state Department of Education argued (see this blog post). I’ve worried that other states that mandate CS would fall into a similar trap (see blog post here on that).

That changed March 28, 2018 with this memo. South Carolina has computer science standards. Keyboarding no longer counts.

It’s an interesting question how this happened.  I know that Pat and others at Code.org have been working a lot in South Carolina.  I know that our South Carolina ECEP collaborators, like Eileen Kraemer, Tiffany Barnes, and Mary Lou Maher, have been working tirelessly on the state. I also know that my involvement from Georgia had limited success.  As one Department of Education official said when I was working in Columbia, “No professor from Georgia Tech is going to tell me about AP CS.”

My suspicion is that this happened because there was significant internal and external pressure.  South Carolina wasn’t going to do much when it was just external pressure. But when it was both, there were changes made.

Pat has promised me that Code.org is going to be helping South Carolina fulfill their plans for new CS requirements.

 

Growth mindset matters for individual human performance, with a more indirect connection to academic success

One of the most talked-about papers at ICER 2018 was this one, “Fixed versus Growth Mindset Does not Seem to Matter Much: A Prospective Observational Study in Two Late Bachelor level Computer Science Courses.” The claim was that fixed and growth mindset did not have much impact on student course performance.  One of the authors wrote a blog post summarizing the paper.

In my opinion, they got growth/fixed mindset theory wrong.  The mistake is in the first line of the abstract, “Psychology predicts that a student’s mindset—their implicit theory of intelligence—has an effect on their academic performance.”  Growth and fixed mindset have an effect on individual student development. There is an indirect effect on academic performance which is more complex. Grades are not the same as measuring learning. Grades are typically a measure of mastery of concepts.

The presentation of the paper had this amazing graph (picture I took below).  Most students fail in the courses they studied.  Look at the big peaks in the distribution on the left. Those are all the fails.

In Freakonomics, there’s a chapter on why, if drug dealers make so much money, why do so many of them live with their mothers?  (The chapter is reprinted here.) The answer is that drug dealing (like professional sports or acting) is a “lottery” — many people try and make no money, and very few people get to the top and make lots of money.  All those high school and college football players who are waking up early to pump weights have a growth mindset — they believe that their effort will take them to the NFL.  However, the vast majority are *wrong*. They won’t make it.  There is no apparent connection between growth mindset and success.

That’s how I saw the ICER paper on fixed and growth mindset.  If the outcome variable is academic success, growth mindset isn’t going to always pay-off. Sometimes the deck is stacked against you, and even if you think you can win, you won’t.

However, if the outcome variable is individual development, growth mindset will likely beat fixed mindset.  If you believe you can get better, you might. If you don’t believe you can get better, you won’t. A good outcome variable would be learning gain, measured pre-test to post-test.  In this study, most students had a growth mindset, so they probably wouldn’t have seen much variation (between growth and fixed) even if they measured learning.

The students thought if they worked harder, they could do better. And they probably did all do better (from a learning perspective). They just weren’t going to win in this lottery.

It’s a different question whether a given intervention to improve mindset might lead to improved academic performance.  If you improve learning, and academic performance is reflective of learning, then there should be a connection IF it’s possible to change mindset with an intervention. Duckworth and Dweck have shown that they have successfully intervened to change students’ mindset and consequently improve academic performance, and that work was recently replicated (see post here).  The efforts to intervene on mindset in CS have had mixed success (see my blog post here on that). But it’s one thing to say that fixed vs growth mindset does not seem to matter much (the title of the paper presented at ICER), and another to say that a given mindset intervention did not result in academic performance increase. The first claim is about theory, and the second is about designing interventions with a multi-step causal chain. I don’t buy the former claim, but completely agree that the latter is a complex and interesting issue to explore.

Join us at the University of Michigan to study Computing Education

As of September 1, Barbara Ericson and I are new faculty at the University of Michigan.  The School of Information has a nice write-up about their new faculty, including Barb here. The Computer Science and Engineering Division (of the Electrical Engineering and Computer Science Department) wrote up their new faculty, including me here.

We are both looking to bring on new students in Ann Arbor.  Readers of this blog can find out a lot about us.  Barb continues to be interested in further developing interactive ebooks as a medium for education and in issues of broadening participation in computing (especially looking to grow Project Rise Up and Sisters Rise Up).  I continue to be interested in how students come to understand program execution (building a mental model of the notional machine) and in the role of programs to be a notation for learning (like in Bootstrap).

Because of how we’re situated at the University of Michigan, there are several avenues for new PhD students:

• I am in the Computer Science & Engineering Division and can advise students in the CS PhD program. https://www.eecs.umich.edu/eecs/graduate/cse/apply/
• Barb will be working with PhD students in the School of Information https://www.si.umich.edu/content/phd, where I can co-advise SI PhD students. Katie Cunningham (whose work is mentioned here) is coming with us to Michigan and will be joining the SI PhD program.
• I will be in the new Engineering Education Research PhD program and can advise students there. https://eer.engin.umich.edu/ph-d-program/

The EER program is hosting a prospective student open house on Oct 22, and there are travel grants available. See https://eer.engin.umich.edu/ for more information, and I have part of the flyer pasted below.

Choosing between the degree options depends on what you want to do with your PhD after you graduate, and what kind of preparation you want during your PhD. You can do CS Ed research via the CS PhD at Michigan (I did), and your preparation will involve classes in CS and a CS qualifying examination. SI is more oriented towards psychological and sociological perspectives on computing. EER will be more about CS and education in an Engineering context. If you want a CS faculty job, the CS PhD is the surest bet, but SI PhDs do get hired in CS departments, and we hope EER PhDs do, too. EER PhD opens up possibilities in Engineering Education departments (like at Purdue, Virginia Tech, Ohio State, and Clemson), where a CS or SI Phd is less common.

Michigan is becoming a seriously interesting place for computing education research.  Elliot Soloway (my PhD advisor) is still an energetic force in CS. SI just hired Ron Eglash (starting the same time as us), who is one of the founders of ethnocomputing and ethnomathematics (see news article here). I’m eager to collaborate with my friends in the learning sciences here, like Betsy Davis, Barry Fishman, and Chris Quintana. Do come join us!