Weekly review through May 27

Self-regulated learning framework

This week I spent a lot of time trying to find a framework for self-regulated learning that would work well for coding my data from interviewing self-studiers. I’ve compiled some stuff on a page called SRL model. I’m learning there are many ways to do a model or theory or framework or measure, and they have varying degrees of usefulness for my purposes. The ideal, as suggested by my advisor, is to have a model that basically shows the different states that a person can be in and then one can identify where there are breakdowns in the process. That’s the approach taken for the help-seeking tutor (Aleven et al., 2006, a diagram of the model), which led to improvements in an intelligent tutoring system in terms of getting people to use the help system in a way that was most beneficial to learning. The closest I could find from self-regulated learning is the model in Winne & Hadwin, 1998. Ideally, this model is also validated and empirically grounded in some way, i.e., the existence and distinctiveness of each stage can be demonstrated.

Other stuff

  • I read Confessions of a Public Speaker and collected some other advice on public speaking. I think the message is: practice your talk. I may even try recording my practices to get some authentic feedback. Which will make me vomit, but I think it’s worth it.
  • Read Clark & Mayer, 2011, which is an excellent overview of research that is important to designing computer-based learning interfaces. It’s also nice for free research questions when they go into questions that still haven’t been answered.
  • Curious what kinds of web-based learning interfaces are out there? I knew you were. Check them out: learning_tools:start.

Weekly review through May 20


I stopped using Shuff. My experience with Shuff has given me the following takeaways:

  1. I need to differentiate between standard to-do items (one-time tasks) and recurring tasks.
  2. The Do 5 system is nice for getting a chunk of one-time tasks done every day.
  3. Timeboxing is very often worthwhile. Especially combined with some data (even just casually observed) about how much can be accomplished in certain amounts of time.
  4. I want to have data that keeps me informed about my progress with things, but I’m not exactly sure what that is.
  5. URL lists are very handy for many of my activities.

For #1, I really like Habit List, which I just started today. You can specify all kinds of different frequencies (e.g. every three days, twice a week, etc.) for a task and then all of your tasks for the day will appear on a list. Each task has a calendar that shows when you’ve completed it, your streak, and so on, so that handles #4 in a new way that I like so far.

For #2, DailyDo.it is a nice clean interface, but it seems unnecessary to keep on top of Habit List. Maybe I’ll just use sticky notes with “Do 5” as a daily task on Habit List.

For #3, I just open a tab with http://e.ggtimer.com/20%20minutes or whatever amount of time I want.

For #5, I put my “articles to read” and “general reading” lists directly on the front page of my wiki, which I’m opening a lot more now.

To combine all of them is the job for a hypothetical Shuff 4, which is essentially a mashup of Habit List and Google Calendar. The main interface is structured around a single day. Like Habit List, recurring tasks appear in a task list in one column. You can also add one-time tasks here. In an adjacent column is a calendar for the day (ideally synced with GCal). You can drag the recurring tasks onto the calendar and adjust their length of time. Finally, at the top of the page is the standard Shuff timer task name and timer countdown, but now it’s running automatically — going through your tasks in real time. It will also continue with the value recording feature from Shuff 3, which allows the data collection mentioned in #3.

After the amount of time I put into Shuff 3 only to quickly abandon it, I’m going to do a lot of thinking before moving forward with it. I encourage my blog readers to be extremely critical and let me know any foreseen problems with this.

Today I also ran into a really awesome app that basically functions as a Shuff-task-but-better for one of the most fearsome recurring tasks of all: checking email. The app is Email Game, and it turns your Gmail inbox into a fast-paced game. Maybe it’s a novelty effect, but I can’t describe how powerful it feels. I think the key may be that it bypasses the inbox list and having to focus on exactly one small thing. That’s kinda what Shuff was supposed to be, but there is a line between processing email and scheduling a whole day.

Individual differences in learning systems

Lately I’ve been referring to my research interests as DIY learning, after inspiration from Stuart Card’s talk at CHI mentioning DIY communities like Make and Quantified Self. My main interest in web tools is not just in how they can replace or supplement classrooms but how they can be employed for individual learning needs. I think that will involve things like self-tracked_learning, goal-setting, and a way of picking and choosing among tons of content.

I’m working towards some broad questions to get started: 1) to describe the space of possible interfaces, 2) to understand the individual differences that might influence how to build these. For the latter I was inspired by a blog post by Mark Guzdial about Turadg’s defense. I think it helps to divide up the individual differences into categories. Where the most controversy exists is probably that of “learning strategies”. In short, my claim is that it’s better to be system-driven than user-driven here. I’ve written some more on my wiki.

Weekly review through May 13

Patterns for writing

According to Thomas Basbøll, an academic writing consultant, the standard scholarly article is about 40 paragraphs, in eight sections of five paragraphs each. One should schedule 30 minutes for each paragraph, and use that time and only that time to write the paragraph. Writing is a process of mastery (or getting into shape), and the path is writing every day, between 30 minutes and 3 hours per day. Given these constraints, one can calculate what writing goals to set over the course of the 16 weeks of a semester.

Since grade school, I’ve forgotten the notion of using the paragraph to express one idea. I just kind of start writing and hit enter a couple times when it feels right. After checking out Basbøll’s blog, I tried writing a class paper by outlining it paragraph-by-paragraph and found the process very satisfying.

I’ll try the goal-setting to plan out my paper due in September. To get to at least 30 minutes a day, I also plan to schedule a 30-minute block to free-write about any topic that crosses my mind and seems to warrant further fleshing out (not limited to a single paragraph in this case).

Another advantage of the 30 minute block is taking breaks from sitting after.

Wiki reboot

I plan to use my wiki as the base for this writing as well as a collection tool to (re-)replace Evernote. As you can see, I redesigned the template for minimalism. Eventually I’ll reinstate the search and a way to get back to the home page.

The advantage over Evernote is mostly in the hypertext nature. It feels much less like a black hole to be able to link to things within a category by writing about them on a category page rather than just having a big list of stuff in a category.

I’ve already been shaping the wiki environment to meet some of my other goals. I moved “to read” papers over from Shuff, as well as links to frequent websites. My “recent changes” is divided into recently read papers and other changes, so that both can be reviewed for the weekly review.

There are still problems. The biggest is finding the balance between having a good structure and being able to quickly add things. Right now I’m trying to keep important things on the home page, but I don’t have process for preventing that from getting overwhelming or outdated.

Persuasion, habits, goals

These have been imminent topics the last couple weeks, especially attending the Personal Informatics workshop at CHI. There are several problems with the persuasiveness of personal informatics tools: 1) the goals of the designer may not match the goals of the individual, or even be socially desirable, 2) persuasion generally requires the user to take action, but habit change is difficult, 3) the suggestions to the individual may not take into account enough local information about the individual.

All three are deep problems. The first and third are related to the idea of legibility in the book Seeing Like a State, which I’m currently reading. A gross simplification of the idea is that optimizing a single variable can ruin a system because it ignores other things that turn out to be important. Brynjarsdottir et al., 2012 had an amazing diagram for this in their CHI presentation, but sadly I can’t find it! Anyway, more later.

The Power of Habit (Duhigg, 2012) addresses the second problem. The first thing to know is that habits, as defined in the book, are unconscious and affect our behavior extensively. Basically, they are anything for which we have an automatic response to a cue. That response generally gives us some kind of reward, which is what ingrains the habit so deeply. (Actually, this sounds a lot like procedural learning in general, where I guess correctness is the reward.)

The book gives a framework for habit change, which I’ve repeated in my notes. (The isolation of cues and rewards is vaguely related to some writing I did today on memory and context.) Habit/routine creation is another means to behavior change (similar to the ideas in that other “Power” book, Loehr & Schwartz, 2006). For myself, I think this is a potentially bountiful avenue since so many things that I do that should be routine aren’t. The writing blocks are a start. Changing the environment seems like a powerful way of changing cues that wasn’t really addressed in the book.

Goal-setting may also be an important mechanism for habit change, and, in some cases, goal achievement could be thought of a habit itself. One of the findings in a major review by Locke & Latham, 2002 is that part of the goal process is not just using the goal as feedback on progress but also setting a further goal once one has been achieved. That seems to be the process of setting up a new reward. In Visible Learning (Hattie, 2009), goal-setting was found to be a powerful intervention for education. I plan to write some more about this next week.