Sunday, June 16, 2013

Instructional Strategies That Work, (and Don't)

In this monograph (link), Dunlovsky, et al discuss 10 learning techniques in detail and offer recommendations about their relative utility. The techniques include elaborative interrogation, self-explanation, summarization, highlighting (or underlining), and the keyword mnemonic, imagery use for text learning, rereading, practice testing, distributed practice, and interleaved practice. The study evaluated whether their benefits generalize across four categories of variables: learning conditions, student characteristics, materials, and criterion tasks. 

Here are five common study strategies that, according to the evidence, really don’t work very well:

Evidently, the value of this scheme in retaining new material is tied up with how you do it. Unless you apply critical thinking, highlighting does not help understanding or retention. In my own experience, underlining and writing marginalia seem to enhance my understanding of the material at the time I'm reading it, but may or may not really boost retention of transfer.

Essentially, they found that unless the second read is more critical than the first, merely revisiting text has little lasting value for retention or transfer.

The keyword mnemonic 
A keyword is one that arbitrarily tags a single mental image with an assigned abstract meaning. This is an instance of tying the new to the familiar. The study shows poor long-term results, particularly in foreign language learning, where it seems to be quite common.

Self-generated Imagery
This consists of students being directed to create their own images of the principle or abstraction under study. The study did find that some domains (math and science) were more amenable to imagery than others.

Two somewhat more effective strategies

Elaborative interrogation
Merely prompting students to answer “Why?” questions about any new material can facilitate learning it. The particular form of the explanatory prompt is “Why would this fact be true of this [X] and not some other [X]? Students posing questions to themselves about similarities, differences and causal links around the unfamiliar evidently activates application of the familiar, weaving the new learning into the old. The more precise and elaborate that weaving is, the more effectively we can retrieve the new content later.
This goes a bit beyond merely asking why about new facts, abstractions or procedures to prompting students to self-generate logical rules to apply to the new material. Part of the value seems to come from speaking the “rules” aloud as one applies them. This strategy might be more helpful in rule-driven domains like math and physical science. 

Three highly effective strategies

Practice testing
This can include any form of low-stakes test from a formal Q & A to flash card practice. The Dunlosky team found substantial support from learning theory (going back to Thorndike) and experimental results to show it works across all ages and learning settings.

Distributed practice
Particularly when compared with “cramming”, practicing new material in short bursts, over time is by ar the better method for retention.

Interleaved Practice
This is an approach covenes with distributed practice (above) when closely-related materials (say, computing volumes of several different solids) are practiced in conjuncture with one another, as opposed to devoting single-topic blocks of time to solutions to a specific type of problem. Interleaving works, they surmise, by reinforcing attention, not to just the subject matter (a given formula) but the distinctions between related issues. In another domain, identifying different artist’s styles of paining, interleaving seems to strengthen discriminatory skill as well as the ability to simply recognize an individual artist’s work. Interleaving even seems to improve performance in motor skills as well.

There is more to learning and transfer than mere repetition. One wonders, thinking about what we typically see in K-12 and Adult Education, just how often we utterly fail to offer students any effective personal study strategies beyond just that!


Monday, June 10, 2013

Eight Ways Of Looking At Intelligence

Anyone concerned with teaching and learning must consider intelligence, not as a fixed capacity, but one always subject to instances and context. To that point, Annie Murphy Paul blogged "Brilliantly" today (link), summarizing and re-organizing the emerging science of learning. I've recapped her categories here with a few thoughts of my own: 

Situations can make us smarter. Intelligence is as intelligence does. Children and adults alike learn and perform in a context, and we, as "context builders" and managers have a lot to do with learning outcomes. As I consider the content I want to "present" to the learners, I must also think about the arrangements I might make for that content to be relevant, actionable, and memorable.

Beliefs can make us smarter. Martin Seligman, Angela Duckworth, and Carol Dweck have taught us that non-cogntive factors, like "grit" and "optimism" hugely affect learning & performance outcomes over entire lives. These are not some vague attributes, nor are they fixed. The self-told story matters. We can help habitually pessimistic students re-frame their experience in a more productive way.

Expertise can make us smarter. Expertly-held knowledge is different from that of the newbie. Experts are those, who over thousands of hours of trial and error, have converted slow-thinking neural paths into fast-thinking, pattern-recognizing raceways to performance. In performance, the marginal differences between experts and nons are largely reflected in unconscious processing. 

Attention can make us smarter. We live in attention-starved times. The ubiquity of screens today is the teacher's greatest challenge. To be a master teacher is an artist at attention-getting and attention-steering.

Emotions can make us smarter.  Just as optimistic renderings of one's life story need not be accurate to be fruitful, the feeling of hope is irrational (given the dark facts of our existence) but it works. Mirth, solidarity and pity can all drive us to learn and perform better. 

Technology can make us smarter. Language evolved to allow humans to store and share knowledge outside of their personal skulls. Writing accelerated this capability and printing put us into orbit. Now we have to watch it, as the proliferation of gadgets (see "screens" above) may make us dumber and less attentive. 

Our bodies can make us smarter. We need sleep. Our brains need food. More more deeply than Paul suggests here, all our knowledge, even the most abstract must be grounded in concrete experience. Even as adults we cannot grasp new ideas without the metaphors based in embodied knowledge. 

Relationships can make us smarter. Humans evolved to dominate the planet by collaborating. Twenty-first century man remains a deeply social animal. Social capital figures more deeply in the differences in achievement that physical capital, and certainly more than sheer indiviudual intelligence. Our institutions define and support us as learners and performers.

"The science of learning suggests that we ought to imagine our roles—as parents, as professionals, as learners. We should aim to be situation-makers—creators of circumstances that evoke intelligence in ourselves and others." -AMP