why we learn

January 3, 2010

I spent 3 months tutoring a middle school student during my senior year of high school, and ever since I’ve believed that the one-on-one, face-to-face environment is uniquely suited for instruction. These feelings have been reinforced, my interest piqued once more, as I have read about cognitive apprenticeship and Rogoff’s Apprenticeship in Thinking. Vygotsky, in his work Thought and Language is fairly direct in his criticism of society’s disregard for what students might be able to accomplish with others

…even the most profoundest thinkers never questioned the assumption; they never entertained the notion that what children can do with the assistance of others might be in some way more indicative of their mental development than what they can do alone (p. 85).

He then instructs the reader as to the qualities of what he deems “good instruction”.

Therefore the only good kind of instruction is that which marches ahead of development and leads it; it must be aimed not so much at the ripe as at the ripening functions (p. 188).

Here he alludes to his well known Zone of Proximal Development, and the fact that students need to be pushed just beyond what they are able to do on their own. Daniel Willingham makes a similar point early in his new book Why Don’t Students Like School.

Working on problems that are of the right level of difficulty is rewarding, but working on problems that are too easy or too difficult is unpleasant (Kindle Edition, 310).

Recently, I’ve made the point that monitoring this sort of thing in a classroom of 20-30 students requires that we classify instruction as complex in nature. My reverence for the tutoring paradigm is so great that I’ve focused my inquiries on ways of using technology to generate time for teachers and students to meet face-to-face. Today I read Open for Learning: The CMS and the Open Learning Network, an article in the first edition of the open journal in education written by Jon Mott and David Wiley. They suggest that social networks might be able to replicate Bloom’s 2-Sigma Effect, which demonstrates that tutored individuals outperform their peers by +2 standard deviations. They begin with a critique of course management systems, predicated on three assertions.

• CMSs reinforce artificial time constraints
• CMSs are teacher-centric, as instructors control both content and direction
• CMSs isolate students from the outside (real) world

Instead of relying so heavily on CMSs, Mott et al. suggest that universities utilize open learning networks, described as the middle ground between PLEs and CMSs. In an OLN,

faculty, students and support staff would reap the benefits of enterprise, networked software for authentication, identity management, integration with SISs, etc. Additionally, they would be able to use a vast range of Web 2.0 apps, integrated into the OLN via web services and other sorts of integrations.

I find this to be a practical, if evolutionary, suggestion. The authors’ argument that students understand that their work within the CMS environment will not be added to or viewed after their course, and that this knowledge depresses the desire to contribute is believable. There is no question that CMSs act primarily as a conduit for the transmission of content, and that teachers make the majority of the instructional decisions. Their failure, however, to acknowledge the existence synchronous/asynchronous forms of communication, collaborative writing modules (wikis), and the ability to share/comment on files, is unfortunate. However, it is their second-to-last sentence that I find most surprising. (The entire paragraph is included for context.)

While the CMS tends to reinforce the knowledge-transfer model, long deemed the best and only way to teach large groups of students, the OLN poses as an intriguing alternative. Instead of limiting ourselves to knowledge transfer, we can leverage the affordances of the web to uncover content, to help students become more than just temporarily knowledgeable about a subject. We can do so by approximating — and perhaps even surpassing — the efficacy of one-to-one learning relationships. As the tradition-preserving CMS gives way to the OLN and the learning affordances it brings with it, Bloom’s challenge may finally be met.

In general, research tells us that tutoring is not homogeneous in application (some tutoring is better than others), that the act of tutoring is beneficial (one learns as they teach), and that tutoring is a skill (it can be learned and improved upon). Yet, here – one sentence from the end of their paper – Mott & Wiley suggest that implementing OLNs will “approximate – and perhaps even surpass – the efficacy of one-to-one learning relationships”? Isn’t this a rather bold statement?

Tutoring’s ability to produce positive results might be attributed to a variety of features. From a behavioral perspective, one might point to the benefits of being able to interpret both verbal (inflection, hesitation) and non-verbal cues as a session unfolds. A cognitivist might suggest that the intimacy of the environment requires that each participant is attentive, this facilitating thought processes (activation of long-term memory, combination of new information with activated LTM in working memory). The teacher might simply say that they are able to monitor the learner’s ability and knowledge more closely and accurately. Regardless of the perspective, tutoring remains unique; a simple and uncommonly effective instructional technique, rooted in the history of the apprenticeship, but generally impractical.

As I said initially, I agree with much of what the authors have to say. I don’t like CMSs, and my vision of what online learning should be is even more liberal than what they suggest. I don’t have a answer to the 2-Sigma question. (I don’t know if there is an answer.) That doesn’t mean that replacing CMSs with OLNs is unwise, or would not be beneficial. But, introducing such a bold, unsubstantiated prognostication two lines from the end of a scholarly work serves no purpose.

Comments and/or critiques are welcomed.

The New York Times has an interesting article – http://j.mp/8aYXO6 – summarizing the extant research on the mathematical abilities of preschoolers. The findings indicate that children’s abilities are greater, and develop earlier, than had been previously thought.

The money quote:

“A crude number instinct us hard-wired into the brain”

Also of note, researchers surmise that math education “sharpens the firing of these quantity neurons”, which has a corollary in the cellular development in the visual cortex.

I know that my boys, ages six and four, have been doing some remedial math for a while now, as I have sometimes given them basic addition problems to pass the time while we drive. Fingers are utilized, but they can be pretty successful in finding the answer to problems such as 3+2.

Posted via email from Andrew J. Cerniglia

As I read chapter 9 in Ormrod¹, I couldn’t help but think of the parallels between the way we represent cognitive structures and computer programming. I’m not a programming guy, but I’ve written enough code and read enough books on the subject to see the similarities. For example, the hierarchical representation of information (p. 241) could be represented vary easily using XML.

<Animals>
    <Vertebrates>
        <Fish></Fish>
        <Mammals>
            <Cats></Cats>
            <Dogs>
                <German_Shepherds>
                    <Rin_Tin_Tin></Rin_Tin_Tin>
                    <Tippy></Tippy>
                    <Anna></Anna>
                </German_Shepherds>
                <Cocker_Spaniels></Cocker_Spaniels>
                <Collies></Collies>
                <Mutts>
                    <Tobey></Tobey>
                </Mutts>
            </Dogs>
            <Bears></Bears>
        </Mammals>
        <Birds></Birds>
        <Reptiles>
            <Snakes></Snakes>
            <Dinosaurs></Dinosaurs>
        </Reptiles>
        <Amphibians></Amphibians>
    </Vertebrates>
    <Invertebrates>
        <Mollusks>
            <Clams></Clams>
        </Mollusks>
        <Insects></Insects>
        <Creey_crawlies>
            <Things_Under_Rocks></Things_Under_Rocks>
        </Creey_crawlies>
    </Invertebrates>
</Animals>

Read more

I downloaded the audiobook of Blink, by Malcolm Gladwell from iTunes last night. I’ve listened to about a third of it today, and it’s wonderful. One of the byproducts of the information contained within is a softening in my position on behaviorism. Some of the illustrations that the author uses to explain unconscious thought scream classical conditioning. Or at least some form of conditioning. For anyone interested, the IAT (Implicit Association Test) is a good example of how, unconsciously most people pair concepts that may be the converse of their chosen or stated beliefs. I was particularly upset to find out that my height is such a demonstrative indicator of upward mobility as it relates to managerial positions. That is to say that out of the millions of CEO’s in the country, less than 10 are 5’5” or shorter (I’m 5’6”, on a good day).

An example of the IAT can be found (taken) here.

Fairly late in chapter 7, Ormrod¹ mentions the work of Dijksterhuis & Nordgren. I did some searching, and found a short summary of their work online. Their work is based on what they refer to as the “Unconscious Thought Theory”. The idea is that humans process information unconsciously, an idea that runs counter to the dual-store memory models which identify working memory, which itself is tied to the conscious, as the area in which processing occurs. In their studies, they investigate the differences in the decisions made by those individuals that consciously deliberate before making a decision, and those that do not. This is “crossed” with the complexity of the decision.

Their findings indicate that when tasks are relatively simple, conscious processing produces better, more satisfying results. Conversely, decisions that require the simultaneous rating, integration, and processing of multiple variables are better resolved by those individuals who do not consciously work towards a solution. Their experiments involved what I would refer to as a more traditional laboratory setting, in which participants were asked to makes decisions about the renting of different apartments and the purchasing of automobiles (two different studies) as well as more informal survey of shoppers exiting two retail stores (furniture – more complex and expensive, department store – less complex and cheaper). Participants in the first two studies made “wiser” decisions when given a more complex set of data when this information was processed unconsciously. Similarly, shoppers reported being more satisfied with their purchases from the furniture store (after several weeks) when they reported making the decision to purchase the item spontaneously.

The summary can be found here. There’s also a larger PDF file summarizing their work that isn’t too hard to find using Google.

1. Ormrod, Jeanne Ellis. (2008). Human Learning. Pearson.