September 23, 2009

Expertise: Who wants some?

Let's talk about expertise and why acquiring more expertise might be a desirable goal of education.

But, rather than start with the premise that having more expertise is preferrable to having less expertise. (And, being an expert is preferrable to being a novice. And, performing more like an expert is preferrable to performing more like a novice (or non-expert)). Let's see if the premise is generally true.

What are the advantages of having expertise or being an expert, where an expert is defined as being more knowledgable than a non-expert?

Such an explanantion assumes three things. The expert has acquired more knowledge in a particular domain and that knowledge is organized and structured. The fundamental capacities and domain-general reasoning abilities of experts and non-experts are more or less identical. And, the differences in the performance of experts and non-experts are determined by the differences in the way their knowledge is represented. See Two Approaches to the Study of Experts’ Characteristics (Chi 2006)

Experts and novices have been studied extensively by cognitive scientists. As a result, we know quite a bit how experts differ from novices.  Chi distills the following seven advantages of being an expert in the above cited article.
The advantages of being an expert and having expertise.

  1. Generating the Best. Experts excel in generating the best solution, such as in solving problems or designing a task, even under time constraints. Moreover, they can do this faster and more accurately than non-experts.
  2. Detection and Recognition. Experts can detect and see features that novices cannot. For example, they can see patterns and cue configurations in X-ray films that novices cannot. They can also perceive the "deep structure" of a problem or situation.
  3. Qualitative Analyses. Experts spend a relatively great deal of time analyzing a problem qualitatively, developing a problem representation by adding many domain-specific and general constraints to the problems in their domains of expertise.
  4. Monitoring. Experts have more accurate self-monitoring skills in terms of their ability to detect errors and the status of their own comprehension.
  5. Strategies. Experts are more successful at choosing the appropriate strategies to use than novices. For example, when confronted with routine cases, expert clinicians diagnose with a data-driven (forward-working) approach by applying a small set of rules to the data; whereas less expert clinicians tend to use a hypothesis-driven (backward chaining) approach. Experts not only will know which strategy or procedure is better for a situation, but they also are more likely than novices to use strategies that have more frequently proved to be effective.
  6. Opportunistic. Experts are more opportunistic than novices; they make use of whatever sources of information are available while solving problems and also exhibit more opportunism in using resources.
  7. Cognitive Effort. Experts can retrieve relevant domain knowledge and strategies with minimal cognitive effort). They can also execute their skills with greater automaticity and are able to exert greater cognitive control over those aspects of performance where control is desirable.
Chi also lists seven ways in which experts do not excel.  One is especially relevent to education policy.  In fact, it's a doozy.

Context Dependence within a Domain.  Expertise is restricted to a specific domain. Moreover, within their domain of expertise, experts rely on contextual cues. For example, in a medical domain, experts seem to rely on the tacit enabling conditions of a situation for diagnosis. The enabling conditions are background information such as age, sex, previous diseases, occupation, drug use, and so forth. These circumstances are not necessarily causally related to diseases, but physicians pick up and use such correlational knowledge from clinical practice. When expert physicians were presented the complaints associated with a case along with patient charts and pictures of the patients, they were 50% more accurate than the novices in their diagnoses, and they were able to reproduce a large amount of context information that was directly relevant to the patient's problem. The implication is that without the contextual enabling information, expert physicians might be more limited in their ability to make an accurate diagnosis. Experts’ skills have been shown to be context-dependent in many other studies, such as the failure of experienced waiters to indicate the correct surface orientation of liquid in a tilted container, despite their experience in the context of wine glasses, and the inaccuracies of wildland fire fighters in predicting the spread of bush fire when the wind and slope are opposing rather than congruent, which is an unusual situation.

To put it more simply: expertise is domain limited. All the advantages of having expertise are limited to the domain in which the expert is knowledgable. In other domains, the expert performs like a non-expert.

This is a robust finding in cognitive science (See the Chi article for numerous research studies).

It seems to me that the advantages of having expertise far outweigh the few disadvantages.  And, I really can't see anyone desiring to engage in a period of study acquiring knowledge and skills and ending up performing no better than an novice.

Further, it seems to me that trying to acquire some enhanced domain-independent skills in lieu of acquiring domain-specific knowledge is a fool's errand and an incredibly risky proposition (not to mention a potentially large waste of time).  I'd like to see the research base for the claims being made by the proponents of  these generic 21st century skills.

Anyway, I'm throwing this out there for you to ponder and I'm going to try to delve a bit deeper into the expert-novice research adn what it means to be an expert and have domain knowledge in the next few posts.  I'm not exactly sure of the right path to take through all this, so if anyone has any suggestions or has already spotted a misstep, feel free to comment.

5 comments:

Dick Schutz said...

So far, so good, Ken! There are other forms of expertise in addition to "knowledge domain" expertise. Outside of academic domains almost all of these involve a combination of knowledge and "how to" skills. But Chi's warning that expertise doesn't transfer very well is generally applicable.

The price of expertise is practice, practice, practice. An advantage of arriving at a "duffer" level, is that you can gain about the same pleasure as an expert in participating without the investment of time and effort. One can be a "jack of a lot of trades," but one can't be "a master of all." And sometimes "expertise" can be deceptively fleeting. Look at Alan Greenspan and the "Bankers."

Downes said...

An interesting turn of discussion. I could comment more on the concept of expertise, but I'd rather raise an interesting point for discussion first.

You, Ken, are a patent attorney, at least, according to your bio. http://www.blogger.com/profile/06853211164976890091 So it follows that in your writings on education you are writing outside your area of expertise.

You write, "Expertise is restricted to a specific domain." This is my domain, Ken. Is it yours?

KDeRosa said...

Stephen, I think you are confusing credentialsim with expertise.

As you know, expertise is a function of domain knowledge which does not necessarily align with the credential conferred by study in a formal education program. Further, in education, I believe there is a bad mismatch between the credential and the real knowledge base. So, at least in the domain of education, I do not believe that the credential necessarily aligns well with expertise.

In any event, I am not an expert in the domain of education. I have acquired quite a bit of domain knowledge after ten years of study. So, I believe I have acquired some measure of expertise, but it does not rise to the level of an expert. Far from it -- which is why I am careful to stick to the knowledge base gleaned from credible sources: valid education and cognitive science research and successful instructional programs.

RMD said...

Ken . . . nice post. I don't know about expertise, but I'd love the schools to focus on "strong proficiency" in key skills: reading, writing, math, with an ease in these subjects that experts have in their areas of expertise.

Instead, from my parental point of view, they focus on "experiential" learning and highly inefficient techniques (i.e., projects) to teach materials. I afterschool my children because of the slow pace of instruction at their school . . . which many regard as a good school!

When I read debates about "21st century" learning and such, it sickens me. The so-called academics and researchers continue to set the bar lower, believing that Google will save our kids somehow. And the educational "experts" talk about the problems with testing and other concerns, instead of just figuring out the best way to teach (through research) and creating an environment where kids learn as much as quickly as they can.

Thanks for your terrific blog . . . I don't know how you continue to do it, but I continue to learn from it (I downloaded the NIFDI study you cited recently).

Robin said...

Ken-

I believe it is Dan Willingham who has said interventions improving a student's knowledge base are much more effective in showing lasting results that working on problem solving strategies.

For example very successful chess players rely primarily on a vast knowledge of successful chess moves and less on reasoning strategies.

Stephen-

Lawyers as a group are much more willing to acknowledge if something is outside their field of expertise or if they regard an analogy to be less than firm. They also tend to be forthcoming about why they believe what they believe. A good one doesn't make assertions they don't have solid evidence for.

Having a law degree is only the beginning of domain knowledge and not a shield to be used to stop any further inquiry. I believe that's why lawyers can be so impatient when others point to credentials to cut off discussion.