---

Please read chapter 2, "How Experts Diffeer from Novices", in the text. It provides some background.

The following synopsis is graciously provided by Dr. Glenn Smith
 
(This is a paraphrasing of the book: Protocol Analysis by Ericsson & Simon)

This is the short version.... try to get what you can from it.... the URL readings should provide more information.

An age-old question is how can you know what people are thinking as they solve problems. 

For example: 
How does a mathematics expert solve a geometry problem?
How does a high school student solve a geometry problem?
How does a chess master approach decide on a chess move? 
How does a beginning chess player decide on a move?
What does a person think as they use a search engine to find something on the net?

Accurate answers to questions like these can help provide better ways to help people learn. They can also provide algorithms so that computer programs can be written to do the same type of problem-solving. There is, in fact, a large group of scientist attempting to write computer programs that do what humans think (based on talk-aloud and think-aloud protocol analysis).These computer programs can then be used for intelligent tutoring. 
 
***
3-7-00

Here is an example a student suggested:
An example of an appropriate study to do with think-aloud methodology would be that of studying the approach used by a preschooler in tieing his/her shoe. The methodology would be to have the child say out loud each step used in accomplishing the task. The researcher can than evaluate the process and formulate a better or easier way for the child to accomplish the task.
***

Such information may also be used to help students to be more aware of their own thought processes.

Sometimes a simple, and seemingly obvious solution, has a lot of merit.

Talk-aloud protocol is based on the simple idea that to get at someone's thought processes while they solve a problem, you get that person to talk aloud what they are thinking, everything that comes into their head, as they solve a problem. This assumes that as someone solves a problem, they retain a small amount of information (7 plus or minus items) in a volatile memory store, their Short Term Memory (STM). The validity of protocol analysis rests on the assumption that if you can get at the contents of someone's STM as they solve a problem, you learn a significant amount about their problem solving process.

Two forms of verbal reports can claim to be closest reflections of the thinking processes (cognitive processes): 1. Concurrent verbal reports (talk-aloud and think-aloud reports) and 2. Retrospective reports.
 
We can speak of 3 levels of verbalization:

Level 1: when information is reproduced in the form that it is heeded…Level 1 verbalization can occur when the problem is a verbal problem. Thus the person can concurrently verbalize the information in the same form as it occurs in STM as they solve the problem. Talk-aloud is the term that describes this.

Level 2: Level 2 verbalizations are concurrent verbalizations where the contents of STM are non-verbal. The person must then translate the contents of STM into words, before verbalizing their thoughts. The contents of STM are not changed but are merely translated into a verbal form. Think aloud is used to describe this.

Level 3: Level 3 verbalizations refer to verbalizations where the content of STM is fundamentally changed before being vocalized. Retrospective verbalizations are level 3 verbalizations. Since the thought process is no longer occurring, the person must remember (recall from long term memory - LTM) and summarize what they were thinking. It certain that the original contents of STM will be altered from their form during the problem-solving, by the remembering process. For example, more than likely, the processes will edited leaving out dead-end reasoning.
 
Methodologically, protocol analysis involves a number of stages.

1. The participants talks aloud while solving some problem. This is recorded by audio and/or videotape.
2. The tape is transcribed
3. Data is segmented (divided up into statements or "thoughts"), usually numbered.
4. The investigator decides on an encoding scheme.
5. The data is encoded, i.e., each segment is translated into a symbol of a category.
6. The encoded data is analyzed. Analysis could be a frequency analysis or even looking for patterns. 
 
Protocol analysis can be educational. Teachers can model their thought processes for problem-solving in a realistic way (not in a cleaned up way, that presents an unrealistic expectation for the student). Students can learn about their own thought processes. Students can do discovery learning projects that are not authentic science (psychology).
Literature: What is a Shakespearean soliloquy, but a fictionalized talk-aloud protocol. Talk-aloud can be used as a tool for students in English to learn how to write realistically about how people think.

See what you can find in the way of articles about think aloud... they are out there.

Please go to these URLs and skim them. Try to get the gist of what they are saying...
 
Scaffolding: A Powerful Tool in Social Constructivist Classrooms
by Laura R. Roehler

Solving problems on the World Wide Web
by Nils Pharo 

Think aloud as a teaching tool: scaffolding: a powerful tool in social constructivist classrooms