models of human knowledge structures. These diagramming techniques can be
used to map cognitive structures and explicitly show their relations to observed
behavior.

Papers in the second part of this section describe actual research studies re-
lated to cognitive process instruction. The papers by Robert Karplus, et al. deal
with tasks for assessing formal reasoning skills. The finding that many college
students are not able to give immediate correct responses to questions on writ-
ten tests designed to tap formal reasoning skills has prompted a growing number
of attempts at interpretation. The Karplus papers provide an excellent in-
troduction to these tests and the issues surrounding their interpretation. They
are the only papers in the book which deal directly with secondary students, and
they are included because they suggest certain limits on the type of preparation
we can expect college freshmen to have. The first study compares the perfor-
mance of students from seven western countries on tasks involving proportional
reasoning and the control of variables and raises some serious questions about
the level of attainment in certain reasoning skills.

Karplus' second paper is a pilot study using one of these tasks and was
conducted in the People's Republic of China. His preliminary findings show
that the performance of Chinese students may be significantly higher than that
of western students on this task.

Jill Larkin provides a concise overview of recent important work in informa-
tion-processing psychology relevant to cognitive process instruction. She
describes a number of ways that this work can be applied to science instruction.

The last three papers in this section by Simon and Simon, Clement, and
Lochhead give the reader a close-up view of individual students thinking out
loud while solving problems. These papers are empirical in emphasis--they in-
clude verbatim sections of interview transcripts and concentrate on identifying
new phenomena to be observed during the process of problem solving. In this
respect they are poised at the frontier of our knowledge about complex cognitive
processes--they exhibit behavior that can be explained currently only in broad
theoretical terms, with many subtle aspects of the behavior being as yet
unexplained.

In the past, most research on problem solving has emphasized relatively
simple tasks in which the problem is unambiguously defined and in which the
required knowledge base is small and specific. Dorothea and Herbert Simon in
their paper, however, examine the protocols of two individuals working on a very
open-ended engineering design problem. They describe striking differences in
their solution approaches to the same problem. The individual differences high-
light some important features of problem solving in open-ended tasks and raise
some important questions concerning the relevance of the engineering curric-
ulum to real-world problems.

John Clement's paper analyzes a student's responses to an open-ended ques-
tion of a different kind. The student is asked a series of questions about the role
of friction and momentum in a simple mechanics experiment. This amounts to a
request not so much to decide on an action which will solve a problem, as to give
an explanation or theory of why the apparatus will behave in a certain way.
Explanation protocols of this kind are particularly useful for analyzing the
structure of the conceptions a student is using to interact with a particular
situation. The causal structures described in this paper are of interest because

-8-