TILE
The Toolkit for an Interactive Learning
Environment (TILE) project, initiated at Berkeley,
involves both the design and
implementation of a new style of interactive
educational modules for teaching introductory
statistics for in a multi-media environment.
Rather than using the computer to teach
statistical procedures and formulae directly,
the focus is on involving the student
in statistical reasoning.
The role of the statistical
procedures is to to allow the student
to explore and justify their thoughts.
Numerous labs have been designed covering
different aspects of the basic introductory
statistics courses.
3 of these labs will be implemented by the 2nd
quarter of 1998.
These are:
-
investigating a selection of experiments
that are described in newspapers and journals to understand
the ideas of bias, controls, randomization, populations,
-
identifying relationships between variables and
confounding factors in different contexts such as
birth weight. The student gathers different types
of information (statistical and non-statistical) from
numerous sources in a virtual hospital
and uses plots, cross-tabs and intuition.
-
probability models and hypothesis testing using
the basic idea of a sampler.
The student controls a particular random
process and then creates a sampler that models
the distribution of outcomes.
The empirical distribution is created
by the student. This same idea is extended to
explore hypothesis testing.
These labs
are being implemented
using the Java based toolkit
being developed as part of this project.
The toolkit provides a flexible and
customizable set of classes
that are used
to create each lab.
These provide a consistent interface
both to the user and developer/instructor team
and can be glued together relatively
simply to create new labs.
Most of the time developing a new lab
is spent in creating non-programmed
input files which govern the behavior of the
objects in the toolkit (e.g. the animation,
text and hints, quizzes, colors, fonts, etc.)
This can be done by instructors
focusing on the pedagogical content
rather than on the programming language.
The toolkit is context-free in the sense
that it can be used in domains
other than statistics.
The design of the labs has been primarily completed
by Professor Deborah Nolan at U.C. Berkeley.
The implementation of the toolkit is being done
by Duncan Temple Lang at Bell Labs.
Duncan Temple
Lang<duncan@research.bell-labs.com>
Last modified: Wed Nov 12 14:09:47 EST 1997