The Carnegie Mellon Curriculum on Causal and Statistical Reasoning (CSR)
concerns causal claims and the scientific process by which they are established,
particularly where it involves the use of statistical evidence. All students,
for example, should be able to critically assess a newspaper report of
a study that "links" the amount of violence watched on TV with
anti-social behavior. Although most colleges and universities offer introductory
courses in statistical or research methods, these courses give little
in the way of a qualitative framework within which students can reason
about causal claims of this sort, claims which must inform our medical
and social policies.
The material in CSR is delivered in three integrated pieces, which we
describe in more detail below:
1
Concept Modules
2 Causality Lab
3 Case Studies
The Causality
Lab (along with an Exercise Builder) can be freely downloaded and
used independently. The Case Studies can be viewed from this guest site
(see the panel at the top), but to sample our concept modules, you must
create an account on the OLI-Jcourse
system, and then use the account to log in. A full
set of instructions for creating the appropriate account are available
if you wish them. Guests can do any of the modules in Area 1: Causation.
Encapsulating the basic results of research done over the last 20 years
by statisticians, philosophers, and computer scientists, our curriculum
presents a simple but rich theory of causation, distinguishes causation
from association, presents the obstacles to establishing causal claims
from associational data and explores the strategies for doing so. The
curriculum is suitable for introductory level courses on critical thinking,
quantitative methods, research methods, or formal reasoning. It is equally
suitable as a first course of a two-semester or two-quarter introductory
statistics sequence, to be followed by more conventional introductory
statistics.
As part of the Open Learning
Initiative at Carnegie Mellon University, our course in Causal and
Statistical Reasoning strives to be an example of widely accessible and
effective online education.
The Concept Modules deliver content on causation, association, how they
connect, and many other topics, but they are nothing like standard textbooks.
Although each module (there are almost 20 total) is approximately the
length of a textbook chapter, after every page or so they present the
student with a few questions, a simulation, an applet, or some other interactive
material.
Most of the material covered in the modules is not traditional statistics,
but rather complementary to it. Instead of teaching statistical formulas,
we motivate the students by putting them in a scientific position where
the need for statistical apparatus is obvious. There is an enormous amount
of web-based material on more traditional statistics education already
out there, much of it quite good. In the Related Links section, we provide
an annotated guide to the major pieces of which we are aware.
The Causality Lab is a simulated laboratory for setting up, carrying out,
and analyzing experiments or observational studies. It is designed to
put the student in the place of the scientist, providing hands on experience
in the scientific pursuit of causal knowledge. Many of the Concept Modules
use the Causality Lab for interactive exercises.
We have now collected almost 100 short case studies in a repository organized
by topic and by the concepts illustrated. One of the goals of our courseware
is to help students become critical consumers of the "studies"
reported in the media. When a newspaper reports that researchers have
established that time spent in day care as a toddler causes more aggressiveness
later in life, for example, we want students to be able to separate causal
theory from associational data, and to be able to critique the move from
data to theory. The case studies in the repository, many of which are
used in exercises within the concept modules, are meant to provide practice
at applying the concepts taught to cases in the real world.
Funding for this project has been provided by the Fund for the Improvement
of Post-Secondary Education, Carnegie Mellon, as well as the Andrew W.
Mellon, the James S. McDonell, and the William and Flora Hewlett Foundations.