Swarna Srinivasan's Public Library

In order to examine the dependence of test calibration on the abilities of
these law students, let us construct the worst possible situation. Into a Dumb
Group we will put the 325 students who did worst on the test. The best of them
got a score of 23. Into a Smart Group we will put the 303 students who did best.
The worst of them got a score of 33. Thus, we have two groups dramatically
different in their ability to succeed on this test of reading comprehension.
There are 10 points difference between the smartest of the Dumb Group and the
dumbest of the Smart Group.

Now for the acid test. How would a test calibration based on the Dumb Group
compare with one based on the Smart Group? To remind us of how things look using
the old way of doing things, I made up these calibrations in terms of sample
percentiles.

Alternative Approaches to Updating Item Parameter Estimates in Tests With
Item Cloning (CT-03-01)
by Cees A. W. Glas,
University of Twente, Enschede, The Netherlands


• Executive
  Summary
  
• Full
  Research Report

A Bayesian Method for the Detection of Item Preknowledge in CAT
(CT-98-07)
by
Lori D. McLeod, Law School Admission Council; Charles Lewis, Educational Testing
Service; and David Thissen, University of North Carolina at Chapel
Hill


• Executive
  Summary
  
• Full Research
  Report

The goal of the Gallant Lab is to understand the structure and function of
the human visual system at a quantitative, computational level, and to build
models that accurately predict how the brain will respond during natural vision.
Predictive models of brain activity are the gold standard of computational
neuroscience, and are critical for the long-term advancement of neuroscience and
medicine.

The research program in my lab reflects a tight integration of three distinct
approaches: neuroscience experiments involving both classical electrophysiology
and functional neuroimaging (fMRI); statistical analysis using methods adapted
from nonlinear system identification and nonlinear regression; and theoretical
modeling. Much of our research uses modern statistical tools to fit quantitative
computational models that describe how visual stimuli elicit brain activity.
Statistical tools drawn from classical and Bayesian statistics and machine
learning are used to fit appropriate computational models to these data. The
resulting models describe how each element of the visual system (e.g., a neuron,
a voxel or an entire visual area) encodes information about the visual world.
Models are evaluated both by statistical significance and by their ability to
predict brain responses under new conditions. This second criterion, accurate
prediction, is the gold standard of science and is fairly unique to our
approach.