Brief overview of DOE research projects

1
Brief overview ofDOE research projects

• Martijn P.F. Berger

Department Methodology and Statistics University
of Maastricht, The Netherlands
2
Applications

• - Educational settings
• Construction of achievement tests and
• questionnaires
• - Clinical trial settings
• Experiments designed to infer treatment effects
• - Observational settings
• Cluster randomized studies
• Neurological experiments
• Scans, EEG, MRI

3
Applications
Goals 1) Minimize the costs of performing the
study 2) Maximize statistical power for finding
real effects
4
Mixed effects models

with
and
(Generalized Linear mixed models)
5
Educational settings
Logistic IRT models
Optimal Test Design Select Items
from an itembank (design space) with maximum
information on (CAT)
Optimal Sampling Design Select students s
from a population of students with maximum
information on set of s.
6
Clinical trial settings
Repeated measures data - Allocation time
points, subjects - Covariance structures (AR1,
CS) - Costs and efficiency loss
7
Observational settings
Robustness against mis-specification - Model
parameters (degree of polynomials -
Covariates - Covariance structures (AR1, CS) -
Model assumptions - Costs and efficiency
loss Efficiency loss unequal cluster sizes
8
Collaboration

• Since optimal design research has many pros
• and can save a lot of money, it would be nice to
• join forces and
• apply for grants (European, national)
• have regular meetings such as these!