B6613 - Optimal Design of Experiments

Learning outcomes

At the conclusion of this course, students will be familiar with the foundations of experimental design and will be able to: -construct optimal or good designs for a range of experiments -evaluate and compare designs using common optimality criteria -understand the challenges and potential practical problems in the implementation of optimal designs -describe advantages and disadvantages of different designs. Examples from several fields will be provided to show the wide applicability of the optimal design techniques.

Course contents

Statistical experiments and the role of the design: outcomes (primary endpoints), factors and their levels.

Fundamental principles: randomization, replication and blocking.

Clinical trial protocols and the phases of clinical trials.

Design of Experiments with one factor:

1. constant (to determine the sample size);
2. dichotomous (balance and Neyman allocation);
3. polytomous (analysis of variance)
4. quantitative (linear regression and factorial design with two levels)
   
   

Design of Experiments with two (or more factors): factorial designs

Optimal design theory

-continuous and exact designs

-the general equivalence theorem

-criteria of optimality: A- optimality, D-optimality and C-optimality

-brief mention on the sequential construction of optimal designs.

Readings/Bibliography

A. C. Atkinson, A. N. Donev e R. D. Tobias (2007). Optimum Experimental Designs, with SAS. Oxford Statistical Science Series 34.

Slides and exercises will be available on the platform "Virtual learning environment".

Teaching methods

Traditional lectures

Lecture attendance is not compulsory but is strongly recommended.

(https://corsi.unibo.it/2cycle/StatisticalSciences/lecture-attendance).

Assessment methods

Non-attending students: oral examination at the end of the course.

Attending students may choose one of the following assessment options:

• an oral examination at the end of the course; or
• the presentation of a paper on a topic related to the course and agreed upon in advance with the lecturer. The paper must be presented with the support of slides prepared by the students themselves.

The paper presentation will be assessed according to the same grading criteria as the oral examination.

Grading scale:

<18 fail

18-23 pass

24-26 satisfactory

27-28 good

29-30 very good

30 cum laude excellent

Students can reject the grade obtained at the exam. To this end, they must send a written communication via e-mail to the lecturer.

For the purposes of learning assessment, the use of AI is prohibited. Any use of AI constitutes a breach of academic integrity.

Teaching tools

Additional teaching material useful for the preparation of the exam will be available on the platform "Virtual learning environment" of the University of Bologna.

Office hours

See the website of Maroussa Zagoraiou