17-810 (A3) - Empirical Methods in Software Engineering Research
Instructor: Jim Herbsleb
University units: 6
Spring 2003, 1st half-term, Ph.D. course.

Description:
The course will focus on several basic types of empirical methods that have proven to be important in advancing the field, including ethnography, interviewing, surveys, experiments, and quasi-experiments. We will cover the methods in roughly that order, moving from more exploratory methods to those suitable for testing causal hypotheses. Students will be required to critique a number of examples of published research, and for their project will propose a program of research to address an important software engineering question.

Text:
Essentials of Behavioral Research: Methods and Data Analysis, 2nd edition, 1991, by Robert Rosenthal and Ralph Rosnow. McGraw-Hill.

Method of evaluation:
Grades will be based on critiques of example papers and the student project.

Topics to be covered:
· Qualitative methods
· Logic of experiments
· Types of Validity
· Designing experiments
· Quasi-experiments
· Multiple methods