Quantitative design and analysis in health science with focus on quasi-experimental research (DHV220)
The main aim of the course "Quantitative design and analysis in health science with focus on quasi-experimental research", is to introduce the candidate to alternative experimental research methods when a strict controlled experiment is not possible. The course covers design of cluster randomized, non-randomized and natural experiments and the statistical analysis of such experiments.
Introduction The main goal of experimental research is to test hypotheses. In medical research, the gold standard is the randomized controlled trial: The only design where causality can truly be established. However, in many instances, such designs are impossible to implement, be it for practical, ethical or other reasons. In this course, alternative designs will be demonstrated, and statistical analytical approaches to these designs will be discussed. The main focus will be on quasi-experiments, where individual randomization is not possible, but where one must study cluster randomized groups, non-randomized as well as natural experiments. The course aims to enable PhD candidates to identify possible solutions to their own research questions, and to better understand threats to internal and external validity in published quasi-experimental research.
Course description for study year 2023-2024
Course code
DHV220
Version
1
Credits (ECTS)
10
Semester tution start
Spring
Number of semesters
1
Exam semester
Spring
Language of instruction
English
Content
The theory of science as related to experimental and quasi-experimental design
- The concept of causality
- Validity: Internal, external, construct and statistical conclusions
Quasi-experimental designs and analysis
- Nonequivalent control groups design
No-treatment control group designs
- nonequivalent dependent variables designs
- removed treatment group designs
- repeated treatment designs
- reversed treatment nonequivalent control groups designs
- cohort designs
- post-test only designs
- regression continuity designs
- Regression discontinuity design
- Statistical analysis of quasi-experimental data.
- Unacceptable or very problematic research designs: Learning to spot bad design and analyses.
Learning outcome
Upon completion of the course, the students will be able to:
Knowledge
- Understand the concept of causation as related to methodology and statistical analysis: when it can be achieved and when it cannot.
- Understand and describe the different forms of validity, such as internal and external validity, and construct- and statistical conclusion validity
- Distinguish between the different forms of experimentation and understand their advantages and disadvantages
- Know the most common and relevant quasi-experimental designs and their associated statistical analyses.
- Know issues related to statistical power and effect sizes
Understand the role of quasi experimental research in a broad health-science research context.
Skills
- Design quasi-experiments that are applicable to the candidates own work
- Analyze data from quasi-experiments
Be able to review and critique published, proposed or submitted quasi-experimental research.
General Competence:
- Hands on training in designing quasi-experimental research, and associated statistical analyses.
- Increased insight in the theory of science as related to different forms of research methodology when the research aim is to test hypotheses and establish causality.
Required prerequisite knowledge
Exam
| Form of assessment | Weight | Duration | Marks | Aid |
|---|---|---|---|---|
| Individual paper | 1/1 | Passed / Not Passed |
An individual paper of 5000 words (+/- 10%) in English.