Design of Experiments > Generalizability and Transferability in Statistics and Research
What are Generalizability and Transferability?
Generalizability and Transferability are two related terms used in research.
Generalizability is a measure of how well a researcher thinks their experimental results from a sample can be extended to the population as a whole. It is usually used in academic research, but it’s sometimes applied to research in other settings. The term usually only applies to specific quantitative (numerical) methods; hypothesis testing is usually a part of this process. Although it’s possible to predict an experiment’s generalizability, it’s not possible to predict it with certainty.
Three types of generalizability are:
- If an experimental treatment produces the same results in different environments.
- If the same experimental results are obtained with different measurements (i.e. measuring depression with a Beck Depression inventory and a Center for Epidemiological Study of Depression Scale).
- If the experiment produced the same results with different groups. Large, randomly chosen experiments can increase generalizability up to about 10,000 subjects.
Transferability is the process of other people transferring experimental results to real life. It’s the ability for a random person to be able to take experimental results and apply them successfully to their own situation. Some examples:
- An experiment shows that 3 hours of tutoring a week increases algebra skills by one grade level. Math teachers take this information and use it (transfer it) to benefit their own students.
- A study shows that lavender plants repel mosquitoes. A random person should be able to transfer this information and use lavender plants in their yard to repel mosquitoes.
Generalizability usually implies that results are transferable. It doesn’t work the other way around: having transferability doesn’t mean that your results are generalizable. Some studies (especially true with case studies) are very detailed and specific about a very small and defined group. The results then, are usually not applicable to the general population. However, the results still have transferability–as long as the results are applied to exactly the same group. For example, a detailed case study on intravenous drug users can be transferable to other intravenous drug users but not the population as a whole.
Firestone, William A. (1993). Alternative arguments for generalizing from data as applied to qualitative research. Educational Researcher, 22(4), 16-22.
If you prefer an online interactive environment to learn R and statistics, this free R Tutorial by Datacamp is a great way to get started. If you're are somewhat comfortable with R and are interested in going deeper into Statistics, try this Statistics with R track.Comments? Need to post a correction? Please post on our Facebook page.