This is for a PhD-level Quantitative Research class, and I am confused over what test to run on the G*Power software.
It has to be an a priori, but would it be bivariate normal modal, multiple regression, t-test, etc...)
Do I need to determine the variables that will be measured, (depression, anxiety, obesity, etc...) before doing the test?
Hello Jacqueline,
In the parlance of G*Power, a priori means that you intend to specify: (a) a specific type of statistical method (for example, independent groups t-test); (b) a specific alpha level/type I risk level to use (e.g., .05 or .01 or whatever fits for you); (c) the desired level of statistical power you'd like the test to have; (d) the target effect size (which varies in G*Power, depending on which statistical method you select, but generally refers to the smallest degree of difference or relationship you'd like to be able to detect, if it in fact exists in the population; and (e) for some tests, additional information (such as, how many groups in an anova, or how many independent variables you intend to include in a multiple regression). Given these, the program will compute the lower bound N of cases that you would need to achieve that scenario. (This is the so-called a priori analysis.)
So, note that you do not need to list specific variables (e.g., "sensation seeking") or measures (e.g., Cattell 16PF) that you intend to use, just the information outlined above.
Do have a look at the simple examples on the UCLA stats page: https://stats.oarc.ucla.edu/other/gpower/
Good luck with your work.
Jacqueline Rich
The majority of the research evaluated indicated unfavorable psychological consequences such as post-traumatic stress symptoms, disorientation, and hostility. Longer quarantine periods, virus worries, frustration, boredom, insufficient supplies, insufficient knowledge, financial loss, and stigma were all stressors.
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion