Given the following fact:
Sample Size = 283
alpha = 0.05
Number of predictors = 1
Effect Size = 0.21
power = 1.0
I also computed effect size using Sensitivity, and it was 0.0462339 given the sample size and effect size
Dear Tahir,
Power is never 1.0. A common choice is 0.8. You also need to descr the design of your study, e.g. an independent samples t test.
I am also waiting the response from expert to learn about effect size.
Dear Tahir,
Power is never 1.0. A common choice is 0.8. You also need to descr the design of your study, e.g. an independent samples t test.
Thanks Peter, here is the T-test result
Variable: Word Count
Group 1 = 58.03±23.3
Group 2 = 42.83±23.4
t (281) = 5.454, p < 0.001
Hello!
If you have the results, I do not understand the question. Your effect size was large enough to be discovered in the experiment, and you already calculated how much it is.
By the way. Can word count be a non-whole number? If not, the t test is inappropriate, you should rather use a nonparametric method.
Dear Richard, The average word count of-course can be non-whole number.
I am still wondering about your question. You ran the experiment already and have the results, you detected a difference, meaning your power was enough to detect said difference. To decide whether such an effect size is enough, you must consider reasons of the field of research (measurement error, data in the literature etc.). This is less of a statistical question
Regarding word count, the average can be non-whole, but can you have a single data of 543.65 word count for example? Seems odd to me, please do clarify.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size