Does anyone have a recommendation to carry out a post hoc sample size?
The data I used I also used two microscopes same brand, but 1 has a higher resolution, I ensured this was spread throughout my data. I also normalised the control averaged to all my data to mitigate any differences of using diff microscopes & to determine the normal condition data, which is the control group. I know there's a slight difference in variance due to this, which I plan to discuss. But wanted to still conduct the sample size needed.
I have collected some data, but most of it is not significant. I have a low sample size and would like to find a rough sample size needed for the assumption of p>0.05 and perhaps a low SEM for a future study of large-scale quantitative data.
I have used G*power software to find out the effect size d= value with this, I used to determine the (1-beta) power I know that anything less than 0.80 is termed very low if you set it as 80%. My (1-beta) power values are also lower than 0.80, meaning this study has low power.
I now want to determine how to calculate sample size with this information for a larger study. Does anyone have any tips on how to do it or any software I could use OR formulas with the variables I have given?
Thank you :) !!
IMO, the best advice about post hoc (or retrospective) power analysis (as it is most typically done) is this: Don't do it. See this short note by Russell Lenth for some nice discussion of this issue.
- https://homepage.stat.uiowa.edu/~rlenth/Power/2badHabits.pdf
The Thomas (1997) article Lenth cites is another good one, if you can track it down. Try a Google Scholar search on . I get a link to a JSTOR PDF, but the link is too long to post here. HTH.
Hello Carla,
I agree with Bruce Weaver about after-the-fact power analysis. However, if what you're asking is, how to use your sample results to estimate what would be a suitable sample size for a future study, then:
1. Identify what, for you, is the smallest effect size that you consider noteworthy (or, of practical significance). This will serve as your target ES. There are a lot of different metrics or scales for ES. If you're planning means comparisons between two groups, then something like Cohen's d might be applicable.(As you probably know, smaller target ES values will result in larger required sample sizes, all other things constant.)
2. Set your target type I risk level ("alpha level"). It sounds as if you have this in mind. A related point is whether the proposed hypothesis test is directional or non-directional. (Smaller alpha level implies larger N.)
3. Identify the desired floor for statistical power. (You mentioned .80, which isn't uncommon in practice, but the real question ought to be: how best to balance type I and type II error risk.) (Higher power implies larger N.)
4. In G*power, specify the particular test statistic family (e.g., F, t, chi-square) and desired test (e.g., independent groups t-test).
That should get you the information you need to plan for the next study.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Engineering
- Petroleum Engineering
- Oil