DanielR ☆ Mexico, 2016-09-14 21:56 (2752 d 13:40 ago) Posting: # 16638 Views: 8,241 |
|
Hello everybody I need to determine a sample size for a parallel design, but only have information from 2x2 designs. I determined the CVintra of a crossover design and used it in the parallel design equation, however, the regulation tells me that the calculation is incorrect since I am using the CVintrasubject, insted of a CVinter What I can do? How I can determine the CVintersubject?: Thank you very much |
mittyri ★★ Russia, 2016-09-14 23:49 (2752 d 11:47 ago) @ DanielR Posting: # 16639 Views: 7,360 |
|
Dear DanielR, I think you can find the answer in this old but very good thread PS: I suppose they meant CVtotal, otherwise — Kind regards, Mittyri |
DanielR ☆ Mexico, 2016-09-15 02:08 (2752 d 09:28 ago) (edited by Ohlbe on 2016-09-18 19:51) @ mittyri Posting: # 16641 Views: 7,254 |
|
thank you very much mittyri. I had read this post, but now the question is: Is it possible to determine the MSs from the confidence interval? (90% CI from a 2x2 design) Thank you very much Edit: Full quote removed. Please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post! [Ohlbe] |
mittyri ★★ Russia, 2016-09-15 02:26 (2752 d 09:11 ago) @ DanielR Posting: # 16642 Views: 7,311 |
|
Dear DanielR, ❝ Is it possible to determine the MSs from the confidence interval? (90% CI from a 2x2 design) From the CI of crossover study you can only calculate the CVintra, you need somehow to get between-subject variance σ²B (ANOVA table or some additional sheets in your software) PS: please follow the forum's policy, remove all unnecessary cites, otherwise yellow card could be given by mighty Moderators/Administrators — Kind regards, Mittyri |
DanielR ☆ Mexico, 2016-09-15 19:45 (2751 d 15:51 ago) @ mittyri Posting: # 16645 Views: 7,036 |
|
Dear mittyry Thank you very much for all your answers |
ElMaestro ★★★ Denmark, 2016-09-19 14:44 (2747 d 20:52 ago) @ DanielR Posting: # 16648 Views: 5,477 |
|
Hi DanielR, ❝ I need to determine a sample size for a parallel design, but only have information from 2x2 designs. (coloring by me) If information includes the data listing then it is actually not difficult: Just extract either the period 1 data or the period 2 data, and do an evalution as though this was a parallel trial. The principle in this approach has a troubled past as it was used to evaluate BE pivotally when there were significant sequence effects in 222BE trials, and it never caught on with regulators. That is fortunately not your dilemma here. For your purpose you can get a reasonably qualified idea of the total variabilities for T and R. — Pass or fail! ElMaestro |
DanielR ☆ Mexico, 2016-09-20 20:24 (2746 d 15:12 ago) @ ElMaestro Posting: # 16651 Views: 5,438 |
|
thank you very much! |