Timeline for Correlation among repeated measures - I need an explanation
Current License: CC BY-SA 3.0
16 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 30, 2013 at 10:02 | answer | added | crsh | timeline score: 3 | |
| Nov 22, 2012 at 14:52 | comment | added | Dr. Manuel Kuehner | let us continue this discussion in chat | |
| Nov 22, 2012 at 14:32 | comment | added | ocram | (0.37+0.13-0.10)/3 might give a rough estimate of the correlation between any two measurements. Another way, if possible, is to fit a mixed model by imposing a compound symmetry structure and to store rho. Anyway, you should probably check to what extend power depends on that estimate... good work ;-) | |
| Nov 22, 2012 at 14:13 | comment | added | Dr. Manuel Kuehner | The motivation is that I need to enter something ;) | |
| Nov 22, 2012 at 14:00 | comment | added | Dr. Manuel Kuehner |
@ocram: Haha - I apparantly made some stupid suggestions - sorry. What would you enter in the software as rho in my case? Would (0.37+0.13-0.10)/3 be ok?
|
|
| Nov 22, 2012 at 9:22 | comment | added | ocram | (|a|+|b|+|c|)/3 --> why do you take absolute values ??? sqrt(a²+b³+c²) --> what is the motivation behind this weird thing? | |
| Nov 22, 2012 at 8:49 | history | edited | ocram | CC BY-SA 3.0 |
deleted 39 characters in body
|
| Nov 22, 2012 at 8:34 | history | edited | Dr. Manuel Kuehner | CC BY-SA 3.0 |
added 441 characters in body
|
| Nov 22, 2012 at 3:32 | answer | added | gung - Reinstate Monica | timeline score: 5 | |
| Nov 22, 2012 at 0:46 | answer | added | RioRaider | timeline score: 5 | |
| Nov 21, 2012 at 20:16 | comment | added | Dr. Manuel Kuehner | Ok. I get it. But G*Power is a well known software and I have to use it. I was wondering if someone can give me a hint what to fill in in the correlation parameter. | |
| Nov 21, 2012 at 19:31 | history | edited | Dr. Manuel Kuehner | CC BY-SA 3.0 |
added 110 characters in body
|
| Nov 21, 2012 at 19:24 | comment | added | ocram | Averaging correlations like you do does not make much sense. In my answer below, I have tried to explain how a correlation matrix can be parametrised so that only a small number of parameters have to be estimated. Anyway, to interpret a correlation matrix, you have to look at the pairwise correlations. | |
| Nov 21, 2012 at 19:11 | answer | added | ocram | timeline score: 9 | |
| Nov 21, 2012 at 18:34 | history | edited | Dr. Manuel Kuehner | CC BY-SA 3.0 |
edited body
|
| Nov 21, 2012 at 18:08 | history | asked | Dr. Manuel Kuehner | CC BY-SA 3.0 |