Timeline for Is there a way to compare linear regression slopes by permutation tests?
Current License: CC BY-SA 3.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 27, 2012 at 13:37 | history | edited | Macro | CC BY-SA 3.0 |
added 33 characters in body; edited tags
|
| Jun 27, 2012 at 7:34 | history | tweeted | twitter.com/#!/StackStats/status/217883577604521984 | ||
| Nov 28, 2011 at 16:10 | comment | added | whuber♦ | Thanks: the problem is getting clearer. I'm still not sure what exactly you mean by "check" the composition change over time and why that requires making all possible comparisons between species. Could you perhaps share a high-level description of your scientific objective for this analysis? Also, there is a preliminary issue concerning what a "population" is. I doubt you truly know the populations: these must either be estimates or summaries of observations, but which are they precisely? That will influence the interpretation of any analyses. | |
| Nov 28, 2011 at 10:38 | answer | added | Manuel Ebert | timeline score: 1 | |
| Nov 18, 2011 at 16:21 | comment | added | user7417 | @whuber: It's population size (=number of individuals) of species per year (and they are bird species, by the way). The aim is to check species composition change over time. When there is directional change, then years further apart should show larger Hellinger distances than years closer together. | |
| Nov 18, 2011 at 15:02 | comment | added | whuber♦ | I cannot offer an alternative because I do not know what's in the data (what exactly has been recorded about these "species"?) and knowing the purpose of the analysis. Your research might be better served by reformulating the original question to provide this information. | |
| Nov 18, 2011 at 14:31 | comment | added | user7417 | @whuber: I'd be eager to get a suggestion... | |
| Nov 18, 2011 at 14:30 | comment | added | user7417 | @onestop: Dissimilarity between years were measured by Hellinger distance, that is, Euclidean distance of Hellinger transformed data. Hellinger transformation: N'ij = √(Nij/∑Nij) where Nij is the population size of species i in year j, and ∑Nij is the sum of individuals across all species in year j. | |
| Nov 18, 2011 at 14:27 | comment | added | user7417 | @psj: My matrices consist of species (several dozens) in rows and years (25) in columns. For the MCPP I shuffle the column positions randomly (10,000 times). | |
| Nov 17, 2011 at 16:05 | comment | added | whuber♦ | Even the ordinary regression fit may be invalid because the $(n^2-n)/2$ points reflect at most $n$ independent observations, causing extreme interdependence. So, testing significance appears to be the least of the concerns here. It looks like you need a different approach to the analysis altogether. | |
| Nov 17, 2011 at 15:44 | comment | added | onestop | ... and clarify what 'distance between years' means? | |
| Nov 17, 2011 at 14:59 | comment | added | psj | can you give examples how your "species x year" matrices look like? | |
| Nov 17, 2011 at 12:48 | history | asked | user7417 | CC BY-SA 3.0 |