[R-sig-eco] transformation of a ratio before analysis

[ZZ]

Hello,
> 2: What transformation would you recommend? I have tried log but also
considering arcsin but any suggestions are most welcome.

Take a look at:

Warton, D.I., Hui, F.K.C. (2011): The arcsine is asinine: The analysis
of proportions in ecology, Ecology 92(1), 3-10.

Best regards
Z.

Dnia 2013-02-02, sob o godzinie 15:53 +0100, Adam Martiny pisze:
> Dear Members,
> I'm interested in understanding the deviations in the ratio of two nutrients in different samples along gradients of temperature and other environmental conditions. The two nutrients are broadly correlated with a correlation coef. r = 0.9. I have a hard time with this analysis and I suspect this in part has to do with this strong correlation as well as the poor statistical properties of a ratio. Thus, I am considering to do a transformation of the ratio before my analysis. I should mention that the denominator is far from zero.
> So my questions are
> 1: Am I on a good path or does anyone have some recommendations about how to analyze ratio data? E.g., should I somehow look at the concentration of the two nutrients and not the ratio?
> 2: What transformation would you recommend? I have tried log but also considering arcsin but any suggestions are most welcome.
> 
> Any input on this matter would be greatly appreciated.
> Thanks,
> Adam
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology