Reply With QuoteIt is common for data to vary more (and thus standard deviations to grow) as the mean increases in size. That is why, for example, heteroscedacity is common when there is such an effect. I am not sure how that applies to your research.
Dear all,
I have been calculating the probability that the form representing a meaning will be replaced in a thousand years in four language families.
For each meaning, I have calculated the mean probability and the standard deviation across the language families. I have found a significant positive correlation between the mean and the standard deviation.
However, it was suggested to me the other day that this correlation is not necessarily meaningful, and that mean and standard deviation tend to correlate positively, under certain circumstances.
After an extensive google search for more information, however, I have not been able to find under what circumstances the mean and the standard deviation are not fully independent of each other.
Could anyone explain when a positive correlation of mean and standard deviation is meaningful, or point me in the direction of somewhere that could help?
Many thanks indeed in advance for your help!
It is common for data to vary more (and thus standard deviations to grow) as the mean increases in size. That is why, for example, heteroscedacity is common when there is such an effect. I am not sure how that applies to your research.
"Facts are stubborn things, but statistics are more pliable." Mark Twain
Thanks very much for your response, noetsi - this looks like it may well be applicable to my research.
Posting Permissions
|
|