In an article in the American Psychologist there is a series of bar graphs with a statistic next each bar. The statistic is described in the text and in the graph captions as being an effect-size measure. The statistics is $\lambda^2$. Now, I can find explanations of $\eta^2$, Wilks $\lambda$, and Cohen's $f^2$ as effect size measures but not of $\lambda^2$. What have I missed? Or is it more likely a typograpical error of some kind?
Asked
Active
Viewed 251 times
3
-
I appreciate that answering with a definite "lambda-squared does not exist" is not a reasonable request. Instead, to copy the approach of Donald Knuth in seeking assistance with the exercises in volume 4 of *Art of Computer Programming*, it would be very helpful if you were in a position to say, for example, "I am familiar with various measures measures of effect size including eta-squared and Wilks lambda but I have not heard of lambda-squared". – user02814 Nov 20 '14 at 09:52
-
Hi if this is in the context of a Gaussian linear model you might find useful information [here](http://stats.stackexchange.com/questions/59559/is-there-a-general-definition-of-the-effect-size) and [here](http://stats.stackexchange.com/a/58133/8402) – Stéphane Laurent Nov 20 '14 at 12:26