0

I have something in between a statistics and R question. I ran a repeated-measures ANOVA with the R package ezANOVA. This package automatically provides eta-squared effect sizes. Furthermore, I'd like to report 90% Confidence Intervals and found a great way to calculate these here: http://daniellakens.blogspot.de/2014/06/calculating-confidence-intervals-for.html

However, some of my effect sizes appear to be outside of these confidence intervals, e.g.: 

F(2, 730) = 20.20, p < .001, eta-squared = .00, 90% CI [.03, .08], 
F(2, 730) = 45.46, p < .001, eta-squared = .05, 90% CI [.08, .15]

These are effect sizes of within-subjects factors (see the second way in Laken's blog post).

Here's my question: Is this statistically possible? And if not, where could be the problem?

Thanks and best, Mathias

MrFlick
  • 195,160
  • 17
  • 277
  • 295
Mathias
  • 47
  • 10
  • When asking for help, you should include a simple [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) with sample input and desired output that can be used to test and verify possible solutions. – MrFlick Mar 12 '18 at 13:13
  • thanks although I think I found the solution – Mathias Mar 12 '18 at 15:24

1 Answers1

1

The answer is that the ezANOVA calculates the generalized eta squared, while the confidence intervals were calculated for the partial eta squared.

Mathias
  • 47
  • 10