Hey,
I'd really appreciate any help you could give me on this.
Basically, my design is w-s, one IV, 3 levels looking at word recall and postive, negative and neutral words, sample size = 24.
I've run a one-way w-s ANOVA and this turned out non-significant:
f(2, 46) = 1.269, p = .291, Effect size = 0.34
I'm trying to do a power analysis on this data, seeing as it's non significant using power tables for F ratios, and this has come up as .995.
If this is so, how come my data isn't significant if i've got a medium effect size and the power is beyond the 0.8 ideal?
The means suggested there was a slight difference in scores of positive (4.13) negative (3.46) and neutral (3.50) words recalled.
So I was wondering if I could get any tips. And also am I doing my power analysis right? Because in class, we've never dealt with a situation where the retrospective power is above the ideal power. And also, if I were to report the ideal power, along with the suggested sample size, this value would be lower than my original sample size. How can this be?
I hope that makes sense, i'm a little worried about it! Any help would be fantastic, thankyou so much!
Lisa.
I'd really appreciate any help you could give me on this.
Basically, my design is w-s, one IV, 3 levels looking at word recall and postive, negative and neutral words, sample size = 24.
I've run a one-way w-s ANOVA and this turned out non-significant:
f(2, 46) = 1.269, p = .291, Effect size = 0.34
I'm trying to do a power analysis on this data, seeing as it's non significant using power tables for F ratios, and this has come up as .995.
If this is so, how come my data isn't significant if i've got a medium effect size and the power is beyond the 0.8 ideal?
The means suggested there was a slight difference in scores of positive (4.13) negative (3.46) and neutral (3.50) words recalled.
So I was wondering if I could get any tips. And also am I doing my power analysis right? Because in class, we've never dealt with a situation where the retrospective power is above the ideal power. And also, if I were to report the ideal power, along with the suggested sample size, this value would be lower than my original sample size. How can this be?
I hope that makes sense, i'm a little worried about it! Any help would be fantastic, thankyou so much!
Lisa.