I am worried that your approach is wrong, but sorry if I'm mistaken. Shouldn't you be using the "as in Cohen (1988) - recommended" setting under "Options"? Especially when using partial eta squared. It seems you used the "as in G*Power 3.0" setting instead, which from my understanding will underestimate sample size in this context.

Thank you so much for the details on this.

Thank you very much for your video. Lights came on!

Excelente explicación, hasta un hispano parlante como yo pudo entenderlo sin ningún problema

Thank you so much for your help. You saved me :D

Thanks, it's Really informative. If we select two tails then the total sample size is for 1 group or for both groups?? @davey

I'm afrai df = 3 may be wrong, isn't it?

Sir, how we know that is with 46 participant, could you explain pleas🙏🙏

Thank you Sir, I was just looking for this type of tutorial

Thank you so much. Was really struggling with this.

I know this is an older video but why is your SD 3.5 ? I'm doing a g power before conducting the study.

Thank u so much. Fricken life saver. Clear and concise!

Hi Mr. Davey! I have a similar problem, except that one of the conditions has an effect size of 0.2. How can I use this number as input since I do not see a specific way to type that number in the software? Thank you for your time!

Great video with brief example

I loved it with the music!!

Thank you - super helpful!!

I do not think this is correct. for repeated measures within subjects, all patients need to experience all conditions.

Thank you for your nice video. I'm trying a post hoc power analysis for goodness of fit tests. This video is really helpful. I have one question about the value for the "p(H1)", though. In Faul et al. (2007) (pubmed.ncbi.nlm.nih.gov/17695343/), the authors mentioned that: "post hoc analyses, like a priori analyses, require an H1 effect size specification for the underlying population. Post hoc power analyses should not be confused with so-called retrospective power analyses, in which the effect size is estimated from sample data and used to calculate the observed power, a sample estimate of the true power." I'm wondering if it is retrospective when I put the observed proportions into the "p(H1)" column. Does this program compute "underlyng population" from the value in the "p(H1)" column? I don't know much about statistics and I'm not a native english speaker. I'm sorry if I'm saying too directly or impolitely. But, If you don't mind, could you tell me why it's OK to put actural proportions into the "p(H1)" column?

hello, how to determine correlation among repeated measure from using data from previous studies, can you explain it, please. Should we run an analysis for it, or find it from their tables.

Never had a peaceful stats video before... Thank you!😄

This track goes way too hard for a simple statistical analysis video, bro. I'm just trying to do my assignment, not walk down the Grand Castle Ballroom through the crowd of citizens and noblemen to accept my lordship alongside my party of other epic heroes.

For the purpose of getting a sample size, can we just assume the observed proportions?

Dear Davey! Thank you for that great explanation, it helped a lot. Do you have any examples how to report the sample size calculation?

Hi Dave. Any ideas or thoughts on how to run a power analysis with multiple predictor variables (some continuous some categorical?) Thank you!

Great!

Hello, I was reading the G-Power manual (and their publications), and it looks that in the "number of groups" input, you have to include the number of cells you have (so in your case, 6 as 3 x 2, not 2). However, I am still confused about what I should include in N of measurements (as this is not mentioned in their manual at all). On the web, I found several sources that performed the power calculation the same way you did (and learned it in the same way). However, I am worried this is not the correct way. Any thought?

Here for the music

thank you, this video was exactly what i needed!

Thank you! I have one WS factor (type of medication: A and B) and two BS factors (gender = 2 levels, age group = 3 levels). I put 2 in the number of measurements and 6 in the number of groups (gender x age group). Is it correct or should I put 2 (gender + age group)?

So helpful! Thank you

This is so helpful, Thanks for sharing!!!

Thank you very much for this video! Really helped me out!

Great sharing! Can the number of groups be more than 2, lets say 3 or 4?

bgm is not very comfortable tbh. content is good

Dear Davey, Can I use this test in G power to calculate the sample size for a three-way mixed ANOVA with two WS and one BS factor? All have two levels. If not, could you please tell me how to do it?

So helpful! Thank u for sharing!

Nice video. Thnx for the good presentation. How did you get the P1, P2 and P3 values? Can you kindly elaborate?

Does this mean the sample size of 36 means 18 for the drug, and 18 for the placebo? or 36 and 36? thank you!

Could you help me explaining how to open this example or tutorial window on the right side

Please help me how to open the guide window showing example and meanings of the symbols

Thank you for this video. Best tutorial I’ve found yet and the Word Document is excellent.

anyone know of a good tutorial for post-hoc power analysis in gpower for repeated measures ANOVA?

I don't think this example (29B in your guide) is correct. Using just one dichotomous explanatory variable to predict a dichotomous outcome, I would be very concerned that an assumption of logistic regression was violated (i.e. That there is a linear relationship between the logit of the outcome and each predictor variable). In this situation, a Chi-sq test would be preferable to the logistic regression.

Hi Dave, we want to examine the relationship between antidepressant medication and depression 'scores' of patients. We want to measure the difference in differences i.e. difference in mean depression scores between medicated and non-medicated patients after x months of time, where the mean scores are the differences between baseline and endline scores of the respective groups of patients. Which formula in GPower will help calculate the sample size required to do so?

Hi. I was wondering if you could clarify how to determine the probability of success (Y=1) when the main predictor is at the mean and when it is one SD above the mean?

Hi Dave, how to compute sample size when there is an interaction between predictor variables? Especially, when one predictor is continuous and the other one is a binary variable.

great!!

It helps me a lot. thank you!

In my G*Power version 3.1 that I have just downloaded I do not have corr among rep measures as input parameters :-/ I have it for the within factors and the between factors but not for the within-between interaction....

This was very helpful. Thank you!