Bora Beran - I've read your blog posts and reviewed your sample workbook on t-tests via R. You will be the ideal person to answer this question. I appreciate any time you can give.
I am trying to implement a wilcoxon rank sum test via R. I can implement this successfully if I am comparing two separate fields/measures as you do in your t-test example file. e.g. if I wanted to compare [Field 1] to [Field 2]. Where I'm running into trouble however is that I need to run a test on the same measure, across a given dimension.
I have attached a sample workbook with real data to illustrate the issue I'm having. There are a lot of variables in the workbook but we can focus in on two in particular: "Raw Bias X (g)" and "Raw Bias Wilcox Test". As can be seen in the table, AVG(Raw Bias X (g)) is shown aggregated across a few dimensions. One of the dimensions being aggregated across is "Rotation Profile ID", which can take on the value of 'default' or 'sphere'. I am trying to test for differences in mean bias between the 'default' and 'sphere' sub-populations.
I can pass the AVG(bias) into R without an issue. It is a numeric array of length 24 that I've checked with write.csv. However since "rotation profile ID" is a dimension, I can't seem to properly aggregate it to get a similar length 24 array. This array should be equivalent to what is seen if you use 'view data' from within the table visualization. It will have 12 instances of 'default' followed by 12 instances of 'sphere'. I have attempted ATTR, but this returns * since there are two values. I have also attempted converting the dimension to a numeric measure and then taking AVG, this failed and appeared to aggregate across an incorrect dimension.
I've spent hours reviewing forum posts and articles online and cannot find clear guidance on this topic. Any suggestions you can provide would be greatly appreciated. I'd love to figure out how to call the dimension directly and aggregate it properly, but failing that if I can just get the unique values of "rotation profile id" I could always reassemble the array in R.
Thank you for the help,
Raw Bias Table_Question.twbx 166.6 KB