I uploaded an excel file with data from 5 studies (N, mean & SD) for tx & control groups. Julius was great at explaining the process, but somehow messed up the data. I have no idea how Julius extracted the data it used from my set up. Data pictured here: top 6 lines are my data, lines 10-16 are from julius. Data used for mean & SD are incorrectly interpreted. How did Julius get these values?
Hi,
How are you formatting your conversation/prompts? There may be something else going on with how the prompts are being understood, which could be causing the output you’re seeing, but I can’t pinpoint it with the info provided.
One thing I think would be really helpful is if Julius could implement a window that tracks all the different spreadsheets you upload during the conversation (similar to how RStudio does it). That way, you’d be able to pull up the adjusted spreadsheets and examine them to see how your dataset is being modified. I think this would help users spot changes earlier in the analysis process. (Sorry, just a little side rant there.)
Tried several ways of phrasing the same thing “please perform hedges g meta-analysis with fixed-effects model”; tried a random-effects model, left out the ‘hedges g’, switched to ‘cohens d’. All returned the wrong variables extracted from the spreadsheet for analysis.
I’m more concerned that no one from Julius.ai seems to be answering my issues in this forum however…