Confused by the discrepancy between StatView and GraphPad analysis

Model Organism: Worms Methods & Reagents
1

Dear all,

I am currently doing some survival assays where i will have to assess the efficacy of the drug treatment given to the worms.
So i have tried using StatView and GraphPad software. But i noticed that, the significant test result from both software are somehow different, which confused me totally.
Allow me to put it simpler here – the Statview revealed that the 200uM drug treatment is not significant different compared to control group; while Graphpad tells it is significant different!

Which one should i believe in then…??
this is bothering me as i am currently using mutant worms hoping to look for the drug target. The significant test of course is going to play a major twist and i do not wish that i made the wrong move.

I have attached along the raw data and analysis files (GraphPad, Statview) if you may help to.
https://www.dropbox.com/s/fswdudal3k659ew/20130802%20Graph%20Pad.pzf
https://www.dropbox.com/s/qrun7gfn5jael5o/20130802%20StatView.xlsx
https://www.dropbox.com/s/xc5t90l601cgl3k/20130802%20StatView.txt

Thanks. Appreciate any inputs too.

2

Hi,

what were the tests for statistical significance you used in each software package then?

I’m guessing Mantel Cox, but both StatView and Graphpad can do Mantel Cox and Gehan-Breslow-Wilcoxon and the answers might vary according to which you use.

Steve

3

Yes, i am performing the Kaplan-Meier analysis and using Mantel-Cox.

Thanks for the heads-up Steve! I will re-analyze the data and hopefully it’s okay then :slight_smile: