How do i plot error bars on log transformed graph?The biological data that i have follows a log-normal distribution, so i have transformed the data to be displayed on a logY axis. However, i am struggling to display the corresponding error bars (2 sd) on the log graph. If plotted directly, the error bars will be asymmetrical and appear misleading.
How can the error bars be better displayed?
Came across this tutorial and would this approach be appropriate?
How can I estimate and plot Logarithmic Error Bars (in Excel, preferably)? | ResearchGate
Data Exploration and VisualizationRe: How do i plot error bars on log transformed graph?Yes, its OK to transform the data.
Just make sure you say that it's transformed. Often in bioscience world, a log10 rather than a natural log transform is used. In JMP "log" is the same as "ln" and log10 is base 10.
but even if you show confidence intervals on the untransformed data using a log scale, the asymmetric appearing CI are still right. It's just that graph axis is scaled. Looks wonky but still right.
