What is the standard convention regarding the error bars of the independent quantity in a graph? In what situations should I show the $x$ error bars? In case both $x$ and $y$ uncertainties are comparable or neither can be disregarded, should I show both or only the greater one?

I know this matter is more related to aesthetics and convenience, but I wonder if there is more straight rule for that.


If you are measuring y at some value x, and both quantities have uncertainty, then in principle you should show the uncertainties on both axes.

In some circumstances you might omit the x error bars. This would be the case if the y value depends on x such that $$\Delta y \gg |dy/dx| \Delta x,$$ where $dy/dx$ is your best estimate of the gradient of $y(x)$. In other words, where y changes by much less than its error bar over a change in x equal to the x error bar.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.