Standard Deviation of Mean Absolute Error

#1
Hi guys,

As basic as this question is, I'm looking for a way to calculate standard deviation of mean absolute error using pivot table or atleast using excel.

I have fitted and predicted values. I calculated abs(y-yhat). and then when i used pivot tables i used "average" value to be displayed for y-yhat which gives me the mean absolute error value for each category.

Now, I also need the standard deviation for this mean absolute error. The value of "stdev" in pivot tables won't work since the value it will work upon is abs(y-yhat) which isnt a mean.

thank you for your help
 
#2
Ok, never mind this. I figured it out

I used error bars (one std deviation) in the plot. I'm trying to understand why the error bars all seem to have the same size?

shouldn't the size/range/extent of these bars be different for different plotted values depending on the values ?