Thread: Max and min values on X axis

May be this link might help?

http://peltiertech.com/Excel/Charts/...OnTwoAxes.html

Hmmm I got the same problem

Its just that one of the values is non-cartesian. Rather its a two numbers ( a min value and a max value for one of the co-ordinates) but the other co-ordinate there its a fixed co-ordinate. For example (0.03-0.09, 0.23)

So at best it is a Scatter chart.. but one of the values is vertial line rather a point so you might see something like this

******************|
**************|***|
**************|***|
**************E***B
***|***************
***|***************
|**|***************
|**D**************
|***************
A*************




And the next part is calcuating the correlation of it..


EDIT: ok figured it out. Just add the minimum and max points as a seperate co-ordinate for the same series. Double click that series and click draw line. and dah!!!!

However if your one is even more complex.. then you would use bounding

Ok next up correlation??? Treat the two points per series seperately?.. or find the mean? or meduim of those two points?

Ok next up correlation??? Treat the two points per series seperately?.. or find the mean? or meduim of those two points?

I suppose it depends on exactly what you are doing, and someone with a real statistical background might have a stronger opinion. When I've faced similar situations, I've done the correlation both ways and found that it made little difference (certainly not significant compared to the uncertainty in the raw data). At this point it became more a question of which approach is easier/cleaner to present and document rather than which is more "correct."

In other cases, using both "end" points would cause that data point to, in essence, carry twice the weight of the other points. This could skew the correlation a little, which could be significant if there is a large uncertainty in the data or a limited number fo data points.