Subscribe Bookmark RSS Feed

How do I add error bars to bivariate fit (y by x)?

I am plotting line graphs using analyze>fit y by x. I get a series of points that I then group by treatment and fit each value. I have been unable to get a nice line graph with error bars. I am currently exporting the data back into excel in order to get this graph. Any advice??
4 REPLIES
Jeff_Perkinson

Community Manager

Joined:

Jun 23, 2011

jlenn,

I'm not clear on what you're looking for.

If you're fitting each value, you are, essentially, connecting the points. There's no error to represent.

What should the error bars represent in this case?

Jeff
-Jeff
If you are fitting a line to a bivariate and want the confidence curves fit, go to the red triangle next to the fit description below the graph. This will give you the option to do either confidence curves fit or confidence shaded fit (same fits, different look).
ms

Super User

Joined:

Jun 23, 2011

To generate a line graph with y error bars you can use the Chart-platform under the Graph-menu. Select the line chart from the options in the lower left corner. As X select your treatments, and as Y select your data with the "Mean" option (if you also want individual points in your graph select your Y's twice: as "Data "and as "Mean").

Then in the resulting graph under the red triangle there is the option to add error bars for the mean. You will be prompted to choose from standard error, standard deviation, confidence limits etc.
da_sola_techy

Community Trekker

Joined:

Jun 23, 2011

I tried this with the bigclass.jmp . If I add Data(Weight) & Mean(weight) as Y in the chart platform and height as X , the chart platform crashes.

It works well with just Mean(Weight) Vs Height).