graphing question

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

graphing question

Maguin, Eugene

A gpl question. The context is interaction plotting. In legacy terms, think of this (GRAPH /LINE(MULTIPLE)=MEAN(yhat) BY ace BY dir.) translated to gpl (of which there is an example to copy!). My question is this. I’d like to shade the region of significance for the interaction. In this specific example the moderator, dir, has 6 discrete values and for each I can compute the ace value at which the interaction slope become significant. Those values when connected by line segments define the boundary for the right side region of significance, which would look nice in an explanatory way, if it were shaded. Doable? If so, I think the key piece of code needed is the components of the element command.

Thanks, Gene Maguin

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Reply | Threaded
Open this post in threaded view
|

Re: graphing question

Andy W
Can you be more specific about the inputs/outputs? You can definitely make shaded regions in SPSS plots (I have a few on my blog) but I can't quite follow what you want based on your description.

A hand-drawn graph illustrating what you want or a link to another example would help.
Andy W
apwheele@gmail.com
http://andrewpwheeler.wordpress.com/