Graphic problem in 16.01

[Mary Oliver]

Hello Everyone,

I would greatly appreciate any help that people can provide with a graphic problem I'm experiencing in SPSS 16.01.

I would like to be able to create a line graph of an interaction in ANOVA based on the emmeans (or adjusted means if I were doing an ANCOVA). Here are additional things:

1) It needs to be in black and white -- journals typically don't print in color. When people print out dissertations, etc., they often have black and white printers.

2) The lines should be differentiated on the basis of markers (not dashes). Why? Because if there are numerous lines, dashes are difficult to distinguish.

3) If it's a 3-way interaction, then it would be good to have panels that are presented side-by-side (rather than as separate graphs) for ease of visual interpretation.

So.... the good news was that all of this was possible in SPSS v. 15.  I defaulted to gray-scale graphs, would obtain the pivot tables, would switch one of the variables from columns to rows, would highlight the emmeans within the pivot tables, would right click, and would request a line graph. From there, I could go about creating the graph I needed.

Here's a URL where I've provided some sample data and have done just this (output provided in the PDF file). On the 3rd page of the PDF is the original graph that was given, and on page 4 I've just changed this to panels.

http://www.personal.psu.edu/mbo1/spssgraphs/

Now, with SPSS version 16, this is no longer an option. Why? I'm finding multiple problems.

Problem #1: If I right click on the means within a pivot table, the graph that is created does not graph the means. It will graph a mean, a SE, and confidence intervals (for example). Basically, it's incorrect. I would love to know if others see the same thing here!

Problem #2: If I use the profile plots, it does not present panels side by side like is optimal (and illustrated on page 4 of the PDF file provided). Rather, it presents one graph after the other.

Problem #3: Even if I can get it to present panels (by going through the painful process of using OMS, using the chart builder, etc.), then I cannot get the lines/legends to differentiate by marker. It wants to include multiple legends, it won't let me do it, etc.

In other words, I'm at the end of my rope on this and am ready to either 1) use excel and forget it in version 16; or 2) forget version 16 (which I'm quickly learning to hate -- for many MANY other bugs in the program other than this one) and go back to version 15. The only problem is that version 16 is now loaded on labs in my university and I'm teaching a stats class. :(

Any help at all would be appreciated. So far the people at SPSS haven't been able to give me a solution to this.

Thanks in advance.

Mary ===================== 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