I have many overlapping data points and I wonder if it is possible to jitter them only in one direction, on the x axis where I have categorical variables. I was trying various commands in the ggraph but it did not work.
I don't use SPSS so not sure if there is a command for this. But you can force your data to jitter by: 1) looking at the number of observations with the same value, and 2) adding or subtracting a small delta value to the points, prior to using a graphical procedure/command.
Paste your data into sigma-plot and increase the resolution of your X-axis....In general, SPSS ain't appropiate for nice art-work. Sigmaplot deals bettern with this type of situations
Alternatively, paste the graph into Powerpoint, ungroup the image and then do what you need to - this has worked for me in the past for graphs produced in SPSS. Just take care not to jiggle your data so that you change the results!
Can also change point(...) to point.jitter(...) for more jitter, but you must must turn off weighting before the GGRAPH command or it will error out (see below)
* Chart Builder.
weight off.
GGRAPH
/GRAPHDATASET NAME="graphdataset" VARIABLES=Age2 MeanRT Sub MISSING=LISTWISE REPORTMISSING=NO