I'm trying to conduct a multidimensional scaling analysis on a dissimilarity matrix using SPSS. I am having no problem importing matrices that are up to 200x200. Over that amount however when using the automatic import wizard to import the data as a delimited text file it limits me to 200 columns (not rows) as it assumes these to be variables.
Is there anyway around this - other than not using SPSS?
Secondly, when I come to plot the 2D data's new coordinates using SPSS's scatterplot plot creator, I don't seem to be able to associate more than one variable to each point. For example I can display information for genera or collection location (as either colour or symbol shape) but not for both e.g. red, star vs. an red circle (Porites, Pacific vs Porites, Atlantic).
Is this possible in SPSS or should I stop fighting the urge to go back to some plotting specific software like sigmaplot or r.
Many thanks,
I have not tried to import this, but maybe you could use the wizard and paste it into a syntax, you might be able to remove the restrictions on the size of the matrices in the syntax.
Do you have the integration of SPSS and R? The graphic in SPSS can not be that flexible or you need to go in the GL language, but that is not straightforward. I use the data-management and analyses of SPSS, but for high grad graphics I use R from within SPSS. You need to integrate SPSS and R (see the added file). Each version of SPSS need it own version of R (SPSS 23 - R 3.1.0 [not 3.1.1 or higher!]). Not only will you be able to include R graphs, but you can have R output with SPSS styled tables to export to WORD or HTML. The connection will also add procedures to your menus of SPSS.
Hi Peter,
Many thanks for your detailed answer. Whilst I'm sure I could continue work with SPSS to integrate R and work with it in a more programmatic sense I think the smart move here will be simply to conduct my analyses in R in the first place.
Thank you again for you help,
Ben
It could be useful to move to R, but remember that R uses only one core and SPSS can use more then one (if the procedure was written for it).
I read that you use a text delimited file to read into SPSS. Did you try to use EXCEL and read the EXCEL file into SPSS? It can read a complete EXCEL sheet with several hundreds of columns. (100.000 x 200). That might be a trick to be used. Another option would be to split the file in two sets or more sets and have id-numbers for the rows (first column). Split the data over two excel sheets (100 columns on each excel sheet, and on each sheet keep the first column, Read each sheet into SPSS, save them and join them again in SPSS with add variables using the identifier for each row,
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS