Repertory grids

Written & mounted by Chris Evans, if you want to be kept updated by Email of changes in this and other psychotherapy resources at this site then Email me!
Mounted 9.i.96, last updated 11.ii.04

You can try the form and program

The program does log your data and host address (to help me deal with it if it is abused) so don't use free text that could identify someone who wouldn't want to be identified. Currently it runs on old hardware on a broadband connection so you'll have to wait a minute or two for the results to be returned to you.

You can see some fictional data analysed if you don't want to overstress the server or waste your own time typing in data but do want to see what you get out.

• http://www.jiscmail.ac.uk/lists/pcp.html and locally I've got some old:
  • archives of messages

I have mounted my own version of the late Patrick Slater's INGRID program for the analysis of single grids by construct correlation, inter-element distance and principal component methods. My version, ingrid2.sas {19kb} uses SAS/IML to do the maths. The program is copyleft: I claim the copyright but you are welcome to copy and distribute it provided you neither change it nor make a profit. The data currently in the program (but easily replaced with your own) are from the grid in the intro. {3kb} noted above and the full listing (edited with the construct names shortened to make things easier to read on screen) is mounted: ingrid2.lis {16kb} together with the ingrid2.gif {3kb} file from the SAS/GRAPH plot of the first two principal components. (This is a very poor GIF, the only way I've found to get from SAS to GIF quickly is to capture the screenshot using Hi-Jaak Pro which, as I've also found when getting from PowerPoint 4.0 to GIF) is a very bad way to do things. If you have SAS/IML and SAS/GRAPH you should get much better plots to screen and hardcopy much more easily.).

I have mounted some free programs:

• The best are definitely Richard Bell's beta release freeware programs for DOS: GRIDSTAT.EXE and GRIDSCAL.EXE. These are simple DOS style menu driven programs that analyse either single or multiple grids (respectively). They run fine under Win95 for me and Richard and I will will be collaborating to keep the latest versions available here. There is a brief summary of the files in /grids/RB/Readme.txt and or everything in a zip archive
• A classic is CIRCUMGRIDS {89kb} written by William Chambers, and some information about it {3kb}. That has a bibliography that is still very useful though ageing.
• Another classic is OMNIGRID v1.5 {105kb} written by Kenneth Sewell sewellk@terrill.unt.edu out of original work for the Apple by Mitterer and Adams-Webber.
• Linked with that, I have Jack Adams-Webber's latest version of omnigrid for the Mac hypercard system, see info. on both versions {7kb}.

• One is James Legg's WinGrid project at http://ingrid.netpedia.net/wingrid_html This is an open source development (i.e. free) though no reference is made to any of the usual open source licences that I could see. One note on something that can otherwise cause confusion: for inter-element distances Jim has subtracted INGRID's "Slater distances" from 1 and multiplied by 100. Slater distances range from 0 (exactly the same ratings on all constructs) to some figure greater than 1 dependent on the size of the grid, rating scheme and distribution of ratings. Jim's rescaled values range 100 for elements which received identical ratings on all constructs (displayed as +99 at least in his earlier program) to some negative number. Jim says he's never seen a raw Slater distance greater than 2 (indicating enormous differences in ratings on all constructs). Hence he regards his rescaled values as going down to a possible -100, (was displayed as -99). I think I've seen distances greater than 2 and I think the largest possible Slater distance is undefined but I'm sure a really good mathematician could tell us a definitive answer here.

I also have some information on packages that are available commercially.

• G-PACK is a package of GWBASIC routines which will run on just about any PC under DOS or in a DOS window under Windows. It's very cheap and strongly recommended by me though it's being superceeded by Richard's more recent freeware programs (GRIDSTAT & GRIDSCAL above). I've mounted a local copy of the summary info. go to original Australian copy if you think you are (topologically) nearer there in the WWW.
• Courtesy of Guillem Feixas gfeixas@psi.ub.es one of its authors, I've mounted some information about the program GRIDCOR and you can look at his original information about the program (in both English and Spanish, the Spanish language version is called RECORD). If you don't have a frames equipped browser you can look at their information on the English language version at: http://www.ub.es/personal/gridcord.htm and at the Spanish language version at: http://www.ub.es/personal/record.htm GRIDCOR offers correspondence analysis, an approach not in the other programs I know of though present in SPSS. The user interface of GRIDCOR is rather more sophisticated and friendly than G-PACK which may make it more robust in routine use. On the other hand, there are analysis options in G-PACK that are not in in GRIDCOR and addicts will want to have both!
• Grid Analysis Package including INGRID. I believe this classic package written by the late Patrick Slater is being ported to the PC in something close to its original form. I'll put up more information here if I can find it.
• Finn Tschudi (and his son?) have written what I believe is an excellent package, FLEXIGRID though I've never used it. Again, I'll put up more info. if I get it.

I'll be mounting more information and probably more software and resources in due course. You might like to look at my review {7kb} of a book on multidimensional scaling (MDS) which is another excellent way of analysing grid data. The book comes with simple but powerful software for MDS which could be applied to grid data and I've written to the authors to find out if they'd be willing for their programs to be placed here as "copyleft". Alternatively, you can use the MDS options in the major mainstream statistics packages. For example, the ALSCAL parts of SPSS and SAS which I have used, and comparable facilities which I know exist in SYSTAT and I think in most of the other major packages now.

• PCP list info. page here at PSYCTC.org
• some decidedly out of date (1995) information about me!