One of the early steps in examining research data is to run a correlation matrix, which is a table that describes the direction and strength of relationships between research variables. Depending on how many variables you're looking at, the thing can be pretty large - not to mention confusing. Since I'm a highly visual person, one of the things I do with one of these is to color-code the cells that show a significant relationship between variables. It's one of the organizational tricks I use that keeps me from blankly staring at these huge sheets of rows upon rows of numbers, which are usually pretty meaningless to me. Since my first forays into research in my undergraduate program, I've used Sharpie markers or highlighters or whatever I can get my hands on to outline the "important" cells with different colors for each variable or specific groups of variables. This practice helps me a) develop a good visual representation of the relationships between the research variables and 2) zero in directly on the important stuff from across the room when I'm attempting to write about what it all means.
There have been a few problems with this procedure. SPSS prints out the variable row headers on every page, which isn't exactly conducive to my organizational style so before I can add color to the cells, this practice has always involved printing out the correlation matrix, then folding, cutting, aligning and taping the sheets together so they form a coherent whole while saving space by eliminating all those extra headers. That's one problem, and it isn't going to go away unless I can obtain one of those really huge printers that print off newsprint-size sheets. The other problem is that it looks pretty messy, what with all those outlines and marker marks all over it. Oh, and then there's the problem of not always having enough different colors and/or the necessity of driving nearly 20 miles to the nearest office supply store to get more, so I usually make do with what's on hand (see pic above). Hideous, but it works for me to begin organizing this kind of information in my head, since I am NOT a numbers person.
I've used Microsoft Excel since first returning to junior college about 10 years ago (Yikes! Has it really been that long!?), so I knew you could add color to cells in that program, but alas, for statistical analyses in the social sciences, we use SPSS when doing quantitative research, and as far as I knew, with the exception of pie charts, graphs, histograms, et cetera, color wasn't possible in SPSS. Grr.
Yesterday was my lucky day. I was futzing about with a random dataset to re-acclimate myself to SPSS and ran a correlation matrix so I could begin the process of taping, cutting, folding and coloring so the relationships between the variables will all make sense to me. Then I accidentally double-clicked on the table and a second window titled "SPSS PIvot Table" opened, with the table in it - only in this second window the cells were editable INCLUDING COLOR! Not only could I color in the cells containing statistically significant information, I could also vary the amount of color dependent upon the strength of the relationship! I just about peed myself. All that was necessary was to click & highlight the cell I wanted to edit, then go to the "format" dropdown menu. It was intuitive from there. The final product ended up looking like this (OooOOoooh, pretty!):
For the uninitiated, in case you're wondering about the black cells, those are the ones that indicate a variable's relationship with itself, which is always going to be perfect, so we don't even consider that information when designing the study. The reasons I only color in the bottom half are 1) because the information in the top half is essentially a mirror image of the bottom half, so I'm not going do twice the work for the same information and b) I'm short, so since I usually tack these up on the wall so I can back up and really look at them, it's easier for me to get to the bottom half ;o)
Granted, it was still necessary to cut, fold, align and tape the sheets after printing, but it just looks so much cleaner. Yeah, good times. (Gawd I'm a nerd...)