Yup, incident meter. Those were scans, so this gets close to stuff we shouldn't talk about, but I think you could easily adjust either to get similar prints. There are also more in that series if you click through to the set. I think I went from -2 to +4.
The point still stands though that I just haven't experienced that Ektar suddenly falls apart with a stop or two of over exposure. I've found it seems to perform pretty much according to Kodak's recommendations, which I believe is an exposure latitude of -1/+2.
In order to make the test give us the info that would help us understand how to approach this film, we'd have to make each print so that the midtones matched. That is the thing that will show a practical analysis of the film's latitude and how its characteristics change when making a real-world use of that latitude.
We know that overexposure of a neg will brighten a normal print, and that underexposure of a neg will darken a normal print. We need to know what happens to a print from an under or over exposed neg when the prints are manipulated to look as normal as possible, despite the abnormal exposure of the neg. That is what we do when printing from real-world negs. We don't just accept the fact that overexposure will make a bright print and underexposure will make a dark print. We manipulate prints away from the norm to adjust for less-than-perfect negative exposure. That is the practical use of latitude, so that is what we need to look at. You match the midtones to the normal print from the normal negative, and it lets us see what happens to the low and high tones.
We'd also have to first neutralize the color balance of the print from the normally exposed neg, and use the same filter pack (though, as was mentioned, not the same print exposure) for all the under and over pix. Then we could see how under and overexposure affect both contrast and color balance. But, most importantly, the midtones must match in density. A computer is probably not the best tool to analyze the results, but it can show us something useful if that is all you have. I think you have the shoot covered, and you can get a good analysis with the film you already have.
Ironically enough, minilab 4x6's are actually a halfway decent way to do this test, especially if the machine makes optical prints directly from the negatives. (Blue Moon is one lab that does this.) The auto exposure of the machine tries to get what its programming says is an ideal print, so a bracketed sequence printed in this fashion can tell you a lot. The auto color balance on the machines does the same. It will not tell you exactly how color balance is affected by exposure changes, but it will tell you to what extent a neg can be improperly exposed and still be color correctable.