The thing is that the MAP and the WJ-III are of the go-until-you-can't-do-anymore variety, so one would expect a GE on one of them to be real--meaning that the 4th grader with a GE of 8.9 (94th percentile as compared with 4th graders) will have taken essentially the same test as the 8th grader with the GE of 8.9 (50th percentile as compared with 8th graders). My kids get the same GEs on these types of tests as they do on the ITBS/ITED. While this is admittedly a small sample size, I do believe that the test makers actually design the tests so that they will correlate fairly well with one another. All of these tests are attempting to measure the same reality and so you would expect the correlation.
Now, I do think that a 3rd grader taking the ITBS intended for 8th graders might run into problems with being able to finish in the time allotted.
I know that psychologists and others who are tasked with explaining GEs to parents always say not to pay any attention to them. And that's true, one shouldn't use GEs for placement because they don't indicate mastery. They only indicate the pathetic state of education in this country.
