Hmmmm. I have to say it's interesting and plausible, but I'm not entirely convinced. I tend to be skeptical of claims of statistical significance if the person doing the analysis is not a statistician. But in any case, it deduces that two of the three sentences that Meier concludes were interpolations really were interpolations, and only the third, longer passage, might have come from the proto-Luke source.
I'd be interested in how this all relates to the Q source, which with Mark is supposed to be the source for Matthew and Luke.