We’ve been working through McLain et al, 2018’s attempt to put feathers on dinosaurs. We’ve elected to skip diagrams 38 and 39 because frankly, they are a mess. They remind me strongly of the initial runs with each dataset, and those are almost exclusively messes of data from which very little can be learned. There are simply too many different kinds and not enough data to form accurate pictures. Therefore this post will focus on figures 40 and 41, the figures that focus on the Tyrannosaur subset of the data.
In preparation for running the data from Lee et al, the dataset was trimmed to match McLain et al’s subset for figures 40 and 41. Upon running with BDISTMDS and BARCLAY, 333 characters were retained for analysis. That sounds like a lot until you realize the Lee et all dataset had over 1500 characters available. Because bootstrapping values were not presented in the main paper, we examined them here for BDISTMDS and BARCLAY. In both cases, bootstrapping values were quite good comparatively to previous examinations. There were a few in the fifties but most were in the mid 70s or higher. While these are not bad numbers, they are not as good as they could be either.
Having examined the bootstrapping values we move into the graphical displays of the values. Immediately we notice a huge difference. The BDISTMDS graphic is very similar (though not identical) to the one produced by McLain et al. My only explanation for the differences is they may have used a slightly lower taxic relevance, though this would require editing the BDISTMDS software and this is not mentioned. Even if they used a lower taxic relevance, it should be the same. However, I could overlook that were it not for BARCLAY throwing up a massively different results. Have a look below.
Again, BARCLAY finds much more continuity and much less discontinuity than BDISTMDS. This result right here could be the poster child for why BARCLAY is actually worse than BDISTMDS by a country mile. 45% of the time according to Todd Wood’s own published work BDISTMDS fails to detect discontinuity. And BACLAY detects even less. The entire BARCLAY thing might as well be one block with Dilophosaurus connecting the two major blocks. Only Majungasaurus is not part of the block. Thing is, I actually think the Tyranosaurus/Allosaurus block is pretty close to accurate. I’ve argued for that before. The problem is, I do not trust BARCLAY because it detects so much less discontinuity than BDISTMDS, when we know BDISTMDS is inaccurate nearly half the time!
Just as a side bar, I tested both of these at 85 and 95% character relevance. Both worked, and produced radically different results. What this tells me is having an arbitrary cut off value makes BDISTMDS/BARCLAY absolutely worthless. I could argue that T rex is related to Allosaurus, or not related to Allosaurus depending on which cut off I chose. The system is completely arbitrary, never mind the veneer of statistical rigor, and that makes it worthless.
Having completed looking at the graphs, I went ahead and ran the completely arbitrary MDS plots because they are part of the system, never mind how you can read them however you choose. Note the MDS plots look fairly similar, even though the graphs do not.
On the plus side, while the graphs look nothing alike, at least the MDS plots look similar and aren’t tetrahedral. So I guess that’s progress?
Repeating figures 40 and 41 has been very enlightening. It has shown that the entire system is completely subjective, hiding behind a veneer of statistical objectivity. Nothing about statistical baraminology is objective. It would, and this is not sarcasm, be far more objective to rely on human intuition, or, shocker, the Biblical definition of a reproductive group. Maybe instead of wasting so much time with an amalgamation of statistics that produce subjective results, the statistical baraminologists should attempt some in vitro fertilization experiments between organisms that could be members of the same kind. At least that would be empirical. What we have now is an unverifiable mass of meaningless statistics that look impressive but have no foundation. All because Kurt Wise got the brilliant idea of importing cladistics into baraminology 25 years ago.
Do you know what’s going to happen when you die? Are you completely sure? If you aren’t, please read this or listen to this. You can know where you will spend eternity. If you have questions, please feel free to contact us, we’d love to talk to you.