Well, should scientists be legally liable for deceiving the public and manipulating the evidence to support their OWN brliefs based on untrue claims and unsupported by scientific evidence?
175 thoughts on “Should scientists be legally accountable for deceiving the public?”
Leave a Reply
You must be logged in to post a comment.
Do you have your own thoughts on that? Why don’t you take a stab at it first? When I’m training someone the rule is they have to think about something for 15 minutes before asking a question. Have you thought about your question for 15 seconds?
OMagain,
I already stated my opinion earlier in the thread. Alan first needs to state his view of the null hypothesis.
Bill,
At the beginning of the school year, when your teachers were taking attendance for the first time, did they sigh when they got to your name?
I’m sure that “null hypothesis” is a term that Bill has heard somewhere, one that he has been told makes you sound very sciencey. Bill, there is no null hypothesis in Theobald’s study. There’s more than one way to do statistical testing in science.
Why? It’s not like you actually listen to anything anyone says. You just carry on regardless.
Why? It’s not like you actually listen to anything anyone says. You just carry on regardless.
Did you? What comment number? I’ve looked and there are only ~50 comments.
ID is the “null hypothesis” because the ID hypothesis is null
John responded “Yes indeed” because he was focusing on what Theobald did, which was to calculate two probabilities and compare them.
But colewd’s question is more general, and the answer to his broader, general question
is “No. You do not.”
Here’s a example.
I deal five cards and roll a fair die.
I want to know the relative likelihoods of
getting 4♠ 4♦ 4♥ 5♥ 4♣ and rolling a six
versus
getting 4♠ 4♦ 4♥ 5♥ 4♣ and rolling a one or a two or a three or a four or a five
Since I know absolutely nothing about the deck (it could be deck #1, or deck #2 or a deck with a million unique cards in it) I cannot calculate either probability. But I do know that the first outcome is five times less likely than the second outcome. See? I can calculate the “difference” (the ratio, actually) without knowing either underlying probability.
Likewise, I can say that three trees is 10^2680 times less likely than it was before I did the analysis.
At some future date, colewd might learn about Bayes, but I doubt it.
colewd,
It’s not difficult, just less likely. There are more places to go that aren’t the same place than are. Unless there is some kind of attractor, it’s generally more likely that 2 similar sequences came from the same place than that they arrived there independently.
Why do you ascribe the figure 10^2860 (from Theobald’s paper) to convergent evolution?
colewd,
Typically, it’s the default position that a distribution is due to chance alone.
Here is Theobald’s own defence against Koonin’s critique. He discusses null hypotheses in depth.
Null hypotheses are not relevant to the genetic code argument, which is qualitative not quantitative.
Theobald’s defence against Koonin’s criticism is that his tests favour trees – phylogenetic structure – over mere similarity, where such structure exists. If all you’ve got to go on is mere similarity – which is all you have in Koonin’s artificial dataset – it picks that.
Allan Miller,
This figure is from a 2 tree analysis. If the two trees rose independently then the sequences that were highly related had to arise separately. Assuming that this is from random change and selection each of the events are not very likely hence the fantastically small number for 2 separate origins versus 1.
As the probability of forming a protein gets smaller then universal common descent becomes a more attractive explanation if you discount the design argument. If we had an identified mechanism that produced new proteins predictably then multiple origins becomes a more viable alternative.
The guys that critiqued Koonin’s work did not believe that Theobald’s paper brought any new insight to the party except the exercising of some new statistical techniques. Do you agree with this?
DNA_Jock,
Hi Jock
Thanks for posting. I have just briefly looked a Bayes. I think that you are right that if you isolate the differences between two events and then you can get a differential probability without knowing the probability of each but just analyzing the differences.
In this case the difference was how a protein arose.
case 1 independently evolved
case 2 inherited from the ancestor
Do you have any thoughts on the value of the paper we are discussing?https://www.google.com/webhp?sourceid=chrome-instant&rlz=1C1CHZL_enUS729US729&ion=1&espv=2&ie=UTF-8#q=theobald+2010&*
colewd,
Note that DNA Jock’s comment is purely a quibble that’s actually irrelevant to Theobald’s paper, because Theobald does in fact compute the probabilities of events. Is it weird that you admit this one irrelevant truth and ignore all the relevant truths that everyone has tried to explain to you?
John Harshman,
OMagain,
keiths,
I want to apologize to you guys. After reading more of the paper and starting to learn that Theobald’s approach is different then I am used to I realize that my approach to the probability problem was wrong. I appreciate all the discussion you guys have provided.
colewd,
Thank you for acknowledging your mistake.
Do you now understand why
a) common descent is a scientific conclusion, not an assumption?
b) separate creation is false, unless you assume (for no good reason) that the Creator acts in a way that makes creationism appear to be false and common descent appear to be true?
colewd,
I am momentarily speechless.
Nice gesture, Bill.
If you do in fact assume this colewd, why?
colewd,
There had to be some kind of mechanism that correlates them in some way. But Theobald is not directly contrasting to convergent evolution as a mechanism in ‘independent origin’ models.
Theobald does not cover how the datasets become so full of aligned proteins – except, of course, for the obvious cause of common descent.
I’m not sure who you mean. In terms of evidence for common descent, I tend to go for things like the genetic code rather than statistical methods, because they are easier to understand (not least for me). It’s hard to see how that can be so widespread if it were not commonly descended. Of course one might say ‘it was designed’. In which Theobald’s analysis of the proteins themselves still answers a question regarding the point of origin. It was prior to the archaeal, bacterial and eukaryotic split. That is, design or no, all of life is commonly descended.
Allan Miller,
We all agree that there is common genetics among all living creatures. I agree that what appears to be non functional sequences that are shared is additional evidence. What I am very skeptical of is that we can account for this by cell division and sexual reproduction.
Theobald appears to be offering an additional way to analyze common descent other then sequence comparison. I am frankly struggling to understand his methods at this point.
colewd,
The mechanism of template copying is overwhelmingly the most likely explanation for sequence commonality and, equally importantly, the tree-like nature of differences across multiple independent gene sets. Common descent predicts precisely the things we find.
Clearly you are struggling, because Theobald is in fact doing sequence comparisons. Perhaps you don’t consider phylogenetic analysis to be sequence comparison, but it certainly arises out of sequence comparison. His methods are not at all unusual, just ordinary likelihood-based phylogenetic analysis followed by likelihood ratio comparisons. It’s the application directly to testing universal common descent that’s the new bit. Otherwise it’s all more or less the same thing as in hundreds of other papers, including the ratite paper you had previously been puzzling over.
And it cross-checks well with the fossil record, to the resolution allowed by the latter.
It could all be a massive coincidence, however, while something with no credible evidence could be true instead. Yeah, let’s go with that.
Glen Davidson
Good form. I’m impressed.