Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

The catch-all forum for general topics and debates. Minimal moderation. Rated PG to PG-13.
Billy Shears
Sunbeam
Posts: 50
Joined: Fri Jul 23, 2021 8:13 pm

Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by Billy Shears »

I just happened to come across the following argument in Steven Pinker's new book Rationality, and noticed that this is precisely the type of argument that Kyler Rasmussen relies on in his Book of Mormon analysis.

Loosely quoting Pinker, "Russell M. Nelson is almost certainly a space alien. The probability that a randomly selected person on earth is the president of the LDS Church is tiny: one out of 7.8 billion, or .00000000013. Russell M. Nelson is the president of the LDS Church. Therefore, Nelson is probably not a human being." (page 128)

This is obviously awful reasoning, but why, precisely? Ironically, Pinker explains that the problem with this approach is that it is relying on traditional statistical testing with p-values, and that the way to correct for it is to use valid Bayesian reasoning. In Pinker's words:
Steven Pinker wrote:And so the convention arose that scientists should adopt a critical level that ensures that the probability of rejecting the null hypothesis when it is true is less than 5 percent: the coveted “p < .05.” (Though one might have thought that the costs of a Type II error should also be factored in, as it is in Signal Detection Theory, for some equally obscure historical reason it never was.)

That’s what “statistical significance” means: it’s a way to keep the proportion of false claims of discoveries beneath an arbitrary cap. So if you have obtained a statistically significant result at p < .05, that means you can conclude the following, right?
  • The probability that the null hypothesis is true is less than .05.
  • The probability that there is an effect is greater than .95.
  • If you rejected the null hypothesis, there is less than a .05 chance that you made the wrong decision.
  • If you replicated the study, the chance that you would succeed is greater than .95.
Ninety percent of psychology professors, including 80 percent of those who teach statistics, think so. But they’re wrong, wrong, wrong, and wrong. If you’ve followed the discussion in this chapter and in chapter 5, you can see why. “Statistical significance” is a Bayesian likelihood: the probability of obtaining the data given the hypothesis (in this case, the null hypothesis). But each of those statements is a Bayesian posterior: the probability of the hypothesis given the data. That’s ultimately what we want—it’s the whole point of doing a study—but it’s not what a significance test delivers. If you remember why Irwin does not have liver disease, why private homes are not necessarily dangerous, and why the pope is not a space alien, you know that these two conditional probabilities must not be switched around. The scientist cannot use a significance test to assess whether the null hypothesis is true or false unless she also considers the prior—her best guess of the probability that the null hypothesis is true before doing the experiment. And in the mathematics of null hypothesis significance testing, a Bayesian prior is nowhere to be found.

Most social scientists are so steeped in the ritual of significance testing, starting so early in their careers, that they have forgotten its actual logic.

Pinker, Steven. Rationality (pp. 224-225)
drumdude
God
Posts: 5214
Joined: Thu Oct 29, 2020 5:29 am

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by drumdude »

“Null hypothesis” is a 4 letter word to LDS apologists.
Philo Sofee
God
Posts: 5015
Joined: Thu Oct 29, 2020 1:18 am

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by Philo Sofee »

Good find Billy! I wonder if Rasmussen will ever grasp the point to this.....
Billy Shears
Sunbeam
Posts: 50
Joined: Fri Jul 23, 2021 8:13 pm

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by Billy Shears »

Thanks Philo.

Some people think Bayesian analysis is a superior method of weighing uncertain evidence. Others think it is algebraic hocus pocus that doesn’t have any real value. After all, if it can be used by multiple people who are ostensibly literate in mathematics to prove the Book of Mormon is true, can’t it be used to prove that any false thing is true? And if that’s the case, isn't "Bayesian reasoning" just a euphemism for obfuscating the evidence to get the results you want?

In an effort to defend Bayesian reasoning from the hacks who have sullied its reputation, I’d like to provide an illustration of valid Bayesian reasoning in a way that is intuitive and illustrates why it is superior to traditional statistics (e.g. hypothesis testing).

Here is the illustration: assume 1% of the population has breast cancer. A breast-cancer screening is 90% accurate—if you have breast cancer, the test will be positive 90% of the time, and if you don’t have cancer, the test will be negative 90% of the time. You take the test and the test comes back positive! Crap! What are the chances you really have cancer?

According to a 2011 article in the British Medical Journal, when this question was given to a sample of doctors, the most common answer they gave was that you’d have about an 85% chance of having cancer. That is an intuitive answer—90% of people who have cancer have a positive test—so it would seem the chances are 90% that you would fit in with this group. Similarly, 90% of people who don’t have cancer show negative test results, so you don’t fit into that group. A traditional statistician would set the null hypothesis to not having cancer, and you’d be able to reject the null hypothesis at the 10% level.

It turns out those interpretations of the data are spectacularly wrong. Here’s why.

Imagine 1,000 people. 10 of them have cancer (that’s 1% of the total population). The other 990 do not. The 1,000 people are all screened. Of the 10 people who have cancer, 9 get a positive test result (90%). Of the 990 who don’t have cancer, 99 get a false positive (10%).

Now let’s revisit the question. If you get a positive test result, what are the probabilities you have cancer? The answer is easy. 108 people get a positive test result. Of those, 9 have cancer. The probability of having cancer if your test results are positive are 9/108 = 8%. Even if your screening test says you have cancer, you can be 92% confident that you really don’t.

That is proper Bayesian reasoning.
User avatar
Physics Guy
God
Posts: 1557
Joined: Tue Oct 27, 2020 7:40 am
Location: on the battlefield of life

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by Physics Guy »

That is a good example.

I like Bayesian inference, but I'm suspicious of Bayesian enthusiasm. Bayes's theorem is just logic by numbers: no less but no more. Spinning straw into gold is Rumpelstiltskin, not Bayes.
I was a teenager before it was cool.
dastardly stem
God
Posts: 2259
Joined: Tue Nov 03, 2020 2:38 pm

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by dastardly stem »

Billy Shears wrote:
Fri Oct 22, 2021 2:27 pm
Thanks Philo.

Some people think Bayesian analysis is a superior method of weighing uncertain evidence. Others think it is algebraic hocus pocus that doesn’t have any real value. After all, if it can be used by multiple people who are ostensibly literate in mathematics to prove the Book of Mormon is true, can’t it be used to prove that any false thing is true? And if that’s the case, isn't "Bayesian reasoning" just a euphemism for obfuscating the evidence to get the results you want?

In an effort to defend Bayesian reasoning from the hacks who have sullied its reputation, I’d like to provide an illustration of valid Bayesian reasoning in a way that is intuitive and illustrates why it is superior to traditional statistics (e.g. hypothesis testing).

Here is the illustration: assume 1% of the population has breast cancer. A breast-cancer screening is 90% accurate—if you have breast cancer, the test will be positive 90% of the time, and if you don’t have cancer, the test will be negative 90% of the time. You take the test and the test comes back positive! Crap! What are the chances you really have cancer?

According to a 2011 article in the British Medical Journal, when this question was given to a sample of doctors, the most common answer they gave was that you’d have about an 85% chance of having cancer. That is an intuitive answer—90% of people who have cancer have a positive test—so it would seem the chances are 90% that you would fit in with this group. Similarly, 90% of people who don’t have cancer show negative test results, so you don’t fit into that group. A traditional statistician would set the null hypothesis to not having cancer, and you’d be able to reject the null hypothesis at the 10% level.

It turns out those interpretations of the data are spectacularly wrong. Here’s why.

Imagine 1,000 people. 10 of them have cancer (that’s 1% of the total population). The other 990 do not. The 1,000 people are all screened. Of the 10 people who have cancer, 9 get a positive test result (90%). Of the 990 who don’t have cancer, 99 get a false positive (10%).

Now let’s revisit the question. If you get a positive test result, what are the probabilities you have cancer? The answer is easy. 108 people get a positive test result. Of those, 9 have cancer. The probability of having cancer if your test results are positive are 9/108 = 8%. Even if your screening test says you have cancer, you can be 92% confident that you really don’t.

That is proper Bayesian reasoning.

No...No...and NO. Proper Bayesian reasoning goes something like this.

If the Book of Mormon was written anciently then Joseph Smith would have had a vision of God and Jesus. If the Book of Mormon was written in the 19th century then Joseph Smith would not have had a vision of God and Jesus. Since Joseph says he had a vision of God and Jesus, then its more probable that the Book was written anciently. And, I mean, the book has Early Modern English in it, so its more likely that it was written anciently since Joseph likely wouldn't have been able to write a phrase here or there in Early Modern English. If it was written in our modern day it would be so unlikely Joseph could have dictated Early Modern English syntax to whatever degree he did, unless God or his mighty powers were feeding him those lines, that we can safely say he didn't. And that makes the Book of Mormon being of ancient origin pretty probable since, you know, it's scripture.

And you can plug in any old numbers you want to show this because we don't have something to compare it to, anyway. It's like 100% likely or 65%. Whatever the case it's more likely that Joseph had it dictated to him by God and it was written anciently originally then Joseph could have written it. Now that's proper Bayesian reasoning. 8-)
“Every one of us is, in the cosmic perspective, precious. If a human disagrees with you, let him live. In a hundred billion galaxies, you will not find another.”
― Carl Sagan, Cosmos
Philo Sofee
God
Posts: 5015
Joined: Thu Oct 29, 2020 1:18 am

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by Philo Sofee »

dastardly stem wrote:
Fri Oct 22, 2021 5:16 pm
Billy Shears wrote:
Fri Oct 22, 2021 2:27 pm
Thanks Philo.

Some people think Bayesian analysis is a superior method of weighing uncertain evidence. Others think it is algebraic hocus pocus that doesn’t have any real value. After all, if it can be used by multiple people who are ostensibly literate in mathematics to prove the Book of Mormon is true, can’t it be used to prove that any false thing is true? And if that’s the case, isn't "Bayesian reasoning" just a euphemism for obfuscating the evidence to get the results you want?

In an effort to defend Bayesian reasoning from the hacks who have sullied its reputation, I’d like to provide an illustration of valid Bayesian reasoning in a way that is intuitive and illustrates why it is superior to traditional statistics (e.g. hypothesis testing).

Here is the illustration: assume 1% of the population has breast cancer. A breast-cancer screening is 90% accurate—if you have breast cancer, the test will be positive 90% of the time, and if you don’t have cancer, the test will be negative 90% of the time. You take the test and the test comes back positive! Crap! What are the chances you really have cancer?

According to a 2011 article in the British Medical Journal, when this question was given to a sample of doctors, the most common answer they gave was that you’d have about an 85% chance of having cancer. That is an intuitive answer—90% of people who have cancer have a positive test—so it would seem the chances are 90% that you would fit in with this group. Similarly, 90% of people who don’t have cancer show negative test results, so you don’t fit into that group. A traditional statistician would set the null hypothesis to not having cancer, and you’d be able to reject the null hypothesis at the 10% level.

It turns out those interpretations of the data are spectacularly wrong. Here’s why.

Imagine 1,000 people. 10 of them have cancer (that’s 1% of the total population). The other 990 do not. The 1,000 people are all screened. Of the 10 people who have cancer, 9 get a positive test result (90%). Of the 990 who don’t have cancer, 99 get a false positive (10%).

Now let’s revisit the question. If you get a positive test result, what are the probabilities you have cancer? The answer is easy. 108 people get a positive test result. Of those, 9 have cancer. The probability of having cancer if your test results are positive are 9/108 = 8%. Even if your screening test says you have cancer, you can be 92% confident that you really don’t.

That is proper Bayesian reasoning.

No...No...and NO. Proper Bayesian reasoning goes something like this.

If the Book of Mormon was written anciently then Joseph Smith would have had a vision of God and Jesus. If the Book of Mormon was written in the 19th century then Joseph Smith would not have had a vision of God and Jesus. Since Joseph says he had a vision of God and Jesus, then its more probable that the Book was written anciently. And, I mean, the book has Early Modern English in it, so its more likely that it was written anciently since Joseph likely wouldn't have been able to write a phrase here or there in Early Modern English. If it was written in our modern day it would be so unlikely Joseph could have dictated Early Modern English syntax to whatever degree he did, unless God or his mighty powers were feeding him those lines, that we can safely say he didn't. And that makes the Book of Mormon being of ancient origin pretty probable since, you know, it's scripture.

And you can plug in any old numbers you want to show this because we don't have something to compare it to, anyway. It's like 100% likely or 65%. Whatever the case it's more likely that Joseph had it dictated to him by God and it was written anciently originally then Joseph could have written it. Now that's proper Bayesian reasoning. 8-)
Just a small nit with your absolutely logical and valid reasoning here. "Now that is proper Rasmussen Bayesian reasoning." There I corrected it for you... :lol:
drumdude
God
Posts: 5214
Joined: Thu Oct 29, 2020 5:29 am

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by drumdude »

Someone really should find Kyler's professors and ask them what they think of his education and what he has done with it.

Oh wait, Kyler didn’t get an education in statistics. So there’s no one to ask.
Billy Shears
Sunbeam
Posts: 50
Joined: Fri Jul 23, 2021 8:13 pm

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by Billy Shears »

Thanks, Physics Guy.

With that example serving as a framework for a proper analysis, I’d like to express my biggest pet peeve with Kyler’s perversion of Bayesian reasoning.

Episode 7 (Nahom) is a good example. https://interpreterfoundation.org/estim ... vidence-7/

Before diving into Yemen, note that in addition to the a priori assumption (i.e. 1% of the total population having cancer), my example above has 4 implicit probabilities:

1- The probability the test is positive if you don’t have cancer (10%)
2- The probability the test is negative if you don’t have cancer (90%)
3- The probability the test is positive if you do have cancer (90%)
4- The probability the test is negative if you do have cancer (10%)

In the Nahom example, the test (like the cancer screening test) is whether or not NHM is found in Yemen right where the Book of Mormon says it should be. We are trying to use this test to figure out if the Book of Mormon is true (like trying to figure out if the patient really has cancer).

Cutting through the obfuscation and rounding the numbers, here are the numbers Kyler came up with.

1- The probability NHM is found if the book is false: 1%
2- The probability NHM is not found if the book is false: 99%
3- The probability NHM is found if the book is true: 100% (!)
4- The probability NHM is not found if the book is true: 0% (!)


Those are Kyler’s numbers. The 1% in item 1 (and its complement in item 2) is fair. Kyler would argue it is generous to the critics. Sure.

The part that is ridiculous and proves Kyler doesn’t understand Bayesian inference is item 3 (and its complement item 4). Imagine getting into a time machine and talking to Hugh Nibley in, say, 1980. Imagine we asked him the following question: “Dr Nibley. We are about to dig in Yemen. Assuming the Book of Mormon is true, what are the chances we will find a place called NHM on the frankincense trail in this expedition?”

Here are two possible answers:

Answer 1: “That was a long time ago, and archeological evidence is scarce. Not finding a sign that says “Welcome to Zarahemla” does not prove the Book of Mormon is false. Likewise, not finding a sign that says “Welcome to NHM” doesn’t prove the Book of Mormon is false. Assuming the Book of Mormon is true, the probability of finding a sign that says NHM is about 1%.”

Answer 2: “The chances are 100%. If the Book of Mormon is true, then we will find a place called NHM. There is no way we could not. By elementary logic, if we don’t find a place called NHM, that definitively proves the Book of Mormon is false. Modus Tollens. Case closed. NHM simply must be there.”

If the honest, competent answer is #1, then NHM isn’t evidence of the Book of Mormon one way or another—finding NHM is just as unexpected if the Book of Mormon is true as it would be if the Book of Mormon is false.

On the other hand, if the honest, competent answer is #2—the answer that Kyler chose—consistency would demand that we also ask other questions such as:

1- What is the probability we’ll find a sign that says “Welcome to the City of Aaron?”
2- What is the probability we’ll find a sign that says “Welcome to the City of Ablom”?
.
.
.
150- What is the probability we’ll find a sign that says “Welcome to the City of Zarahemla”?
151- What is the probability we’ll find a sign that says “Welcome to the City of Zeezrom”?

If we can be 100% certain that a true Book of Mormon means we’ll find NHM as Kyler claims, why can’t we also be 100% certain that a true Book of Mormon means we’ll also find all of these other cities?

If Kyler treated all geographical places listed in the Book of Mormon in the same way he treats NHM, not finding a single one of the hundred-plus cities in the Book of Mormon would be definitive proof the Book of Mormon is false.

In general, Kyler uses exaggerated estimates of the unlikeliness of seeing the evidence we see under the assumption the Book of Mormon is a 19th century production (e.g. the probability of early modern English in a 19th century book is 5.24 x 10^-24). My pet peeve is what he does next. He assumes that what we see is precisely what we'd expect if the Book of Mormon is historical (e.g. according to him, the probability of seeing early modern English in a book written by a 5th century Mesoamerican is precisely 100%). That second half is ridiculous. But like a magician using misdirection, he puts all of his focus on the first half.
User avatar
Res Ipsa
God
Posts: 9568
Joined: Mon Oct 26, 2020 6:44 pm
Location: Playing Rabbits

Re: Kyler Rasmussen Proves Russell M. Nelson Is An Alien!

Post by Res Ipsa »

Billy, I think your explanation of Bayesian analysis is quite good, but I don’t see what your paraphrase of Pinker’s argument has to do with the shortcomings of p values or even frequentist statistics. It’s just an improper application of frequentist statistics. You didn’t draw Nelson’s name at random. You choose him because he is President off the Church. And he wasn’t chosen as President randomly either. When his predecessor passed, the odds of him becoming President were nearly 100%.

There are inherent problems with the use of frequentist statistics, including reliance on p values, but I don’t see Pinker’s modified example illustrates those problems.
he/him
When I go to sea, don’t fear for me. Fear for the storm.

Jessica Best, Fear for the Storm. From The Strange Case of the Starship Iris.
Post Reply