Is “good evidence” a subjective or objective determination?

[fromderinside]

Geez. Back to the basics.

 Bayes theorem

One of the many applications of Bayes' theorem is Bayesian inference, a particular approach to statistical inference. When applied, the probabilities involved in the theorem may have different probability interpretations. With Bayesian probability interpretation, the theorem expresses how a degree of belief, expressed as a probability, should rationally change to account for the availability of related evidence. Bayesian inference is fundamental to Bayesian statistics.

 

[steve_bank]

...
Demonstrating Newtonian gravity is an objective fact regardless of who observes the expermet.
...

A priest, sitting under another apple tree, says, "Thank you God for this apple you have provided".

s there a Greek god of sophistry? It fails.

Someone who sees scifi everywhere and believes in ET visitant might say it was caused by ET in another dimension trough a cross dimensional vortex. A Wican practioner might say she did it with a spell.

Objective science and agency are mutually exclusive.

The fact that gavitational acceleration near the Earth's surface is 9.8m/s^2 does not depend on who measures it.


Agency as causation is subjective.

 

[steve_bank]

Geez. Back to the basics.

 Bayes theorem

One of the many applications of Bayes' theorem is Bayesian inference, a particular approach to statistical inference. When applied, the probabilities involved in the theorem may have different probability interpretations. With Bayesian probability interpretation, the theorem expresses how a degree of belief, expressed as a probability, should rationally change to account for the availability of related evidence. Bayesian inference is fundamental to Bayesian statistics.

Given a system with a demonstrated Mean Time Between Failures(MTBF), give an addition to system what is the new MTBF without having t rerun reliability demonstrations?

Bayesian techiqus can be used to draw comcl;usions with scant data.

 

[steve_bank]

Geez. Back to the basics.

 Bayes theorem

One of the many applications of Bayes' theorem is Bayesian inference, a particular approach to statistical inference. When applied, the probabilities involved in the theorem may have different probability interpretations. With Bayesian probability interpretation, the theorem expresses how a degree of belief, expressed as a probability, should rationally change to account for the availability of related evidence. Bayesian inference is fundamental to Bayesian statistics.

I am glad to see you are still rolling around.

 

[Swammerdami]

What's with the "Geez"? Did somebody say something stupid? I thought Marvin's "For example, if 1% of people get prostate cancer, that would actually be 2% of men" was an excellent example to show that Bayesian analysis is, fundamentally, simple arithmetic.

 

[steve_bank]

The use of statistics can be objective or subjective depending on assumptions and data.

Bayesian or any statistical technique has to be applied rationally.

"1% of people get prostate cancer, that would actually be 2% of men"

Sample size? Data? If...hen is simple logic.

Applying Bayes' Theorem in itself does not ensure truth. Same goes for any statistical method. Applying statistics with an understanding of the dta especialy with small samples is risky if yiu are making consequential decisons.

 

[Marvin Edwards]

The use of statistics can be objective or subjective depending on assumptions and data.

Bayesian or any statistical technique has to be applied rationally.

"1% of people get prostate cancer, that would actually be 2% of men"

Sample size? Data? If...hen is simple logic.

Applying Bayes' Theorem in itself does not ensure truth. Same goes for any statistical method. Applying statistics with an understanding of the dta especialy with small samples is risky if yiu are making consequential decisons.

Right, and I think that is one of the things that Bayes was concerned about, the factors that were not taken into account. But I don't really know that much about it. I think it is something that an intelligent person would normally do, prior to Bayes.

 

[steve_bank]

The use of statistics can be objective or subjective depending on assumptions and data.

Bayesian or any statistical technique has to be applied rationally.

"1% of people get prostate cancer, that would actually be 2% of men"

Sample size? Data? If...hen is simple logic.

Applying Bayes' Theorem in itself does not ensure truth. Same goes for any statistical method. Applying statistics with an understanding of the dta especialy with small samples is risky if yiu are making consequential decisons.

Right, and I think that is one of the things that Bayes was concerned about, the factors that were not taken into account. But I don't really know that much about it. I think it is something that an intelligent person would normally do, prior to Bayes.

Statistics applied to general things in physical reality is straightforward. I have a few thousand widgets and I want to estimate the weights of the objects. I take a few samples of 100 and average the sample averages yielding an estimate of the average weights. It gets a little more complcated but that s the idea.

In political or sales polling each human sample is weighted by demographics and location. One sample is multiplied by some number which can be subjective. The choice of weights while not always outright fraud can be intentional or unintentionally biase. Where the samples are taken can be biased. Ethnic groups may be excluded or minimized, and so on.

So when I hear a commercial say the majority of people prefer Colgate toothpaste proven statistically I think
yea right...'


There is a book that may still be in publication, How To Lie With Statistics.

Over the years I pissed off a few people when not accepting a conclusion without seeing data, I got burned once and never again.

Over on religion Drew is using subjective probabilities as a proof of a creator.

 

[Marvin Edwards]

The use of statistics can be objective or subjective depending on assumptions and data.

Bayesian or any statistical technique has to be applied rationally.

"1% of people get prostate cancer, that would actually be 2% of men"

Sample size? Data? If...hen is simple logic.

Applying Bayes' Theorem in itself does not ensure truth. Same goes for any statistical method. Applying statistics with an understanding of the dta especialy with small samples is risky if yiu are making consequential decisons.

Right, and I think that is one of the things that Bayes was concerned about, the factors that were not taken into account. But I don't really know that much about it. I think it is something that an intelligent person would normally do, prior to Bayes.

Statistics applied to general things in physical reality is straightforward. I have a few thousand widgets and I want to estimate the weights of the objects. I take a few samples of 100 and average the sample averages yielding an estimate of the average weights. It gets a little more complcated but that s the idea.

In political or sales polling each human sample is weighted by demographics and location. One sample is multiplied by some number which can be subjective. The choice of weights while not always outright fraud can be intentional or unintentionally biase. Where the samples are taken can be biased. Ethnic groups may be excluded or minimized, and so on.

So when I hear a commercial say the majority of people prefer Colgate toothpaste proven statistically I think
yea right...'


There is a book that may still be in publication, How To Lie With Statistics.

Over the years I pissed off a few people when not accepting a conclusion without seeing data, I got burned once and never again.

Over on religion Drew is using subjective probabilities as a proof of a creator.

The one thing that irks me is when someone reports a percentage increase or decrease without providing the relevant population data. For example, if the prevalence of a disease has increased by 100% (effectively doubling), but the increase is actually only going from 1 in a million to 2 in a million, then it would be nice to know the whole story.

 

[fromderinside]

Geez. Back to the basics.

 Bayes theorem

One of the many applications of Bayes' theorem is Bayesian inference, a particular approach to statistical inference. When applied, the probabilities involved in the theorem may have different probability interpretations. With Bayesian probability interpretation, the theorem expresses how a degree of belief, expressed as a probability, should rationally change to account for the availability of related evidence. Bayesian inference is fundamental to Bayesian statistics.

Given a system with a demonstrated Mean Time Between Failures(MTBF), give an addition to system what is the new MTBF without having t rerun reliability demonstrations?

Bayesian techiqus can be used to draw comcl;usions with scant data.

Time passes systems age ...