Scientifical accuracy not actually that high?

[JoeP]

Good. As you were, then.

[Clutch Munny]

Originally Posted by fragment

I have at times wished people would stop getting het up about saying what is and isn't science and just look at whether the reasoning makes sense or the observations were accurate. Too many internet arguments have foundered on cries of "Not Science!" & "Is too!"

I was saying this about Intelligent Design before it was cool.



Wait, what? It actually isn't cool, and I never, ever will be?

[lisarea]

This is pretty good, even though it's Irish:

How Scientific Inquiry Works

[The Man]

I don’t really feel like starting a new thrad and this seems like as good a place to put it as any, so: Statisticians have finally found an area of agreement regarding the misuse of p-values. I don’t want to summarise too much of the article in case I get something wrong, but essentially, there is a popular misconception about what p-values actually are. It’s popularly believed that they represent a the chance that a hypothesis is true given certain results, but it’s actually the reverse; it’s the chance of getting certain results given that a hypothesis is true. If this confuses you, compare the statement “The Pope is Catholic” (guaranteed to be true, unless there’s a big secret about the Pope of which no one’s been informed) to “This Catholic is the Pope” (highly unlikely to be true).

I haven’t read this thread in its entirety in awhile but I’m pretty sure this dovetails with a lot of what TLR and others were talking about earlier in the thread.

[Dragar]

Originally Posted by The Man

If this confuses you, compare the statement “The Pope is Catholic” (guaranteed to be true, unless there’s a big secret about the Pope of which no one’s been informed) to “This Catholic is the Pope” (highly unlikely to be true).

Better: consider the hypothesis (model) that this person is a woman. The data is that this person is pregnant. Then it is very clear that the probability of the data given the model is nothing like the probability of the model given the data.

The probability the person is pregnant, given the person is a woman? Quite low.
The probability this person is a woman, given the person is pregnant? Quite high.

[JoeP]

[thanks]ed to read later

[The Lone Ranger]

I've argued before that we really should be doing more to encourage replication studies. One of the biggest problems with modern science is that there's tons of funding/pressure to do original research, but very little funding, support, or incentive to do replication studies to test the results of previous research.


But then, every working scientist knows that you put very little trust in the results of preliminary studies with small sample sizes, especially if they report results which are just barely statistically significant. (Studies like this are often done so that the researchers can decide whether it's worthwhile to sink the time, effort, and money into doing a proper study.)


Your average reporter ... doesn't seem to understand this. And that's a big part of the reason why the popular media tend to do such a terrible job when it comes to science-related journalism.


John Oliver has a great take on that: Last Week Tonight with John Oliver: Scientific Studies (HBO) - YouTube


The last portion is a great dissection of the reasons why I don't particularly like TED talks.

[-FX-]

Originally Posted by erimir

The motivation to subtly fudge things is definitely there.

It also affects what types of things you even investigate, as the article notes. Replications aren't likely to get you anywhere (career-wise) unless you can show that some important finding is bunk.

Then there is the real big problem with "science".

Money. And power.

Protective immunity
The study suggests pre-existing cross immunity from the antibodies – which can last for one year or longer – may partly explain why Africa’s Covid-19 death toll has been lower than expected.

Cold viruses may have helped curb Africa's Covid deaths | Mirage News

Fewer deaths
Despite Covid-19 infection rates in Africa being very high, the number of deaths has been comparatively low. Possible explanations include the continent's lifestyle and climate, and its relatively youthful population.

A team led by Edinburgh researchers investigated another possible reason - whether antibodies produced in response to exposure to the six other types of common coronaviruses that infect humans react with the Covid-19 virus.


In 339 pre-pandemic blood samples collected between 2000 and 2019 from people in Zimbabwe - where flu and pneumonia cause almost one in 10 deaths - the team detected antibodies against Covid-19 in 32 per cent of the population.

Previous infections
Detailed molecular and computational analysis of the antibodies' targets indicates many were produced by immune responses to infection with one of the six other types of coronavirus, including common cold viruses.

Some also appeared to have been produced to fight other pathogens, included flu viruses and parasites that cause malaria and sleeping sickness.

The antibodies detected included those targeting parts of the Covid-19 virus - for example, the spike protein - used in some vaccines to induce protective immunity.

Archived samples
Some studies in Europe have found Covid-19 antibodies in blood samples taken before the first cases were detected. However, these cases have been attributed to early, undetected spread of the virus before the pandemic was declared.

This is not the case in Zimbabwe because the team identified antibodies in archived samples collected from as early as 2000.

The world has a lot to learn from researching reasons underlying the way the pandemic unfolded in Africa. Only through studies such as this one can we derive actionable knowledge to better prepare for the next pandemic.

Professor Francisca MutapiSchool of Biological Sciences
The study, published in The Lancet Microbe journal, was supported by the Scottish Funding Council and National Institute for Health Research (NIHR).

Cold viruses may have helped curb Africa’s Covid-19 deaths | The University of Edinburgh

[-FX-]

During the Covid-19 panic/pandemic, efforts were made to get a billion people in Africa to take the mRNA treatments, with all the dire warnings of massive death and horror.

Most Africans refused. And almost no Africans died from Covid.

Meanwhile, in the highly vaccinated countries, the death rate was enormous. According to a lot of Doctors and researchers, the mRNA treatments causes a huge amount of deaths, and serious health problems.

Because of money, power and the need to control, the truth about this is not easy to know.

[Crumb]

Take your anti-vax bullshit elsewhere.

[-FX-]

Originally Posted by The Lone Ranger

I've argued before that we really should be doing more to encourage replication studies. One of the biggest problems with modern science is that there's tons of funding/pressure to do original research, but very little funding, support, or incentive to do replication studies to test the results of previous research.

When it comes to serious as fuck shit, like your life and health, there should be an extreme level of replication required, along with all the skeptical tools you can muster.

Especially if the "science" is only coming from those who will profit from the "conclusions" published.

[LarsMac]

Originally Posted by Crumb

Take your anti-vax bullshit elsewhere.

Why stop there?

[-FX-]

[LarsMac]

Self Portrait?

[Ari]

Ah yes the contrarian waveform.
In this problem the question of 'serious or trolling' will always collapse to the opposite view you take. If the bullshit he posted is taken serious, then you're obviously easily triggered, if it's taken as trolling then you obviously don't know the real truth and I feel sorry for you pathetic sheep. Whatever answer you choose, the contrarian collapses to the opposite answer.

[JoeP]

Originally Posted by Ari

Whatever answer you choose, the contrarian collapses to the opposite answer.

Oh no he doesn't!

[Kamilah Hauptmann]

Originally Posted by JoeP

Oh no he doesn't!

Wrong!