Lewandowsky doesn't like us either

[jimells]

Take having an age of 32757 in a data set I suspect some would say that is one data point in many and therefore a minor error (although outliers heavily skew parametric stats).

Would a credit card company hire the Lewandowsky crew to handle their data?

 

[CFS_for_19_years]

All of the comments are excellent and at least three are quite long and very excellent. Even though I cringe at the idea of giving him the traffic, the comments are worth reading.

 

[JaimeS]

The comments are gold. Didn't bother to read the article itself.

 

[Luther Blissett]

A researcher had a 32,757 year old in his data. Therefore, ALL researchers have 32,757 year olds in their data.

A researcher used minors without consent in his work. Therefore, ALL researchers use minors without consent in their work.

Professor Stephan Lewandowsky is wrong. Therefore, ALL Stephan Lewandowskys are wrong.

This logic thing is fun

 

[TiredSam]

The comments are gold. Didn't bother to read the article itself.

x2

 

[Yogi]

Dr Harold Shipman was a serial killer doctor. Therefore ALL doctors are serial killers

Sir Jimmy Saville was a predatory paedophile. Therefore ALL those knighted (Sir) are predatory paedophiles

Psychiatrist Hannibal Lecter was a cannibalistic mass murderer. Therefore ALL psychiatrists are cannibalistic mass murderers.



Professor Sir Simon Wessley is a doctor who is also a psychiatrist and has been knighted by the Queen.

 

[Forbin]

I wonder if the 32,757-year-old person in the data set could have come from Mount Dora, Florida, because that's the zip code of Mount Dora, Florida.

 

[Skippa]

I wonder if the 32,757-year-old person in the data set could have come from Mount Dora, Florida, because that's the zip code of Mount Dora, Florida.

Winner.

Fat fingers and small digital form fields.

 

[JaimeS]

I wonder if the 32,757-year-old person in the data set could have come from Mount Dora, Florida, because that's the zip code of Mount Dora, Florida.

A very reasonable conclusion. Equally likely is that this person

or this person

Actually have M.E.

Let's go with that.

-J

 

[Freddy]

I have got a rather boring explanation for the 32757 year old minor. This seems to be a typical EXCEL error. 32768 is 2^15 so 32757 is 2^15-11, so the minor is actually of age 11. Or the whole is to be understood as two's complement. The number is just too close to 2^15. The p-value for my theory is p<0.001, so in the range of PACE. If this has been said already, I apologize one more time for boring.

Anyway, I don't oppose to the title of this thread.

 

[Luther Blissett]

I have got a rather boring explanation for the 32757 year old minor. This seems to be a typical EXCEL error. 32768 is 2^15 so 32757 is 2^15-11, so the minor is actually of age 11. Or the whole is to be understood as two's complement. The number is just too close to 2^15. The p-value for my theory is p<0.001, so in the range of PACE. If this has been said already, I apologize one more time for boring.

Anyway, I don't oppose to the title of this thread.

I have nightmares of learning to use EXCEL, but wouldn't it make sense to have some kind of specified range of values for the field? (like 18 -99 years) If you're doing research and want to include an age variable in the analysis - wouldn't you want to make sure your ages are at least likely?

 

[Forbin]

A very reasonable conclusion. Equally likely is that this person or this person actually have M.E. Let's go with that.-J

I think the 32,000-year-old person was actually a member of a study that looked for a correlation between climate change denial and belief in conspiracy theories.

http://websites.psychology.uwa.edu....yetalPsychScienceinPressClimateConspiracy.pdf

"Uhn! Thag think statistical models fail to account for cosmic ray influence on cloud formation.
Also think suspicious that "Martian" win Golden Globe for best comedy or musical."​

 

[Riley]

Here's what I think of that "article."

 

[Freddy]

I have nightmares of learning to use EXCEL, but wouldn't it make sense to have some kind of specified range of values for the field? (like 18 -99 years) If you're doing research and want to include an age variable in the analysis - wouldn't you want to make sure your ages are at least likely?

I fear I have to defend Leandowsky in this case: these kind of errors happen anywhere, even in serious science...
However age of 32000 should skew results noticeable.

 

[user9876]

I fear I have to defend Leandowsky in this case: these kind of errors happen anywhere, even in serious science...
However age of 32000 should skew results noticeable.

These type of errors happen which is why checks need to be done on the data and results. If academics don't have a decent set of test errors will creep in and the results won't be trustable

 

[jimells]

I have nightmares of learning to use EXCEL, but wouldn't it make sense to have some kind of specified range of values for the field? (like 18 -99 years) If you're doing research and want to include an age variable in the analysis - wouldn't you want to make sure your ages are at least likely?

Since they were creating propaganda with a predetermined conclusion, there was simply no need to bother with tedious details like editing the data, or even looking at it, apparently. Their bonehead programmers need to venture out of the ivory tower once in a while, although they probably wouldn't last a day in an office where software is expected to, you know... actually work.

 

[jimells]

I fear I have to defend Leandowsky in this case: these kind of errors happen anywhere, even in serious science...
However age of 32000 should skew results noticeable.

Now that you mention it, I do recall a certain NASA mission to Mars that disintegrated because someone mixed up metric and standard measuring systems...

 

[Luther Blissett]

I fear I have to defend Leandowsky in this case: these kind of errors happen anywhere, even in serious science...
However age of 32000 should skew results noticeable.

These type of errors happen which is why checks need to be done on the data and results. If academics don't have a decent set of test errors will creep in and the results won't be trustable

Yes, that's sort of why I asked. I would've thought that entering vast amounts of numbers into a spreadsheet is just asking for errors to be made.

I used to work with a massive filing system where people were sorted by name and identifying number and these kind of errors happened all the time.

I'm guessing that specialist data entry clerks are not used in this kind of work, so I wonder what type of system is used to try to ensure the integrity of the data. I've made 3 errors just typing this message, but If it was just a list of numbers I would make a lot more, but I would hope that I would notice a neolithic age in a list of them.

 

[biophile]

The error itself is only part of the story. The other part is the reluctance to do much about it once discovered.

https://jcoynester.wordpress.com/20...rs-in-stephen-lewandowskys-paper-in-plos-one/

It appears the only reason something was eventually done about it was because someone was persistent and PLOS is "ethically distinctive".

 

[Woolie]

A researcher had a 32,757 year old in his data. Therefore, ALL researchers have 32,757 year olds in their data.

A researcher used minors without consent in his work. Therefore, ALL researchers use minors without consent in their work.

Professor Stephan Lewandowsky is wrong. Therefore, ALL Stephan Lewandowskys are wrong.

This logic thing is fun

!!!