The statistics proving anything comment is correct. All you have to do is pick and choose which data set you use as well as any of a number of other variables. You can even make numbers up. There are so many statistical manipulations one can do with numbers it makes the head spin. I watched that sort of thing happen in the medical resear…
The statistics proving anything comment is correct. All you have to do is pick and choose which data set you use as well as any of a number of other variables. You can even make numbers up. There are so many statistical manipulations one can do with numbers it makes the head spin. I watched that sort of thing happen in the medical research department I worked in. I was to told to chill when I objected because if the numbers weren’t crunched “ just so” then no one would get any grant money. Their justification. I finally left that field and it’s only gotten worse as we can see, today.
Yea, I've been there too. I left a position doing research for a non-profit after I was told just to "Make the numbers look right" in a report, along with "You know what we want to say, just start writing the report now and finish the analysis when you have time."
Basically researchers are one step up from examining chicken entrails to tell the future, that step up being that some don't lie instead of everyone.
A good description of researchers. And yes, there are some honest ones out there but that honesty used to be the norm. I fear it’s more about chicken entrails now, than not.
Are you certain that all of the chicken entrailers were wrong?
If so , may I see your data and calculations?
There was a native lady in a news article a few months ago who described a method for predicting the weather. If you killed a deer and it was quite fat, you were in for a hard winter. I wasn't able to find out how that worked for chickens, however.
It seems to me that statistics is about probability, and probability is about how to bet under conditions of partial knowledge, when you don't actually know. Ultimately, there is a subjective element to it that allows you to change the numbers according to what conditions of partial knowledge you assume.
"It seems to me that statistics is about probability, and probability is about how to bet under conditions of partial knowledge, when you don't actually know."
On my first day of Stats, the professor stated at the beginning: "Statistics is the collection, manipulation, organization, and presentation of numerical data."
Even in the collection - for example, which data should be ignored - is susceptible to, generously, _interpretation_.
Climate 'data' is a good example. They come up with ways to justify ignoring the data that does not support their desired result -- the removal of the Arctic weather stations from the data set, for instance.
Or that most of the weather stations are in the Northern Hemisphere, in mostly non-arid regions that are largely populated so have high density development including pavement and structures that puts most thermomostats in the system in what I like to think of as a Heat Oasis.
There is some truth to this. However, statistics are mostly used in a manipulative manner to make things appear to be fact. Take the global warming stuff for instance. If you look at an overall graph of the planet’s history regarding CO2 levels and temps you see one picture. Mostly showing we need more CO2 for hapoy plant growth and that higher CO2 levels are often associated with COLDER temps vs warmer ones. Plus, the planet fluctuates from one extreme to another quite ofte, over time. I’m sure She understands that but so far, we clearly do not. If you cut out a part of that overall graph - one that shows only a small view and seems to support your effort to panic people - you get a picture that is correct but inflated and mis-used. It does not represent the full, numerical data set. Only a piece of it. And that is only one way to fool people with those numbers. Statistics also are used to show what happened in a study. 100 people are looked at. They are all given an orange and blue ball. They measure how often an orange ball is played with vs the blue ball. If 54 people choose the orange ball and 46 people choose the blue one, you could say people prefer orange over blue. What if the orange ball bounces better than the blue one? Or fits most people’s hands better. Those parameters aren’t mentioned in the study but could be relevant. And is 54% vs 46% really statistically significant? These kinds of games are played out all the time in an effort to fool a public who does not understand statistics, data sets, pieces of data vs an entire data set over time, relevancy, significance, etc. Some folks do. Most don’t.
I agree, and I think that's what I was getting at in the above comment. There is reality, which is objective, and there are our respective mental models, or understandings, of that reality. We "know" some facts about the reality, and from those facts we try to extrapolate the full model. Changing the set of facts we supposedly know changes the assumptions on how to bet, and what model we project to explain the reality. In a formal study, we get to choose our assumed fact set. So if we look at only a small portion of an historical graph, or consider only certain of a set of factors that might be relevant to which ball a set of people use to play with, we change the conditions of the bets we make to extrapolate that data into an explanatory model. The math itself might be fine, but it is that ability to cherry-pick the operative assumptions that makes it so easy for a study to lie with statistics.
Yes to this. It’s good to talk this out for clarification, though. Most people don’t realize just how badly their minds are being messed with. All some journalist or news article has to do is toss some numbers or bogus statistics out there and people nod their heads and say, “well, that must be so, then.” I’m afraid numbers are our last bastion of both truth and manipulation with the later now the norm.
Agreed. Talking it out helps to clarify our own thoughts, and then we become better at communicating with other people who haven't realized yet the nature of what is messing with them.
The statistics proving anything comment is correct. All you have to do is pick and choose which data set you use as well as any of a number of other variables. You can even make numbers up. There are so many statistical manipulations one can do with numbers it makes the head spin. I watched that sort of thing happen in the medical research department I worked in. I was to told to chill when I objected because if the numbers weren’t crunched “ just so” then no one would get any grant money. Their justification. I finally left that field and it’s only gotten worse as we can see, today.
Yea, I've been there too. I left a position doing research for a non-profit after I was told just to "Make the numbers look right" in a report, along with "You know what we want to say, just start writing the report now and finish the analysis when you have time."
Basically researchers are one step up from examining chicken entrails to tell the future, that step up being that some don't lie instead of everyone.
A good description of researchers. And yes, there are some honest ones out there but that honesty used to be the norm. I fear it’s more about chicken entrails now, than not.
Are you certain that all of the chicken entrailers were wrong?
If so , may I see your data and calculations?
There was a native lady in a news article a few months ago who described a method for predicting the weather. If you killed a deer and it was quite fat, you were in for a hard winter. I wasn't able to find out how that worked for chickens, however.
It seems to me that statistics is about probability, and probability is about how to bet under conditions of partial knowledge, when you don't actually know. Ultimately, there is a subjective element to it that allows you to change the numbers according to what conditions of partial knowledge you assume.
"It seems to me that statistics is about probability, and probability is about how to bet under conditions of partial knowledge, when you don't actually know."
On my first day of Stats, the professor stated at the beginning: "Statistics is the collection, manipulation, organization, and presentation of numerical data."
Even in the collection - for example, which data should be ignored - is susceptible to, generously, _interpretation_.
Climate 'data' is a good example. They come up with ways to justify ignoring the data that does not support their desired result -- the removal of the Arctic weather stations from the data set, for instance.
Or that most of the weather stations are in the Northern Hemisphere, in mostly non-arid regions that are largely populated so have high density development including pavement and structures that puts most thermomostats in the system in what I like to think of as a Heat Oasis.
There is some truth to this. However, statistics are mostly used in a manipulative manner to make things appear to be fact. Take the global warming stuff for instance. If you look at an overall graph of the planet’s history regarding CO2 levels and temps you see one picture. Mostly showing we need more CO2 for hapoy plant growth and that higher CO2 levels are often associated with COLDER temps vs warmer ones. Plus, the planet fluctuates from one extreme to another quite ofte, over time. I’m sure She understands that but so far, we clearly do not. If you cut out a part of that overall graph - one that shows only a small view and seems to support your effort to panic people - you get a picture that is correct but inflated and mis-used. It does not represent the full, numerical data set. Only a piece of it. And that is only one way to fool people with those numbers. Statistics also are used to show what happened in a study. 100 people are looked at. They are all given an orange and blue ball. They measure how often an orange ball is played with vs the blue ball. If 54 people choose the orange ball and 46 people choose the blue one, you could say people prefer orange over blue. What if the orange ball bounces better than the blue one? Or fits most people’s hands better. Those parameters aren’t mentioned in the study but could be relevant. And is 54% vs 46% really statistically significant? These kinds of games are played out all the time in an effort to fool a public who does not understand statistics, data sets, pieces of data vs an entire data set over time, relevancy, significance, etc. Some folks do. Most don’t.
I agree, and I think that's what I was getting at in the above comment. There is reality, which is objective, and there are our respective mental models, or understandings, of that reality. We "know" some facts about the reality, and from those facts we try to extrapolate the full model. Changing the set of facts we supposedly know changes the assumptions on how to bet, and what model we project to explain the reality. In a formal study, we get to choose our assumed fact set. So if we look at only a small portion of an historical graph, or consider only certain of a set of factors that might be relevant to which ball a set of people use to play with, we change the conditions of the bets we make to extrapolate that data into an explanatory model. The math itself might be fine, but it is that ability to cherry-pick the operative assumptions that makes it so easy for a study to lie with statistics.
Yes to this. It’s good to talk this out for clarification, though. Most people don’t realize just how badly their minds are being messed with. All some journalist or news article has to do is toss some numbers or bogus statistics out there and people nod their heads and say, “well, that must be so, then.” I’m afraid numbers are our last bastion of both truth and manipulation with the later now the norm.
Agreed. Talking it out helps to clarify our own thoughts, and then we become better at communicating with other people who haven't realized yet the nature of what is messing with them.
"When mathing lacks integrity"...