So, here is the link to a paper which Keiths claims says something about income inequality, and I say is another example of the proliferation of shoddy science.
http://pss.sagepub.com/content/early/2015/03/19/0956797614567511.abstract
The highlight of the paper is this claim:
“We found that of the 40 search terms used more frequently in states with greater income inequality, more than 70% were classified as referring to status goods (e.g., designer brands, expensive jewelry, and luxury clothing). In contrast, 0% of the 40 search terms used more frequently in states with less income inequality were classified as referring to status goods.”
Where does one begin to critique the ridiculousness of this claim? 70% of the majority of searches are for luxury goods in some states, 0% of the most searched items in other states?
If one claims the difference in search patterns from one state to another is that dramatic, shouldn’t ones bs detector already be ringing alarm bells?
And what is considered a luxury good? What is the cut-off for equal states and unequal states? Did they decide the luxury terms before or after they viewed the data? Who do they claim is doing all this searching for luxury, the haves or the have nots?
The red flags are everywhere. Isn’t it likely that they had a conclusion that they wished to reach, and that they fulfilled their own prophecy?
phoodoo,
Have you read the paper? Can you identify a flaw in it?
Also, Walasek and Brown have made their data available on the internet. See the ‘Open Practices’ section of the paper.
I look forward to your demonstration that their conclusions are not supported by the data.
phoodoo:
That’s not what the paper claims. Did you even read the abstract? It says:
Have you read the paper (not the abstract), Phoodoo? Because I don’t think you have. (I haven’t, but then I’m not writing a post about a paper I haven’t read..)
Richardthughes,
Keith presented the paper. Has HE read it??
keiths,
Keiths,
Thank you very much for confirming exactly what I have written.
Now, have you read the paper?
Don’t dodge the question, Phoodoo. The only information you’ve disclosed is in the abstract. How pathetic would it be to ‘critique’ a paper on the abstract alone? I don’t think you have read it, and I think KeithS has. If you’d have read it your post would have been more insightful.
Fondle your watch and think on.
There is no such thing as income inequality. left wing commie jazz.
People get the income based on abilities or lack of it.
if there is a interference with what a income should be relative to the market then okay.
Yet that would be minor cases relative to hundreds of millions.
Why is this on a origins forum? Id evolution tied to left wing commie worldviews? Say it ain’t so? Could explain why it stucjk around and why its going along with commie to the ashheap of history as Reagan.
Richardthughes,
I don’t think keiths has read it.
But if he has, he can either submit a link to it, or he can explain HOW they determined these insane statistics.
Yes, phoodoo, I’ve read the paper. Have you?
As Rich says:
Byers:
That’s right, Robert. Everyone makes exactly the same amount of money.
phoodoo,
As Creodont suggested, you could always hock your watch to pay for the paper.
If you’re worried about the loss of status, then just strap the paper to your wrist, where the watch used to be.
Have you read it Phoodoo? As you’re critiquing it an all..
“But if he has, he can either submit a link to it,”
Imagine a guy who wears an expensive watch but wont pay to read a paper and tries to get others to find a free copy for him. How sad would that be?
blah blah blah, something off topic.
blah blah blah, you are wrong.
blah blah blah, there is a better way but you can’t see it.
blah blah blah.
And to answer your question
I’m not surprised you are at a loss. Where does one begin indeed? Answer that and you’ll understand why your previous “critiques” are not taken seriously.
Now, apply that thinking to ID….
Out of interest, have you ever heard of “projection”? It’s where, in essence, you project the flaws you have onto other people.
So, please please please, do more OP’s. As far as “projection” goes, this is a solid gold classic.
Tell me, did you decide the universe was designed before or after having “evidence” for that?
keiths,
You claim to have read it…so I claim the results are bogus.
How did they determine what was a luxury search item?
What was the dividing line between an “equal” state and an “unequal state?”
What ranking did they use to decide which states were equal and which weren’t?
Have you read it Phoodoo? Pathetic.
Claims based on nothing are worth precisely that.
Claims based on nothing are worth precisely that.
OMagain,
I agree, keiths claim that he read it is worthless until he proves otherwise.
Very good point Omagain!
I’m with OMagain: given phoodoo’s continued inability (despite repeated correction) to understand what the abstract actually says, rather than the caricature he wants it to be,
is deee-elicious.
I don’t hold out much hope that the situation would improve if he actually read the paper.
It’s the “Barry Arrington Quotemine incident” all over again. We gave Phoodoo authroship to talk about a paper he hasn’t read. ID-tastic.
Nice watch though.
Another phoodoo flameout.
Pathetic. Well done ID; you’re full of ’em.
Keiths,
You have the nerve to claim it is a flame-out from me??
You post an abstract to a paper which you claim says something about your topic. When you are challenged on that information, all you have done is run away and hide. You claim to have read the extended paper, and yet you refuse to comment on the paper.
This forum exists as a way for outsiders to see the point of view of those who claim to be scientific. I post here for one reason-to show just how poorly informed, biased, and uninterested in truth, those who claim to be “scientific skeptics” actually are.
I am perfectly happy for you to leave a record of dubious postings, and uncritical thought-for anyone who wishes to take an honest view of. I don’t mind if your obvious intellectual dishonesty is pleasing for the Omagains and DNA Jocks of the world. There are others will actually have an interest in truth, who will read what you write and see that the truth is, the science skeptics are full of hot air and no substance. For that cause your style is very enlightening.
DNA_Jock,
Has someone prevented you from saying what I stated that was wrong?
Or are you just another keiths, a guy who throws out bs science but is unable to defend it.
Do you believe the figures in this paper? Is it possible that the difference in search patterns from state to state are so dramatic and so tied to the income inequality of the state?
Do you want to buy some land in Florida. Or maybe a watch from Keiths? Just how gullible are you science skeptics?
What the authors did (explanation without commentary):
They determined the income inequality level for each state using the Gini coefficient, which is a commonly used measure. It provides a metric for how distributed or concentrated income is in a population (e.g., if 20% of people make 80% of the income, the Gini value is high).
The goal was to see if interest in “status” goods is affected by inequality. But since interest in status goods may also be affected by other factors as well, the authors did a regression analysis to control for overall income levels, population, urban population, and numbers of domestic vs. foreign born residents, in order to isolate the effects of inequality from these other factors.
They thus derived a metric for each state that represents income inequality adjusted to account for those confounding factors.
To answer one of Phoodoo’s questions, there was no cut-off for equal or unequal states, rather each state was assigned a value that could be higher or lower.
To answer another of the questions, the authors did not decide what was a luxury good, nor did they determine what Google search terms to look for.
They used Google Correlate. Google Correlate allows you to upload a set of data by state with which you can then correlate search terms. Google Correlate doesn’t tell you the most frequently searched for terms by state; that is a misinterpretation of the results. Rather, it tells you the search terms that are most highly correlated with the data you uploaded.
For example, if you had data for how many members of the Star Trek fan club resided in each state, and uploaded that data to Google Correlate, it will tell you which search terms are most highly correlated with that data – that is, those search terms for which the distribution by state of the frequency with which people search for them most closely matches the distribution by state of your original data. So, in our Star Trek fan club example, we might find that states with higher concentrations of fan club members than other states will have a correspondingly relatively higher frequency with which people search for terms like “Kobayashi Maru”, “stardate”, etc. It doesn’t mean that “stardate” is one of the most searched for terms in a state, just that the frequency of search on that term is most highly correlated by state with the original data, in this case the number of Star Trek fans. From this we might surmise that Star Trek fans have a higher than average interest in Star Trek terminology. (This may seem obvious, but the point of the method is to discover correlations that may not be obvious.)
Google Correlate will also give you the search terms that are most negatively correlated with the original data. For instance, the term “romance” might be the top term negatively correlated to Star Trek fan club membership (that’s a joke, by the way, for the sake of the hypothetical example; sorry Star Trek fans!). We might surmise from this that Trekkies have little interest in information about romance.
Anyway, back to the study. The authors submitted the control-adjusted values of inequality by state to Google Correlate, which then output the top terms that were most positively correlated to that data and the top terms that were the most negatively correlated. The authors didn’t know in advance what those terms would be.
The next step was to determine if those terms were indicative of interest in status goods. The authors didn’t make that determination. They got 60 volunteers to participate in the study. They gave them the following definition:
“Some things that people are interested in, or like to buy or find information about, are things that show how rich or successful they are compared to other people. These are sometimes called “positional goods” or “status goods”. Someone who buys such goods may be particularly concerned to demonstrate their social status”
They then gave them the lists of the search terms positively and negatively correlated with income inequality (without telling them anything about the study, the distinction, or the how the terms were generated) and asked them to determine for each term whether it indicated an interest in status goods based on the above definition. They could respond yes, no, or not sure.
The result was that of the 40 terms most positively correlated with inequality distribution by state, 29 (72.5)% were viewed as referring to status goods by the majority of the test subjects. Of the 40 terms most negatively correlated with inequality distribution by state, none of them (0%) were viewed as referring to status goods by the majority of the test subjects.
This doesn’t mean that status goods were the most searched for terms in high inequality states, nor does it mean that status goods were never searched for in low inequality states. It means that most of the search terms most closely correlated with inequality make reference to status goods (e.g., people in relatively higher inequality states seem to be proportionately more interested in “David Yurman earrings”), while none of the search terms least closely correlated with inequality make reference to status goods (e.g. people in relatively lower inequality states seem to be proportionately more interested in “pirate talk”).
phoodoo,
And you fail miserably. However, you do succeed in showing just how poorly informed, biased, and uninterested in truth you are.
And you amuse us, which is worth something.
Still waiting for answers to the questions I posed in the very first comment:
Hobbes,
Thanks for your summary. I’m afraid it will be wasted on phoodoo, however.
You are thanking him for showing that you have not read the paper , and that you are incapable of commenting on it? For exposing that you are a fraud?
What were the terms keiths?
phoodoo,
Hobbes has read the paper, therefore I can’t have? What is this, the Law of Conservation of Readers?
Keep it up, phoodoo. You’re doing great.
Hobbes,
I am glad you put in a little more effort than Keiths (none) but lets look at bit more closely at what is being said here. First:
“The goal was to see if interest in “status” goods is affected by inequality. But since interest in status goods may also be affected by other factors as well, the authors did a regression analysis to control for overall income levels, population, urban population, and numbers of domestic vs. foreign born residents, in order to isolate the effects of inequality from these other factors.
They thus derived a metric for each state that represents income inequality adjusted to account for those confounding factors.”
I see one problem here already. If they already had the Gini index to work off of, then skewing the results of this by separating out foreign born, from non-foreign born, high population, rural areas, etc..you are now creating a separate index of people that doesn’t necessarily reflect their income equality as much as it does their lifestyle choices.
Problem number one, but there are more that I will continue with later…
Anyone care to write the last sentence of the paper here?
Rich:
Let’s let phoodoo do it. Maybe in the process of typing it in, he’ll actually ponder what it says.
Hobbes,
Secondly, You say that, “The result was that of the 40 terms most positively correlated with inequality distribution by state, 29 (72.5)% were viewed as referring to status goods by the majority of the test subjects.”
Yet the abstract says this:
“We found that of the 40 search terms used more frequently in states with greater income inequality, more than 70% were classified as referring to status goods
So I am not clear, are you saying this is a state by state comparison, or a comparison of inequality within the state?
Richardthughes,
Well, your boy keiths would like to be able to, but since he hasn’t read it…
I’m afraid you have it exactly backwards here. The purpose of the regression to control for confounding variables is to eliminate the effects of what you refer to as “lifestyle choices”, and leave a measure of inequality that is independent of those effects. For example, a population that has income distributed in a particular way between a high of 100k per year may have the same Gini index as a population whose incomes are distributed in a similar way, but between a high of 50k. But the former population is likely to have much different lifestyles and attitudes than the latter population because of the difference in absolute income levels. The regression is eliminate that part of the results that is due to the “lifestyle” variable (in this case absolute income level) and leave only that part of the result that is due to the variable of interest, inequality.
Hobbes,
Its looking to me more and more that what we really are saying is that in those populations where the population is relatively well off, they look at luxury items more.
Isn’t that was is really being explained here?
phoodoo,
You crack me up. If I haven’t read the paper, how did I know about the ‘Open Practices’ section?
And if you’ve read it, why do you keep quoting the abstract and nothing but the abstract?
As Rich says:
You’re terrible at bluffing. Keep up the good work.
Now go hock that watch and buy the paper.
Hobbes,
Of course the claim is that is what is being done. The question is what else are you changing by weeding out variables like where someone was born and if they live in a rural versus city environment.
This is their formula, and yet they are testing to prove their hypothesis. Whose to say that their formula is the correct one?
Hobbes,
So again, are we looking at state by state comparisons of the inhabitants search habits, or are we looking at the search habits within the state given variables of where the people live, where they were born, if they prefer city living compared to rural life….
phoodoo, to Hobbes:
No, phoodoo, and this part is even in the abstract, which you supposedly have read:
It was not a comparison of inequality within a state. Every state was assigned an inequality value. For simplicity, imagine there are only three states, and those three states had inequality values of 1, 2, and 3, respectively. That data was uploaded to Google Correlate which then determined what search terms were distributed by frequency in a similar manner across the states. So, for example, if the term Gucci was searched for 100 times in the first state, 200 times in the second, and 300 times in the third, then the distribution (1x, 2x, 3x) is exactly the same as the inequality values, hence perfect correlation and this term would be flagged. It the term “backache” was searched for 500 times in each of the three states, then there would be no correlation between this term and the inequality distribution across the three states, so this term wouldn’t be flagged (even though it was searched for more times that Gucci). If the term “picnic” was searched for 300 times in the first state, 200 times in the second state, and 100 times in the third state, this would be a perfect negative correlation and the term would be flagged because it’s distributed across states in the exact opposite way as the inequality.
Hobbes,
Can you first show us the inequality rankings they gave each state? Or maybe keiths can if he read the paper?
If we have an environment where most of the residents are wealthy, is that state a high equality state? And does that state demonstrate a low propensity to search for items of luxury status?
I am pretty sure the answer is not that the places where all residents have high income do little searching for luxury items.
No. The purpose of controlling for confounding variables is to do exactly the opposite of what you say. As in my previous example two populations with the same number of people, one wealthy and one poor, can have similar Gini indices if income is distributed similarly within the two populations. Clearly, as you suggest, the wealthier population is more likely to search more frequently for luxury items, so just using the Gini index would give you a potentially false conclusion that searching for luxury items is not related to inequality since we have a difference in frequency of searches between the populations, but similar Gini indices. The point of controlling for income is to eliminate that part of the search difference that is due purely to income level differences and to leave only that part of the search difference that is due to the difference in income distribution.
The values are the “inequality residuals” after the regression to control for confounding variables. The larger negative values mean lower inequality. The larger positive values mean higher inequality. So the list goes from lower to higher inequality.
State Residual_GINI
Utah -0.02769
California -0.02413
Alaska -0.02396
Idaho -0.02376
Wyoming -0.02121
Hawaii -0.0204
Iowa -0.0188
Wisconsin -0.01693
Nevada -0.01616
Indiana -0.0146
Nebraska -0.01277
Washington -0.01073
New Hampshire -0.01015
Vermont -0.0087
Montana -0.00778
Texas -0.0076
Maine -0.00758
Delaware -0.00565
Minnesota -0.00551
Florida -0.00443
South Dakota -0.00372
Kansas -0.00357
Ohio -0.00335
Arizona -0.00298
Michigan -0.00289
Oregon -0.00249
Maryland -0.00137
Missouri 0.0026
Pennsylvania 0.00328
North Dakota 0.00543
Illinois 0.00594
Oklahoma 0.00597
North Carolina 0.00648
Arkansas 0.0065
Colorado 0.00711
West Virginia 0.00766
South Carolina 0.01014
Virginia 0.01107
New Mexico 0.01111
New Jersey 0.0114
Georgia 0.012
Kentucky 0.01286
Rhode Island 0.01408
Tennessee 0.01539
Alabama 0.01693
Mississippi 0.02257
Massachusetts 0.02431
New York 0.02584
Louisiana 0.02771
Connecticut 0.04252
If everyone in a state makes the same income, 1M per year, then they are all wealthy but the state also has zero inequality. If half the people in a state make 500k, then there is high inequality. If half the people in the state make 2.5M, then there is also high inequality. This is why researchers control for confounding variables: states can have vastly different levels of wealth, yet similar levels of inequality, either high or low. Those different levels of absolute wealth can greatly affect the results and drown out any effect caused by inequality. This is what researchers painstaking design their methods to avoid.
Yes, this is their formula. The goal of research is confident understanding. Toward that end they constructed a model that attempts to determine whether income inequality is correlated to interest in status goods. They made choices in their analysis. They chose the Gini index, which is not the only measure of inequality that people use. They chose to control for four other confounding variables using a particular statistical method. They could have used other methods and they could have included other possible confounding variables. Their choices were no doubt based on their experience, expertise, and reasoning as to which confounding variables were likely to have the most impact on the results.
Who is to say their formula is the correct one? Actually who is to say it is “a good one” – there is no “correct one” since all models are approximations that are either better or worse? Anyone who wants to challenge, verify, or improve upon their work is free to do so. They published their method and explained it. They made their data available for others to see, download, and manipulate. If someone can show any error in their methods or calculations, the same journal would no doubt be happy to publish that analysis. If someone can add to, modify, extend, or completely supplant their analysis with a better one, and explain why it is better, in order to improve upon their results, they are free to do so and the same journal would no doubt be happy to publish that as well. That is the way science works.