If there’s one field of mathematics that social scientists have grabbed onto, it is statistics. Many sociologists, for instance, love the use of statistics so much so that the top sociology journals in the United States can’t help but publish any highfalutin nonsense based on regression analysis that is simply wrong.
Let’s examine one form of the incorrect regression analysis based on race. Social scientists would use race as a variable in numerous published works. They would say the following in the form:
Y(Income) = a. f(race) + b. f(class) + c. f(educational status) + ….
That is, income, they may say, is a function of race, class and educational status. What is wrong with this thinking is quickly obvious to a mathematician. Not to an American sociologist. Take any unit of analysis of the form above. Since social scientists deal with societies, or group effects, the unit of analysis here is an individual human being. Let’s glean some rudimentary information from this expression as follows: (1) When the class of an individual human being changes, the income level will change. That is if an individual human being can change his class (from poor to elite, for instance) the income level of that individual will invariably change. Imagine a poor man winning a mega million lottery and now hangs around New York’s elite. (2) If an individual can change his educational status, the income level of that individual will invariably also change. A man can go back to college and obtain post-graduate degree in order to boost his level of income. (3) The third issue, however, is what is worrying.
Now, how does a Black man change his race so that he can attain a higher level of income in the United States? By bleaching? Or by undergoing profound surgery? In what way can a white woman change her race in order to earn income as low as a Black woman? How can a Black child change her race in order that she can one day sit at home, like a white woman, and collect ninety bucks an hour of consulting fees from KPMG? By undergoing a kind of post-racial surgical operation? Race does not vary for an individual. Another way to put this is that race is not a variable. Independent or otherwise.
But in Europe, Canada, Australia and the likes of the United States of America, race has always been a variable. One way they like to satisfy themselves with this falsehood is to tell themselves the unflattering lie that race is actually an “independent” variable. By independent what they mean is that the variable is only dependent on itself (on its own choice). How is that? Does one determine himself to be Black the first thing in the morning, and then change it to white after lunch in preparation for that impending interview at KPMG? Absolutely not! But the lie pervades all aspects of social science work, and the lie has crept into genomics, biostatistics and such dubious fields of pseudoscience. The basic form of that expression above, no matter its renewed embellishments in journals, can be read almost anywhere one looks in social science.
Where social scientists derived this nonsense from is actually hidden right under their noses: the field of Statistics itself in the works of Darwinists. Evolutionary theorists adore the idea that an individual specie can actually metamorphose into another specie given just enough time. This metamorphosis has never been actually demonstrated in the field or in the lab. Not yet, at least. However that statistics describing how the form of a species depends on time remains insidious in all aspects of pseudoscience and social science:
Y(species form) = a. f(time) + ….
Almost readily, one can see why it is tempting to rewrite this in the following form (after all in many aspects of biological-speak, race and species are almost interchangeable):
Y(race) = a. f(time) + …..
That is, if the form of a species is dependent on time, then the race of an individual could also be dependent on time. Put another way given enough time a Black man for instance can become a white man and apply for that job at KPMG. Racist theorists in American and European sociology, for instance, still publish works of this form. For instance, the idea that Jews weren’t (regarded as) white, or that the Irish or the Italians weren’t (regarded as) white until much later in American society is used as evidence of trans-racial theorization. The idea then goes that given enough time Africans (people of African descent in the Americas) too will be considered as white in American society. Don’t laugh! This sort of white supremacist incantation remains insidious and pervasive in American sociological thought because at its first assumption is that it regards whiteness as a thing for a person to aspire to become. Race then must be aspirational, not inherited.
Of course, a white supremacist incantation such theories remain, but once the species, and then the race, had been used as variables (dependent on time) in white supremacist evolutionary theories, the idea of using it as an independent variable in other analyses (not involving time) became a straightforward adaptation. Never mind that it is wrong mathematics.
There’s motivation for continuing to use race as a variable in social science research. Sheer willful ignorance is part of it for many social scientists who simply do not have enough nerve cells in the brain to understand this article, for instance. The bigger reason exceeds all: Racism. White supremacism. This has no regard for logic. This has no regard for rationality. This has no regard for mathematics.
The great Narmer ? ????
What an article Narmer! What sheer simplicity of the mind. This is elegance. I hope that every social scientists will grow some balls to read this… but I fear they will shy away and bury their heads in the sand. People do what pays. White supremacism pays. Irrationality pays. Wrong mathematics pays, especially in academia.
You probably need to read about dummy variables. https://dss.princeton.edu/…/analysis/dummy_variables.htm
The idea of dummy variables applies mathematically to such a category as political affiliation, and such things as class, for instance, but not to race. The race of a person will never change. Never! The political affiliation of a person can change! People can change certain things about themselves, but not their race! The class of a person can change. Race is not a variable and insisting that it is is beyond the scope of statistical analysis. The Princeton University article you mentioned is foolish to have mentioned race as an example of a dummy variable. This is the kind of drivel they teach in the Ivy League, for instance. In fact, the Ivy League is the correct source of most of the confusion around the misuse (incorrect use) of race as a variable. The pseudosciences of Craniology and Phrenology all begun and were advanced at such schools as the University of Pennsylvania, just 45 minutes from Princeton University (both Ivy League Schools). Most researchers in these schools, and I have met plenty, are below average! Don’t read their incorrect use of mathematics. You will become brainless!
Race is a categorical value. It’s not dynamic, therefore it’s represented by a dummy variable. If you read any regression analysis and the author modeled race as a dynamic variable, you would have every reason to pillory that study..
Yaw, you are not reading what I am writing. Please let us debate clearly what I am saying without resorting to fancy statistical terms. Whether the thing is called “Categorical value” or “Categorical variable” or “dummy this” or “dummy that” is not the point. The point is that most of all categorizations in which humans belong, and as such most of the ones that lead social scientists to model them, are variables. Race is not one of them. I have explained why!
In the Princeton article, the author used Political Affiliation, Year In College, as examples in the “categorical or dummy” modeling of his choice. The author should have left race alone. But he doesn’t know the difference between “race” and “Year In School” because he is of average intelligence. Any human being can change their year in school by graduating to the next classroom. No human being can change their race through any such graduation! Do you get it? Race is not a variable. It never changes! Not even once, massa!
Yaw, let me explain even further so that next time you see this somewhere in some Ivy League classroom, you can stand up to the Professor and say that Narmer taught you the correct use of statistical modeling in regression analysis.
The Princeton author himself said: “Results only have a valid interpretation if it makes sense to assume that having a value of 2 on some variable is does indeed mean having twice as much of something as a 1, and having a 50 means 50 times as much as 1.” Now I ask you, in what instance where race is a variable that this can be said to be true? Zero. Is a result of 2 by some variable called “Asian” make that individual now twice as more “Asian” than he was before. Or does it? Now you see the nonsense of using race as a variable?
But, for do the exercise for Year In School, or for Political Affiliation, and the results might be interpreted, in some cases, correctly. The tendency to use these as variables can be spared. But not for race. The race of a person never changes. It does not vary (dynamically or not). Race is not a variable! Full stop, massa.
It boils down to understanding what a dummy is. It’s not a fancy terminology. It’s very basic.
My friend, they can embellish their stupidity in English. What they can’t do is proper mathematics. Real mathematicians admit it when they are wrong. They don’t double down. The Ivy League social scientists are largely correct when they work with dummy variables as Year In School and Political Affiliation. They are wrong when they use race in the same way that their Ancestors were foolish enough to use it in the fanciful statistical justifications of the Craniology and Phrenology era. Again, they can use their English to embellish simple ideas, but they can’t run away from the mathematics in English. Their interpretation of race as a variable in regression analysis is patently stupid, hell, even still racist. But ignorance is bliss, especially in the Ivy League.
Yaw Amponsah, a “Variable” is something that can change. “Race” does not change, how then can one even begin to make assumptions about, let alone model, something that does not change?
There’s a specific definition for a dummy variable which is a statistical terminology. It’s not the typical English definition of variable you are familiar with.
Yaw, please enlighten me. What is this specific definition for a dummy variable (that does not include an attribute of change)? Please tell me.
In regression analysis, a dummy variable (aka, an indicator variable) is a numeric variable that represents categorical data, such as gender, race, etc. Therefore, by using dummy to represent race, data can be categorized. It doesn’t mean that race changes over time.
Thanks Yaw, so in essence what you are saying is that the Indicator (which might be Year In School) on a given unit of study (student, for instance) might switch from “Freshman” to “Sophomore” to “Junior” to “Senior” depending on the “Level of depression.” (This is one example in the Princeton article). Another generic example is “0” and “1”. An indicator such as this on a given circuit can switch from “0” to “1” depending on whether there’s current or not.
Now, how does an indicator (a dummy) variable switch from “Black” to “White” given a certain student? I mean under what circumstances, say, might one student’s indicator variable switch from “Black” to “white?”
Dade, it will not switch. For instance, all individuals who identify themselves as blacks can take on specific dummy code. For example 0. Then all the individuals who identify themselves as whites will take code 1. At no point will the code change for the two races. Thus every black will take on 0 and every white will take on 1. I’m going to run.
The last time I commented on your wall or Narmer’s wall (I’ve forgotten who specifically), I couldn’t disappear. I remember having to come back and engage with you guys for hours on end. I was catching flight from Seattle to DC, and for more than 6 hours, I had to answer a question or respond to a question. Lol.
If every Black is “1” and every white is “0” at all times, then the “0” and the “1” are no longer (dummy) variables. This is perhaps where your misunderstanding lies. Not all categories can be dummied. If a student’s category for another instance cannot change from “Freshman” to “Sophomore” there will be no such thing then as a (dummy) variable called “Year In School.” Models are no devised from things we already know. Models are for prediction. If we already know what we know, then it is useless to write a model to predict it. “Blacks” remain blacks, and “whites” remain whites, all we need do is look, not regress race.