Allegedly Team which I presume is black runs IQ test in Nigeria and presenter claims they came up with similar results to previous test giving studies
Screenshot from video at link
If the test would have a finite countable and whole number of questions equal to N and you can only get 1 or 0 points for each question towards the raw test score this would result in N + 1 combinations for raw test scores. If N + 2 or more people took the test there would be a 100% chance that at least two people got exactly the same test score.
These raw test score could be converted to an IQ score based on whatever bad statistical assumptions are assumed like normal distribution
If you have a continous variable with a unlimitted number of possible values the probability of getting exactly the same quantity with measurements of 2 different things by random chance should approach zero
Usually for continous variables measurements are put in categories within a certain range
for example instead of counting the number of outcomes that are exactly 60 inches one could count the number of outcomes between 59.5 & 60.5 inches
But if it is a discrete variable it is not 0
If someone is saying there's some population somewhere, they did an IQ test, and 26% of the population matches ANY number, there's something very wrong with their math.
Maybe they meant <= 69, but they didn't say it, and they did use <= in other places so it looks shoddy...
Thdy might have nust gave them a 10 question IQ test or some othdr small number of questions instead of giving them hundreds of questions that take an hour or more to answer
I would doubt that most Nigerians would have the patience for a long test
It also do not expect competant design or statistical analysis from most social scientists
I even less suspect competant design or statistical analysis from most Nigerian social scientists
Because this is the default assumption, serious researchers always describe their process in great detail so that it can be independently validated and gut-checked against their results. So if you see what looks like a mistake, and they don't explain why it's like that, that's red flags all around.
Where did you get the raw data from to plot the graph if it is called a graph in this context
Maybe a histogram?
The point was mainly that the original figure was kind of garbage because it doesn't intuitively show the data in a sane way since they showed the IQ value and population percentage as two different colored bar graphs in the same figure. Normally you want it to look like the figure I generated because then you can actually see the actual distribution (which shows the relationship between the random variable measured, here IQ, and the probability/percentage of different values).
