This article makes a statistical error. It implies that an "aggregate" measure necessarily has greater predictive power than each of its component measures. For example, the author says:
"Stated more formally, the predictive power of a perfect aggregate of many small direct signs will be of necessity greater than any of the aggregated signs considered individually, even if the aggregate sign is indirect."
The author concludes that one's astrological sign would be a better predictor of health outcomes than one's geography, the climate in one's area, the season, etc.
To see why this is a mistake consider an analogy:
A group of high school students takes the SAT. Answers to each individual question of the SAT are correlated with being admitted to a college. That is, if a student gets Question 1 correct, they are more likely to be admitted, and likewise with Question 2 and Question 3 and so on. Does it follow that the overall SAT score is a better predictor of getting into the college than every individual question is?
The correct answer is no!
It’s entirely possible that one specific question, say Question 3, is *perfectly* correlated with being admitted, while at the same time, the overall SAT score might only modestly predict being admitted. The reason is because the overall SAT score would also be composed of answers to other questions that are imperfectly correlated with college admission, and these other questions could introduce an arbitrarily large amount of noise into the prediction.
So even if astrological signs are 1) correlated with variables like one's geography, the climate in one's area, the season, etc. and 2) these other variables are correlated with health outcomes, it doesn't follow that astrological signs are better predictors of health outcomes than any of these variables taken individually.
And indeed, one recent study examined the predictive power of astrological signs on a host of life outcomes (e.g. life satisfaction, suicidality, BMI, household income, number of deep emotional connections). The authors conclude that astrological signs "had no predictive accuracy whatsoever."
So in sum, the problem with astrology is not merely that there is no causal mechanism by which the positions of stars would affect our destinies. It's that there is no correlation to begin with!
This article makes a statistical error. It implies that an "aggregate" measure necessarily has greater predictive power than each of its component measures. For example, the author says:
"Stated more formally, the predictive power of a perfect aggregate of many small direct signs will be of necessity greater than any of the aggregated signs considered individually, even if the aggregate sign is indirect."
The author concludes that one's astrological sign would be a better predictor of health outcomes than one's geography, the climate in one's area, the season, etc.
To see why this is a mistake consider an analogy:
A group of high school students takes the SAT. Answers to each individual question of the SAT are correlated with being admitted to a college. That is, if a student gets Question 1 correct, they are more likely to be admitted, and likewise with Question 2 and Question 3 and so on. Does it follow that the overall SAT score is a better predictor of getting into the college than every individual question is?
The correct answer is no!
It’s entirely possible that one specific question, say Question 3, is *perfectly* correlated with being admitted, while at the same time, the overall SAT score might only modestly predict being admitted. The reason is because the overall SAT score would also be composed of answers to other questions that are imperfectly correlated with college admission, and these other questions could introduce an arbitrarily large amount of noise into the prediction.
So even if astrological signs are 1) correlated with variables like one's geography, the climate in one's area, the season, etc. and 2) these other variables are correlated with health outcomes, it doesn't follow that astrological signs are better predictors of health outcomes than any of these variables taken individually.
And indeed, one recent study examined the predictive power of astrological signs on a host of life outcomes (e.g. life satisfaction, suicidality, BMI, household income, number of deep emotional connections). The authors conclude that astrological signs "had no predictive accuracy whatsoever."
https://www.clearerthinking.org/post/how-accurate-are-popular-personality-test-frameworks-at-predicting-life-outcomes-a-detailed-investi
So in sum, the problem with astrology is not merely that there is no causal mechanism by which the positions of stars would affect our destinies. It's that there is no correlation to begin with!