Thread: Ratings Thread
View Single Post
Old 06-26-2019, 05:01 PM   #727
Destor
I am the cheese
 
Destor's Avatar
 
Posts: 51,037
Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)Destor makes a lot of good posts (200,000+)
Lets say we did a survey on whether people like or do not like wrestling and the results are that 90% of children like wrestling however only 60% of adults like wrestling. So, can we claim that 75% of the population ( average = (90%+60%)/2 = 75% ) like wrestling?

This cannot be correct; we do not know anything about the sample size. Let's say there are 100,000 children and 400,000 adults surveyed. From those, 170,000 people do not like wrestling, while 330,000 like wrestling. Here, we can confirm that 90,000/100,000x100% = 90% of the children, and 240,000/400,000x100% = 60% of adults like wrestling.

So we are left with two answers. 66% (accurate calculation), and the 75% (inaccurate calculation). The averaging percentages can provide inaccurate results, and this is exactly what the tweet tries to do. Its awful math being masqueraded as a rebuttal when if you follow the math (and weight it against comparable data) it proves the opposite is true.
Destor is offline   Reply With Quote