Study Shows Low Average of Books Read, Man Debates Its Validity

[Artus]

I was in a discussion with a man. I've read a study about reading habits. The study shows a low average of read books for some countries and told to a man "most of people doesn't read books" because the average was 6 books a year for a sample taken from a population of 100 millions of people.

The man told me, you don't know everyone of the 100 millions, so you can't tell most people don't read. I told him I don't need to know every person, just read the study and see the low average of read books to infer the people needs to read more books. He says the study is incomplete because they don't made the survey on the 100 millions. Is there a kind of fallacy behind his statement?

I think the survey is accurate in statistic terms, the study says, "in a country of 100 millions of persons, there is an average of 6 read books a year", I think that's low for this country so I say "most of people don't read books". He for the other side says "you need to know every person to make that statement".

Who is right?

 

[whybother]

It's not a meaningful study unless you give the sample size for the population. For a population that large, 100 million people, you need to have a very large sample size to say it's statistically a good representation of the people. If they only interviewed 1000 people to generate that number, of 6 books per year, then he is right, you really can't make statements about all 100 million people. Without the sample size, it's a meaningless statement.

 

[CRGreathouse]

It's not a meaningful study unless you give the sample size for the population. For a population that large, 100 million people, you need to have a very large sample size to say it's statistically a good representation of the people. If they only interviewed 1000 people to generate that number, of 6 books per year, then he is right, you really can't make statements about all 100 million people. Without the sample size, it's a meaningless statement.

No.

For a given confidence level and tolerance (confidence interval), there exists a sample size which is suitable for a population of any size. For +/- 3%, to a 95% confidence, that size is just over a thousand. (That's why statistical surveys of large groups like the US use n = 1067 or so.)

 

[Artus]

It's not a meaningful study unless you give the sample size for the population. For a population that large, 100 million people, you need to have a very large sample size to say it's statistically a good representation of the people. If they only interviewed 1000 people to generate that number, of 6 books per year, then he is right, you really can't make statements about all 100 million people. Without the sample size, it's a meaningless statement.

No.

For a given confidence level and tolerance (confidence interval), there exists a sample size which is suitable for a population of any size. For +/- 3%, to a 95% confidence, that size is just over a thousand. (That's why statistical surveys of large groups like the US use n = 1067 or so.)

It was a large sample (about 15 millions) even if it isn't necessary. My argument is correct based on the size of the sample, although I was doubtful about it. My proposition is "most of people -of this country- don't read books" & I think is correct because of the low average they have.

Thank you guys for your help.

 

[CRGreathouse]

It was a large sample (about 15 millions) even if it isn't necessary. My argument is correct based on the size of the sample, although I was doubtful about it. My proposition is "most of people -of this country- don't read books" & I think is correct because of the low average they have.

Your argument is correct, your friend's has no merit. Five thousand would be plenty for a study like that; 15 million is just overkill (but still works).

 

[rbeale98]

i think you are wrong. first, 6 books a year is not low enough to say a person doesn't read. To say someone doesn't read is to say they read 0 books per year. you need to know the percentage of people that read 0 books per year to say "most people don't read books." the average doesn't tell you anything about how many people read 0 books. if every person in the country reads 6 books per year, would you say that most people don't read books?