So Eric, when you read an article that says, you know, more than half of all Americans

say they regularly experience ex-emotion or only 12% of Americans feel such and such,

what does that experience like for you as a sociologist?

Right, so about half the time I think, wow, that's pretty interesting and about half the

time I'm pulling out my hair thinking, no, don't say that!

Eric Kleinemberg is a professor of sociology at New York University.

Unfortunately, what I find is that journalistic reporting will use survey data when it's

useful for the story, and they kind of don't care that much about whether the data underlying

it is reliable.

And what's wrong with survey data?

A lot of survey data is based on a sample that's not really worth generalizing from a lot

of surveys, ask questions that will work for a particular time and place, but might

not work very well after that, which means you can get a snapshot of a moment in time,

but not really a dynamic portrait of something over time.

Would you like an example of how survey data gets used in the media?

Okay, here's an example.

A top doctor calls it a national health crisis, not obesity or heart disease.

A condition that is so common, you actually may not think of it as a mental health problem.

Loneliness.

That's right, loneliness.

People who struggle with loneliness end up living a shorter lives, and they also are

to increase risk for heart disease, depression, dementia, anxiety, and a host of other conditions.

...