The facebook emotion study – dubious graphs

So if you hadn’t heard, some researchers at Facebook and US universities ran a study on around 689,000 Facebook users. The aim of the study was to see if, by changing how often happy/positive words and sad/negative words appeared, they could change how happy or sad users felt – measured by how often they used positive or negative words in their posts.

The s#!+ has hit the fan over this because the closest anyone came (and it could have been you or me) to opting into this study was clicking ‘I agree’ to Facebook’s user agreement. Most people see that as falling well below the standard of ‘informed consent’ required for psychological studies on humans. Even the journal that published the paper has now backed away, issuing an ‘Expression of Concern‘ (which is journal-editor-speak for serious unhappiness). Just to say, legally, the authors seem to be ok. The study was approved by their university’s internal research board.

My beef, though, is with their graphs. If you look at their paper, they have this graph, where to the casual reader, it looks like there’s a huge effect.

Figure reproduced from Kramer, et al, PNAS 1111, 8788 (2014)

But read the axes… that’s right, they suppressed zero. Here’s a quick redo of their graph in a more transparent way:

facbookExpBarGraphHere, I’ve made the axes start at 0, and both the ‘positive’ and ‘negative’ axes have the same scales.

So don’t worry that facebook can manipulate your emotions.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s