Note: This is part 3 in a series for high school students about reading and interpreting science on the internet. Read the intro and get the index here, or go back to part 2 here.

As any good internet person knows, you can’t have a good story without a featured image…pictures being worth a thousand words and all that jazz. Nine times out of ten if you see a story pop up on Facebook, it’s going to have an image attached.  Those deceptive and possibly false memes I was talking about in Part 1? All basically words on images. For most stories, there are really two different types of images: graphs or technical images and what I’m going to call “narrative” images.  In this section I’m going to cover the narrative images or what I call:

Narrative Images: Framing the Story

Okay, so what’s the problem here?

In Part 2 I mentioned a study that was mostly about headlines, but that had a really interesting point about pictures as well. In that study, they purposefully mismatched the headline and the picture in a story about a crime. Basically they had the headline read something like “Father of two murdered” and showed a picture of the criminal, or they had it read “Local man murders father of two” and showed a picture of the victim. Later, people who had read a “victim” headline with a picture of a murderer actually felt more sympathy towards the murderer, and those who read a “criminal” headline with a picture of a victim liked the victim less. That’s the power of pictures. We know this, which is why newspapers can end up getting sued for putting the wrong picture on a cover even if they don’t mention any names.

Any picture used to frame a story will potentially end up influencing how we remember that story. A few weeks ago, there was a big kerfluffle over some student protests at Oberlin. The NY Post decided to run the story with a picture of Lena Dunham, an alum of the college who is in no way connected to the protests. In a couple of months, my bet is a good number of people who read that story will end up remembering that she had something to do with all this.

What should we be looking out for?

When you read an article, at the very least you should ask how the picture matches the story. Most of the time this will be innocuous, but don’t forget for a minute the picture is part of an attempt to frame whatever the person is saying.  This can be deviously subtle too.  One of the worst examples I ever heard of was pointed out by Scott Alexander after a story about drug testing welfare recipients hit the news.  The story came with lots of headline/picture combos like this one from Jezebel:

Now check that out! Only .2% of welfare applicants failed a drug screening! That’s awesome.  But what Scott Alexander pointed out in that link up there is that urine drug testing actually has a false positive rate higher than .2%.  This means if you tested a thousand people that you knew weren’t on drugs, you’d get more than a .2% failure rate.  So what happened here? How’d it get so low?

The answer lies in that technically-not-inaccurate word “screening”.  Once you saw that picture, your brain probably filled in immediately what “screening” meant, and it conjured up a picture of a bunch of people taking a urine drug test. The problem is, that’s not what happened. The actual drug screening used here was a written drug screening. That’s what those people failed, and that’s why we didn’t get a whole bunch of false positives.  Now I have no idea if the author did this on purpose or not, but it certainly leaves people with a much different impression than the reality.

Why do we fall for this stuff?

Every time we see a picture, we’re processing information with a slightly different part of our brain. In the best case scenario, this enhances our understanding of information and engages us more fully. In the worse case scenario, it skews our memory of written information, like in the murderer/victim study I mentioned above. This actually works in both directions….asking people questions with bad verbal information can skew their visual memory of events.  Even people who are quite adept at interpreting numbers, words and data can forget to consider the picture attached.

So what can we do about it?

Awareness is key. Any time you see a picture, you have to think “what is this trying to get me to think?”  You have to remember that we are visual creatures, and if text worked better than visuals commercials wouldn’t look the way they do.

Now, before I go, I have to say a few words about infographics. These terrible things are an attempt to take entire topic and make them in to nothing but a narrative photo. They suck. I hate them. Everything I say in this entire series can be assumed to go TRIPLE for infographics. Every time you see an infographic, you should remember that 95% of them are inaccurate. I just made that up, but if you keep that rule in mind you’ll never be caught off guard. Think Brilliant has the problem summed up very nicely with this very meta infographic about the problem with infographics:

Think Brilliant has more here.

The key with infographics is much like those little picture memes that are everywhere: consider them wrong at baseline, and only trust after verifying.

If you want more, read Part 4 here.