How to trick people using.. correct data!

Its unbelievable how easily we can be tricked! Even the most perceptive amongst us, even the rational amongst us, even the cautious amongst us.

I was reading an article that explained how hotels use different kinds of messages to get guests to reuse the towels. Most hotels place a well crafted message urging us to reuse the towels. Some message use environmental friendliness as their message while some others talk about how other guests reuse – hoping to create some social pressure. This article talked about which of these is more effective. They showed this graph that makes it clear that we respond better when there is a social pressure (description norm message).

There is no question about what this data is telling – that we respond far more to social pressure.

Only when I finished the article and read few comments did I realize how cleverly this graph has been manipulated to exaggerate the difference. I was stunned that even I Рwho works with presentations for a living, who sees hundreds of presentations every month, could be misled so easily.

I have been studying for a while on how our mind looks for a pattern when none exists. This is a classic example of how our mind can be tricked to see a story when none exists.

So, what’s happening here?

Typically, the way your mind reads a chart like this is to compare the sizes of each bar to figure out the relationship between each. In this case, your mind sees the chart like this. It mentally compares the sizes of the bars to figure out the difference. In the picture below, I have added the grey boxes to the chart:

So, its pretty logical to see that the bar on the right side is 5x taller than the left. The conclusion you’d come to is that message with social pressure is 5x more effective. That’s pretty powerful, right?

But the key difference is that your mind has been tricked about the scale of the difference. In typical charts, the height of the bar is the absolute height. But in the graph above, if you see the scale on the left hand side, unlike typical charts, it doesn’t start at zero. They are showing a zoom view where the graph itself starts at around 30 (not 0). The value of the left bar is 35 and the right bar is 44. The difference between them is a meagre 9. Its not 5x by any stretch of imagination.

You’ll appreciate this only when you actually see the zoomed out version of the graph. I made one up in excel. Here it is.

Now you can see that the difference isn’t as startling as it seemed.

In fact even this chart is misleading because the scale on the left goes only from 0 to 50. So, even though you can now tell that there is no real difference between the two messages, you are still tricked to think that both are pretty effective – both seem to almost hit the top.

The graph that tells the real story is the one below where the scale goes from 0 – 100.

Now we know what exactly is going on.

Less than half the guests even respond to either message – which tells that both messages are not effective. So, it really doesn’t matter how one is better than the other – both are pretty ineffective. And frankly between the messages, there’s isn’t much difference.

If you were a hotel manager, and you didn’t see this difference, based on the first graph, you’d be right now rushing to change the message to incorporate social pressure on your guests. But in reality, either message didn’t do well – so you should be coming up with some other idea.

Before someone sends me a nasty message, let me clarify that the author of the original article didn’t “hide” any information. The tampered scale is visible. But because it was zoomed in, our mind is tricked to see the cropped bars as the absolute bars and that exaggerated the difference. I believe that there was an intention to exaggerate this difference. Because, had they shown the last graph (scale going from 0-100), the whole point of the article would become invalidated.

Well, it really doesn’t matter in this article because readers are not going to make any important investment decisions based on this chart (unless you are a hotel manager), but I think there is a responsibility that each author must bear – and clearly, this one fell short.

PS: Here’s the original article



Comments are closed.