Paging Mr. Edward Tufte!
January 12, 2014 8:31 AM   Subscribe

What is the lie factor of this graph?
posted by Tom-B to Science & Nature (4 answers total)
 
Best answer: This graph is all kinds of weird (why is 6.5% shorter than 5.91%?!). I guess that's why you're asking about it.

I'm not an expert, but an initial calculation based on the formulas on your link would be:

The lowest bar in the graph measures 0.45 units, and the tallest bar measures 2.31 units.

Then, the size of effect in graphic is: (2.31-0.45)/0.45 = 4.13

Meanwhile the size of the effect in the data is: (5.91-4.31)/4.31 = 0.37

Then the Lie Factor would be: 4.13/0.37 = 11.16
posted by Wulfhere at 9:00 AM on January 12, 2014


If anyone would like to do calculations for the other bars, here is the graphic I used to measure them.
posted by Wulfhere at 9:06 AM on January 12, 2014


For starters I'm going to guess that the typo factor of this chart is "the last column was supposed to read 6,91%".
posted by ook at 9:08 AM on January 12, 2014


For starters I'm going to guess that the typo factor of this chart is "the last column was supposed to read 6,91%".

I think the typo was in the data entry rather than the labeling of the 2013 value, as Brazil's 2013 inflation has been reported at 5.91%.

Perhaps that was unintentional. And sometimes, not showing a zero line on a bar chart is an innocent mistake. But allowing for those errors, the line labeled "4,5%" is either mislabeled or misplaced. Looks more like it is located at around 4.8% on the vertical axis.

Here's an article
including a better chart of inflation in Brazil, with a longer-term context. The article explains that "The government considers prices within two percentage points of 4.5% to be within its target." An honest charting of that target area would produce a very broad horizontal stripe encompassing all the values.
posted by Snerd at 9:49 AM on January 12, 2014


« Older What is this movie?   |   Resources on understanding female power Newer »
This thread is closed to new comments.