Religion again. We had that last time. Often, such data is scaled like that. Have a look.
Data from the FASZ, 2013-03-10, page 22. Redesign: me. Scaled globally.
You might think scaling is fair. Because all the axes look the same. Same for different is rarely fair. You want to compare trends. Different ones. On different levels. Look at the people with no religion (in German: “religionslos”). Starting at zeropointsomething up to 11 percent. Which is much more than the muslims: starting at 13.1, growing up to 22.5 percent. But both look more or less alike.
This is better. Have a look.
The same, scaled comparably.
You calculate like this: What changes the most? In percent. That gets the full height. The rest gets less. In relation to the one that has changed the most. Like that you get them all in. More steep means more change. Less steep means less change. You read the magnitude from the numbers.