Calling Bullshit 6.3: Manipulating Bin Sizes

[@kOaMaster2]

I love your whole course, definitely going to watch it until the end, it's a great and important knowledge you are teaching here. I wish I had something like this at my former university. However, I do not agree with your "fair" bin size conclusion at the end. The bin size shouldn't be equal to the income stretch but it should include to the percentage of people belonging to it. Since the goal of the graph is to show the main targeted "tax payer class" this is only defined by a part/percentage of the whole population, not by the absolute income. And to include a little bit more insight about the amount of taxation, one could add "average tax-% from income" in each bin (or as line graph).

[@antistone349]

It seems to me that the intuition behind those unequal income bins was probably to do the bin equivalent of a log scale, with bins getting exponentially bigger in a somewhat-regular fashion (though distorted to make the boundaries fall on round numbers). One could argue about whether a log scale is appropriate here and whether that distortion is acceptable, but if you believe that log-scale graphs are sometimes OK and that bins are sometimes OK then I don't see why one should *categorically* object to log-scale bins.

[@Innengelaender]

I would agree with that in general. In this case you can still clearly see how the sudden jump distorts the image and that is "Bullshit". Without pointing out that jump that can be very misleading for someone who doesnt look out for that (At first I did not notice that either). As someone else here commented when you talk about "middle class" you shouldnt talk about absolute income at all but equal percentiles of population to get a better picture. So approaching the topic by absolute income-numbers is already distorting the picture. In the end one graph is always a simplification and will never tell the whole truth. Some are just better than others for a given situation.