Net favorability is approval-disapproval. It doesn’t necessarily say anything about a majority of everything. It depends on the study design.
You probably arent’ going to be convinced this isn’t “lying with data”™, but in your example, if you were a car guy, which figure would you prefer, the left or the right? The right would be basically useless if you were someone who was in the space and had the context to interpret the figure.
I would prefer a table if I needed the figures that I could use. A bar chart is for a quick comparison by its very nature, and the left one suggests that the Toyota is using four times less fuel than the Honda.
Toyota is using four times less fuel than the Honda.
Only if you didn’t read the figure. Which is the point. A car person would never make that mistake.
And I promise, in figure making, we are almost always adjust axis to represent the range of the data. It would be ridiculous not to. It would be a waste of real-estate for any journal or paper. If you are writing a Science paper, you get 5 figures, total. You have to do as much with them as you can. There are plenty of times when we might only adjust in 1/2 of a dimension, but usual its the whole
Thanks for admitting that you make useless charts on which the reader has to read the figures anyway. Hopefully there comes a time when you realize you should’ve just used tables instead.
You’re really saying that you like the crap here on the left?
There is lying with data and then there is lying with data. This ain’t lying with data.
The axis are labeled and uniformly scaled. It’s fine.
They’re just as labeled and ‘uniformly scaled’ as in my example.
Musk’s line looks like he’s supported by the majority of Republicans.
Net favorability is approval-disapproval. It doesn’t necessarily say anything about a majority of everything. It depends on the study design.
You probably arent’ going to be convinced this isn’t “lying with data”™, but in your example, if you were a car guy, which figure would you prefer, the left or the right? The right would be basically useless if you were someone who was in the space and had the context to interpret the figure.
I would prefer a table if I needed the figures that I could use. A bar chart is for a quick comparison by its very nature, and the left one suggests that the Toyota is using four times less fuel than the Honda.
Only if you didn’t read the figure. Which is the point. A car person would never make that mistake.
And I promise, in figure making, we are almost always adjust axis to represent the range of the data. It would be ridiculous not to. It would be a waste of real-estate for any journal or paper. If you are writing a Science paper, you get 5 figures, total. You have to do as much with them as you can. There are plenty of times when we might only adjust in 1/2 of a dimension, but usual its the whole
Thanks for admitting that you make useless charts on which the reader has to read the figures anyway. Hopefully there comes a time when you realize you should’ve just used tables instead.