The scale in this chart is so bizarre—it is straight out of that old classic: “How to lie with statistics”.
Small moves are small moves—sticking them on a big graph is not helpful.
The scale in this chart is so bizarre—it is straight out of that old classic: “How to lie with statistics”.
Small moves are small moves—sticking them on a big graph is not helpful.
The Scaling it fine. Two months ago I wrote an email to the LAtimes Poll Survey Director Jill Darling because the scaling was biased toward Hillary, making some things appear that her margins were bigger then Trumps. I ask if they could correct the scaling.
Jill’s Reply to me is below:
TWO DAYS LATER TRUE TO HER WORD THEY CHANGED THE SCALING TO FAIRLY REPRESENT THE DATA.
Hello John,
Your point about the need for consistent scales is very well taken! We are in the process of working on making sure we are presenting those subgroup numbers in a way that makes good visual sense. What you pointed out about inconsistent scales is an artifact of the automated process of pulling in data and creating the graphs, and an oversight on our parts, not anything anyone here is doing to try to influence others. We just care about reporting our results!
Appreciated your thoughtful feedback, and wanted to let you know we’re already on the case. Should see an improvement in the information presented on that page in the next couple of days.
Regards,
Jill