This is a minor rant, since I'm at home this Monday on a holiday, so I'm less irritated by life in general than I normally am on a Monday.
In marking hordes of undergraduate formal lab reports, one thing that gets fucked up on a regular basis is any graph. For some reason, the repetitive lessons everyone gets in junior high school, high school, and intro-level science courses about appropriate labels and axes in graphs are easy to forget. Look, you're supposed to pay attention to the axes labels on a graph because that's how you know some slimy marketer-type is lying to you.
This concept must be foreign to both the hordes of SFU undergrads and MicroSoft's software engineers, given the default (shitty) settings of MS Excel.
You MUST label your graph axes.
You MUST include a numerical scale
You MUST make the scale make sense - drop the 120% mark on your "percent survival" curve!
If your numbers are all much closer together than any is to zero, you have to either use a ridiculously long vertical axis, or include a visible break line.
These are not difficult rules - so why do I see so many Watson-damned broken figures?
Monday, October 09, 2006
Subscribe to:
Post Comments (Atom)
2 comments:
"If your numbers are all much closer together than any is to zero, you have to either use a ridiculously long vertical axis, or include a visible break line."
This one is my pet-peeve. What the hell is wrong with people an axes. For example, they'll frequently make histograms around a distribution of between say, 120 to 240 (in some numerical scale) but start the 'bins' at 0 so that all info is lost. Ah yes, ALL the values are in 101 - 175 and 176 - 250 bins... Great job asswad.
Sorry for butchering the spelling and punctuation on that last comment... Trying to type too quickly = bad.
Post a Comment