Sociological Images has a guest post today about how data presented in graphs can be misleading to the unobservant viewer. Take this graph for example, which supposedly correlates lower GPAs with increased alcohol consumption among undergraduates. 
The trend looks pretty damning indeed until you realize that the y-axis only ranges from 3.0 to 3.5, which means the difference between a student who consumes zero drinks of alcohol per week and those who have 6+ drinks a week is far less than a full letter grade. At that point you kind of have to wonder if you can even really call it a trend anymore. I for one don't really have much faith in GPAs as an indicator of intelligence or performance because of the wide-spread proliferation of grade inflation (or maybe I'm just bitter because most of my engineering classes were curved to a C average, which honestly is as it should be). Anyways, check out more examples in the full post.
The trend looks pretty damning indeed until you realize that the y-axis only ranges from 3.0 to 3.5, which means the difference between a student who consumes zero drinks of alcohol per week and those who have 6+ drinks a week is far less than a full letter grade. At that point you kind of have to wonder if you can even really call it a trend anymore. I for one don't really have much faith in GPAs as an indicator of intelligence or performance because of the wide-spread proliferation of grade inflation (or maybe I'm just bitter because most of my engineering classes were curved to a C average, which honestly is as it should be). Anyways, check out more examples in the full post. 
No comments:
Post a Comment