I wonder how to interpret this plot. I know, in general, that a qqplot depicts the quantiles of the data against the quantiles of the normal function.
How do I apply this to this plot?
Asked
Active
Viewed 1,517 times
4
-
1Might be worth comparing to an exponential distribution – Henry Mar 07 '22 at 23:28
-
@Henry In which way? You mean instead of comparing it with a normal distribution? – Ben Mar 08 '22 at 06:31
-
I suspect it might be informative – Henry Mar 08 '22 at 08:36
-
sure, I just wonder in which way I could compare it? – Ben Mar 08 '22 at 09:14
-
@Ben, you could use a qq-plot against an exponential instead of against a normal. There are lots of posts about these topics on the site. You should search around & read some. – gung - Reinstate Monica Mar 08 '22 at 12:06
1 Answers
8
I explain how to read qq-plots in general here: QQ plot does not match histogram, and walk through constructing one here: PP-plots vs. QQ-plots. Those posts may help you.
Because your data are on the vertical axis, when we see the top right points above the line, we can conclude that they are too far out relative to a true normal, whereas the lower left points aren't far enough out. That is, your data are positively skewed relatively to a true normal. In addition, we can see that the points at the bottom turn perfectly horizontal. That means there is a floor that your data don't drop below. These could be data that cannot be negative, for example.
gung - Reinstate Monica
- 145,122
-
Thanks a lot! In case it would be negatively skewed, would the data need to be mirrored along the red line? – Ben Mar 08 '22 at 06:33
-
1@Ben, if the data were normal, they would follow the line. If they were negatively skewed, the top & bottom points (at the ends) would be below the line. – gung - Reinstate Monica Mar 08 '22 at 12:05