0

enter image description here

Hey guys, I am researching a social science topic on fertility rate per county. However when I finish my linear model and begin to plot my QQ plot, I got a curve that is curving up. Is it a heavy tailed one?

Kristian Richthoffen
  • 1
  • 1
    See the answer [here](http://stats.stackexchange.com/questions/101274/how-to-interpret-a-qq-plot/101290#101290) – Glen_b Apr 26 '16 at 02:03
  • So it's heavy tailed. right? – Kristian Richthoffen Apr 26 '16 at 02:05
  • 1
    I'd describe it first as right skewed. On the left it is mildly heavy-tailed. Below about -1/2 the relationship is straight with slope close to 1, which suggests the left tail is close to normal, but with a larger standard deviation than that in the middle (between -1/2 and 1/2 on the actual quantiles) where the peak is sharper (the standard deviation is smaller). Above that there's a distinct curve pulling the tail far up on the right, which dominates the impression of shape. So right skew, somewhat heavy tailed/peaky. – Glen_b Apr 26 '16 at 03:39

1 Answers1

1

It looks both heavy tailed and somewhat skewed to me. I think the best way to use a qq-plot is just to see if the distribution looks normal overall (follows the line). Then you can interpret the shape with other visualizations such as histograms and density plots.

gung - Reinstate Monica
  • 132,789
  • 81
  • 357
  • 650
  • 1
    Seems worth noting that fertility data may be naturally right skewed owing to the impossibility of going below zero while large variables compared to the mean may easily be found. – Robert de Graaf Apr 26 '16 at 04:49