I know the word normal distribution
. But what does a statistician call it when the curve of the distribution tails less steeply (ie falls more gradually) to the left or right?
-
5Do you mean "a positive/negative skew"?– CentaurusOct 31, 2015 at 16:53
-
1I agree with Centaurus -- "skewed curve".– Hot LicksOct 31, 2015 at 20:08
-
Yes, the pictures tell me that I absolutly mean that. :D Please transfrom your comment to an answer.– buhtzNov 1, 2015 at 10:26
2 Answers
Skew is the term that I learned in engineering school, 40-odd years ago. If the plot of a "smooth" frequency distribution is not perfectly symmetrical it is said to be skewed (or possess skewness) in one direction or the other. (I'm a little fuzzy on what goes on if the distribution "curve" is ragged.)
A distribution can be skewed due to a longer "tail" on one side or a fatter tail or both.
Skew is a convenient term since it's basic meaning is intuitive to anyone with good familiarity with English, even though the detailed technical definition is incredibly Greek.
It's simply referred to as a long-tail[ed] distribution.
Wikipedia has:
In statistics, a long tail of some distributions of numbers is the portion of the distribution having a large number of occurrences far from the "head" or central part of the distribution. The distribution could involve popularities, random numbers of occurrences of events with various probabilities, etc.
A probability distribution is said to have a long tail if a larger share of population rests within its tail than would under a normal distribution. A long-tail distribution will arise with the inclusion of many values unusually far from the mean, which increase the magnitude of the skewness of the distribution. A long-tailed distribution is a particular type of heavy-tailed distribution.
-
"Long-tailed" is referring to the deviation from "normal" ("normal distribution", that is), not the degree of skew. A normal distribution can still be skewed. Oct 31, 2015 at 20:11
-
1No it can't. The equation demands that there is a line of symmetry in the graph: it's a translated even function. You're talking about a 'skew normal distribution' (a three word compound); there's no such thing as a skewed 'normal distribution' {Wikipedia}. Are you a mathematician? Oct 31, 2015 at 21:25
-
1@HotLicks I am a mathematician and I work with probability and I confirm that the normal distribution cannot be skewed. The probability distribution function is a Gaussian curve. Nov 1, 2015 at 4:36
-
1Certainly a distribution can be skew, that's a fine term. But then it's not a normal distribution anymore (there may be such a thing as a "skew normal distribution", but that's different from a normal distribution; everybody knows what you're talking about when you say "normal distribution."). Nov 1, 2015 at 7:07
-
1@Hotlicks anyway, I would probably use "asymmetrical" or "skew distribution." As far as I know this answer is fine though. A statistician has more use for precise names of distributions. I was just putting in my two cents to confirm that normal distribution has a standard meaning. Nov 1, 2015 at 7:16