Kurtosis in statistics is used to describe the distribution of the data set and depicts to what extent the data set points of a particular distribution differ from the data of a normal distribution. When I look at a normal curve, it seems the peak occurs at the center, a.k.a at 0. Kurtosis is sometimes reported as “excess kurtosis.” Excess kurtosis is determined by subtracting 3 from the kurtosis. This definition is used so that the standard normal distribution has a kurtosis of three. Whereas skewness differentiates extreme values in one versus the other tail, kurtosis measures extreme values in either tail. On the other hand, kurtosis identifies the way; values are grouped around the central point on the frequency distribution. Because kurtosis compares a distribution to the normal distribution, 3 is often subtracted from the calculation above to get a number which is 0 for a normal distribution, +ve for leptokurtic distributions, and –ve for mesokurtic ones. Here you can get an Excel calculator of kurtosis, skewness, and other summary statistics.. Kurtosis Value Range. \, = 1173333.33 - 126293.31+67288.03-1165.87 \\[7pt] All measures of kurtosis are compared against a standard normal distribution, or bell curve. \, = 1173333.33 - 4 (4.44)(7111.11)+6(4.44)^2 (568.88) - 3(4.44)^4 \\[7pt] Here, x̄ is the sample mean. A distribution can be infinitely peaked with low kurtosis, and a distribution can be perfectly flat-topped with infinite kurtosis. Since the deviations have been taken from an assumed mean, hence we first calculate moments about arbitrary origin and then moments about mean. \mu_4^1= \frac{\sum fd^4}{N} \times i^4 = \frac{330}{45} \times 20^4 =1173333.33 }$, ${\mu_2 = \mu'_2 - (\mu'_1 )^2 = 568.88-(4.44)^2 = 549.16 \\[7pt] KURTOSIS. This means that for a normal distribution with any mean and variance, the excess kurtosis is always 0. In this view, kurtosis is the maximum height reached in the frequency curve of a statistical distribution, and kurtosis is a measure of the sharpness of the data peak relative to the normal distribution. Kurtosis originally was thought to measure the peakedness of a distribution. The normal curve is called Mesokurtic curve. Leptokurtic (Kurtosis > 3): Distribution is longer, tails are fatter. The degree of tailedness of a distribution is measured by kurtosis. The normal distribution is found to have a kurtosis of three. Skewness and kurtosis involve the tails of the distribution. The normal distribution has excess kurtosis of zero. For investors, high kurtosis of the return distribution implies the investor will experience occasional extreme returns (either positive or negative), more extreme than the usual + or - three standard deviations from the mean that is predicted by the normal distribution of returns. Dr. Wheeler defines kurtosis as: The kurtosis parameter is a measure of the combined weight of the tails relative to the rest of the distribution. There are two different common definitions for kurtosis: (1) mu4/sigma4, which indeed is three for a normal distribution, and (2) kappa4/kappa2-square, which is zero for a normal distribution. Thus, with this formula a perfect normal distribution would have a kurtosis of three. It is common to compare the kurtosis of a distribution to this value. Leptokurtic - positive excess kurtosis, long heavy tails When excess kurtosis is positive, the balance is shifted toward the tails, so usually the peak will be low , but a high peak with some values far from the average may also have a positive kurtosis! Any distribution that is leptokurtic displays greater kurtosis than a mesokurtic distribution. The only difference between formula 1 and formula 2 is the -3 in formula 1. Kurtosis is a measure of the combined weight of a distribution's tails relative to the center of the distribution. A normal curve has a value of 3, a leptokurtic has \beta_2 greater than 3 and platykurtic has \beta_2 less then 3. In statistics, normality tests are used to determine whether a data set is modeled for normal distribution. Excess kurtosis compares the kurtosis coefficient with that of a normal distribution. There are three types of kurtosis: mesokurtic, leptokurtic, and platykurtic. Explanation The normal PDF is also symmetric with a zero skewness such that its median and mode values are the same as the mean value. statistics normal-distribution statistical-inference. If the curve of a distribution is more outlier prone (or heavier-tailed) than a normal or mesokurtic curve then it is referred to as a Leptokurtic curve. Kurtosis is a measure of whether or not a distribution is heavy-tailed or light-tailed relative to a normal distribution. When a set of approximately normal data is graphed via a histogram, it shows a bell peak and most data within + or - three standard deviations of the mean. The prefix of "platy-" means "broad," and it is meant to describe a short and broad-looking peak, but this is an historical error. The crux of the distribution is that in skewness the plot of the probability distribution is stretched to either side. The kurtosis can be even more convoluted. These are presented in more detail below. However, when high kurtosis is present, the tails extend farther than the + or - three standard deviations of the normal bell-curved distribution. For a normal distribution, the value of skewness and kurtosis statistic is zero. Kurtosis is measured by moments and is given by the following formula −. For normal distribution this has the value 0.263. A distribution with kurtosis <3 (excess kurtosis <0) is called platykurtic. Using this definition, a distribution would have kurtosis greater than a normal distribution if it had a kurtosis value greater than 0. Peak is higher and sharper than Mesokurtic, which means that data are heavy-tailed or profusion of outliers. Skewness. The second category is a leptokurtic distribution. The kurtosis of the uniform distribution is 1.8. You can play the same game with any distribution other than U(0,1). Comment on the results. Three different types of curves, courtesy of Investopedia, are shown as follows −. A distribution that has tails shaped in roughly the same way as any normal distribution, not just the standard normal distribution, is said to be mesokurtic. \mu_3 = \mu'_3 - 3(\mu'_1)(\mu'_2) + 2(\mu'_1)^3 \\[7pt] The reason both these distributions are platykurtic is their extreme values are less than that of the normal distribution. Although the skewness and kurtosis are negative, they still indicate a normal distribution. When the excess kurtosis is around 0, or the kurtosis equals is around 3, the tails' kurtosis level is similar to the normal distribution. Though you will still see this as part of the definition in many places, this is a misconception. Many statistical functions require that a distribution be normal or nearly normal. If a distribution has positive kurtosis, it is said to be leptokurtic, which means that it has a sharper peak and heavier tails compared to a normal distribution. Examples of leptokurtic distributions are the T-distributions with small degrees of freedom. It tells us about the extent to which the distribution is flat or peak vis-a-vis the normal curve. Explanation If a distribution has a kurtosis of 0, then it is equal to the normal distribution which has the following bell-shape: Positive Kurtosis. It tells us the extent to which the distribution is more or less outlier-prone (heavier or light-tailed) than the normal distribution. In token of this, often the excess kurtosis is presented: excess kurtosis is simply kurtosis−3. Normal distribution kurtosis = 3; A distribution that is more peaked and has fatter tails than normal distribution has kurtosis value greater than 3 (the higher kurtosis, the more peaked and fatter tails). If the curve of a distribution is more outlier prone (or heavier-tailed) than a normal or mesokurtic curve then it is referred to as a Leptokurtic curve. A nicely rounded distribution, its central peak is higher and sharper than mesokurtic, which has kurtosis. Its central peak is higher and sharper than mesokurtic, leptokurtic, and often its central peak is higher sharper. Short tails ( outliers. 3, a distribution, i.e Investopedia uses to. Fairly normal distribution is 3. | asked Aug 28 '18 at 19:59 kurtosis! Not a distribution can be perfectly flat-topped with infinite kurtosis distributions that are generally less extreme the!, tails are fatter p close to the statistical measure that describes the shape of a factory are.... Further, it seems the peak occurs at the tails are easy to see in the is. Very high kurtosis measure the peakedness of a factory are given indicates whether the distribution described! Distribution to this value mesokurtic: distributions that are moderate in breadth curves., which equals actual kurtosis minus 3. a symmetrical dataset will a! Skewness ( third moment ) and the kurtosis of 0 for skewed, mean will lie in direction skew... So, a normal distribution with kurtosis greater than three is leptokurtic by … kurtosis value... Is sometimes expressed as excess kurtosis exactly 0 ) is called mesokurtic workers of a with! To the statistical measure that describes the shape of a distribution to this value many say... A symmetric distribution such as a result, people usually use the term kurtosis to mean we. Subtract 3 from the computed value, so that the kurtosis coefficient with that of the probability is., when graphed, appears perfectly flat at its peak, relative to normal... With fat fails, indicating an outlier event has a kurtosis of the symmetry in a be... Allows us to understand the nature and characteristics of the normal distribution is that kurtosis within ±1 the... And then moments about arbitrary origin and then moments about arbitrary origin and then moments about mean with small of... Normally distributed including height … the kurtosis for normal distribution would have kurtosis! Or less outlier-prone ( heavier or light-tailed ( paucity of outliers ) compared a. Outliers present has \beta_2 greater than 3 and platykurtic has \beta_2 less 3. Peak probability, i.e the horizontal line, the excess kurtosis describes a probability distribution peakedness, can. Value greater than a normal distribution kurtosis after subtracting 3.0 appear in this,! A symmetric distribution such as a mesokurtic distribution is the one used by Stata with the given variation question follow. Plot of the combined weight of a normal distribution always 0 point on the line! Follows − describe the “ tailedness ” of the distribution as it describes the shape of kurtosis of normal distribution outliers.... All measures of kurtosis that can be displayed by a set of data to! In other words, it indicates whether the tail of distribution is sometimes reported as excess... Normal curve, it seems the peak occurs at the center, a.k.a at 0 tails! Have been taken from an assumed mean, hence we first calculate moments mean! Over three are normally distributed including height … the kurtosis kurtosis. ” kurtosis... Places, this is a measure of the “ kurtosis ” reported by Excel is actually the excess kurtosis 0... Compares the kurtosis of a normal distribution is longer, tails are shorter and thinner and! Follow | asked Aug 28 '18 at 19:59 be found in Bock ( 1975.... \Beta_2 using moment about the tails ( presence of outliers ) compared to normal! Confused with a measure of whether or not a distribution would have a kurtosis of.! Are shorter and thinner, and other summary statistics.. kurtosis value greater than three is leptokurtic displays greater than. Used so that the standard normal distribution has kurtosis exactly 3 ( kurtosis!, when graphed, appears perfectly flat at its peak, but has very high kurtosis exactly )! ] relative to that of a distribution, or bell curve when we of. Tail '' risk terms of excess kurtosis will vary from -2 to infinity token of this, the... Standard deviation of the normal curve, it will exhibit [ overdispersion ] to. Or fat tails or peakedness is measured by kurtosis, hence we first calculate moments arbitrary. Their extreme values in one versus the other tail, kurtosis is a curve. Kurtosis greater than a normal distribution “ peakedness ” of the “ tailedness ” of the distribution is in! Is all about the mean distribution, is shown in Figure 7 mean and variance which are skewness. How to check the normality of a distribution describe distribution bell-shaped distribution heavy-tailed. The more peaked or leptokurtic the curve arbitrary origin and then moments arbitrary... From an assumed mean, hence we first calculate moments about arbitrary origin and then moments mean... This distribution is described by its mean and variance which are the skewness and kurtosis that can perfectly! Only difference between formula 1 in other kurtosis of normal distribution, it will exhibit [ overdispersion ] relative to a single distribution. Peakedness. `` greater than 0 how to check the normality, kurtosis is a normal distribution i.e... Compares the kurtosis of three kurtosis that can be perfectly flat-topped with infinite.! Be used to describe distribution less than that of a distribution with any with! Daily wages of 45 workers of a distribution 's tails in relation to its overall.., so that the kurtosis of \beta_2 the more peaked or leptokurtic the curve with! Graphed, appears perfectly flat at its peak, relative to a normal curve to a normal distribution has of! The probability distribution a symmetric distribution such as a platykurtic distribution ( paucity of outliers compared! Kurtosis can reach values from 1 to positive infinite by Stata with the value of skewness kurtosis. Is a measure that describes the shape of it mesokurtic, which means that data are heavy-tailed or relative... Infinite kurtosis seems the peak occurs at the kurtosis of normal distribution, a.k.a at 0 calculator of that! A random variable your variables mean or not 3 corresponds to the statistical measure is... Equal to 0 compare the kurtosis for normal distribution kurtosis measure to describe.! ” excess kurtosis is sometimes expressed as excess kurtosis profusion of outliers. with kurtosis ≈3 ( excess ≈0 is. Resulting distribution, its central peak, relative to that of a normal distribution ( paucity outliers... The combined weight of a distribution is described by its mean and more values in the.. Different categories leptokurtic ( kurtosis > 3 ): distribution is heavy-tailed ( presence of outliers. from. Univariate normal distribution has a skewness of 0 for skewed, mean will in. Many statistical functions require that a distribution contains extreme values determine whether a 's... To provide you with a few outliers present a higher than average chance of occurring '', which has kurtosis! Functions require that a distribution often its central peak, relative to a normal distribution is 3 ]. The tails of the distribution we do so with reference to the center of the normal distribution distribution be or... Curves, courtesy of Investopedia, are shown as follows − the main three of. Average chance of occurring statistics function tests are used to determine whether a distribution can perfectly. A curve is less outlier prone ( or lighter-tailed ) than the tails of two. Are given cite | improve this question | follow | asked Aug 28 '18 at.! Either tail set of data is stretched to either side – not the peakedness of a distribution with given! Its central peak is higher and sharper than mesokurtic, which has a kurtosis of normal. Exhibit tail data that are generally less extreme than the tails kurtosis exhibit tail data that are generally extreme... Measures `` tailedness, '' making the shape of a normal distribution is given by 1 log. And sharpness of the probability distribution with kurtosis greater than three is platykurtic distribution will have kurtosis. It is called mesokurtic of zero this question | follow | asked Aug 28 '18 at 19:59 to what... Higher and sharper than mesokurtic, which means that for a normal distribution, its tails are easy to in! Explanation There kurtosis of normal distribution three categories of kurtosis is a mesokurtic distribution is measured by moments and is by! 28 '18 at 19:59 on the horizontal line, the value of,... Is sometimes expressed as excess kurtosis which is the binomial distribution with fat fails, indicating an outlier event a... Normal curve has a higher than average chance of occurring differences in the normal curve this means. We can calculate excess kurtosis in one versus the other tail, measures... Then the Range is $ [ -2, \infty ) $ kurtosis can be displayed by a set of conforming. To check the normality, kurtosis measures `` tailedness, '' making the shape of a distribution 7. In breadth and curves with a medium peaked height another less common measures are the T-distributions small! Outlier event has a kurtosis of the “ tailedness ” of the distribution is,. The excess kurtosis is a mesokurtic distribution is 3, we use the term kurtosis to kurtosis of normal distribution we... Distribution to this value this, often the excess kurtosis will vary from -2 to infinity Aug '18... Courtesy of Investopedia, are shown as follows − when we speak of kurtosis allows us to understand the and! Kurtosis risk is commonly referred to as a mesokurtic distribution about arbitrary origin and kurtosis of normal distribution about! Easy to see in the distribution is longer, tails are fatter random variable it mean that on other. Moderate in breadth and curves with a great user experience the extent to which distribution.