The following is a basic explanation of commonly used summary statistics (including mean, median, variance, standard deviation, skewness, kurtosis, and percentiles), which you can calculate in Excel using the Descriptive Statistics Calculator. Follow the links for more details concerning calculation and interpretation of individual measures.
The size of the data set (sample or population) is simply how many observations (values) you use to calculate all the summary statistics. It is usually marked as N (or n). In Excel you can easily get N using the COUNT function. In the Descriptive Statistics Calculator, N is in cell D8.
Arithmetic mean is the most commonly known and used of all the summary statistics. It is the arithmetic average of all values in a data set. Arithmetic average is calculated as the sum of all values divided by the data set size (N). In Excel, you get arithmetic mean using the AVERAGE function. It is cell D9 in the calculator.
The simplicity of its calculation and interpretation is the main advantage of artihmetic mean. However, there are also drawbacks. In some cases (for example when there are extreme values in the data set), other measures tell a better story about the data (one of them is median – see below).
More about arithmetic average:
Median often complements arithmetic mean as a measure of central tendency (average or most “typical” value) of a data set. Median is the value which is greater than half of all the other values and smallar than the other half. If you sorted all the values in a data set from smallest to greatest, median would be ranked in the middle. In Excel, the function is MEDIAN. The cell in the Descriptive Statistics Calculatoris is D10.
You can see more detailed explanation of median and its calculation here:
Variance is a measure of dispersion or variability in the data. It is defined as the average squared deviation from the mean and therefore it’s always non-negative. Low variance is interpreted as little variability in the data set (the values tend to be very similar), while high variance indicates a very diverse data set with big differences in individual values.
Standard deviation is a more common measure of dispersion (or volatility). It is directly derived from variance – in fact standard deviation is just a square root of variance. Its advantage and the main reason why standard deviation is used more frequently than variance is that it is measured in the same units as the underlying data, while variance is measured in the units squared.
For example, standard deviation of a set of prices is measured in dollars, while variance is measured in dollars squared. Standard deviation of investment returns (historical volatility) is measured in percent, while variance would be measured in percent squared.
Variance is still very useful for various tests and analyses (e.g. Analysis of Variance – ANOVA) or as an input for calculation of other measures or indices. With variance, you don’t need to be bothered with the square root as with standard deviation.
The exact calculation of variance and standard deviation is slightly different depending on the data set being a sample or the entire population. The respective Excel functions are VAR.S, VAR.P, STDEV.S, and STDEV.P. Variance is in cell D11 and standard deviation in D12 in the Descriptive Statistics Calculator. You can set whether your data set is sample or population in the combo in cell D6.
Here you can see more detailed explanation of variance and standard deviation and their calculation:
Skewness measures the symmetry of the distribution – most importantly it compares relative frequency of extreme low (left tail) and extreme high (right tail) values. Normal distribution, which is perfectly symmetric, has skewness of zero. Positive skewness means that extremely high values are relatively more common (right tail is fat), while negative skewness means that extremely low values are more common (left tail is fat). In finance and investing (and even more so in options pricing and trading), knowing skewness of return distributions is very useful, as it may indicate frequency or probability of huge gains and (more importantly) huge losses.
The calculation of skewness may look complicated at first, but as soon as you get the underlying logic, it is quite straightforward. It is not unlike calculating variance and standard deviation. Skewness too has slightly different formula for population and sample. Here you can see a detailed explanation and derivation of skewness formula.
In Excel, you can calculate sample skewness using the built-in SKEW function. There is no built-in function for population skewness, but you can easily calculate it through a small adjustment of the SKEW function (see how). In the calculator skewness is in the cell D13.
Kurtosis measures how fat tails the distribution has. High kurtosis means that there are many extreme values far from the mean to either side; low kurtosis means extreme values are relatively less common.
Kurtosis is often quoted in the form of excess kurtosis, which equals kurtosis less 3. The lowest possible value of excess kurtosis is negative 2; the highest is infinite. Normal distribution has excess kurtosis equal to zero. This is why excess kurtosis is often used rather than basic kurtosis – excess kurtosis is the kurtosis relative to, or in excess of, normal distribution kurtosis.
As with skewness, there is a small difference in calculation of population and sample kurtosis. See detailed explanation of formulas for sample and population kurtosis and excess kurtosis.
In Excel, the function for sample excess kurtosis is KURT. There is no built-in Excel function for population kurtosis (you need to calculate it manually by adjusting from KURT – see how to do it here). You can calculate either population or sample excess kurtosis in the Descriptive Statistics Calculator in cell D14.
Percentiles are used to describe the structure of the data in greater detail. They use the same logic as median (above on this page), which in fact is the 50th percentile, as there are 50% values in a data set greater and 50% values smaller than median. Analogically, for 5th percentile there are 95% values greater and 5% smaller, for 99% percentile there are 1% values greater and 99% smaller – and for Nth percentile there are 1-N% values greater and N% values smaller.
The Excel function for percentiles (exclusive or inclusive) is PERCENTILE.EXC and PERCENTILE.INC. In the Descriptive Statistics Calculator, percentiles are in cells C18-C28.
The extreme values – minimum and maximum provide a very useful information that the other summary statistics don’t. In many cases (and especially in finance and investing), it is very useful to know the actual smallest and largest value (e.g. portfolio return in a period). Nevertheless, if your data set is only a sample (e.g. past returns), you should still be prepared for the population (which also includes unknown future returns) to contain even more extreme values.
In Excel, the functions are MIN and MAX. In the Descriptive Statistics Calculator, maximum value is at the top of the percentiles table (cell C17) and mimimum at the bottom (cell C29).
You can find screenshots and download the calculator here.