5.2.4. Random Numbers

A random variable is generated with a specified length and with distribution parameters approximate to those specified by the user.

First select a distribution from the distributions menu. There are three Distribution Functions for which the Random Numbers procedure is not implemented: Weibull, negative binomial and hypergeometric distributions. After choosing any other distributions, the program will ask for the length of the random variable. The number suggested by the program is the maximum row number which is currently in use; i.e. the Used row number as displayed on the Status Panel of the Data Processor. Next the distribution parameters are entered. The numbers suggested by the program are the ones last entered by the user. A table containing the column of generated Random Numbers is displayed.

There are two alternative ways of comparing the generated Random Numbers with the theoretical distribution function. The first is the Expected Frequencies procedure which is explained below. This procedure will display the Expected Frequencies on the basis of the generated Random Numbers either in the form of a Table or a Character Histogram. The second alternative is the Histogram procedure. This will display a high resolution histogram of the generated Random Numbers and optionally superimpose up to six estimated theoretical distributions simultaneously (see 5.3.3.3. Fitting Distribution Functions).

Example

Select Statistics 1 → Distribution Functions → Random Numbers, from the distributions list Student’s t and enter 10 for the Variable Size. At the parameter dialogue enter:

·      Mean = 0

·      Standard Deviation = 1

·      Degrees of Freedom = 5

When you reproduce this example you will obtain different numbers as they are randomly generated.

Random Numbers

Student’s t Distribution

Mean: 0.0000

Standard Deviation: 1.0000

Degrees of Freedom: 5.0000

 

	T-Rand
1	0.0971
2	-1.0758
3	0.1968
4	-0.0896
5	-1.2651
6	0.3970
7	-1.5811
8	-2.0857
9	-3.0010
10	0.3280