You will need to know the frequency for each class. (A class could consist of nominal variables like Ford,Audi,Toyota etc.., or discrete data like ages of presidents, or continuous data. In each case you can identify a class, and then find the frequency for that class.)

You determine the percentage for each class by dividing the frequency by the total frequency, usually written `f/n` .

You then take each percentage and multiply by `360^@` ; this gives you the central angle for each class. Note that since the percentages sum to 100, the degrees of the angles will sum to 360.

Now draw a circle. Using a protractor measure out a central angle for each class -- the angles should have a radius in common. Then label each sector (pie wedge) with the appropriate data (either the frequency or percent along with the class name)

------------------------------------------------------------------

Ex: Suppose in a certain town there were 1000 people. 400 were democrats, 400 were republicans, and 200 were independent. Then the classes are d,r,and i with respective frequencies of 400,400,200.

The percentages are democrats `400/1000=.4` or 40%, republicans 40% and independents `200/1000=.2` or 20%.

Now we take `(.4)(360)=144` and `(.2)(360)=72` .

Draw a circle: using a protractor measure an angle of `144^@` with the vertex of the angle at the center of the circle. From one of the sides of the angle, measure another angle of `144^@` again with the vertex at the center. The remaining angle should measure `72^@` (You should measure it just in case).

Now the circle is divided into sectors (wedges): label one of the large sectors democrats and 40%, the other large sector republicans and 40%, and the small sector independents and 20%.

The same method holds true if the data are not nominal ;e.g. you make a pie chart to depict the heights of students in your school. In this case you would have identified classes (5' to 5'4",5'4" to 5'8", etc...) and these classes are treated the same way as we treated democrats, etc...