Feinstein listed the different shapes commonly seen in data distributions. The type of shape determines the mathematics of the analysis.
Parameters:
(1) symmetrical vs nonsymmetrical
(2) convex (open down) or concave (open up)
(3) defining features
Symmetry |
Concave/Convex |
Features |
Term |
symmetrical |
convex or concave |
square wave or pulse function |
uniform |
symmetrical |
convex |
smooth with no distinct peak |
convex curve |
symmetrical |
convex |
smooth with central peak |
Gaussian |
symmetrical |
convex |
2 peaks |
bimodal |
symmetrical |
concave |
smooth with no distinct peak |
concave curve |
nonsymmetrical |
convex |
peak at left |
skewed left |
nonsymmetrical |
convex |
peak at right |
skewed right |
nonsymmetrical |
NA |
curved line with right end higher than left |
J (upward slope) curved line |
nonsymmetrical |
NA |
curved line with left end higher than right |
L or reverse-J (downward slope) curved line |