ABSTRACT
Some data analysis difficulties can be traced to curves themselves and oth ers to the models chosen to represent the curves. Although what is meant by an incorrect model is easy to explain, problems related to curve choice per se, and not model choice, are more difficult. For instance the distinc tion is made below between continuous and discrete variates in terms of what are referred to as types of curves. The difficulties that result from this distinction are harder to study than difficulties that are due to the incorrect choice of a model, like that of the Normal curve.
