Three-Factor Cartographic Representation

Abstract

The conventionil shaded statistical map has one fundainental weakness. The shading of the geographic divisions emphasizes their surface areas rather than their statistical populations under study. Since there is no uniform relationship between the two, erroneous and distorted concepts of the relative importance of conditions prevailing in a division are often conveyed. When large divisions are in fact sparsely populated and sinall divisions densely populated, as is often the case, the visual imnpression is, of course, seriously misleading. Researchers sometimes eliminate this bias by supplanting the geographic divisions as the units for shading with soine form of area diagram. Circles, with areas proportionate to one of the variables under analysis, seeem generally suitable for this purpose. Their use not only removes the surface-area bias, but also permits the simultaneous consideration of two pertinent factors instead of one. For example, the distribution of radio ownership by states might be presented by showing in each state a circle proportionate in area to the number of dwellings and shaded to represent the percentage having radios. Such a two-factor representation is analytically superior to the two conventional shaded maps which it replaces. Spatial variation of many phenomena is influenced by more than one basic factor. In the distribution of radio ownership, for example, type of residents as well as number of dwellings has an inmportant effect. Even in the two-factor type described above, the question remains unanswered as to whether observed differences are actually geographic differences or are merely reflections of the varying proportion of rural and urban residents in the population. To solve such a three-variable problem, the analyst ordinarily resorts to a multiplication of conventional one-factor maps. The use instead of a series of two-factor maps, although more advantageous, also requires of the reader the difficult and