Other writers have also focused on the meaning of the ranking task used to

Other writers have also focused on the meaning of the

This preview shows page 5 - 6 out of 10 pages.

are nearly equal. Other writers have also focused on the meaning of the ranking task used to generate preference data. “One of the most common criticisms pertains to the expectation...to rank geographic entities in order of preference” (Thill and Sui, 1993, p. 264). Tuan’s (1974) opinion was that the ranking task is a game that is unusual or unrealistic to subjects, though he believed “the maps...are an appealing way to display data” (p. 591). Gould and White themselves noted, in the preface to their second edition, that “criticism [of the first edition] focused on the viability of the ranking of spatial preferences and the representativeness of the sample of respon- dents” (1986, p. ix). The bulk of these criticisms are based on intuition rather than evi- dence, and we find them a bit overstated. Instead, we find it surprising that so little criticism has focused on the cartographic display of the preference rankings, including the way interpolation of data at the state level has been used to construct continuous isoline maps of preference. Our primary concern with the isoline representation of ranked prefer- ence data is that the “landscape” of underlying preference is not suitably continuous, particularly given the coarse resolution at which such data are collected. Such isoline maps communicate invalid impressions of continuous variation and intraregional variation to viewers. As is well known, interpolating a continuous surface from discretely sampled data is valid insofar as the underlying variable being mapped varies continu- ously across space. Put another way, the underlying variable must ex- hibit considerable autocorrelation, typically positive, across the surface: “The critical assumption is that events in one geographical area influ- ence those in adjacent areas” (Tobler, 1979, p. 526). Existing literature on cartographic interpolation focuses primarily on questions of the rela- tive validity of various sampling schemes and interpolation techniques in constructing isoline maps (for example, MacEachren and Davidson, 1987). As Lam (1983) put it in her overview of interpolation methods: “The fundamental problem underlying all these interpolation models is that each is a sort of hypothesis about the surface, and that hypothesis may or may not be true” (p. 130). Of course it is at least as important to consider when continuous mapping techniques are valid in the first place as it is to choose the right interpolation technique. Our main argument is that the hypothesis of an underlying continuum is sufficiently untrue in the case of regional prefer- ences as to make isoline interpolation a poor cartographic choice for their display. There has been research on the variables underlying regional preference, much of it discussed by Gould and White in their second edition. Some of these variables (climate, topography) do vary more or less continuously. We concede that states discriminated primarily on these variables are validly mapped with isolines, although one could still criticize the sparseness of sampling only once per state when the isolines are based on state rankings (cf. MacEachren and Davidson, 1987). Other
Image of page 5
Image of page 6

You've reached the end of your free preview.

Want to read all 10 pages?

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

Stuck? We have tutors online 24/7 who can help you get unstuck.
A+ icon
Ask Expert Tutors You can ask You can ask You can ask (will expire )
Answers in as fast as 15 minutes