Dealing with Non-Areal Proportions

Measure data include Areal Proportions, which are composed of two count values, one of which is an area value, and Non-Areal Proportions which are composed of two count values, neither of which are area values.

Areal Proportions such as the number of cows per hectare, can be adequately interpolated, since the Volume Preserving interpolation method incorporates the source area in the calculation, and this has the effect of adjusting the proportion as if it were a count value.

In contrast, Non-Areal Proportions such as the number of cows per farm, cannot be adequately interpolated without additional data, even when using area preserving interpolation techniques. Without having additional count data for the denominator of the proportion (i.e. the number of farms in cows/farm) it is not possible to correctly convert the measure value into a count value before interpolation.

Procedure

• If both the numerator and denominator are also stored within the source database as count values, they should be interpolated separately as count values and subsequently converted to measure values within the target polygons.
• If the denominator is also stored as a count value in the database, it can be used to calculate a count value for the numerator.
• If only the measure type proportion is available, there are two approaches that can be used:
  1. treat the values as if they are areal proportions, and accept the increased potential for error, or
  2. treat the measure proportions as if they were nominal data.

Source: modified from Ballard and Schut, 1995 by Peter Schut
Contact: Head, CanSIS