A researcher is using PCCF+6A1 to assign health regions to the postal codes listed in the CCHS 1.1. She's noticed that for several postal codes, not all cases are assigned to the same health region. Most cases within the postal code are assigned to one health region, with a minority being assigned to another. Some examples are B0E 1J0 and K0K 2C0. They would like clarification as to whether this is normal behaviour.
The description of the PCCF+ mentions that it “uses population-weighted random allocation for Postal CodesOM that link to more than one geographic area.” Would these cases represent situations where postal codes and health region boundaries do not neatly line up?
Answer
The postal codes in question are more rural in nature and tend to cover a large area. Health regions do not necessarily follow postal code boundaries, however, so there are instances where a postal code could include portions of two or more regions. In cases like that, the PCCF+, assigns the records probabilistically with a greater chance of it getting assigned to a region with a greater population (so the "population-weighted random allocation for Postal Codes that link to more than one geographic area" that the client had already identified). In any case, yes, I expect these cases do represent situations where postal codes and health region boundaries do not line up.
