Unemployment: The unemployment factor is calculated by calculating unemployed residents as a percentage of economically active residents. This factor is assigned a weight of 3.34

Children under 5 years old: This factor is calculated as the percent of residents who are younger then 5. This factor is assigned a weight of 4.64.

Residential mobility: This factor is calculated as the percent of the residents who changed their home address in the previous year. This factor is assigned a weight of 2.68.

Elderly living alone: This factor is calculated as the percent of residents over 65 years old who are living alone out of the total number residents over 65 years old. This factor is assigned a weight of 6.62.

Lone parents: The lone parent factor is calculated as the percent of residents living in lone parent households. This factor is weighted by 3.01.

It was necessary to modify the original index in order to include census data from the GVRD. This was a sensitive procedure, as the weights associated with the various factors must be adjusted accordingly when the structure of its input factors are altered. These changes have often been performed in a somewhat arbitrary and ambiguous fashion as many census databases vary from the original factor scores used by the index (Morris, 1991).  A typical example refers to the original classifications based upon “per household” data and the Canadian census data counted on a “per person” basis. Additionally, this increases the effect of the original weighting criteria. As a general rule, this analysis reduced the weights that were originally assigned for “per household” factors in half.

In addition, we were unable to adapt the following factors directly as they are related to per-household counts rather than per-person counts:

·      Overcrowding.

·      Ethnicity

·      Low Social Class

 Although it is possible to calculate average households, and henceforth per-average household’s values, this is an unappealing choice to employ as the values produced would be highly statistical in nature and misrepresentative of the actual distribution of the areal data.

After the changes and adaptations outlined above were carried out the following three new factors were established:

Overcrowding: The percentage of residents living in overcrowded conditions; were overcrowded conditions are defined as more then one person per room. This factor is assigned a weight of 1.44

Ethnicity: The percentage of residents that are born outside of Canada. This factor is assigned a weight of 1.25.

Low Social Class: This factor is calculated as the percentage of the workforce employed in unskilled labor. This factor is weighted by 1.87. The original Jarman index is defining low social class as people employed in the labor category IV (Human Resources Canada). According to Human Resources Canada, low-skilled workers are defined as:

·      Cashiers;

·      Occupations in food and beverage service;

·      Occupations in protective services;

·      Occupations in travel and accommodation, including attendants in recreation and sport;

·      Trades helpers, construction and transportation laborers and related occupations;

·      Primary production laborers;

·      Laborers in processing, manufacturing and utilities;

 The defense for this classification was the assumption that the above skills could be taught on the job and without additional schooling beyond secondary. The data fields were summarized and applied towards the “Low Social Class” factor.

 In addition to these changes, the weight for elderly living alone was also divided by two. The original weight assigned in the Jarman index is 6.62. Because of the high weighting and the amount of skewness that was being directed toward areas with relatively large amounts of elderly (thus entering into the ‘house rich, but income poor’ dilemma) it was recommended to cut the weight in half so as not to overshadow the other factors. The weight applied for elderly living alone in the final analysis was 3.32