Exploring the Link between Crime and Socio-Economic Status in Ottawa and Saskatoon: A Small-Area Geographical Analysis

6. Study # 1 Findings: Dissemination Areas of Ottawa

6. Study # 1 Findings: Dissemination Areas of Ottawa

6.1 Descriptive Statistics

Table 6.1 shows the descriptive statistics for the 6 crime and 26 census variables used in the analysis. The table indicates that all 6 crime variables have high coefficients of variation (the standard deviation divided by the mean) indicating significant dispersion of individual values around the mean. In particular, minor property and drug offences have values that fluctuate greatly among the 1187 DAs examined in Ottawa, pointing to substantial geographic disparity within Ottawa. Similarly, several of the census variables also have high coefficients of variation including those denoting recent immigrants, people living in low income, youth unemployment and apartment high-rises suggesting a significant geographic disparity of disadvantaged residents in the city.

6.2 Principal Components Analysis

As Table 6.2 shows, the analysis produced an 8-component solution, accounting for 78.5% of the total variance in the dataset. An examination of the component loadings in Table 6.3 reveals that the 5 crime variables (violent, major and minor property, drugs and disturbance/other) are highly inter-correlated with one-another (Component 2) but are not significantly associated with any of the 26 socio-economic variables. The other components identify a number of dimensions of socio-economic status in Ottawa, including "mobility and housing" (Component 1), "income and education" (Component 3), "immigrants and visible minorities" (Component 4) and "youth and unemployment" (Component 7). As can be seen in Table 6.3, none of these dimensions (by way of component loadings) is significantly related to any of the 5 crime variables, suggesting that, overall, there is a weak association between crime and socio-economic status in Ottawa at the intra-urban scale – at least with the use of data at the DA level. These findings essentially confirm the output from the correlation matrix (31 variables x 31 variables) which displayed relatively low correlation coefficients between the crime and socio-economic variables, in the range of r = -0.35 to r = 0.35.

6.3 Multiple Regression

Table 6.4 shows the results of the regression analysis between each of the 6 crime variables and 6 selected census variables characterizing disadvantage. An effort was made to select independent variables that were not highly correlated with one another but, nevertheless, demonstrate a range of factors associated with disadvantaged communities (youth, recent immigrants, low-income, mobility, apartment high-rises and low education). The table demonstrates that in Ottawa there appears to be a rather weak statistical association between crime and factors related to socio-economic disadvantage with low multiple correlation coefficients{R} and coefficients of multiple determination {R2}. In fact, each crime variable has an R2 below 0.11, clearly demonstrating that the 6 socio-economic variables, when taken together, are not strong predictors of increased levels of crime. In other words, no more than 11% of the variation in any of the crime indicators can be explained by the socio-economic variables at the level of the DA.

Despite the overall weak relationship, however, several of the independent variables, when examined individually, had significant beta coefficients (transformed partial regression coefficients) at the 95% confidence level (p<0.05). The recent immigrant variable recorded significant negative betas on 5 of the 6 crime variables pointing to an inverse relationship between the two indicators - the higher the level of crime in an area the fewer recent immigrants living there. Table 6.4 also indicates that people living in low-income (LOW_INC) and residential mobility (MOVERS_1_yr) were the best predictors of crime. In fact, mobility was the only socio-economic indicator to record significant betas on all of the crime variables. Interestingly, the presence of young people (TOT_YOUTH) was not a significant predictor of crime.

The 1187 DAs in Ottawa were ranked from highest to lowest according to their Z-scores on the variable "Total Offences" with the top quintile (20%) of DAs being identified as "High Crime Areas" (n=237). A second multiple regression was performed on the data for these areas. Table 6.5 shows the results and demonstrates that in Ottawa's High Crime Areas (HCAs) there appears to be a weak statistical association between crime and factors related to socio-economic disadvantage with low values of R and R2 for each of the six crime variables. Furthermore, only one independent variable, mobility (MOVERS_1_yr), had significant betas on the crime indicators related to violence and drugs. A third multiple regression analysis was performed on the data set. This time, the 1187 DAs were ranked from highest to lowest according to their Z-scores on the variable "Low-Income" with the top quintile (20%) of DAs being identified as "Disadvantaged Areas" (n= 237). Table 6.6 shows the results of the analysis and again reveals a weak overall association in these areas between crime and factors related to disadvantage.

6.4 Cartographic and GIS Analysis: Examining Spatial Patterns of Crime and Disadvantage

High Crime Areas (HCAs)

Figure 6.1 is a map illustrating the location of 'High-Crime Areas' (HCAs) with respect to total offences in 2001. The map inset clearly illustrates that these areas are concentrated within the built-up central core and suburbs of Ottawa, with very few HCAs visible in the outer and rural parts of the city. (Two HCAs are present in the southwest corner of Cumberland and another in the southeast portion of Goulbourn). However, the enlarged section of the map shows a dispersed pattern of HCAs at the three levels (elevated, high and highest) within the urban core, including sections of the inner city and a prominent visibility of large DAs (in terms of area) in suburban locations. Compared to central Ottawa, these suburban locations have lower residential densities and larger spaces devoted to commercial activity and industrial parks.

There are several clusters of DAs at the "highest" crime level (above 1 SD) in downtown Ottawa (the Central Business District and 'Market' area), the east-central part of the city (Vanier, Overbrook and Ottawa North-East) and several communities west of downtown including Carlington. An examination of the raw data reveals that the higher crime rates in these areas are attributable to large numbers of incidents relating to minor property offences (particularly 'theft under $5000' and 'theft from vehicle') and to a lesser extent to major property offences (most notably 'residential break and enter' and 'auto-theft'). In addition, the downtown HCAs (including the Market) have a much higher than average number of violent offences, particularly 'threats' and 'assault'. The map also highlights what appears to be 'corridors' of HCAs adjacent to major transportation routes such as Highways 417 and 17 (east-west) and Highway 16 (north-south) indicating a spatial relationship between crime and mobility/accessibility. These 'corridors' contain some of the city's highest residential densities and significant commercial activity, including several of the city's largest shopping centres such as Place d'Orleans, St. Laurent, Pinecrest and Bayshore.

Figures 6.2 to 6.5 consist of a series of maps displaying the geographic distribution of HCAs according to four offence types: "Violent", "Major Property", "Minor Property", and "Drugs". "Violent" HCAs (Figure 6.2) are somewhat more concentrated within the central core of Ottawa (with the exception of several DAs in the rural portion of Cumberland) and are particularly noticeable in the inner city (Center Town and the Market) as well as in Vanier/Overbrook and parts of Ottawa North-East. While the large majority of "Major Property" HCAs are located within the central core of the city, Figure 6.3 reveals a more dispersed pattern with several HCAs also visible in suburban and rural communities, where rates of residential and commercial break and enters tend to be higher. Figure 6.4 shows that "Minor Property" HCAs have a more compact distribution, particularly in the inner city and along transportation corridors where higher densities and the concentration of commercial activity likely present greater opportunities for offences such as theft. Figure 6.5 illustrates that HCAs associated with drug offences are the most dispersed geographically in Ottawa with a number of these areas visible in the rural parts of the city. Drug crimes, however, accounted for only 2% of all offences in Ottawa in 2001 (Table 4.2).

Disadvantaged Areas

Figure 6.6 consists of a map showing the geographic distribution of disadvantaged DAs in Ottawa according to their Z-scores on the variable 'Low-Income'. In light of the serious socio-economic problems associated with people living in low-income, including higher unemployment, lower rates of labour force participation and educational attainment and greater dependency on social assistance, it was felt that this variable would be the most appropriate composite measure of disadvantage. Similar to the crime classification, the Z-scores for low-income were ranked from highest to lowest with the top quintile (20%) of DAs being identified as "disadvantaged" (n=237). The map clearly displays a very tight spatial concentration of disadvantage at all levels ('elevated', 'high' and 'severe') within Ottawa's central core, particularly the inner city neighbourhoods of Dalhousie, Centre Town, Sandy Hill and Lower Town as well as a large cluster (including DAs at a 'severe' level of disadvantage) in the east- central portion of the city, comprising the communities of Vanier, Overbrook and Ottawa North-East.

Other pockets of disadvantaged DAs are present in the south central part of the city including several neighbourhoods in Riverview, Alta-Vista and Hunt Club and in west central Ottawa and Carlington. Further west, similar conditions are present in several DAs in Pinecrest/Queensway, Nepean North and Bells Corners. It is also evident that across the city, individual areas of 'severe' disadvantage are in most cases bordered by areas of 'elevated' and 'high' disadvantage. All of these areas are characterized by significantly lower family and household incomes, higher rates of unemployment, lower levels of educational attainment and higher proportions of recent immigrants, visible minorities, lone-parent families and people who are single.

The Intersection of High Crime and Disadvantaged Areas

The "intersection" operation in ArcGIS was used to create a series of maps showing the location of DAs in Ottawa that are both disadvantaged and have a high crime rate. Figure 6.7 is a map displaying the intersection of HCAs (by total offences) and disadvantaged areas in the city. In total, 98 of the 1187 DAs (8% of the total and 41% of HCAs) were found be to experiencing both conditions. Geographically, these "Hot-Spots" are clustered tightly in Ottawa's inner core (Dalhousie, Centre Town, Sandy Hill, and Lower Town), as well as large sections of Vanier, Overbrook and Ottawa North-East. The map also reveals several small, isolated "Hot-Spots" within suburban communities surrounding the core including Riverview, Alta-Vista, Hunt Club, Pinecrest/Queensway and Nepean North. Figure 6.8 illustrates the intersection of HCAs (by violent offences) and disadvantaged areas and reveals a very similar spatial distribution. In this case, 103 DAs (9% of the total and 43% of HCAs) were found to be both violent and disadvantaged.

Table 6.7 provides a summary of criminal offence and socio-economic conditions in the four main spatial groupings presented thus far in the report:

  1. High Crime Areas (n=237)
  2. Disadvantaged Areas (n=237)
  3. "Hot-Spots" A – Intersection of Total Offence HCAs/Disadvantage (n=98)
  4. "Hot-Spots" B – Intersection of Violent Offence HCAs/Disadvantage (n=103).

It lists the mean Z-scores on the 6 crime variables and 23 selected census indicators for the DAs in the four groups. Scores higher than 0 indicate conditions above the citywide average while scores less than 0 signify conditions below the citywide average. It is important to note that the figures in this table do not necessarily imply cause and effect between crime and socio-economic status but rather present a general picture of conditions in these areas.

The table reveals that HCAs, on average, have higher proportions of people living in low income (0.671) as well as singles (0.831) and have more rented dwellings (0.673) and apartment low rises (0.599). In addition, they have larger proportions of residents who have not finished high-school (0.589). "Disadvantaged Areas", on the other hand, have only slightly higher than average rates of total criminal offences (0.279) but moderately higher rates of violent crimes (0.495). Table 6.7 also shows that the "Hot-Spots" (A & B) are characterized by significantly higher proportions of recent immigrants (0.841, 0.822), visible minorities (0.978, 1.039), residents who had moved during the past year (0.850, 0.749), low rise apartments (0.938, 0.907) and substantially higher rates of residents without a high-school diploma (1.042, 1.174). With respect to the incidence of crime, the "Hot-Spots" had markedly higher rates of violent offences (1.284, 1.310).

6.5 Discussion

Ottawa is a relatively safe city with a low crime rate. Overall, it has an affluent population and a strong economy; but serious social problems persist in a number of disadvantaged communities. In 2001, minor property crimes were the most prevalent offences, accounting for 54% of the total. Two minor property-related crimes - 'theft under $5000' and 'mischief' together accounted for 40% of all offences. Violent crimes accounted for 16% of total offences in 2001.

The study found that there is a weak statistical association in Ottawa between crime and socio-economic disadvantage. The results of the analysis (principal components analysis and multiple regression) revealed that, overall, there are no clear social 'predictors' of crime in the city at the level of the dissemination area (DA). For example, when examining the city as a whole, DAs with higher proportions of youth, unemployed people, recent immigrants, visible minorities, renters or high-school dropouts are not more likely to be areas with higher crime rates.

The mapping of crime variables was effective in discerning geographic patterns of criminal activity in the city. In 2001, "High Crime Areas" (HCAs) were largely contained to the built-up urban core of Ottawa (including the suburbs) with very few HCAs evident in outlying and rural areas. While this was also the case with 'violent' and 'minor property' offences, the maps showed that 'major property' and 'drug' offences had a more dispersed pattern with several HCAs located in suburban and rural communities. The GIS analysis found that there was a moderate geographic relationship between crime and socio-economic status in the city, with 40% of disadvantaged DAs also being HCAs. Places where these two conditions intersected were labeled as "Hot-Spots" and represented just 8% of all DAs in the city. These "Hot-Spots" were found primarily in inner city communities but several were also visible in suburban neighbourhoods with low-income subsidized housing projects.

While the relationship between crime and socio-economic status was found to be tenuous at the citywide level, a number of characteristics did emerge when specific areas were examined more closely. HCAs, for example, exhibited certain conditions consistent with the ecological approach to criminology and social disorganization theory, most notably above average levels of low-income and transient residents. In addition, the "Hot-Spots" were found to have higher rates of violent crime and significantly larger proportions of recent immigrants, visible minorities and residents living in apartment buildings.

The concept of criminal opportunity is clearly applicable to the situation in Ottawa as the majority (60%) of HCAs in the city are not socially disadvantaged. Suitable targets for crime are found in areas where commercial, institutional and recreational activities are located, such as shopping centres, offices, transit-way stations, warehouses and recreational spaces. In addition, unguarded homes in suburban communities are targeted for their valuable and easily transportable goods. Furthermore, routine activities theory helps to explain the high rates of violent crime evident in areas with a concentration of bars and restaurants such as in Ottawa's Market district.

Table 6.1 - Descriptive Statistics (n=1187)

Table 6.2 - Explanatory Power of the Principal Components
Component Eigenvalue % Total Variance Cumulative %
1 9.8 31.6 31.6
2 3.8 12.3 43.9
3 2.7 8.8 52.7
4 2.3 7.6 60.2
5 2.0 6.3 66.5
6 1.5 4.7 71.3
7 1.2 3.7 75.0
8 1.1 3.5 78.5