Crime rates--calculated by dividing a measure of crime occurrence by units of risk for a specific time period--have long been established as standard metrics in criminological research and have been recognized for their utility in many different contexts. Sometimes referred to as the 'crude crime rate' when residential population is used in the denominator, this metric has been said to provide a general depiction of risk for criminal victimization (Silverman, Teevan, and Sacco 2000). In practical terms, the crime rate offers a way to measure changes to the risk of victimization over time, compare the risk of victimization among jurisdictions, and correlate independent or explanatory variables for empirical analyses (Sparks 1980). Further, crime rates have been noted for being "particularly useful in planning prevention campaigns and in assessing the impact of changing social conditions on the risk of crime" (Brantingham and Brantingham 1998: 266). The utility of crime rates, however, depends largely on the accuracy of data that are used to operationalize them.
Both the numerator and the denominator may be criticized for inaccuracies related to how data are compiled. With respect to the numerator, the dark figure of crime refers to criminal events that go unrecorded. Sources of error for this variable may include the exercise of discretion on the part of victims, witnesses, complainants, and police with respect to how criminal events are defined and reported. With respect to the denominator, the dark figure of population refers to persons who go unrecorded. Sources of error for this variable differ by the type of data but may include the inability to count persons who do not maintain a permanent residence and the failure to accurately track population turnover (i.e., residents who move into or out of a jurisdiction).
Given that federal censuses provide the most comprehensive data on population counts at a variety of levels of aggregation, these measures are easily accessible and, therefore, are often adopted as default metrics for units of risk in crime rate calculations. There are, however, specific concerns regarding the accuracy of federal census population data. As Boydell (1969) pointed out, census population figures are only indicative of the sleeping population in a geographic area. As a result, they do not report on populations that may engage in other activities (e.g., work or leisure) in that same space. Although in Canada, Statistics Canada has taken steps to account for this issue by, for example, calculating populations for census metropolitan areas (CMAs) and census agglomerations (CAs) (2) in place of municipal jurisdictions, there are still some jurisdictions that contain large errors in reported populations.
One of the sources of such error is the inability of federal censuses to accurately capture shadow populations. The term shadow population may refer to different groups, depending on the context of the discussion. In much of the criminological literature to date, the term has been used to identify groups of illegal immigrants or illegal migrants who go unrecorded in population counts (Bauder 2013; Mandel 1978; Schlegel 2011; Zatz and Smith 2012). In Canada, however, the term is most often used to refer to temporary worker populations. Alberta Municipal Affairs (2013), for example, defines the shadow population in the following way:
The shadow population refers to temporary residents of a municipality who are employed by an industrial or commercial establishment in the municipality for a minimum of 30 days within a municipal census year. These persons reside in the municipality for a given period of time, but do not consider this to be their usual residence. Note that post-secondary students are not considered part of the shadow population. (Alberta Municipal Affairs 2013: 3)
Because this group is identified as "transient workers" by Statistics Canada, it is not counted in the federal census. It is important to recognize, however, that in many jurisdictions the shadow population is not characterized as a group that holds a transient status. Instead, these shadow populations are a constant...