Ambiguities and Irregularities in Census Data
Both during the creation of the census data and the subsequent use of these data, several possibilities exist for ambiguities and irregularities. Ambiguities and irregularities can be generated by the respondent, enumerator, commissioner and researcher. Some examples include the following: enumerators could have systematically undercounted a portion of the resident population; enumerators could have misspelled or phonetically spelled surnames (which is more problematic if in a later census or other documents, the same person’s surname is spelled differently).
Despite the details captured by the hundreds of nominal entries made in the course of the historical censuses, the information on the census schedules should not always be taken at face value. Unlike recent censuses, the 1901 Census was not a self-administered survey that most people could read, complete and return by mail. Few archival records relating to the early censuses have survived and published sources do not provide many details on how the schedules were completed, especially in the unorganized districts. Whether many of the families enumerated actively provided the information requested by the census schedule, or whether well-informed community residents, along with priests and fur traders, provided the information, some anomalies and irregularities can be expected. Clearly the respondents, or whoever provided the information on the individual, could have unknowingly provided inaccurate information.
Another source of possible irregularities could derive from using various types of enumerators. North West Mounted Police were used to complete enumerations in some cases. Also, individuals residing in institutions such as prisons, asylums, schools, convents, etc. were often enumerated by the officials within the institutions. This sometimes resulted in the collected data appearing visually different from the data collected by other sources of enumeration.
Perhaps the most common irregularities in the nominal lists concerned the ages provided for enumerated individuals. The relative completeness of the schedules for this particular category indicates that someone put in considerable effort. It would appear, however, in the case of ages, respondents or enumerators often provided estimates. As such, it is not un-heard of for individuals to age more or less than ten years between the censuses. In these cases, the age or date of birth estimates in at least one of the censuses would seem to be inaccurate. Scrip applications and church records, however, can sometimes be used to cross-reference the ages of Métis residents.
A common ambiguity in the census return schedules stems from the inconsistent spelling of surnames. A variety of Aboriginal languages, French and English were used throughout the Métis homeland. Literacy was not universal in Canada at this time, especially in the Unorganized Territories, and consequently consistent spelling of surnames cannot be expected. Oftentimes, enumerators would record names as they heard them or knew them. For example, “Marie” was sometimes recorded as “Mary,” while another individual was recorded as “Iron” in one enumeration and as “Pewapiskoussa” in another. Thus, enumerators may have sometimes correctly known the surnames of individuals, but their renditions on the schedules were at times ambiguous and can make it difficult for readers unfamiliar with census procedures to track individuals across the census years.
Ambiguities and irregularities can be introduced from a number of different sources, including the respondents, the enumerators and the census officials. For example, the respondents could have introduced mistakes if they did not understand the question, did not have the information, or were reluctant to provide the information. The enumerators would have had to either calculate the information, make an educated guess, or leave that particular category blank on the census schedule. Similarly, enumerators could have introduced mistakes if they did not understand the question, did not have the information, did not follow the instructions, completed the schedules inconsistently, chose not to enumerate inaccessible families, or simply recorded information in the wrong column or inaccurately. Other census officials who tabulated and compiled these schedules sometimes altered the original schedule. These officials checked, verified and calculated rough tallies on the enumerators’ original schedules. At times, these officials may have even changed some of the originally recorded information. Such subsequent modifications of the nominal lists can be a source of ambiguity and can also make the data even harder to read on the microfilm.
The 1901 Census nominal returns are important and detailed sources of information for historical communities. However, users of this information must be aware of the potential for irregular or ambiguous data. Steps can be taken to identify and reduce the potential for these problems: it may be helpful to examine nominal returns from several censuses and not just one in isolation; other sources of information, such as church records, Hudson’s Bay Company records, and scrip applications, can be used to check and verify information recorded on the nominal returns. With careful examination of the data, the nominal returns provide a rich data set that reveals important—but often hidden—details about historical communities.