Greetings All. In Bruce Phillips' annual report to the Government he had some comments regarding the upcoming Census of 2001. For those interested in his full report it can be found on the internet at http://www.privcom.gc.ca/english/02_04_08_e.htm Gordon --------------------------------------- Counting Canadians--Keeping Promises, Building Trust 2001 Census--enhancing transparency in the census collection process On May 15, 2001, Statistics Canada will ask some 31 million people in about 12.8 million households to complete and return their census questionnaires. Collecting the data will require 40,000 field staff, working from five regional offices. The total projected cost for the 2001 Census is $400 million. The census is the federal government's largest collection of personal information and arguably its most detailed for the 20 per cent of Canadians who receive the long form. Naturally, Statistics Canada's conduct of the census interests the Privacy Commissioner. As in previous censuses, 80 per cent of Canadian households receive the short questionnaire. The short form normally contains basic demographic questions, such as date of birth, sex, marital and common-law status and the relationship of persons living in the household. It could also include a question on the language first learned at home. The remaining 20 per cent of Canadian households receive the long form. In 1996, in addition to the basic demographic data, the long form asked 47 additional questions on physical limitations, language knowledge, education, work and household activities, immigration, ethnicity and aboriginal status, housing, shelter costs and income. Some respondents consider many of these questions very intrusive, sensitive or even offensive. Although Statistics Canada provides Canadians a good deal of information about the process of completing and returning their census forms, it does not adequately inform Canadians that local census representatives in each community examine their completed questionnaires before sending them to Statistics Canada in Ottawa. Thus, respondents are not warned that someone they know could examine their completed form. Of all the privacy complaints the Office received following the 1991 and 1996 censuses, those that generated the strongest negative reactions concerned respondents and census takers knowing each other. In most cases, complainants were both shocked and angry to learn that neighbours serving as census representatives reviewed their completed questionnaires; they assumed an unknown bureaucrat in Ottawa reviewed the information. In summary, complainants felt the process had betrayed the promise of confidentiality and were outraged that friends, neighbours and others whom they know could have access to such financial information as the family members' income, mortgage payments, retirement savings and utility bills. The great majority of complainants drew little comfort from Statistics Canada swearing census workers to secrecy or the possibility of fines and/or jail terms if they revealed personal information. Neither did much to remedy the resulting embarrassment and invasion of their privacy. Having their completed census forms reviewed by some unknown civil servant in Ottawa mitigated the intrusion to some extent. The Privacy Commissioner observed that allowing collection by neighbours who know the respondents demonstrates a complete lack of understanding of what privacy means. To resolve the problems, Statistics Canada advised the Commissioner that it was developing a centralized edit methodology to replace the current system. Rather than returning completed questionnaires from an enumeration area to the local enumerator for editing and follow-up, all census questionnaires would be sent to district offices. Local enumerators would deal only with households that had not returned the form or with problems that district office staff could not resolve by telephone. In those cases, Statistics Canada could ensure that the assigned field enumerators were not local. The "Centralized Edit System" was tested during the 1996 Census and again during the October 1998 "National Census Test" to prepare for the 2001 Census. Unfortunately, Statistics Canada found the tests did not yield the anticipated results. Centralized editing led to increased risks that respondents would not complete the forms, complete them only partially, and not return them. This caused more contacts with respondents than needed in the traditional method, increasing the risk of friction between census staff and respondents. This problem, combined with Statistics Canada's incomplete and sometimes inaccurate household address file, convinced the agency not to use the centralized edit methodology for the 2001 Census. Statistics Canada continues reviewing options for the 2001 census, such as computer-assisted telephone interviews in two regional offices, and collection over the Internet. Apparently the agency will test the Internet option on two Web sites during the next census. Respondents will be assigned a Personal Information Number (PIN) and their response data will be encrypted, thus eliminating the need to mail back census questionnaires to local census enumerators. Statistics Canada is also considering cutting the number of times census staff have to go back to respondents. In fact, the agency would like to reduce the rejection rate for the long questionnaires from 55 per cent to 35 per cent, significantly reducing the number of contacts with households and thus the friction between respondents and census staff. Two tests will be conducted in the 2001 census and the results compared; one using a sample of 125,000 long questionnaires (approximately 5 per cent) for which there will be no edit and no follow-up, and the other using a sample of 325,000 long questionnaires (approximately 14 per cent) involving only telephone follow-up. Statistics Canada will also assign census takers in urban areas to neighborhoods outside their own, thus reducing the risks of their collecting information about someone they know. This requirement will be considered when the agency hires staff for the next census. However, in rural areas and small towns it is not possible to guarantee that respondents and census representatives will not know one another; the pool of available staff is not large enough to avoid the situation. In addition, Statistics Canada finds that the only way to ensure all households are enumerated in rural areas is by assigning someone thoroughly familiar with the area. However, Statistics Canada will attempt to alleviate the problem with several steps. It will print on the back of both the questionnaire package and the return envelope an advisory that a "Statistics Canada representative responsible for your area" will review the questionnaires. Respondents who object to providing their completed census form to their local enumerator will be told by the enumerator or the Census Help Line that they can have a census commissioner collect the information or they can mail their completed form to the regional office. The agency will also provide census staff additional training and procedures to emphasize the importance of protecting collected information and heighten census enumerators' awareness of privacy concerns. Although these steps might address some aspects of the problem, they do not resolve it. The Commissioner is concerned that the process lacks transparency. For example, the proposed message on envelopes advising that an agency official will review their questionnaire, does not alert respondents to the possibility that it might be someone they know--a neighbour or friend. Since Statistics Canada recognizes that it is not uncommon for residents in a collection area to know the enumerator (and particularly true in rural areas), it must clearly inform respondents about the probability and offer them options for returning the questionnaire. And these measures should apply to both the short and the long questionnaires because the short form will ask a question on sexual orientation (same-sex partner). The Office suggested wording to include in both the census guide and on the front of both census forms that would meet the transparency requirement: Although Statistics Canada is taking measures to avoid having census enumerators work in areas close to where they live and/or to ensure that enumerators do not know any of the respondents in their collection area, residents in a collection area might know their local enumerator. If you are personally acquainted with the local enumerator and feel uncomfortable giving information to this person, please call our 1(800) Census Help Line to find out about the alternate arrangements for returning your completed questionnaire without having the local enumerator see it. The Office also believes that part of the problem could be avoided by clearly instructing census representatives to actively offer alternate arrangements when they know the householder. It is best to offer this option at the outset rather than waiting to have the respondent object. Census representatives must also be instructed to turn over the completed questionnaires of anyone they know to the area census commissioner.