Keyword search

Hidden human populations, the Internet, and other networked structures conceptualized mathematically as graphs are inherently hard to sample by conventional means, and the most effective study designs usually involve procedures that select the sample by adaptively following links from one node to another. Sample data obtained in such studies are generally not representative at face value of the larger population of interest. However, a number of design and model based methods are now available for effective inference from such samples. The design based methods have the advantage that they do not depend on an assumed population model, but do depend for their validity on the design being implemented in a controlled and known way, which can be difficult or impossible in practice. The model based methods allow greater flexibly in the design, but depend on modeling of the population using stochastic graph models and also depend on the design being ignorable or of known form so that it can be included in the likelihood or Bayes equations. For both the design and the model based methods, the weak point often is the lack of control in how the initial sample is obtained, from which link-tracing commences. The designs described in this paper offer a third way, in which the sample selection probabilities become step by step less dependent on the initial sample selection. A Markov chain "random walk" model idealizes the natural design tendencies of a link-tracing selection sequence through a graph. This paper introduces uniform and targeted walk designs in which the random walk is nudged at each step to produce a design with the desired stationary probabilities. A sample is thus obtained that in important respects is representative at face value of the larger population of interest, or that requires only simple weighting factors to make it so.

This publication presents a series of tabulations produced from the General Social Survey on time use of Canadians. It includes information on average amounts of time spent on various activities by sex, by age, by selected role groups.

This document summarizes the results of content analyses of the 2004 Census Test. The first section briefly explains the context of the content analyses by describing the nature of the sample, its limitations and the strategies used to evaluate data quality. The second section provides an overview of the results for questions that have not changed since the 2001 Census by describing the similarities between 2001 and 2004 distributions and non-response rates. The third section evaluates data quality of new census questions or questions that have changed substantially: same-sex married couples, ethnic origins, levels of schooling, location where highest diploma was obtained, school attendance, permission to access income tax files, and permission to make personal data publicly available 92 years after the census. The last section summarizes the overall results for questions whose content was coded and evaluated as part of the 2004 test, namely industry, occupation and place of work variables.

This article presents an analysis of Census at School survey results that shows what students themselves had to say about their reading and associated daily habits. It was written by Statistics Canada analysts as an example to students of the type of detailed analysis that can be made using Canadian Census at School results. The article uses data that were collected from over 22,000 students across Canada during the 2004-2005 academic year. Census at School is an international classroom project that teaches students aged 8 to 18 about statistical enquiry and census-taking. Students anonymously fill in an online questionnaire about themselves - their height, time use, eating habits and much more - and then use their class results to learn statistical concepts, practice data analysis and explore social issues. Their responses also become part of national and international project databases that are used for teaching statistics.

The Census at School statistical literacy project is not an official Statistics Canada survey conducted under the Statistics Act. Schools; students participate on a voluntary basis and the data collected are not representative of Canada's student population. This is clearly stated with the summary Canadian results that are made available on the project website for the benefit of participating students.

Hot deck imputation is a procedure in which missing items are replaced with values from respondents. A model supporting such procedures is the model in which response probabilities are assumed equal within imputation cells. An efficient version of hot deck imputation is described for the cell response model and a computationally efficient variance estimator is given. An approximation to the fully efficient procedure in which a small number of values are imputed for each nonrespondent is described. Variance estimation procedures are illustrated in a Monte Carlo study.

The Ethnic Diversity Survey generated methodological challenges like choosing the sampling plan, developing the questionnaire, collecting the data, weighting the data and estimating the variance.

For a study on smoking cessation programs, respondents were found via referrals from key informants. A challenge was tracking the calling process and keeping records of information obtained during telephone calls.

The paradigm shift in disability limits participation of persons with disabilities in the survey process. Evidence of barriers to survey participation is reviewed and approaches to reducing barriers are discussed.

I will try to look at the future of survey research in the light of the incredible developments in information technology. I will speculate on what new technologies might contribute to doing surveys differently.

The testing of questionnaires with specialized populations such as Aboriginal people, homosexuals, bisexuals, children, victims of abuse brings challenges: identifying respondents, testing methodology, location of testing, respondent rapport and trust.

Series Description - The Nation Series (1996 Census of Population) is the first released series where basic data at a high level of geography are presented on variables collected by the 1996 Census.There are a total of 143 tables in the Nation Series which cover all census variables.The Complete Edition CDROM, Catalogue number 93F0020XCB96004 contains the cumulative set of all data tables from all Nation Series CDROMs.This comprehensive CDROM provides a full range of statistics on characteristics of the population which includes:Demographic information (100% data only for Age and Sex, Marital Status and Common-law Unions); Families (Number, Type and Structure); Structural Type of Dwelling and Household Size; Immigration and Citizenship; Languages; Aboriginal Origin, Ethnic Origin and Visible Minorities (Population Groups); Labour Market Activities and Household Activities (unpaid work); Place of Work and Mode of Transportation; Education; Mobility and Migration; Family, Dwellings and Household Information; as well as Individual and Family Income. Selected variables, such as occupation, are available to illustrate the analytical potential of the data based on cross-tabulations (i.e. sex by age and occupation).These data are national in coverage and provide information for Canada, provinces and territories and, in some tabulations, census metropolitan area levels. Some tables include comparisons with data from earlier censuses to provide an historical perspective.A variety of Nation Series data table extracts presenting social and economic characteristics of the Canadian population are available at the Statistics Canada Census Web site (www.statcan.gc.ca).

The National Population Health Survey (NPHS) uses the Labour Force Survey sampling frame to draw a sample of approximately 22,000 households. The sample is distributed over four quarterly collection periods. In each household, some limited information is collected from all household members and one person, aged 12 years and over, in each household is randomly selected for a more in-depth interview.

The questionnaire includes content related to health status, use of health services, determinants of health and a range of demographic and economic information. For example, the health status information includes self-perception of health, a health status index, chronic conditions, and activity restrictions. The use of health services is probed through visits to health care providers, both traditional and non-traditional, and the use of drugs and other medications. Health determinants include smoking, alcohol use, physical activity and in the first survey, emphasis has been placed on the collection of selected psycho-social factors that may influence health, such as stress, self-esteem and social support. The demographic and economic information includes age, sex, education, ethnicity, household income and labour force status.

This public use microdata file provides unaggregated data on the Aboriginal adult population - those who identify with their Aboriginal origin(s) and those who do not. For persons who identify, it contains almost 700 variables from the 1991 survey, such as, the group with which they identify, language proficiency, disability, chronic health conditions, schooling, work experience and the 1991 Census variables such as, income levels, marital status, fertility. The same census variables are provided for the population who does not identify.

There are many issues to consider when developing and conducting a survey. Length, complexity and timing of the survey are all factors that may influence potential respondents' likelihood to participate in a survey. One important issue that affects this decision is the extent to which a questionnaire appears to be an invasion of privacy. Information on income and finances is one type of information that many people are reluctant to share but that is important for policy and research purposes.

Collecting such information for the Survey of Consumer Finances (SCF) has proven difficult, and has resulted in higher than average non-response rate for a supplemental survey to the Labour Force Survey. Given the similarity between the SCF and an upcoming survey, the Survey of Labour and Income Dynamics (SLID), it is important to examine the reasons behind the SCF's higher non-response rate and obtain suggestions for increasing response rate and gaining commitment from respondents to the 6-year SLID.

Statistics Canada asked Price Waterhouse to conduct focus groups and in-depth interviews with respondents and non-respondents to the SCF. The objectives of these focus groups and in-depth interviews were to explore reasons for response and non-response, issues of privacy and confidentiality and understanding of the terms used in the survey, and to test reactions to the appearance of a draft SLID package.

A 24-hour dietary recall from 16,190 respondents aged 12 or older to the Canadian Community Health Survey - Nutrition was used to determine energy and nutrient intake. To identify plausible respondents, a confidence interval was applied to total energy expenditure derived from equations developed by the Institute of Medicine. Estimates of energy and nutrient intake for plausible respondents were compared with estimates for all respondents.

The 2006 Census hit the streets during the hottest labour market in a generation, with many regional unemployment rates at long-term lows and wage increases outstripping price hikes. While technological advances reduced manpower needs, tens of thousands of temporary workers were still required which proved to be a problem in many areas. Census managers extended the collection period and moved an unprecedented number of enumerators across regions to attain acceptable data quality while remaining within budget.

Based on a representative sample of the Canadian population, this article quantifies the bias resulting from the use of self-reported rather than directly measured height, weight and body mass index. The analysis is based on 4,567 respondents to the 2005 Canadian Community Health Survey, who provided self-reported values for height and weight and were then measured.

General population health surveys often include small samples of smokers. Few longitudinal studies specific to smoking have been carried out. We discuss development of the Ontario Tobacco Survey (OTS) which combines a rolling longitudinal, and repeated cross-sectional components. The OTS began in July 2005 using random selection and data-collection by telephones. Every 6 months, new samples of smokers and non-smokers provide data on smoking behaviours and attitudes. Smokers enter a panel study and are followed for changes in smoking influences and behaviour. The design is proving to be cost effective in meeting sample requirements for multiple research objectives.

Using survey and contact attempt history data collected with the 2005 National Health Interview Survey (NHIS), a multi-purpose health survey conducted by the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention (CDC), we set out to explore the impact of participant concerns/reluctance on data quality, as measured by rates of partially complete interviews and item nonresponse. Overall, results show that respondents from households where some type of concern or reluctance (e.g., "too busy," "not interested") was expressed produced higher rates of partially complete interviews and item nonresponse than respondents from households where concern/reluctance was not expressed. Differences by type of concern were also identified.

As SLID is a longitudinal survey, it is desirable to retain respondents in the sample for as long as possible. However, sample attrition and changes in the popu lation result in the sample becoming less representative of the population as time passes. To balance these factors, the sample for a longitudinal survey may be comprised of panels, with each panel being representative of the target population. Starting with a fixed sample size, the paper examines feasible options for the number of panels in the sample and the length of time which each panel remains in the survey. The rationale for the selected option is reviewed.

Labour force status will be an important analytical variable for many users of SLID data. The document discusses the issues involved in deriving this variable, and details the approach to be adopted.

Briefly, a value will be assigned for every one-week period, with three possibilities: employed, unemployed and not in the labour force. To a large extent, concepts used in the Canadian Labour Force Survey will be used. Since there are several situations where a straightforward approach to the classification is not possible, additional information will be available to data users who wish to adjust the definitions used.

This paper, as part of the design development process for the Survey of Labour and Income Dynamics (SLID), examines the implications for data quality of accepting proxy responses. While it is apparent that rules that minimize proxy responses add to data collection costs, a review of the existing evidence suggests the effects of proxy reporting on data quality are less clear. The general conclusions of this review are that proxy respondents tended to underestimate participation in government income support programs, be subject to higher item non-response rates and lower rates of personal and household interviews, offer more consistent responses to sensitive subject matter and have greater difficulty in reporting detail and events of short duration. As a result of shortcomings in the design of the research into this question, the evidence is not conclusive. While there is no firm basis for the rejection of proxy responses, the maintenance of some control over proxy responses would be a reasonable strategy.

This article summarizes the background, history and rationale for the Canadian Health Measures Survey, and provides an overview of the objectives, methods and analysis plans.

Starting in 1994, the Survey of Labour and Income Dynamics (SLID) will follow individuals and families for at least six years, tracking their labour market experiences, changes in income and family circumstances. An initial proposal for the content of SLID, entitled Content of the Survey of Labour and Income Dynamics : Discussion Paper, was distributed in February 1992.

That paper served as a background document for consultation wit h interested users. The content underwent significant change during this process. Based upon the revised content, a large-scale test of SLID will be conducted in February and May 1993.

This document outlines the current demographic and labour content, leading into the test.

When a survey respondent is asked to recall various events, it is known that the quality of the responses diminishes as the length of recall increases. On the other hand, increasing the frequency of data collection increases both the costs of collection and the burden on the respondents. The paper examines options which attempt to strike a reasonable balance between these factors. As it relates to this decision, the paper also describes how the sample has been designed to ensure that it remains representative of the target population, both for a given year and over time.

The conclusion is that, at this time, SLID should collect labour data in January to cover the previous calendar year and to collect income data in May, again to cover the previous calendar year.

This report describes the comments received as a result of the second round of the 2006 Census consultations. As with the previous 2006 Census consultation, this second round of consultations integrated discussions on the dissemination program, questionnaire content and census geography. However, the focus of this second round of consultations was placed on the 2001 Census of Population dissemination program and proposed directions for 2006 geography. Consultations were held from January to June 2004. Approximately 1,000 comments were captured through written submissions and the organization of over 40 meetings across Canada.

This report describes users' feedback on dissemination and geography issues received through this second round of consultations. In addition to user's comments, web metrics information serves as a valuable tool when evaluating the accessibility of public good data tables. Therefore, page view counts have been integrated in this report.

Some general planning assumptions that focus on the production and dissemination of 2006 Census products are also included in this report.

The National Longitudinal Survey of Children and Youth (NLSCY) is the first Canada-wide survey of children. Starting in 1994, it will gather information on a sample of children and their life experiences. It will follow these children over time. The survey will collect information on children and their families, education, health, development, behaviour, friends, activities, etc. This document describes the survey instruments of cycle 4.

This guide provides general information on the ethnic origin concept in the census and how the question of ethnicity has changed over time. In addition, the guide discusses the historical comparability of the ethnic origin data.

This guide will be of assistance in understanding the concepts, methodology and data quality of the surveys conducted as well as the data analysed by the Pensions and Wealth Surveys Section of the Income Statistics Division. It covers the following surveys/programs:- Pension Plans in Canada;- Trusteed Pension Funds (Census and Quarterly);- Survey of Financial Security; and- Pension adjustment/registered retirement savings plans data file provided by Canada Customs and Revenue Agency.

This paper discusses in detail issues dealing with the technical aspects of designing and conducting surveys. It is intended for an audience of survey methodologists.

In surveys with low response rates, non-response bias can be a major concern. While it is not always possible to measure the actual bias due to non-response, there are different approaches that help identify potential sources of non-response bias. In the National Center for Education Statistics (NCES), surveys with a response rate lower than 70% must conduct a non-response bias analysis. This paper discusses the different approaches to non-response bias analyses using examples from NCES.

The 2001 Census Handbook is a reference tool covering every aspect of the 2001 Census of Population and Census of Agriculture. It provides an overview of every phase of the census, from content determination to data dissemination. It traces the history of the census from the early days of New France to the present. It also contains information about the protection of confidential information in census questions and variables, along with information about data quality and the possible uses of census data. Also covered are census geography and the range of products and services available from the 2001 Census database.

This series includes six general reference products: Preview of Products and Services, Census Dictionary, Catalogue, Standard Products Stubsets, Census Handbook and Technical Reports.

Cycle 1.1 of the Canadian Community Health Survey (CCHS) will provide information for 136 health regions. A brief overview of the CCHS design, sampling strategy, interviewing procedures, data collection and processing is presented.

This paper describes the collection method and content of the 1999 Survey of Labour and Income Dynamics (SLID) income interview.