Analysis

Description: Quality of life and well-being research often involves survey content that is subjective in nature, for example questions pertaining to life satisfaction. Two phenomena impacting responses to self-reported life satisfaction are studied across a range of social surveys: the framing effect, where a respondent’s answer is influenced by the theme of the survey or its content; and the mode effect, where a respondent’s answer is influenced by the method in which survey data is collected (with an interviewer, through an online collection portal, etc.). The objective of this paper is to document the effect that survey collection and survey content have on Canadians’ self-reported satisfaction with their lives. The impact of these effects on life satisfaction responses is measured across three Statistics Canada survey series: the General Social Survey, the Canadian Community Health Survey, and the Canadian Social Survey.

Description: This series provides detailed documentation on income developments, including survey design issues, data quality evaluation and exploratory research.

Description: Papers in this series provide background discussions of the methods used to develop data for economic, health, and social analytical studies at Statistics Canada. They are intended to provide readers with information on the statistical methods, standards and definitions used to develop databases for research purposes. All papers in this series have undergone peer and institutional review to ensure that they conform to Statistics Canada's mandate and adhere to generally accepted standards of good professional practice.

Description:

Conceptual arguments and examples are presented suggesting that the Bayesian approach to survey inference can address the many and varied challenges of survey analysis. Bayesian models that incorporate features of the complex design can yield inferences that are relevant for the specific data set obtained, but also have good repeated-sampling properties. Examples focus on the role of auxiliary variables and sampling weights, and methods for handling nonresponse. The article offers ten top reasons for favoring the Bayesian approach to survey inference.

Description:

We provide a critical review and some extended discussions on theoretical and practical issues with analysis of non-probability survey samples. We attempt to present rigorous inferential frameworks and valid statistical procedures under commonly used assumptions, and address issues on the justification and verification of assumptions in practical applications. Some current methodological developments are showcased, and problems which require further investigation are mentioned. While the focus of the paper is on non-probability samples, the essential role of probability survey samples with rich and relevant information on auxiliary variables is highlighted.

Description:

Non-probability surveys play an increasing role in survey research. Wu’s essay ably brings together the many tools available when assuming the non-response is conditionally independent of the study variable. In this commentary, I explore how to integrate Wu’s insights in a broader framework that encompasses the case in which non-response depends on the study variable, a case that is particularly dangerous in non-probabilistic polling.

Description:

This discussion attempts to add to Wu’s review of inference from non-probability samples, as well as to highlighting aspects that are likely avenues for useful additional work. It concludes with a call for an organized stable of high-quality probability surveys that will be focused on providing adjustment information for non-probability surveys.

Description:

Strong assumptions are required to make inferences about a finite population from a nonprobability sample. Statistics from a nonprobability sample should be accompanied by evidence that the assumptions are met and that point estimates and confidence intervals are fit for use. I describe some diagnostics that can be used to assess the model assumptions, and discuss issues to consider when deciding whether to use data from a nonprobability sample.

Description:

Non-probability samples are deprived of the powerful design probability for randomization-based inference. This deprivation, however, encourages us to take advantage of a natural divine probability that comes with any finite population. A key metric from this perspective is the data defect correlation (ddc), which is the model-free finite-population correlation between the individual’s sample inclusion indicator and the individual’s attribute being sampled. A data generating mechanism is equivalent to a probability sampling, in terms of design effect, if and only if its corresponding ddc is of N-1/2 (stochastic) order, where N is the population size (Meng, 2018). Consequently, existing valid linear estimation methods for non-probability samples can be recast as various strategies to miniaturize the ddc down to the N-1/2 order. The quasi design-based methods accomplish this task by diminishing the variability among the N inclusion propensities via weighting. The super-population model-based approach achieves the same goal through reducing the variability of the N individual attributes by replacing them with their residuals from a regression model. The doubly robust estimators enjoy their celebrated property because a correlation is zero whenever one of the variables being correlated is constant, regardless of which one. Understanding the commonality of these methods through ddc also helps us see clearly the possibility of “double-plus robustness”: a valid estimation without relying on the full validity of either the regression model or the estimated inclusion propensity, neither of which is guaranteed because both rely on device probability. The insight generated by ddc also suggests counterbalancing sub-sampling, a strategy aimed at creating a miniature of the population out of a non-probability sample, and with favorable quality-quantity trade-off because mean-squared errors are much more sensitive to ddc than to the sample size, especially for large populations.

Description:

Statistical inference with non-probability survey samples is a notoriously challenging problem in statistics. We introduce two new methods of nonparametric propensity score technique for weighting in the non-probability samples. One is the information projection approach and the other is the uniform calibration in the reproducing kernel Hilbert space.

Description:

On 1 January 2016, the world officially began implementation of the 2030 Agenda for Sustainable Development - the transformative plan of action based on 17 Sustainable Development Goals - to address urgent global challenges over the next 15 years. This report presents an overview of the 17 Goals using data currently available to report on the global indicator framework.

Description: Release published in The Daily – Statistics Canada’s official release bulletin

Description:

Gathering, exploring, analyzing and interpreting data are essential steps in producing information that benefits society, the economy and the environment. In this video, we will discuss the importance of considering data ethics throughout the process of producing statistical information.

As a pre-requisite to this video, make sure to watch the video titled “Data Ethics: An introduction” also available in Statistics Canada’s data literacy training catalogue.

Description:

In this video, you will learn the answers to the following questions: What are the different types of error? What are the types of error that lead to statistical bias? Where during the data journey statistical bias can occur?

Description:

Gathering, exploring, analyzing and interpreting data are essential steps in producing information that benefits society, the economy and the environment. To properly conduct these processes, data ethics ethics must be upheld in order to ensure the appropriate use of data.

Description:

This video will break down what it means to be FAIR in terms of data and metadata, and how each pillar of FAIR serves to guide data users and producers alike, as they navigate their way through the data journey, in order to gain maximum, long term value.

Description:

By the end of this video you will understand what confidence intervals are, why we use them, and what factors have an impact on them.

Description:

This American Sign Language video provides an introduction to the Canadian Survey on Disability. Specifically, it includes a brief description of the benefits of participating in the survey, what participating in the survey involves, how respondents were selected to participate, and information on privacy and confidentiality.

Description:

This video is part of the confidentiality vetting support series and presents examples of how to use SAS to perform the dominance and homogeneity test while using the Census.

Description:

This video is part of the confidentiality vetting support series and presents examples of how to use SAS to create proportion output for researchers working with confidential data.

Description: Quality of life and well-being research often involves survey content that is subjective in nature, for example questions pertaining to life satisfaction. Two phenomena impacting responses to self-reported life satisfaction are studied across a range of social surveys: the framing effect, where a respondent’s answer is influenced by the theme of the survey or its content; and the mode effect, where a respondent’s answer is influenced by the method in which survey data is collected (with an interviewer, through an online collection portal, etc.). The objective of this paper is to document the effect that survey collection and survey content have on Canadians’ self-reported satisfaction with their lives. The impact of these effects on life satisfaction responses is measured across three Statistics Canada survey series: the General Social Survey, the Canadian Community Health Survey, and the Canadian Social Survey.

Description:

Conceptual arguments and examples are presented suggesting that the Bayesian approach to survey inference can address the many and varied challenges of survey analysis. Bayesian models that incorporate features of the complex design can yield inferences that are relevant for the specific data set obtained, but also have good repeated-sampling properties. Examples focus on the role of auxiliary variables and sampling weights, and methods for handling nonresponse. The article offers ten top reasons for favoring the Bayesian approach to survey inference.

Description:

We provide a critical review and some extended discussions on theoretical and practical issues with analysis of non-probability survey samples. We attempt to present rigorous inferential frameworks and valid statistical procedures under commonly used assumptions, and address issues on the justification and verification of assumptions in practical applications. Some current methodological developments are showcased, and problems which require further investigation are mentioned. While the focus of the paper is on non-probability samples, the essential role of probability survey samples with rich and relevant information on auxiliary variables is highlighted.

Description:

Non-probability surveys play an increasing role in survey research. Wu’s essay ably brings together the many tools available when assuming the non-response is conditionally independent of the study variable. In this commentary, I explore how to integrate Wu’s insights in a broader framework that encompasses the case in which non-response depends on the study variable, a case that is particularly dangerous in non-probabilistic polling.

Description:

This discussion attempts to add to Wu’s review of inference from non-probability samples, as well as to highlighting aspects that are likely avenues for useful additional work. It concludes with a call for an organized stable of high-quality probability surveys that will be focused on providing adjustment information for non-probability surveys.

Description:

Strong assumptions are required to make inferences about a finite population from a nonprobability sample. Statistics from a nonprobability sample should be accompanied by evidence that the assumptions are met and that point estimates and confidence intervals are fit for use. I describe some diagnostics that can be used to assess the model assumptions, and discuss issues to consider when deciding whether to use data from a nonprobability sample.

Description:

Non-probability samples are deprived of the powerful design probability for randomization-based inference. This deprivation, however, encourages us to take advantage of a natural divine probability that comes with any finite population. A key metric from this perspective is the data defect correlation (ddc), which is the model-free finite-population correlation between the individual’s sample inclusion indicator and the individual’s attribute being sampled. A data generating mechanism is equivalent to a probability sampling, in terms of design effect, if and only if its corresponding ddc is of N-1/2 (stochastic) order, where N is the population size (Meng, 2018). Consequently, existing valid linear estimation methods for non-probability samples can be recast as various strategies to miniaturize the ddc down to the N-1/2 order. The quasi design-based methods accomplish this task by diminishing the variability among the N inclusion propensities via weighting. The super-population model-based approach achieves the same goal through reducing the variability of the N individual attributes by replacing them with their residuals from a regression model. The doubly robust estimators enjoy their celebrated property because a correlation is zero whenever one of the variables being correlated is constant, regardless of which one. Understanding the commonality of these methods through ddc also helps us see clearly the possibility of “double-plus robustness”: a valid estimation without relying on the full validity of either the regression model or the estimated inclusion propensity, neither of which is guaranteed because both rely on device probability. The insight generated by ddc also suggests counterbalancing sub-sampling, a strategy aimed at creating a miniature of the population out of a non-probability sample, and with favorable quality-quantity trade-off because mean-squared errors are much more sensitive to ddc than to the sample size, especially for large populations.

Description:

Statistical inference with non-probability survey samples is a notoriously challenging problem in statistics. We introduce two new methods of nonparametric propensity score technique for weighting in the non-probability samples. One is the information projection approach and the other is the uniform calibration in the reproducing kernel Hilbert space.

Description:

This response contains additional remarks on a few selected issues raised by the discussants.

Description:

Multiple imputation (MI) is a popular approach for dealing with missing data arising from non-response in sample surveys. Multiple imputation by chained equations (MICE) is one of the most widely used MI algorithms for multivariate data, but it lacks theoretical foundation and is computationally intensive. Recently, missing data imputation methods based on deep learning models have been developed with encouraging results in small studies. However, there has been limited research on evaluating their performance in realistic settings compared to MICE, particularly in big surveys. We conduct extensive simulation studies based on a subsample of the American Community Survey to compare the repeated sampling properties of four machine learning based MI methods: MICE with classification trees, MICE with random forests, generative adversarial imputation networks, and multiple imputation using denoising autoencoders. We find the deep learning imputation methods are superior to MICE in terms of computational time. However, with the default choice of hyperparameters in the common software packages, MICE with classification trees consistently outperforms, often by a large margin, the deep learning imputation methods in terms of bias, mean squared error, and coverage under a range of realistic settings.

Description: This series provides detailed documentation on income developments, including survey design issues, data quality evaluation and exploratory research.

Description: Papers in this series provide background discussions of the methods used to develop data for economic, health, and social analytical studies at Statistics Canada. They are intended to provide readers with information on the statistical methods, standards and definitions used to develop databases for research purposes. All papers in this series have undergone peer and institutional review to ensure that they conform to Statistics Canada's mandate and adhere to generally accepted standards of good professional practice.

Description:

The journal publishes articles dealing with various aspects of statistical development relevant to a statistical agency, such as design issues in the context of practical constraints, use of different data sources and collection techniques, total survey error, survey evaluation, research in survey methodology, time series analysis, seasonal adjustment, demographic studies, data integration, estimation and data analysis methods, and general survey systems development. The emphasis is placed on the development and evaluation of specific methodologies as applied to data collection or the data themselves.

Description:

This report summarizes the annual achievements of the Methodology Research and Development Program (MRDP) sponsored by the Modern Statistical Methods and Data Science Branch at Statistics Canada. This program covers research and development activities in statistical methods with potentially broad application in the agency’s statistical programs; these activities would otherwise be less likely to be carried out during the provision of regular methodology services to those programs. The MRDP also includes activities that provide support in the application of past successful developments in order to promote the use of the results of research and development work. Selected prospective research activities are also presented.

Description:

Since 1984, an annual international symposium on methodological issues has been sponsored by Statistics Canada. Proceedings have been available since 1987.

Description:

The Geography working paper series is intended to stimulate discussion on a variety of topics covering conceptual, methodological or technical work to support the development and dissemination of the division's data, products and services. Readers of the series are encouraged to contact the Geography Division with comments and suggestions.

Description:

Historical works allow readers to peer into the past, not only to satisfy our curiosity about “the way things were,” but also to see how far we’ve come, and to learn from the past. For Statistics Canada, such works are also opportunities to commemorate the agency’s contributions to Canada and its people, and serve as a reminder that an institution such as this continues to evolve each and every day.

On the occasion of Statistics Canada’s 100th anniversary in 2018, Standing on the shoulders of giants: History of Statistics Canada: 1970 to 2008, builds on the work of two significant publications on the history of the agency, picking up the story in 1970 and carrying it through the next 36 years, until 2008. To that end, when enough time has passed to allow for sufficient objectivity, it will again be time to document the agency’s next chapter as it continues to tell Canada’s story in numbers.

Description:

The Record Linkage Project Process Model (RLPPM) was developed by Statistics Canada to identify the processes and activities involved in record linkage. The RLPPM applies to linkage projects conducted at the individual and enterprise level using diverse data sources to create new data sources to meet analytical and operational needs.

Description:

This document briefly describes Demosim, the microsimulation population projection model, how it works as well as its methods and data sources. It is a methodological complement to the analytical products produced using Demosim.

Description:

This publication is a catalogue of strategies and mechanisms that a statistical organization should consider adopting, according to its particular context. This compendium is based on lessons learned and best practices of leadership and management of statistical agencies within the scope of Statistics Canada’s International Statistical Fellowship Program (ISFP). It contains four broad sections including, characteristics of an effective national statistical system; core management practices; improving, modernizing and finding efficiencies; and, strategies to better inform and engage key stakeholders.