ESRA 2019 Draft Programme at a Glance

Innovations in Sampling Methods

Session Organiser Mrs Denise Abreu (USDA/NASS)
TimeFriday 19th July, 13:00 - 14:00
Room D17

This session includes papers that address challenges in sampling and coverage using different innovative techniques.

Keywords: sampling, coverage, dual frame, nonprobability

The Coverage of Survey Population Using Overlapping Dual Frame Survey Design

Ms Lidija Gligorova (Croatian Bureau of Statistics)
Ms Marijana Kožul (Croatian Bureau of Statistics) - Presenting Author

The sample design for the statistical survey Tourist Activity of the Population of the Republic of Croatia is the overlapping dual frame sample design. The sampling frame is the annually updated phone book divided in two parts: landline and mobile phone numbers of phone users with their names and addresses. The paper gives the assessment of the coverage of survey population and the coverage of all landline and mobile users in phone book. In addition, differences in socio-demographic characteristics of respondents reached by landline and mobile phones are presented. In spite of undercoverage of the survey population, using both phone types improves the total coverage of the survey population and relieves a problem of coverage and response bias.

Understanding the Characteristics of Unresolved Matched Records in Capture-Recapture Methodology

Mrs Denise Abreu (USDA/NASS) - Presenting Author

The National Agricultural Statistics Service (NASS) conducts a Census of Agriculture every 5 years, in years ending in 2 and 7. The census uses a list frame. For the 2017 Census of Agriculture, capture-recapture methods were used to adjust the Census of Agriculture for undercoverage, nonresponse, and misclassification of farms/non-farms. NASS's June Area Survey (JAS) was used as the independent survey in the capture-recapture approach. The JAS is conducted annually in June. It is based on an area frame and the data are collected via in-person interviews.

The capture-recapture framework requires a matched dataset consisting of all matches of a Census of Agriculture record to a JAS record. This dataset is the foundation for modeling the probability that a JAS farm is captured by the Census of Agriculture. A farm is a place with $1000 or more of sales or potential sales based on the reported data. In the dataset, the farm status based on the JAS and the Census of Agriculture agree in most cases. However, in other cases, a record is identified as a farm (non-farm) on the JAS and a non-farm (farm) on the Census. These records have unresolved farm status. Resolving the farm status is an important to the accuracy of the Census of Agriculture published estimates. The characteristics of the records with unresolved farm status are described. The approach to resolving farm status is discussed.

Community-Driven Health Assessment and Respondent-Driven Sampling: Methods to Collect Population-Based Health Information among Urban Indigenous People in Canada

Ms Chloe Xavier (Well Living House, St Michael's Hospital) - Presenting Author
Ms Kristen O'Brien (Well Living House, St Michael's Hospital)
Dr Michelle Firestone (Well Living House, St Michael's Hospital and Dalla Lana School of Public Health, University of Toronto)
Dr Raglan Maddox (Well Living House, St Michael's Hospital and University of Canberra, Faculty of Health)
Ms Sara Wolfe (Seventh Generation Midwives Toronto)
Ms Gertie Mai Muise (Indigenous Primary Health Care Council)
Dr Janet Smylie (Well Living House, St Michael's Hospital and Dalla Lana School of Public Health, University of Toronto)

Background: The absence of representative population data contributes to a significant underestimate of inequities in health determinants, health status, and healthcare access between Indigenous and non-Indigenous people in Canada. This results in the masking of critical unmet health and social needs. Available data sources are often biased, non-random samples, have high levels of non-response or misclassification, and tend to lack Indigenous identifiers. Such limitations systematically undermine the development, implementation, monitoring, and evaluation of community health programs, services, and policies.
Objectives: Our Health Counts (OHC) aims to address the gap in health and wellbeing data for Indigenous people in urban settings in Canada. Utilizing Indigenous community-driven methods, OHC generates population-based health status and health care utilization datasets using respondent-driven sampling (RDS).
Methods: OHC’s Respectful Health Surveys are developed in collaboration with Indigenous researchers, Indigenous community partners, and a reference group composed of local service providers. This collaboration fosters the creation of locally relevant health assessment surveys. OHC also employs RDS to recruit participants; a statistical method developed to sample ‘hard-to-reach populations’ using the inherent social networks and Indigenous kinship lines within communities to collect and generate representative population data.
Findings: Our demonstrated approach attempts to balance the power relationships between Indigenous community partners, academics and additional governmental, data, and public health stakeholders throughout the research process, while maintaining rigour and policy relevance. This underpinned the unprecedented participation of over 3,000 urban Indigenous adults and children across four major cities in Canada. It also demonstrated that data collection tools, such as the census, chronically undercount Indigenous populations in urban areas. The success of OHC can be attributed to the community-centered approach, including an Indigenous data and research governance model; engagement with Indigenous researchers and community partners; and methodologies which informed OHC implementation and highlighted the need for meaningful community partnerships.