GenPopWeb2 reports
The GenPopWeb2 network has published the below reports:
2024
Strategies to improve response rates in probability-based online surveys: A Systematic literature review
The purpose of this systematic literature review is to: 1) classify the different strategies available to maximise response rate and minimise bias and describe the characteristics of each strategy; and 2) assess existing evidence regarding the impact of each strategy on nonresponse and selection bias. Consequently, the authors aim to providing guidance to better understand available strategies and their effect on response rates and selection bias.
This report includes: details of the methodology used; the results of the literature review, comparing the results found for the different strategies; a discussion of the practical implications of the review; and recommendations of the ways in which researchers can use the evidence emerged by the review to improve response rates when conducting research with probability-based online surveys.
2023
Adjustments for Mode Effects
This report summarises a discussion of survey methodology experts at a closed event organised by the ESRC-funded project GenPopWeb2. The meeting, which took place on 23rd September 2020, addressed issues associated with adjustments for mode effects.
Read the report, Adjustments for Mode Effects
The utility of probability-based online surveys – literature review
This paper reviews the literature on probability-based online surveys to 1) provide an up-to-date compilation of their general utility and 2) compare the various dimensions in which they can vary from a data quality perspective. The aim of the paper is to provide a guide that enables researchers to better understand the specific errors associated with different design choices.
Read the report, The utility of probability-based online surveys – literature review
The utility of probability-based online surveys – a non-technical summary
Probability-based online surveys can be defined as surveys which use an online mode to administer the questionnaire to all or most of the sampled participants selected using a probability-based sampling approach (i.e., every unit from a sampling frame of target population has a known and non-zero probability of inclusion). Nevertheless, probability-based online surveys are not homogeneous. They can vary in different dimensions: they can be one-time surveys or can have a panel nature, they can implement different recruitment strategies, the individuals who do not have access to the internet (offliners) can be treated differently and different mobile devices could be allowed for survey completion.
In this report, we provide an up-to-date and comprehensive compilation of existing evidence exploring the utility of probability-based online surveys. We also compare different approaches to probability-based online surveys across a number of dimensions from a data quality perspective (e.g., the effect on nonresponse errors of providing internet to offliners or allowing them to answer using offline modes).
Read the report, The utility of probability-based online surveys – a non-technical summary
2022
Transitioning from Interviewer-Administered Surveys to Online Data Collection: Experiences, Challenges and Opportunities (GenPopWeb2 final report)
This report summarises the GenPopWeb2 project. It includes information on the project's background and aims, as well as overviews of its literature reviews. The report also discusses each GenPopWeb2 event and summarises their main recommendations for future research and implications for survey practice.
Within-household selection for push-to-web surveys
Random selection of one respondent at each address is not straightforward for push-to-web surveys that use address-based sampling. As the sampling frame contains no information about who lives at the sampled address, the letter needs to include an instruction on who in the household is selected to take part in the survey. If only one adult is being selected, then a prescribed method for selecting the one adult is crucial as most households (about 65 per cent) consist of more than one adult and self-selection is likely to bias the survey results.
Read the report, Within-household selection for push-to-web surveys