Relevance & Research Question

Key words: time use survey, online diary data collection

Time-use surveys provide invaluable insights into how individuals allocate their time across various activities. Traditionally, paper-based diaries, often administered through face-to-face interviewers, have been the primary method for gathering such data. Advances in technology have enabled the development of online diary tools, but while digital tools offer advantages like reduced costs, cleaner data, and increased flexibility, they also raise concerns about data quality and inclusivity.

In recent years, two time use surveys using online diary tools have been developed and used in the UK, providing a unique opportunity to explore the benefits and challenges of this methodology.

This study assesses the impact of design choices on the quality and efficiency of online time-use surveys. By examining factors such as incentivisation, completion mode, use of support materials, and user interface, we identify strategies for minimising respondent burden while maximising data quality and response rates.

Methods & Data

Key words: time-use data, respondent burden, respondent journey, desk review, probability-based panel

This study analyses eight waves of time-use surveys conducted in the UK between 2020 and 2023, using sample from the NatCen's probability-based panel. A desk review of the online diary tools, participant materials, and analysis of available data (response rates, sample profiles, split experiments, paradata, and respondent feedback) was conducted to assess the performance of the fieldwork design and evaluate the respondent journey.

Results

Key words: mobile-first design, incentives, response rates, representativeness

Several key factors influence the success of online time-use surveys. Invitations and reminders must be timely and effective. Online tools should be designed with a mobile-first approach. Printed support materials can guide participants and improve data quality. A telephone fieldwork option can significantly boost response rates and improve sample representativeness. While higher incentives can increase response rates, their effectiveness diminishes with increasing amounts. However, these benefits must be weighed against additional costs.

Added Value

Key words: practical considerations, time-use online tool, data quality

Practical considerations and recommendations for the design of online time-use studies, ensuring they meet the diverse needs and maximise data quality and sample representativeness.

Relevance & Research Question

Social research is increasingly recognizing the relevance of spatial context to social phenomena. Thus, many surveys ask respondents about neighborhood characteristics. An open question, however, is the appropriate scale of the spatial context: Is it the immediate neighborhood, the municipality, or the county that matters? We investigate the consistency between respondents' subjective perceptions and objective characteristics across these spatial scales. Additionally, we examine how these perceptions differ between socio-demographic groups.

Methods & Data

The analysis is based on data from the GLEN study, a nationally representative panel study started in 2024. In the self-administered push-to-web survey, respondents were asked to rate their place of residence on a number of characteristics, including green space, labor market, mobility, and health care accessibility. We test whether the respondents’ subjective ratings are consistent with administrative data from official statistics at different spatial scales. Differential effects are tested for dimensions such as age, employment status, and urban vs. rural residence.

Results
[The survey is currently in the field. Results will be updated by end of February 2025]
Added Value

We contribute to the understanding of spatial mechanisms by identifying the spatial scale that best aligns with respondents’ subjective perceptions, and by highlighting how the area of reference may vary across demographic groups. Our results can give guidance for survey practitioners regarding the design of neighborhood questions as well as the selection of spatial context data as auxiliary variables.

Relevance & Research Question

Survey researchers often wish to combine survey data from multiple sources to create longer time series. One central problem when combining survey data is that variables of interest are measured with similar but not identical questions. Differences in the question text or the response options make data from questions incomparable, even though they measure the same construct. In this study, we evaluate the harmonization method Observed Score Equating in a Random-Groups Design (OSE-RG) for time series harmonization.

OSE-RG aligns differences between survey questions by transforming the response scale of one question into the format of another question requiring data from the same point in time. The challenge with time series is that we often do not have access to data from both questions at all points in time. To still use OSE-RG for time series harmonization, we need to re-use existing harmonizations at times where only data from one question is available. Thus, our research question is as follows: Is it possible to re-use existing OSE-RG harmonizations over time?

Methods & Data

To explore the re-usability of OSE-RG harmonizations over time, we harmonize time series of ten pairs of survey questions from three German general population survey programs over a period of 14 years. The central idea is that we create a response scale transformation in one year and then re-use the response scale transformation to harmonize data from other years, tracking harmonization error in the process.

Results

We find that OSE-RG harmonizations are re-usable over time for some questions, for example questions measuring general health, but not for others, for example political interest questions. We conclude that OSE-RG is a viable choice for harmonizing time series survey data. However, researchers need to be aware of reduced re-usability of OSE-RG in time series harmonization as a potential source of bias.

Added Value

The added value is threefold: First, we empirically demonstrate how time series survey data can be harmonized using OSE-RG. Second, we pinpoint the re-usability of harmonizations over time as a potential source of bias. Third, we discuss consequences for different patterns of data availability.