Relevance & Research Question

A large corpus of studies across various domains has demonstrated a birth month effect, wherein individuals born in specific months display distinct outcomes compared to those born in other months with respect to areas such as health, socioeconomic status, and behavior. In contrast, the common use of birthday-based sampling methods (e.g., selecting respondents in a household whose birthday is next or was last) in large-scale surveys assumes that birth month is uncorrelated with outcome variables. If a birth month effect exists, this assumption may introduce bias, particularly when comparing groups with systematic differences in household size, such as non-Western immigrants and majority populations in Western Europe.
Methods & Data

We first develop a theoretical framework to explore the relationship between birth month effects and potential biases in birthday sampling designs. We conduct a preregistered empirical analysis using the LISS panel (Longitudinal Internet studies for the Social Sciences), a probability-based online panel of Dutch households. In the LISS panel, data is collected from all individuals within a household aged 16 and above. Through simulations across 12 different fieldwork periods (i.e., months of the year), we assess the extent of bias that might arise if the LISS panel employed the next-birthday sampling method instead. We examine 35 variables, including personality traits, health outcomes, and socioeconomic status, to evaluate the potential impact on research outcomes.
Results

Our analysis does not reveal evidence for a strong birth month effect for the selected variables. The simulations show that, for the Dutch context, the next-birthday sampling method does not introduce substantial bias for the variables of interest.
Added Value

Though a null finding, our study provides important insights for survey methodology. It suggests that, in the Netherlands, next-birthday sampling is unlikely to produce bias related to birth month effects, at least for the way how these variables are commonly measured in social science surveys. This contributes to the ongoing discussion on sampling methods and enhances the reliability of results in large-scale surveys.

Relevance & Research Question

As reported by the UNHCR for 2023, the majority of refugees are hosted in low- and middle-income countries (75%) and countries that neighbor their country of origin (69%). For refugees who are particularly vulnerable and unable to return to their country of origin, resettlement programmes aim to provide long-term prospects by resettling them to Western countries. And despite the fact that only a small proportion of those in need for resettlement are actually resettled (8% according to UNHCR), little is known about the living situation of those who are left behind. This study addresses this gap by conducting a web-survey targeting potential resettlement refugees using social contacts of already resettled refugees.

Methods & Data

The study uses addresses of all refugees resettled to Germany since 2013 (i.e., approx. 17,000) from the German Central Register of Foreigners (AZR) to invite participants via postal mail to a web-survey. This survey marks the start of a snowballing approach, where refugees in Germany function as seeds and are asked to forward a survey-link (with a mobile “share”option) to up to three contacts, who still reside in countries of first refuge and may be eligible for resettlement (target population A). To reduce the risk of realizing too few cases – especially in the first step from Germany to abroad –, participants are also asked to share the survey to refugees who may not be eligible for resettlement (target population B). In the subsequent steps, this snowballing process continues in the countries abroad.

Results

Although data collection is ongoing at the time of the GOR conference, findings from cognitive pretests and from data already collected will be presented, offering insights into the potential but also challenges of surveying refugees via international snowball sampling.

Added Value

This study contributes to the understanding of the living conditions of refugees while offering methodological insights into sampling hard-to-reach populations. By demonstrating how snowball sampling with multiple target populations can mitigate recruitment challenges, it provides valuable lessons for researchers focusing on vulnerable groups.

Relevance & Research Question

Traditional online survey panels often lack coverage in smaller or less digitally integrated countries, limiting researchers' ability to collect reliable data from these regions. Social media sampling presents a promising alternative for quantitative surveys in such contexts. This study investigates the feasibility, reliability, and potential biases of social media sampling as a data collection method. Using a multi-country survey spanning 18 nations—including Zimbabwe, Kazakhstan, Costa Rica, and Iceland—we address the question: Can social media sampling provide reliable insights for brand perception and societal measures, in countries where online panels are unavailable?
Methods & Data

The study deployed a quantitative survey during June and July 2024 through targeted social media advertisements, optimized to recruit representative samples across diverse demographics. In total n9.000 Participants responded to a standardized questionnaire including brand image measures and selected questions from the World Happiness Report. Sampling quotas and algorithmic targeting ensured coverage of gender, age, regions, ethnicities and income levels each country. Reliability was assessed through comparative analysis against external data sources, where available, and consistency checks within the datasets.
We compared metrics like available audience in Meta, click-rates, completion rates, drop-outs per questions, differences in response behaviour between countries and between ethnic groups within countries.

Results

The results demonstrate that social media sampling can effectively generate diverse, balanced samples in countries lacking established online panels. Across the 18 countries, response rates and sample representativeness varied but were sufficient for robust analysis. Insights from brand image and happiness metrics revealed consistent trends across nations and offered valuable local context, while some limitations, such as underrepresentation of older rural populations, and low-income groups and some ethnicities in African countries were noted
Added Value

This research highlights the untapped potential of social media as a viable sampling solution for most of the investigated countries. By showcasing a rigorous approach to survey design, execution, and evaluation, this paper contributes a practical framework for using social media to extend the reach of quantitative research globally. The findings are particularly relevant for researchers, seeking solutions for data collection in emerging and underrepresented markets.