Non-response bias was the major culprit driving survey error in the 2020 election cycle. After extensive research on this topic (read more here), we established methods and procedures to ensure our surveys are representative. We continue to explore additional approaches to strengthen representation in our survey design.
Accounting for Turnout Uncertainty
Another potential source of error in political polling is turnout – the inability to know precisely who plans to vote. This type of error will always exist, but we believe it is important to communicate about and account for this uncertainly through scenario modeling of different electorates.
Survey Design Innovation
A well-written survey is vital for effective research results. We continually review and optimize the way we ask our questions – in what order, angle of approach, word choice, etc. – to evoke the most revealing responses. Recently, we presented findings on The Relationship Between Turnout and Self-reported Motivation to Vote at the American Association for Public Opinion Research (AAPOR) conference.
Understanding Key Audiences
Inclusive Data Collection & Analysis
Unconscious bias and exclusivity can radically impact the quality and accuracy of research data. We look for the presence of bias and for where inclusivity might be missing across our research process — in our research plans, the way we word questions, the way we categorize groups of people, etc. — as part of our ongoing commitment to inclusivity.
Going Beyond Demographics
Demographics will always be an important tool for analysis, but in a society increasingly divided by attitudinal and cultural lines, it can’t be the only way to analyze data. That’s why we segment voters more frequently by psychographic composition and affinity groups, and not just traditional demographic categories.
Audience Deep Dives
We believe dedicated surveys or oversamples of key subgroups or demographics should be considered in any research plan to ensure confidence in findings among key audiences like African American voters, AAPI voters, or Hispanic voters. Such deep dives help us to understand nuances within those groups so they are not misunderstood as having monolithic attitudes.
We now offer ad testing through a new platform that surfaces data to guide decision-making. What differentiates this from other platforms is that it brings together GSG’s consulting acumen and holistic knowledge of campaigns with innovative ad-testing technology to support you in making sound strategic decisions.
Expanding Qualitative Research
We believe that qualitative research is central to an effective research program. We look for new approaches to bring qualitative research to the fore and to adapt traditional tools, like focus groups, to identify blind spots and uncover effective language and values-based messages approaches that were previously unknown.
Advanced Statistical Analysis
Beyond simply analyzing shares of responses, when appropriate, we will recommend incorporating statistical techniques in our analysis: regressions, TURF (Total Unduplicated Reach and Frequency) analysis, scenario analysis, and MaxDiff. And, we’ll also explain these techniques and their results to you in plain language so you don’t have to be an expert in statistical analysis methodologies.