The 81st Annual AAPOR Conference was an amazing opportunity to learn, connect, and share findings from our recent work across a wide range of topics.  We’ve gathered the key takeaways from select sessions to share with those who were unable to attend the event.

Panels They Are a-Changin’: Exploring Factors Other Than Attrition That Impact Panel Composition In a National Probability Panel

In the world of panel management, we spend a great deal of time thinking about who enters our panel and who leaves it. We build complex models for recruitment and attrition, striving to keep our probability panel representative of the population. That said, there is a subtle and profound force at work: the people who stay are not the same people they were when they first signed up.

The research presented at this year’s AAPOR national conference isn’t a story about attrition—the panelists who leave. This is a story about the people who remain, and how their individual evolution shapes the very nature of the data we collect.

Dynamic Panel Variables

When you ask an audience of conference attendees to raise their hand if they are the exact same person they were five years ago, no hands go up. Life happens. The same is true for the roughly fifty thousand individuals on the SSRS Opinion Panel.

  • Financial Lives: Income is one of the most dynamic variables. While it’s expected that incomes generally rise over time due to career progression or inflation, a significant number of panelists also experience income decreases or fluctuations in both directions. This reflects the real-world volatility of employment, household changes, and economic uncertainty.
  • Social Connections: How often do you talk to your neighbors? For 42% of panelists, the answer changed over a 12-month period. People moved between more and less interaction, a quiet indicator of changing neighborhoods, new local relationships, or shifts in personal priorities.
  • Company Size: In the age of the gig economy, the definition of “employment” itself is fluid. The data shows high variability in how people report the number of employees wor at their company, especially among non-white panelists. Is an Uber driver a part-time employee or a contractor? A person’s answer might change not because their job did, but because their understanding of it did. This ambiguity is a critical area for future research.
These changes are not isolated events. They are ripples from a larger story unfolding in a person’s life.

When one core aspect of a panelist’s life shifts, other things often shift with it.

One of the most compelling aspects of the research is its focus on correlated changes—the way one life event often cascades into others. The data suggests that when one core aspect of a panelist’s life shifts, other things often shift with it.
  • When Education Changes: An update to a panelist’s educational attainment was frequently accompanied by changes in their political ideology (40% of the time) and employment status (31% of the time). This aligns with what we know about the profound impact of education on a person’s worldview and career path.
  • When a Household Grows (or Shrinks): A change in the number of adults in a household often coincided with shifts in religious identity, political ideology, and even whether a gun was kept in the home. One can easily imagine the stories here: a marriage brings a new partner with different beliefs, or the arrival of a child changes a parent’s perspective on safety and community.
  • When Health Changes: The onset of a reported disability correlated with changes in religious identity, volunteer activity, and household gun ownership.

These correlations are not just statistical curiosities; they are the data-driven echoes of human experience. They show that behind every updated variable is a person navigating life’s milestones—marriage, a new job, parenthood, illness, or a crisis of faith.

At a macro level, a large panel might appear remarkably stable. The overall percentage of Democrats and Republicans may not shift much month to month. But this stability is an illusion, masking a turbulent reality underneath.

This has critical operational implications for anyone who manages or uses panel data. If a panel’s profile data isn’t refreshed regularly, researchers are essentially working with an outdated map of their sample. The panel’s composition is not what it was a year ago, or even six months ago. This can subtly skew targeting, weighting, and ultimately, the conclusions drawn from a study.

Refresh Profile Data Regularly

The solution is straightforward but resource-intensive: refresh profile data regularly. SSRS, for example, conducts a comprehensive refresh annually, incentivizing panelists to provide updated profile information. This isn’t just a best practice; it’s a necessity.

Ultimately, this research is a powerful reminder of a simple truth: “Behind every data point is a beautiful, messy, wonderful person.”

In our quest for scientific rigor and statistical significance, it’s easy to forget this. We look at aggregates, at trends, at the clean lines on a chart. But the real story is often in the individual stories that those clean lines obscure.

The goal of research is to understand the diversity of human experience. To do that well, we must acknowledge that experience is not a fixed point. People evolve. Their beliefs shift, their households change, and their lives move in directions we can’t always predict.

The challenge for researchers is to build systems that honor this complexity. We must not only recruit a representative sample but also diligently track its continuous, inevitable transformation. Because the people in our panel are not just subjects; they are protagonists in their own ongoing stories. And those are the stories we are here to understand.