Different Presentations of the Same Data can Lead to Opposing Inferences

This talk is fully online! You will receive a password for the Zoom session on the day of the talk after signing up.

We are excited to host a talk given by Dr. Stephen Spiller (UCLA) and Dr. Nicholas Reinholtz (CU Boulder). This will be the last talk of the academic year! More info below:

In a world full of bias, it is tempting to view data as a neutral arbiter of truth. In this talk, we offer a prospective of caution: Presenting data entails making choices on *how* to present these data. And, unfortunately, these choices on how to present data can affect how people interpret it and the judgments they make about it. We demonstrate this in the domain of time-series data. We show that in certain cases the choice of presenting data as "stocks" (absolute levels over time) versus "flows" (change in absolute levels over time) can lead to opposing inferences. For example, when employment data from 2007 to 2013 are shown as flows (jobs created or lost), President Obama’s impact on the economy during his first year in office is viewed positively, whereas when the same data are shown as stocks (total jobs), his impact is viewed negatively.