The Corona Insights team wanted to know how Coloradans were holding up during shelter in place, so we conducted a survey! We asked residents about a broad range of topics, including their overall wellbeing, challenges, concerns, community response, and others.
All posts in this series:
- Colorado’s COVID-19 Experience Part 1: How CO is Coping
- Colorado’s COVID-19 Experience Part 2: Community Response
- Colorado’s COVID-19 Experience Part 3: Past & Future Plans
- Colorado’s COVID-19 Experience Part 4: How COVID-19 Impacts Quality of Life
- Colorado’s COVID-19 Experience Part 5: Types of Worry
We’ve discovered a lot of interesting things in our COVID-19 Experience Survey data. For my own analysis, I wanted to better understand how worried Coloradans are about health and the economy during this pandemic. During an emergency, officials often want the public to stay calm so that people don’t engage in rash behavior. Understanding what people are worried about can be useful when figuring out how to keep everyone calm.
To get this information from our data, I want to segment the data by a variable that captures worries. We often segment data by demographic variables (e.g., age, race and ethnicity, gender, etc.), which can be very illuminating. But sometimes we want to look at the data by some psychographic variable (e.g., whether people do or do not support a policy, whether people have knowledge of a topic, etc.). At the beginning of our questionnaire, we asked people to rate on a scale of 0 to 10 how concerned they were about their personal health, personal finances, family’s health, health of those in their community, and the economy. This set of questions seemed like a good choice for this analysis.
For this analysis, I chose to use clustering to create a set of personas. I explain personas further in this post, but essentially, this type of analysis lets me group respondents based on their answers to a set of questions. I could have created a “Worry Scale” in a sense by adding up how concerned people said they were about each of the topics. Then, we could have compared people who are very worried to those who are not as worried. However, that would have hidden any differences in what people were worried about. By clustering the data into personas, I can see whether people are naturally grouped by how much they worry about everything or by how much they are worried about specific topics. As is often the case in the real world, you see both.
I ended up choosing to create four personas based on the cluster analysis. As you can see below, Coloradans differed both in terms of what they were worried about and how much they are worried in general.
The first group on the left I’ve named “Everything but My Wallet”. They are the second most concerned overall about most of the items we asked about, but they are not very concerned about their own finances. The second group, “Just Worried. Period.,” is very concerned about everything. The third group, “It’s the Economy, Stupid,” is only worried about the economy. And in general, they aren’t very worried. The last group, “Money > Health,” is very concerned about their own finances and somewhat concerned about the economy. They are less concerned about health. The Just Worried. Period. group is almost half of our sample, and the Everything but My Wallet group is more than a quarter. The remaining two groups combine to about a quarter of the sample, with the It’s the Economy, Stupid group being the smaller of the two.
Now that I’ve identified my four personas, I can now examine how they differ from each other. First, I looked at whether the four personas were experiencing negative emotions to a differing degree (depicted to the right). Perhaps not surprisingly, Coloradans in the Just Worried. Period. group have been experiencing more negative emotions during the pandemic compared to the other groups. This was especially true for feeling bored and feeling anxious.
So it’s clear from the data that the personas differ in terms of what they are feeling, but do they differ in terms of how the pandemic is touching their lives? I first looked at whether people in each group were facing different challenges during the pandemic. As shown below, Coloradans in the Just Worried. Period. group were more likely to be facing challenges brought about by COVID-19. Those in the Money > Health group are having difficulty making ends meet. Those in the It’s the Economy, Stupid group were the least likely to report challenges.
Additionally, I examined whether the different personas were aware of other people who had been affected by the virus (depicted below). Coloradans in the It’s the Economy, Stupid group were the least likely to know of someone who had been impacted by the virus, either in terms of their job or their health. Those in the Just Worried. Period. group were the most likely to know someone impacted by the virus. Thus, it seems like the personas are having different experiences during the pandemic.
Finally, I looked at demographics to get a different sense of who these personas are. Those in the It’s the Economy, Stupid group were disproportionately more likely to live outside of the Denver metro area, to be male, to be older, and to be retired. For this group, it might make sense not to be that worried because they are already retired (thus no job concerns), and they are less likely to know of someone who has been impacted by COVID-19. Those in the Money > Health group were more likely to be out of work and looking or to be unable to work. They were also more likely to be low income and younger than 40. So for them, the economic impact might be more painful right now than the health impact. Coloradans in the Everything but My Wallet tended to have higher levels of education. They may be in jobs that are more insulated from the economic impact so far. Because the Just Worried. Period. group is large, they were more defined by their experiences with the virus than by any one demographic.
Overall, this analysis reveals how our personal experiences with the virus are shaping both our personal and our global concerns. It will be interesting to see how this pattern may change over time. Unfortunately, the longer the pandemic continues, the more likely it is that people will know someone impacted by the virus.
The survey was fielded between April 7-9, 2020. Resulting data were cleaned and coded. Responses were weighted to represent the state of Colorado by geography and key demographic characteristics from the 2018 American Community Survey.
For further inquiry, please contact Corona Insights.