Megan Brenan reports from Gallup News:
Amid the ongoing COVID-19 pandemic and record-high inflation, Americans’ satisfaction with the direction of the country has fallen to 17%, the lowest in a year. At the same time, Americans’ satisfaction with their own lives has ticked up to 85%, just five points shy of the 2020 record-high point.
Lots more at the link. People who attend religious services, richer people, and better educated people are all more likely to be satisfied with their lives.
Regarding satisfaction with the direction of the country, we see the usual partisan pattern:
I don’t know why they only show these data since 2020. I’d like to see the plot going back to 1980, just as in the graph at the top of this post.
Yair showed me the top plot and noticed how the different time series seemed to track each other.
First, he de-trended each series by taking the residuals of each from a simple linear regression over time. After subtracting off the overall time trends, there’s a strong correlation of 0.6 between personal satisfaction and satisfaction with the way the country is going:
A correlation of 0.6 might not sound like much, but it’s pretty impressive given that these are averages of noisy survey responses.
Yair then standardized each de-trended series by dividing by its standard deviation, and he plotted them both on the same graph:
Wow—that’s pretty striking! As Yair says, the biggest pattern does not seem to be in the year-to-year variation but rather in decades:
– From 1980-1991 there was a general upswing in both series.
– Then a reset in 1992 and another gradual increase until about 2002.
– Then a steady decline until 2010 or 2012.
– Then a steady rise until 2020.
– Then a drop and a rise since then.
Not sure how “important” this is, just an interesting little data story. First glance = the two series don’t look related and imply a certain story: that the public’s views on politics are pessimistic and not related to personal satisfaction. Second look = more nuanced: personal satisfaction going up over time (slightly) and USA satisfaction going down, but after accounting for that, the two do seem to be related.
My mental model now is one where people (in the US) are more or less conditioned to say, “things are good” when asked this question, so it inherently has less signal than the question about USA satisfaction. But I don’t know that literature at all.
I’m not sure either!
A research project for you
I think that in any serious analysis of these time series would want to break both responses down by party identification. It should be possible to get these raw data from old polls, so this could be a good research project.
It is striking to me that the time variation in personal satisfaction is so low and yet it tracks satisfaction with the country so well.
I’d also like to compare to time series of business cycle and consumer confidence.