A big chunk of Biden's approval slump could be partisan non-response 📊 August 29, 2021

Is Biden losing supporters or poll respondents? It’s probably a bit of both.

Note: Subscribers received an earlier version of this post last Sunday. The charts and numbers have been updated to take recent data into account.

The fall of Kabul, Afghanistan to the Taliban two weeks ago exacerbated a long-term decline in Joe Biden’s job approval rating. But by how much? Let’s review the data:

First up is Biden’s average rating over time as per FiveThirtyEight:

And RealClearPolitics:

Then, there are trends from pollsters who conduct regular interviews and smooth their data over time. The following trends from Civiqs and Yougov also weight their data to be representative of their respondents’ political attributes, not just their demographic breakdown as most pollsters do. That’ll be important later. Both trends here are among registered voters.



And here is a table showing the decline in Bidens net approval rating in each aggregate since July 31st:

  • FiveThirtyEight: +8 to +0 (a decrease of 8 points)

  • RealClearPolitics: +7 to -1 (-8)

  • Civiqs: -7 to -8 (-1)

  • YouGov: +4 to +1 points (-3)

Let’s focus on three things for today.

First, that there has been a long-term decrease in Biden’s approval that has only gotten worse — but did not start — in the last two weeks. The charts show Biden’s net job approval dropping steadily almost every week since his inauguration. That’s due in part because of a decline in his approval rating, but mostly due to a steeper rise in his disapproval rating. On average across aggregates (NB: a meta-average, perhaps?) the president’s approval dropped 6 points from January to August 22nd while his disapproval rose by 12. This decline has been predictable, roughly matching the shift for presidents Obama and Trump (see this chart from The Economist for parallel trends):

Second, despite the severity of the long-term decline, the aggregates do show more movement against Biden on average in the last two weeks than there has been any week in the last 7 months. The president’s net approval dropped two points on average over the last two weeks, compared to an average of 0.5-1 point each week prior. I think it’s pretty clear the fall of Kabul two weekends ago, as well as the media coverage of the fiasco, has impacted Biden negatively.

Thirdly, however, it’s worth quibbling for a moment about the differences between surveys. YouGov and Civiqs, which both balance their samples to be representative of the registered voter population politically (either past vote or current partisanship) have picked up only a modest move against Biden: maybe two points or so on net approval. We could infer from this that the last week has also dampened the share of Democrats willing to answer polls, or perhaps produced a surge in response rates among Republicans and Independents who want to express their dissatisfaction with the president.

In fact, there is some outright evidence for this: Johannes Fischer, who runs polls for the think tank/polling outlet Data for Progress, said on Twitter (I know, I know) that they’ve seen an increase in response rates among GOP voters recently. Last week’s new CBS News/YouGov poll also found a smaller prorated decline in Biden’s net rating than other pollsters or the aggregates. And, when read literally, a poll from Suffolk University finds a decrease in the share of Democrats in the electorate — which I doubt:1

Your interpretation of the data will come down to which pollsters you trust and whether you think pollsters that balance their data by party are more accurate than the ones who don’t. For what it’s worth, both YouGov and Civiqs had higher numbers than other pollsters for Trump’s approval rating throughout 2020, and slightly better pre-election forecasts. I, for one, think the aggregates are still meaningfully biased towards Democrats.

But no matter how you slice it, the public opinion has soured a bit on Biden over the last few months. He is probably net positive still among adults and slightly underwater or even with voters — a much worse position than three months ago. Perhaps, over the next few weeks, we will get more data and be able to reevaluate these hypotheses.

Posts for subscribers

Subscribers received one post this week, on mid-terms forecasting and what it means for the Democrats’ policy goals:

G. Elliott Morris's Newsletter
Saturday subscribers-only thread on the Democrats' midterms calculus
Read more

On deck: I’ll be sharing a part of the second chapter of my book about what fish can tell us about democracy. (Yes, fish.)

Assorted links

My writing:

Recommended reading:

  • Something fun: I have tried and failed over the last three years to finish reading the Lord of the Rings trilogy. At first, progress was steady. I would steal moments on the train, waiting for planes or sitting at parks to fully take in Middle Earth. But then 2020 happened, and my transit-oriented reading plan went up in smoke. Well, I’m now starting the final book and might spend less time here until I’m done. My motivation is two-fold: First, the sunk cost fallacy; and second, I really want to be able to watch the films. I plan to put up a short review on my website (as I do for most books) once I’m done!

Thanks for subscribing

That’s it for this week. Thanks so much for reading. If you have any feedback, please send it to me at this address — or respond directly to this email. I love to talk with readers and am very responsive to your messages.

If you want more content, you can sign up for subscribers-only posts below. I’ll send you one or two extra articles each week, and you get access to a weekly gated thread for subscribers. As a reminder, I have cheaper subscriptions for students. Please do your part and share this article online if you enjoyed it.

Share G. Elliott Morris's Newsletter


Credit for this goes to Jonathan Robinson on tweeter dot com