Congress Atlas

What the public thinks

Gallup has asked Americans the same question every month since 1974: do you approve or disapprove of the way Congress is handling its job? Here is every year of answers, next to what Congress was actually doing.

Approve today
15 %

2026 average so far — roughly 1 in 10 Americans. The year is not over, so this figure will move.

Average since 1974
29 %

Across 50 years of polling.

Best year on record
57 %

2001. Approval hit 84% in October 2001, the single highest monthly reading Gallup has recorded.

Worst year on record
14 %

2013. The lowest single month was 9%, in November 2013.

Fifty years of approval

Each point is a yearly average of Gallup's monthly polls. The spike is the months after September 2001; the long slide begins around 2005.

DecadeAverage approvalYears measuredIn plain terms
1970s30%6Roughly 3 in 10 approved
1980s36%7Roughly 3 in 10 approved
1990s32%10Roughly 3 in 10 approved
2000s39%10Roughly 4 in 10 approved
2010s17%10Fewer than 2 in 10 approved
2020s20%7Roughly 2 in 10 approved

Does passing more laws make people happier with Congress?

On the evidence we have, no. Across the 7 years where we have both figures, the relationship between substantive laws passed and public approval is essentially nonexistent (correlation -0.08). Approval moved far more with events — September 11, the 2013 shutdown — than with legislative output. This covers 2013–2015 onward only, which is a short window and a real limit on what can be concluded.

Both series scaled to a shared 0–1 range — approval as a share of 60%, laws as a share of the record high (346) — so the shapes, not the units, are what you compare.

Why be careful with a correlation like this?

Two things that both drift in the same direction over time will correlate strongly even when neither causes the other. Public approval has fallen fairly steadily since 2005 and partisan division has risen fairly steadily over the same period, so almost any pair of these measures will show a strong-looking relationship.

The safeguard is to look at whether the two move together year to year, not just whether both trend. When a relationship only shows up in the trend and disappears in the year-to-year changes, it is not evidence of anything.

The other limit here is length: complete bill records exist from 2013, giving about a dozen years of overlap with the polling — too short to settle a question this size.

Data generated 2026-08-21 21:14:06. Total years of polling held: 50.