Every CoreOpinion test asks key positions twice, once in a sympathetic frame and once in a critical one, and measures how far the answer moves between the two. Split by age, that movement traces a U. It is highest among people 65 or older (27.6) and those under 18 (25.2), and lowest among people aged 35 to 44 (21.5).
The score runs from 0 to 100, where 0 means your answers did not move at all when the wording flipped and higher numbers mean they moved more. It is the same number shown as a percentage on your own result.
Show the data
| Age | People | Average | 95% interval | Spread (SD) |
|---|---|---|---|---|
| Under 18 | 2,914 | 25.2 | 24.6 to 25.8 | 16.2 |
| 18 to 24 | 2,465 | 22.9 | 22.3 to 23.4 | 14.3 |
| 25 to 34 | 779 | 22.8 | 21.7 to 23.9 | 15.5 |
| 35 to 44 | 399 | 21.5 | 20.2 to 22.7 | 12.8 |
| 45 to 54 | 311 | 23.6 | 21.8 to 25.4 | 16.2 |
| 55 to 64 | 227 | 26.8 | 24.1 to 29.5 | 20.7 |
| 65 or older | 153 | 27.6 | 24.3 to 30.9 | 20.6 |
The gap between people aged 35 to 44 and those 65 or older is 6.1 points, with a 95% interval of 2.6 to 9.6. The gap at the young end is smaller, 3.7 points (2.4 to 5.1), and its interval also sits clearly above zero.
Is this just Finland?
It was the first thing we checked. Finland accounts for 3,374 of our results, and 85% of Finnish respondents who gave an age are under 25. About 50% of everyone in the under-18 group took the Finnish test. If Finnish teenagers happened to score high, they alone could have produced the left side of the U.
Show the data
| Age | All results | Excluding Finland | People excluding Finland |
|---|---|---|---|
| Under 18 | 25.2 | 26.0 | 1,468 |
| 18 to 24 | 22.9 | 23.0 | 1,324 |
| 25 to 34 | 22.8 | 23.8 | 538 |
| 35 to 44 | 21.5 | 22.2 | 315 |
| 45 to 54 | 23.6 | 24.1 | 234 |
| 55 to 64 | 26.8 | 27.1 | 199 |
| 65 or older | 27.6 | 27.6 | 132 |
It isn't Finland. Taking every Finnish result out moves no age group by more than 1.0 points, and the U is still there.
Older respondents disagree more with each other
An average can hide very different people. Among those aged 35 to 44, scores cluster fairly tightly, with a standard deviation of 12.8. Among those 65 or older it is 20.6. So the higher average for the oldest group is not a case of everyone moving a little more. It is closer to a mix: many older respondents barely move at all, and a minority move a great deal.
Why might this happen?
Our data can show the pattern. It cannot show the cause, and we would rather lay out the plausible explanations than pick one.
Views still forming. For the youngest respondents, many of these questions may be ones they have not had much reason to think about yet. Without a settled position to anchor to, the wording of the question does more of the work.
A style of answering, not persuasion. Survey researchers have long documented acquiescence, a tendency to agree with statements whatever direction they point, and have linked it to several factors including age. A test that asks each position both ways would register someone who agrees with both versions as having moved. If that is part of what is happening at the older end, some of the score reflects how people answer questionnaires rather than how easily they change their mind.
Who turns up. The people over 55 who find and finish an online political quiz are unlikely to be typical of their age group, and there are few of them in our data: 153 people 65 or older, against 2,914 under 18.
Age or generation. This is a snapshot taken over a few months. It cannot tell apart an effect of being older from an effect of belonging to a particular generation, and it will take years of data before we can.
What it does not mean
It does not mean people in midlife are immune to framing. The lowest group still averages 21.5 on a 0 to 100 scale, and the most common result across all ages is a score between 20 and 30. Nobody in any age group is typically unmoved by wording. The difference between the groups is real but much smaller than what they have in common.
Find out your own score: take the test.
About these figures
- Based on 7,248 results where the person gave their age. Age is optional and self-reported in bands.
- Age groups are shown only with at least 100 people. Intervals are 95% intervals for the mean.
- Different country tests use different questions, and people choose tests of different lengths, so scores are not perfectly like-for-like between individuals. With groups this size those differences largely average out, but they add noise.
- This is a self-selected online sample and is not weighted to any population. Full detail is on the methodology page.