The Red Flag Observatory

Why this page exists

Most red flag lists you find online are one person's opinion. This one is different: it doesn't tell you what you should think, it measures what people actually judge, by making them pick between two behaviors, thousands of times over.

So far, 119,107 votes have been cast, and 183 everyday behaviors are analyzed here. Behaviors from the “Love & Sex” category, played between adults, are not published.

The most interesting finding isn't the ranking itself. It's how big the disagreements are: on some behaviors, the gap in judgment between two groups tops 200 points — more than what separates a harmless behavior from a problematic one in the overall ranking.

Men vs. women: the behaviors that divide the most

A higher score means the group sees the behavior as more problematic. The gaps below are the widest in the whole set.

16-18 vs. 27 and over: the generation gap

Same reading, this time between the youngest and the oldest age groups.

The most divisive behaviors

Here we stop comparing two groups and look at all six: gender and age groups. The spread is the gap between the harshest group and the most lenient one. The wider it is, the less shared norm there is around that behavior — these are, quite literally, the topics where a conversation can go south because nobody's on the same page.

1. Putting your partner down in front of your friends

   Spread across groups: 466 points · overall score 1526

2. Hitting your dog

   Spread across groups: 452 points · overall score 1666

3. Dumping your pet right before vacation

   Spread across groups: 411 points · overall score 1635

4. Scratching your crotch in public

   Spread across groups: 347 points · overall score 1199

5. Voting right-wing

   Spread across groups: 334 points · overall score 1280

6. Forbidding your partner from seeing their friends

   Spread across groups: 328 points · overall score 1503

7. Starting a GoFundMe for a fake illness

   Spread across groups: 309 points · overall score 1450

8. Not paying child support

   Spread across groups: 292 points · overall score 1317

9. Voting far right

   Spread across groups: 285 points · overall score 1455

10. Mocking people’s accents

    Spread across groups: 285 points · overall score 1169

11. Being a trader / investment banker

    Spread across groups: 276 points · overall score 913

12. Judging everyone on their looks

    Spread across groups: 274 points · overall score 1358

The ones everyone agrees on

On the flip side, these behaviors are judged the same way by every group. Broad agreement doesn't mean “serious”: it means the verdict doesn't depend on age or gender. The overall score next to it tells you whether that consensus is about something harmless or something serious.

1. Wearing clothes once, then returning them for a refund

   Spread across groups: 59 points · overall score 986

2. Ordering for the whole table without asking

   Spread across groups: 61 points · overall score 1116

3. Never working out

   Spread across groups: 63 points · overall score 928

4. Never brushing your teeth in the morning

   Spread across groups: 67 points · overall score 1141

5. Doing drunk karaoke

   Spread across groups: 68 points · overall score 807

6. Only listening to techno

   Spread across groups: 69 points · overall score 907

7. Putting ketchup on everything

   Spread across groups: 71 points · overall score 932

8. Showing up late every single day

   Spread across groups: 72 points · overall score 1012

9. Working in HR

   Spread across groups: 73 points · overall score 1045

10. Refusing to separate the art from the artist

    Spread across groups: 76 points · overall score 1191

11. Dipping your fries in your milkshake

    Spread across groups: 77 points · overall score 915

12. Being a stand-up comedian

    Spread across groups: 78 points · overall score 787

How to read these numbers

Every behavior has a score that goes up when it's picked as the worse one in a duel, and down otherwise. A separate score is kept for each declared group, which is what makes these comparisons possible.

Two caveats. First, scores by subgroup rest on a fraction of the votes: we only show behaviors that have passed a participation threshold, but a gap is still a trend, not a precise measurement. Second, our sample is self-selected — it's the people who chose to play. These numbers describe our community, not the general population.

The full calculation is explained in the methodology.

Explore it yourself

The full rankings can be filtered by category, gender and age group. Two articles dig into the gaps between men and women: at work, and in everyday life. You can also add your own votes — every duel you settle moves the scores.