Hacker News new | past | comments | ask | show | jobs | submit login

>There are plenty of people who like each other that have nothing in common. The idea that the reason people like people is because they are like them just feels ridiculous to me, but maybe I am an outlier. Personally, the more someone is like me, the less I like them.

Yes, you are an outlier. There is a very large amount of studies that show that people like people that are like them.

[1] https://journals.sagepub.com/doi/abs/10.1177/026540750809670...




Yes, you are an outlier.

Fair enough

There is a very large amount of studies that show that people like people that are like them.

The link you provide shows a coefficient of .47 for actual similarity, and .39 for perceived similarity. I'm not good at statistics, but that seems like a far cry from "the reason why people like other people is that they are like them." which is the line I was replying to. If I'm understanding it correctly it would be more accurate to say "people are somewhat more likely to like someone who is like them." If that's what was originally written, then I probably wouldn't have replied.


> The link you provide shows a coefficient of .47 for actual similarity, and .39 for perceived similarity. I'm not good at statistics, but that seems like a far cry from "the reason why people like other people is that they are like them." which is the line I was replying to. If I'm understanding it correctly it would be more accurate to say "people are somewhat more likely to like someone who is like them." If that's what was originally written, then I probably wouldn't have replied.

r, or Pearson's Correlation Coefficient [1], is a value from -1 to 1. for r < 0, the variables are negatively correlated, for r > 0 the variables are positively correlated, and for |r| ~= 0, the variables are uncorrelated. 0.47 and 0.39 are high values of correlation.

[1]: https://en.wikipedia.org/wiki/Pearson_correlation_coefficien...


This is just my understanding, but I believe in social sciences r values are usually quite deflated and a coefficient of .47 is extremely high.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: