Hacker News Reader

Are you saying that larger sample sizes lead to greater variance in results? Because I don’t think statistics usually works that way.

That wouldn't skew the resulting average/mean that much, it would just be less accurate for specific localities...

I don't understand, are you arguing that maybe if they'd measured the right July day in 1951 then we'd see that that was the warmest month?

If we have more data points in cities, this could skew the results. I live in a forested area and going into the city can easily add 3 or 4 degrees Celsius.

I did some reading and this is actually happening. [0]

Sensors are increasingly being surrounded by urbanization and this influences readings by the heat island effect.

[0] https://www.researchgate.net/publication/262564641_Modelling...

Why is this observation relevant?