Dialects of the USA

[vor_lord]

We must sound almost exactly opposite.

What we can't tell from the graph is the strength of the similarity/difference. My assumption is that they take the whole 'similarity index' and normalize it onto the complete color scheme. For example if my most similar to least similar "range" is very wide, i.e. I have a very strong dialect, then my 'red' and 'blue' are much stronger indicators than someone without a strong dialect. While Boston might be your most similar dialect, it might not be a very good match. It's still listed as red, i.e. most similar. I assume that is true because it makes for a lot more interesting graph than an absolute similarity mapping. No one wants to see a bland graph of mostly yellow everywhere.

But I guess I can't argue, they nailed me despite the fact that I know I have picked up dialect items that don't match my origin (I say soda for example).