They define in-party and out-party very weirdly for multiparty systems like the Netherlands, Denmark and Norway. In NL, if I'm a leftie, I have 3-4 reasonable parties to choose from. Most people voting for any of these parties won't deeply hate any of the other. Similar stuff on the right, or along a conservative/progressive axis. But the authors do this:

> Thus, we follow these studies and define the in-party as the party receiving the highest score among all parties rated by the respondent on the party feeling thermometers.4 The out-party measure is drawn from the average of all other parties’ feeling thermometer scores, weighted by their respective vote shares.

By mixing all the other parties that aren't the in-party, they mix parties the voter is quite OK with, with parties that the voter deeply and passionately detests. I can't imagine this not skewing the graphs enormously.

For example, if you're right wing in NL you get to choose between the anti-Muslim party, the conspiracy theory anti-Muslim party, the farmers-against-immigrants party, the God-wants-the-death-penalty party, and the regular oldschool conservative right wing party (who call themselves "liberals", by the way). Most right-wing voters will have a strong preference among this lot, but at the same time they won't mind deeply if one of the other options wins elections instead.

But they hate hate hate the progressives! A few weeks ago right wing protesters attacked and trashed the HQ of one of the more progressive parties (D66). The hatred is very deep (and mutual), totally unlike what the graphs in this article show you.