> I took a quite specific discussion ("Sweden has turned dangerous because of crimes by immigrants") and in that discussion the fact(?) that total crime numbers have not increased is very relevant.
Not necessarily the case because demographics are never homogeneously distributed across territory. If the rape/crime numbers didn't fall uniformly across the whole population, then you could have scenarios where crime fell by 20% in some regions and increased 25% in others, and overall it would be constant. I'm making up the numbers to illustrate.
More data would make your point much more convincing.