I don't think you can really conclude that it works based on the linked research data. I'm assuming you're referring to table 12, where they compare the percent

I don't think you can really conclude that it works based on the linked research data. I'm assuming you're referring to table 12, where they compare the percentages of incidents with the pre-Nordic model study. I think you cannot compare the percentages (something that the author acknowledges) because in the 2012 study a time limitation was set when reporting on the occurrence of an incident (only report it if it occurred since introduction of Nordic model), but this time limitation was not set in the other study (report it if it happened at any point in time in the past). You cannot compare victim-percentages of incidents that happened over a period of three years with victim-percentages of incidents that happened over an indefinite period of time. Claiming that the Nordic model works based on such a comparison is like claiming that the birth rate has dropped in the last three years because the total number of children born since 2013 is less than the total number of children born before 2013.