Gfyuki

I'm currently writing a thesis based on Cyber Crime, however I'm unsure of the proper to compare/analyse my data sets to talk about them in my thesis.

One piece of data (https://www.pandasecurity.com/mediacenter/src/uploads/2014/07/Pandalabs-2015-anual-EN.pdf on page 9) it states that the 'infection rates' of Sweden is 20.88% (bottom 3 ranking), USA at 29.48% (middle ranking) and China (first rank) having 57.24%.

Another (http://www.virusradar.com/en/home/world) , uses a different measurement to define the 'threat rates', which is different to the one above, which has Sweden at 2.13%, US at 2.87%, and China at 15.17%

Another piece of data (https://www.symantec.com/content/dam/symantec/docs/reports/istr-22-2017-en.pdf
on page 50) states 'identity theft' in Sweden is 6 million(low ranking), US 791 million (high ranking), and China 11 million (middle ranking)

I'm unsure how to compare these, because by just looking with our eyes we can see Sweden is the lowest in all these statistics which I can use to further discuss my argument. However, these are clearly different measurements (one is overall infections, one is just identity theft, etc), and I have much more similar data which provides numbers/percentages of certain types of cyber crime by country.

My goal is to analyse the results so I can compare them to further the goals of the thesis about where cyber crime is most/least prominent and to create visualisations of this.

So would just simply ranking them be fine? (Sweden 3 because it was lowest all times?, USA 2 because it was second twice?, China 1 because it was first twice?) but that sounds very incorrect. Should I convert all my numeric data sets (like the identity theft one) to percentages then compare them all by percentage to rank/discuss accordingly? And compare them then even though they're different measurements of the same thing(country)?

Thanks for any advice.

You're right, you cannot directly compare the scores, since they are extracted from different samples, using different metrics and parameters, etc. Is exactly as comparing apples with airplanes.

But is not so wrong to play around with the placement in the rank: you are comparing placements with placements, so it makes perfectly sense. If Sweden is always at the last place, this definitely means something.

Maybe you can also think of weight those placements by some parameter/set of parameters which takes in account the subject of the ranks.
As example, seriousness of the threats (ex. infection by AdWare can be less important than identity theft, etc), coverage of the population by the authority which published the results, reputation of the authority, etc.

You can also extract uncertainties on those ranks, maybe based on the statistical samples taken in account by the authorities, and so on.

Cheers