>>13324734I wouldn't assume your conclusion about suppression is correct. It is likely both months have suppression on them to varying degree, although granted, only big companies/organisations do fancier meddling. and you'd be surprised how quickly you can identify things with only a little bit of data and a little bit of motivation.
had to nearly destroy half the dataset once to keep enough anonimity to not identify some people.
ONS for instance are quite open about the fact they randomize the data at a more granular level (making sure it stays the same at higher levels), so it's possible something similar could have happened in your case.
if you want to do academic research, you have to explain exactly what you did and how you inferred the data. again, keep in mind, you're making assumptions. the reasons for differences between months could be due to one month needing some anonimization, or it could be both had same levels, but one month the reporting was delayed and some numbers rolled over, or there could be something completely different going on.
whatever you do, explain what you did with the data, and allow for variations.
you're fine to use it, because it's out in the open. if anyone gets in trouble, it will be the person making it public, which is why they went through trouble of making it anonymous in the first place. just don't be facebook/google about it, and you'll be fine.
