0

I am trying to find a statistical method to identify anomalies in votes. I have a dataset including all the vote for every party in every voting table in every polling station. These votes are counted manually and therefore, subject to many human errors. Additionally, there are many parties to be counted. I would like to identify in which voting tables I could find anomalies in votes for a single party comparing it to the total votes that the party got in that polling stations. For example, if a party got 100,000 votes in the whole voting station, it could be strange to see 0 votes in one of the voting tables. Which method could I used to identify when the total votes for one party is anomalous

user2246905
  • 223
  • Not sure if this suits your case, but would it be reasonable to assume a uniform distribution across the tables at a polling station? In this case, you could use a chi-squared test (or some other test) for goodness-of-fit. – num_39 Mar 17 '22 at 16:27
  • Could you please define "voting table" and "polling station"? Is there just one vote per person? – frank Apr 08 '22 at 07:49
  • A polling station is a group of voting tables, and yes only one vote per person – user2246905 Apr 26 '22 at 19:59

0 Answers0