I've recently been quite shocked to learn that in the US there is a serious amount of election fraud - at least in the presidential elections. Are there any numbers on the magnitude of this?
I am deliberately not limiting this question to specific types of fraud, since there is quite a number of them, and I would be grateful to know about at least semi-serious calculations to estimate their effect, whether individually or all together. Examples are as follows.
- There is the design problem that it is completely unimportant whether 51% or 100% of the voters of a state vote for a candidate. I am aware that it is quite debatable to call this "fraud", but I am inclined to include this since it contradicts the equality of all votes as demanded by the human rights charta.
- Perhaps the worst problem: the introduction of loads of intransparent voting computers. These are often manufactured by Republican controlled companies, and there is not much of a chance of noticing flipping or miscounting votes. It is hard to verify those (but should not be!), but there are hints that there is something seriously wrong. The "red shift" - a statistically highly significant discrepancy between the official election results and polls done on the exit of the voting place - could prove this, but such polls are partially deliberately blocked or performed by the media without reliable disclosure if the results. (See also this question.)
There are also some troublesome things that do have an influence on the outcome, but cannot exactly be called fraud unless you somehow manage to prove malicious intent. The magnitude of their effect would be very interesting, nevertheless.
- The voting is held on tuesday such that many poor people have problems attending. Similar effects might have the sometimes several hours long queues at voting booths in poorer districts.
- Gerrymandering, Caging
- sometimes loosing hundreds of thousands of absentee ballots
- Super PACs that can intransparently spend unlimited money on political campaigns.
EDIT: This question was heavily edited to try to accomodate comments