Reply #29: Thanks, but

[Febble]

where do either of those plots tell me which states had paper trails and which didn't?

New York certainly had a very large exit poll discrepancy - and no paper trail (all levers). New Hampshire also had a large exit poll discrepancy, and as far as I know was mostly optical scanners, AND it had a recount in which the recounted precincts were selected on the basis of aberrant results - and those who requested the recount were satisfied with it. Florida was half paperless DREs, probably crooked, but didn't have a particularly large exit poll discrepancy; Georgia was DREs, but again, the redshift was fairly small. Delaware was DREs AND had a large exit poll discrepancy. But as far as I can see it's the only state that makes the point, and I haven't seen a great deal of interest in investigating fraud in Delaware. New Mexico had DREs of course, and I happen to think, for reasons other than the exit poll discrepancy that Kerry won it. But the exit poll discrepancy was not particularly large. Tennessee and Colorado both had a substantial proportion of DRE precincts, but both had a slight but insignificant blueshift in the exit poll.

My concern, Botany, is not to downplay the problems with which US democracy is riddled, but to get a decent handle on what those problems actually are, and what there is good evidence for. I don't think any case is best made with faulty arguments, let alone faulty data, and I don't think problems are easily solved without a good problem statement.

In the analysis I did for Mitofsky, at precinct level, where it matters, there was no tendency for the exit poll discrepancy to be higher where DREs were used. As I've said elsewhere, the highest discrepancies were in lever precincts, followed by punchcards. When HCPB precincts were compared with precincts in similar neighbourhoods (rural, suburban, small town) there was no statistically significant difference between paper and other voting technologies. This finding was reported in the Edison-Mitofsky report, and I repeated it, rather more rigorously, myself.