Much election law interpretation is grounded in unreality. For instance, lawmakers, voters, and election workers commonly behave and react from the foundational premise that aggregated printed or electronic data are presumptively free of error. So, for example, lawmakers will specify that the statewide voter registration database should be compared to the Social Security Administration’s “death index” in order to identify voters who are possibly deceased. Or, proceeding from a similar institutional faith in data, a poll worker may presume that a voter who is listed as having already voted in the election is ineligible, even to the point of treating that voter’s option to vote provisionally in order to clear up the mistake as a kind of morally suspect technical evasion.
This is all notwithstanding our intellectual awareness of the gross errors that riddle large collections of data. Every item of data in a large electronic database was typed into that database by a fallible human being, often as one step in a long chain of transactions that magnify errors. If one compares the statewide voter registration list to the death index, and if one takes the product of that tragicomic Kafkaesque collision of typos, misprints, electronic hiccups, syntax errors, missing fields, and general noise as being somehow “accurate,” then one lacks a common-sense understanding of information theory.
The legal presumption in circumstances where data is being compared, and where discrepancies are weighed, should be this: Presume in favor of the accuracy of the original source. Unless rebutted, the statements of the person that supplied the data that was ultimately aggregated into some sort of database should carry more weight than the database itself.