President Trump’s announcement that he will launch an investigation of voter fraud is interesting for many reasons. Some of these have been well-documented, such as that he continues to believe massive voter fraud caused his popular-vote loss, and that the main “evidence” cited for such fraud has been thoroughly debunked.
In the context of other recent announcements, it’s also interesting because it may offer an opening for demonstrating the value of evidence-based, systematic inquiry: that is, of science as a basis for policy.
The state of the debate right now is this. Trump and conservatives believe that many people ineligible to vote do so. They are ineligible because they are “illegal immigrants,” registered in two or more states, dead, fictional, etc., and for whatever reason and through whatever mechanism they voted. That threatens the principle of popular sovereignty because the parameters of the populace granted sovereignty don’t adequately resemble the parameters of the legitimately-voting populace.
Meanwhile, Democrats and liberals point out (correctly) that there is no existing evidence of actual fraudulent voting on any significant scale. Furthermore, they argue, any policy remedy that would address fraudulent voting would likely reduce the ability of legitimate voters to vote. That, too, threatens the principle of popular sovereignty for the very same reason: because the parameters of the populace granted sovereignty would not adequately resemble the parameters of the legitimately-voting populace.
However, without a systematic study, absence of evidence is not evidence of absence; furthermore, without a systematic study of the effects of voter ID and similar on reducing voting by legitimate voters, it is impossible to know the extent of that effect.
Right now, the “debate” is largely on the terrain of opinion and anecdote. Generally, neither side provides evidence that is convincing to the other. (See below for a version of the anecdote approach sent to me today.)
So what is one to do with competing claims about events in the real world — events that are systematically counted and recorded, in particular? Science, that’s what! Here are a few hypotheses:
- Enough ineligible voters voted in the 2016 election to constitute a difference in election outcome.
- Enough ineligible voters voted in the 2016 election to constitute a difference in the presidential popular vote direction (e.g., a Trump popular-vote win instead of the documented Clinton win).
- There were significantly fewer ineligible votes cast in jurisdictions having Voter ID and similar voting restrictions than in jurisdictions lacking these restrictions.
- There were significantly more eligible, would-be voters prevented from voting in jurisdictions having Voter ID and similar voting restrictions than in jurisdictions lacking these restrictions.
My proposal is that the Trump Administration commission a large-scale, national study by a board of professional social scientists registered in both political parties. The study should address these or similar hypotheses using the best possible data. It can then be used to settle policy questions about election integrity and put to rest an ongoing dispute that is very damaging to democratic governance.
This approach, if successful, would have some important advantages. First, it could be used in other areas–particularly for congressional districting, which also goes directly to the question of popular sovereignty. Second, it would be an opportunity for the practice of science to demonstrate its value by providing reference to evidence, inquiry, and democratic standards that lie outside partisan positioning.
I recognize that this is unlikely in the current environment. But if, in fact, everyone actually believes that their claims are empirically correct, there should be no barrier to testing those claims rigorously and thoroughly.
PS – here’s the anecdotal graphic I mentioned above:
(Spoiler: all the assertions in the graphic have been debunked, which is precisely why viral anecdotes should not be used as evidence.)