In March, 1954, newspapers in Seattle reported that some car windshields were damaged in a city 80 miles away. Vandalism was suspected. But then something strange happened.

People started to find car windshields speckled with tiny pits. Reports multiplied. Within a couple of weeks, the police had taken 242 calls from concerned citizens reporting damage to more than 1,000 cars.

The United States had recently detonated the first hydrogen bomb in the South Pacific. Could nuclear fallout be doing this? Fear mounted. The mayor of Seattle declared that local police could not cope. He called on president Dwight Eisenhower to take charge.

Then it occurred to some people that they were looking at windshields, rather than through them, for the first time. Maybe those tiny pits had always been there. Maybe they'd just never noticed them before. Maybe they were fooling themselves.

When this rather more mundane theory appeared in newspapers, calls to the police about pitted windshields abruptly declined. And then stopped altogether.

What became known in sociology textbooks as The Seattle Windshield Pitting Epidemic is a classic illustration of how easily mass delusion can set in if we don't think carefully about evidence. We need to remember that warning now, as reports of electoral fraud and chicanery multiply rapidly.

No, I am not saying there is nothing to those reports. On the contrary. There is solid corroborating evidence in enough cases that we can now be confident that phone scams were real and widespread.

But since this story was broken last week by the intrepid journalists Glen McGregor and Stephen Maher, reporters, politicians, and party officials have been working feverishly to find more incidents. In that atmosphere, the risk of something like a Seattle Windshield Pitting Epidemic becomes very real.

The Conservatives now claim that they, too, were victims of phone scams. Is this credible? It's hard to say. Of course these claims are blatantly self-interested at a time when the opposition and many Canadians are pointing the finger of blame at the Conservatives. And yet we can't arbitrarily dismiss claims simply because they're self-interested. A self-interested claim can be a true claim. And remember that the opposition's proliferating claims of victimization are also self-interested.

But it's also important not to take any of these claims at face value. There's the self-interest, obviously. But there's also the "pitted windshield" phenomenon.

If we look very closely at the 2011 election and discover countless examples of inappropriate or illegal use of phone calling, we may conclude there was an unprecedented effort to manipulate voters and rig the election. And given the novelty of the tactic, and the scale of its deployment, we could also conclude that a sophisticated organization with extensive resources was responsible for most or all of it.

But we didn't make a similar examination of past elections. What if we had? Maybe we would have discovered a similar level of inappropriate or illegal use of phone calling. And if that were the case, it would suggest that this behaviour is endemic to Canadian electioneering. And it would be much less plausible that a single sophisticated organization was responsible in 2011.

Adding to the uncertainty is the simple fact that we are now dealing with memories.

"Did you receive a #robocall or harassing phone call during the election?" a reporter asked on Twitter. "Tweet me!" Seems a reasonable thing to ask. But the election was almost a year ago. And the media are now saturated with the narrative about phone scams during the election. Inevitably, many people's memories of phone calls received during the election will adjust to fit their current perception of what happened during the election. That's Psych 101. They will have recollections that are clear, compelling, and wrong.

And then there's the feedback effect, in which reports encourage more reporting, which strengthens misperceptions, which encourages more reporting, which pumps up the political volume, which encourages more reporting, and so on. It's a common phenomenon. And as I documented in a book several years ago, it can, taken to the extreme, firmly convince most people of what just ain't so.

Add this all up and the danger is obvious: A real and important phenomenon that needs to be dragged into the light could instead be buried under an avalanche of faulty recollections, self-serving illusions, and plain old lies.

The only way to avoid that is painstaking investigation. One by one, allegations must be carefully examined. Corroboration and context are essential. Subsequent analysis must be done solely on the basis of verified facts.

It would be slow, difficult work. It requires major resources and legal power. It also requires independence, both in reality and in perception.

Elections Canada isn't up to that job. Neither is the RCMP. The Conservatives ensured that with their controversial centralization of political control over the Mounties. And remember that in 2005 Stephen Harper personally accused the RCMP of not investigating the Liberals because the Mounties were under the government's political control. Whatever the reality, the RCMP simply wouldn't have the appearance of impartiality that is essential to do the job.

That leaves only one option.

The Conservatives insist they want the truth to be exposed. If that's true, they must appoint a fully independent, fully empowered judicial inquiry.

And why shouldn't they? To paraphrase what many Conservatives said about warrantless Internet surveillance, they have nothing to fear if they have nothing to hide.