OVV Understands it Blundered: Time to Retract Publicly

The Observatorio Venezolano de Violencia's deeply flawed estimate of violent deaths in Venezuela is damaging our public sphere. Time to retract and correct.

The violent death rate estimate published by Observatorio Venezolano de Violencia (OVV) must be one of the two or three most cited figures about Venezuela. OVV’s contention that 27,875 Venezuelans died violently in 2015 has been cited by virtually everyone, from Time magazine and Associated Press to Venezuelanalysis, Wikipedia, Vice News and The Miami Herald: everyone.

The OVV estimate is a totem, a touchstone of the way sophisticated analysts understand Venezuela.

Later today, Caracas Chronicles will publish the result of a detailed analysis by Professor Dorothy Kronick of the University of Pennsylvania that demonstrates that figure is very likely wrong. And not by a little, by a lot. As best as we can tell, OVV overestimates the violent death rate in Venezuela by a staggering 6,000 victims or so. That makes their 2015 estimate a whopping 29% higher than ours (90 per 100,000 versus 70 per 100,000).

Over the last four months, I’ve had front row seats to Professor Kronick’s investigation into what exactly went wrong with the OVV estimate (and yes, calling a close friend like Dorothy “Professor Kronick” sure feels weird, but in this context, she is very much Professor Kronick.)

Her deep dive into how this happened had all the hallmarks of a good detective story: the careful sifting through clues, the false starts, the doubts, the hunt for corroborating details, the vital push-along by a well-placed source, and finally the break in the case, that a-ha moment when it all falls into place.

The culprit in the OVV case, we’ve come to see, was not so much malice as carelessness. It’s clear, of course, that if CICPC simply did its job and published the statistics it already collects, none of this need have happened. But official opacity doesn’t excuse sloppiness: just the opposite. In a data poor environment, researchers need to be more scrupulous, not less.

You can read Dr. Kronick’s report for a detailed explanation of what went wrong, and how. Scrambling for data that was getting harder and harder to obtain, OVV researchers made a major blunder. They built their estimate on the basis of data whose definition was unclear, whose provenance is still unclear, and from a source even they now admit they didn’t entirely trust. As a result, they double-counted a substantial number of violent deaths in 2013, showing a big upswing in homicides at a time when the murder rate was actually falling somewhat.

It was the kind of mistake that would be perfectly understandable and not a big deal in the hands of anyone else. For an academic research body whose whole purpose is to quantify violence in Venezuela, though, it amounts to serious malpractice.

But it didn’t end with the blunder over the 2013 data, because OVV’s subsequent use of projections ensured that the 2013 mistake was amplified over time, making the in 2014 estimate more wrong than that of 2013, and the and 2015 even wronger.

Each January OVV puts out a new murder-rate estimate for the previous year, to a flurry of domestic and international media attention. The fast-growing trend contributed to a sense of an out-of-control society. It fueled political tension, as opposition political leaders brandished it as more evidence of government lies. In reality, lost in the hyperpolarized melodrama, the violent death rate has actually leveled off somewhat since 2012.

Professor Kronick was adamant that OVV needed to be given multiple chances to review and comment on her findings.

“Maybe it’s all just a misunderstanding,” she kept telling me, “maybe I made some mistake.” She checked and rechecked, going far, very far, to give OVV the benefit of the doubt, to allow it to put forward a defense, or failing that, a correction.

Their response, when it came, was hard to figure out.

OVV lead statistician Alberto Camardiel reviewed Professor Kronick’s work, agreed she had accurately portrayed OVV’s methodology…and left it at that.

I find that response unacceptable.

If OVV understands and accepts that its murder-rate estimate is profoundly flawed, surely it follows they must issue a public retraction along with a correction.

Millions of people in Venezuela and around the world have been misinformed by OVV’s blunder. The dozens of well intentioned journalists who have propagated a mistake in repeating OVV’s deeply flawed estimate are entitled to an explanation about what happened, a retraction and a correction.

None of this is to suggest that the violence epidemic in Venezuela is anything less than catastrophic. It is. But we can’t begin to face up to that catastrophe without a careful understanding of its scale, and in that quest, OVV has done our country a real disservice.