Transcript (Approximate)

I have a confession to make. I am obsessed with measurement. Totally and completely obsessed with measurement. I know what you are thinking now. “Man, measurement? That is probably the most boring topic in the history of human kind” well, I have heard that before. Mostly from my wife. 

I have decided to spend my life measuring first and foremost because I think we do a lousy job at it. We measure infrequently, late, and mostly extreme outcomes. For example, we do not measure sadness, we count suicides and cases of depression. We do not measures how upset a society is, we just observe them rioting. We do not evaluate the quality of a relationship, we only know when they break up. We do not measure the impact of pollution on a lake, only when the lake is lost. We are repeatedly late, too late. I think it is a miracle we actually make good decisions.

At MIT I am trying to change the way we measure many aspects of life. I am working on more than 20 things… measurement stress at the households, inflation, GDP, human trafficking, opioids consumption, discrimination in the labor force, the quality of a political system, team chemistry, etc. In all these areas I tend to use similar principles. To explain what I do let me use an example that we can all relate too. Women Empowerment: which I am working with Sandy Pentland at the media lab.

First, the bad measurement: how do we measure women empowerment in organizations? We use two variables: first we count women and sexual misconduct claims. 

Yes, we count women. For example, here we have about 50 percent! Much better than SF. You know, Miami rules!! Well, I think that makes no sense. Since when increasing the proportion of women from 30 to 40 percent guarantees better treatment? This statistic is only meaningful at the extremes: when you are close to zero or a hundred.

Our second measurement is counting the number of sexual harassment cases. I don’t know what you think, but to me this is the definition of too late. Furthermore, we only know about it when the media is outraged by it. 

I am not saying we should not pay attention to these statistics. We should. Because if you are not doing a good job in these variables you are for sure discriminating women. But the fact that these statistics are “great” means nothing. Three years ago, half of the actors in Hollywood were actresses, and there were no complains reported in the news. It is not the case that this got bad in three months - we just did not measured correctly.

How do I want to change the measurement. There are five pillars of the new measures I want to pursue. This is work a group of obsessed young people are conducting at MIT. We call it the Aggregate Confusion Project. They do all the hard work. Think about how to measure, and then go out and find ways to do so. Of course, I do absolutely the hardest part. I take credit for all their hard work. After all, I provide 90 percent of the motivation and 125 percent of the entertainment. 

The first two principles are about the collection of the data. I would like it to be non-intrusive and continuous. So, we can measure without mayor disruptions in the organizations. Today, most communications in business are electronically - conference calls and video conferences. And we can measure some interesting statistics such as interruptions and air time, and provide immediate feedback. Most of research has proved that men interrupt women 4 and 7 times more than men interrupt men. And it is also the case that women in a business environment have less airtime. We would like to verify and account for this in real time.

The dissemination of these measurements follows two principles: privacy protection and open source.

I want these apps to be distributed for free, and adopted by many. And even though I am working on identifying perpetuators with my research on human trafficking, for most of what I want to do I would like to capture the organizational and cultural behavior. The reason is that I believe that 99.9 percent of the people in the world are good. And if we measure their misbehavior I believe they would immediately change. 

A consequence of these principles is that the measures will be imperfect. I am fine with imperfection. Look at me! I prefer to imperfectly measure something relevant, than perfectly measure something completely irrelevant. 

Let me finish by highlighting your role. This research can only happen at MIT. My papers and the papers written in this research will be bad papers in economics, bad papers in psychology, bad papers in sociology, bad papers in computer science. Bad papers at everything and probably it will be under appreciated research. But I believe we need to do it. In sum,  this research is multidisciplinary, important, and thankless - that tends to spell MIT. 

We have been able to conduct this research the last 15 years thanks to the support of people like you. So, thanks for all the help you have provided us so far. We are not done, and we need more. So, let me thank you also for the creativity, inspiration, encouragement, and support you will continue to give us in the future. We desperately need your help. We can’t wait for the public sector to get their act together - and even if they get their act together it is not clear that the values of such organizations are aligned with the needs our society has. Therefore, it is our responsibility to create awareness and lead the changes we need. 

I believe that what most call unconscious biases, are indeed unmeasured biases. If this is the case, I am certain we can make our lives, our business, and our societies much better. To do so we need to understand them better. So, get your rulers out and dusted, and lets start measuring every conceivable thing. Thanks.