I met David for the first time at a scientific meeting in Dublin in 1995. That serendipitous encounter was one of the greatest blessings of my life and marked the beginning of a long-lasting friendship with David. At the time, I was an unknown junior researcher. Nevertheless, David, with his characteristic warmth, care for young investigators, and unassuming humility, showed interest in learning about my background, my life in Argentina, and my career journey in the United States.
My presentation at the Irish meeting focused on semiparametric analysis with missing data. During my talk, I posed a question for which I had no answer and I illustrated, through an example, that the problem could also arise in a parametric analysis. I argued that, in its simplest form, the question boiled down to how to make inferences about a smooth function of a parameter indexing an identifiable parametric model with a singular information matrix at a given point. David, who was in the audience, approached me after my talk and told me that he would think about my problem. Just 2 short weeks later, I received David's handwritten notes in the mail, detailing his initial thoughts on my problem. In them, he showed the bimodality of the likelihood in a neighborhood of the singular law, under certain scenarios, and explained the unusual consequences that this bimodality had on the power for detecting the directionality and magnitudes of nonstandard local departures from the target parameter at the singular law. His analysis was simply exquisite; he offered a solution to a complicated technical problem that made it seem easy and intuitive. This initial communication sparked a collaboration that proved to be one of the most enriching and educational experiences of my life. Not only did I have the privilege of interacting with, and learning from, a brilliant mind and a giant in the field of statistics, but even more significantly, I had the chance to develop a cherished and long-lasting friendship with a truly special person: generous, kind, sensible, and with a delightful sense of humor.
David always made me feel what I could never be: his intellectual peer. But his support and encouragement were a pillar in which I sought refuge every time things in my academic life didn't go as well as I had hoped.
I have an abundance of cherished memories from our encounters in both Oxford and Boston, where I was a faculty member at Harvard Biostatistics. Of all the moments I recall, one in particular stands out for its illustration of David's loving and kindhearted nature. It was during one of his visits to Boston when a ceremonial dinner was being held in his honor at Harvard. Despite the festivities, David requested that I secretly accompany him to stealthily slip away on a detour to MIT. There, he extended his heartfelt congratulations to a Venezuelan scientist friend who had just won a prestigious prize. The joy and pride David radiated for his friend's achievement was a true testament to his caring and selfless character. Following this touching moment, we continued our work over burgers and fries at a local diner. These lighthearted moments paint a picture of David's character and persona: he found greater joy in attending a ceremony in honor of a friend than his own, and he preferred work over ceremonies.
In 2000, my relocation to Argentina from the United States limited our opportunities for face-to-face interactions. Despite this, we kept in touch through email. Whenever there was an extended period of silence, David always took the initiative to reconnect by sending the first email. Our conversations delved into our passion for statistics, but also included personal anecdotes, some of which dated back to our upbringings. David's life stories underscored the exceptional individual beyond his remarkable intellect, deepening my admiration and affection for him.
In the winter of 2020, just before the start of the pandemic, I met David in Oxford one last time and spent a full day with him. As always, on that day we engaged in lively and intellectually stimulating conversations. We exchanged thoughts and ideas about our current projects, delved into some of the principles of inference, and shared insights on the evolving field of data science. As it often happened during our encounters, we agreed to disagree on the role of causality in statistics. However, David did mention, as he had done on several occasions in the past, that he deeply regretted not paying attention to Jamie Robins when he visited him in Oxford in 1983 to explain his article on the foundations of confounding that had been rejected from Biometrika.
I was struck by David's vitality, concentration, and boundless enthusiasm for learning new ideas, despite being in his mid-90s. In fact, his eagerness to explore and discover was greater than mine, even though I was more than three decades younger! As I bid farewell to him that evening, I knew it would be a while before I could return to the United Kingdom, but I was confident I would see David again. I had a feeling that he was an immortal being, and our paths would cross once more. Alas, that was not to be.
Andrea Rotnitzky has no financial or non-financial disclosures to share for this article.
©2023 Andrea Rotnitzky. This article is licensed under a Creative Commons Attribution (CC BY 4.0) International license, except where otherwise indicated with respect to particular material included in the article.