Domain experts are your friend

by Billy Caughey

Domain experts... what do they know???

Starting out as a statistician, I thought I was big stuff. I had an ego from studying an advanced topic. My math skills made me powerful and desired as a members of various teams. I was just looking for that one opportunity!

While a study coordinator at the Rocky Mountain Center for Occupational and Environmental Health, I had such an opportunity. I was asked by one of the researchers to look into something. I fiercely dug into the data. I found associations and made great calculations. I put my findings into spreadsheet and, with a chip on my shoulder, fired off my fingers.

I can't begin to stress the amount ego I had at this point. I was a study coordinator and a statistical research assistant. I was the man, or so I thought. Soon after I fired off my findings, I received an email back. The researcher was confused by my findings and not sure if they were correct. I tried to implement some level of composure, and decided to redo my analysis. I knew my numbers were right, but I was gonna double check my numbers. 

After redoing my numbers, I sent them back off to the researcher. To my dismay, this researcher still believed my numbers were off. Then, my desk phone rang. It was this researcher. He had to audacity to say my numbers were wrong and then call me to tell me about it?! I have to admit, I was not happy about.

The researcher calmly told me what he was concerned about. We talked about the calculation. The researcher didn't raise his voice at all. In fact, the researcher took the time to teach me a five minute crash course on the measure, why we used it and its application in medicine. After the phone call, I recalculated the numbers, sent them off, and the researcher approved of the numbers. In the end, it was my calculations that were off, not his knowledge.

Turns out... they know a lot!

I marvel at that experience to this day. That was a great learning moment in my career as a statistician. I learned I am not all that. In fact, I am very much powerless in certain domains because I just don't have the experience or knowledge of what is going on.

Stephen R. Covey spoke of the power of "interdependence" in his book, The 7 Habits of Highly Effective People. I have watched a lot of data folks come and go with the attitude of 'I'. Some of my favorite phrases are:

• I'm the one in charge of the data!
• I look at the data every day.
• I put this table to together, and what do you know?

All of these statements from the statistician, data scientist, or data engineer make it quite clear: they haven't figured out the power of 'we'. When statisticians and domain experts are able to sit down together, the world takes on a new view. The domain expert can bring the science and knowledge together with the analytical methods of the statistician. Together, they tell a beautiful story. This story usually goes a long way to teaching the world, expanding a topic, supporting findings of another research group, or challenging the norm. Whatever the story is, the domain expert and the statistician have to work together.

And they want to work with you!

In closing, I want to share a very personal experience I had with someone I consider a mentor. I won't share names to ensure anonymity. On one occasion, I spoke to this mentor about a PhD program he was the director of. Some time previous to this occasion, we had talked about a statistics focus in his program. He spoke highly of the need of statistics in his field and how skilled practitioners of statistics could improve the field. 

I decided to press my chance to press for entry to this program. I told him I wanted to be the first student through this statistics focus. I was ready to set the stage and the bar for future students to come. After all, I was the one he spoken to about this topic.

This mentor sat back in his chair and said no. I was devastated! How could he say no? I pressed him for his reasoning. He simply said that he had worked with me long enough to trust me. I had a good understanding of the field, but a great understanding of statistics. He wanted me to sit opposite of him as the statistician to him, the domain expert. He knew I would be honest with him and he with me. He didn't need me to be the domain expert, he needed me to help explain his domain with statistics, mathematics, and machine learning. That conversation continues to ring in my mind today as it did that day.

Caution though...

This doesn't give you license to not learn about the domain you operate in. Your calculations will mean nothing if you don't know what they are trying to model. Become the generalist. Gaining a general knowledge of the topic area will go a long way. Your general understanding of the field will provide insight to the models and analyses you are performing. Then, what becomes fun is when something happens that isn't in your general knowledge. When you loop in an expert, they may say, "Oh yeah, let me explain" and your knowledge grows. On the other hand, the expert may not know what is going on. At that point, you are beginning a journey to add to the knowledge of the field!