Behind the expert testimony in the ‘Fitbit murder’ trial

A Connecticut jury found Richard Dabate guilty of murder in the killing of his wife, after a five-week trial that centered on data from her device. Richard said a man dressed in camouflage broke into their home and shot a woman. After Richard said the break-in happened, data from the device showed movement for around an hour.

Prosecutors called on a professor at Columbia University Medical Center to make the case that the data from the fitness tracker helped prove Richard's guilt. Studies have been done to verify the accuracy of Fitbits.

In criminal trials, Diaz often testifies as an expert witness. It's gratifying to participate in trials, but it's also a challenge, he told The Verge. The way he thinks about the data as a scientist is different from the way he thinks about it in a courtroom.

The devices are generally accurate, but they aren't perfect. It's rare to get an answer that comes with 100 percent certainty for scientific research. Science has some amount of statistical error. The law wants to know if something happened beyond a reasonable doubt. That has been one of the challenges and the hurdles to clear.

In an interview with The Verge, he talked about how he navigates the tensions between science and the legal system in testifying in trials.

The following was lightly edited for clarity.

What is the difference between what a scientist wants to know about a device and what a jury wants to know?

We care about things like if I only took 92 steps, or if I only took 100 steps. How much does that matter in the criminal cases? It's a matter of if the device can detect movement. That is good enough for the case. If they moved, it doesn't matter if it's 100 steps or 92. There was a lot of back and forth about the error rate with this case. The error rate was on the number of steps someone took, not the rate if they were moving. That is a very different story.

Large sample sizes are used in a lot of the studies I work with. There is so much noise with so many people. With a criminal case, there is only one person, so noise can bring in doubt about what really happened. There is a level of precision you want to have around in criminal cases. If there is an error because you have a large sample size, it can be removed from science. It has been difficult to figure out how you convey that these devices can tell you what's happening.

How do you make the science work in a criminal case?

You can use the things that science sees as errors. If you wear a device on your wrist and brush your teeth, there are cases where the device misregisters that as steps. That is an error in science. It shows that it is a high-sensitivity device, which can be helpful with some of these cases. It is a strength to use these devices in this context because of their high sensitivity. We should be pretty sure that a person is not moving when there is no motion detected or recording steps.

We had to shift what we looked at in the science to say, "Well, what is the most important thing in the context of this criminal case?" What does that mean in this situation?

In a criminal trial, you are asked to testify for either side or the other. What is it like to be a scientist?

I have never been hired by the defense. Some prosecutors tried to bias my interpretation. They will have the perspective that they need someone to come in and support the case if they think this person is guilty. I want to give a balanced interpretation of what the science means and how it can be applied in the situation. I could say things that would support the case for the prosecutor. How do you balance that with acknowledging that there is error and that it isn't a perfect device?

The prosecution doesn't want to highlight the errors. From the defense side, I've noticed they aren't asking the right questions. They don't know what to say about the limitations of the device. I can only answer the questions that are posed to me.

How you think about your research has changed after this experience.

I would want to conduct more scientific studies if this is something I go further down the road into. The science that exists is not designed to answer the important questions in a lot of the situations that I have seen. Is it possible to get from a bed to a bathroom without a device? I think it would strengthen my ability to give expert testimony. It would be helpful to have better answers to the questions that come up in a criminal case that don't involve science. It is a side gig for now. This is outside of what I do on a day-to-day basis.