In the world of high-impact research, the most important discoveries often hinge on a single question: Are we looking at the data the right way?

For a group of world-renowned University of Minnesota (UMN)  lion researchers, the answer came only after they sat down with Aaron Rendahl, associate professor of statistics and informatics. The team was investigating why lions attack humans more frequently at certain times of the month, suspecting it was linked to the brightness of the moon. But the data was noisy until Rendahl spotted a layer of nuance that had been overlooked.

"The amount of the moonlight is the same whether it's waxing or waning, but the timing of it is different," Rendahl explains. While one moon phase might be bright all night, another is dark in the early evening—exactly when humans are most active. "We did this analysis, and it turned out to really matter."

The resulting study from the University’s College of Biological Sciences was featured in National Geographic. That project, which happened shortly before  Rendahl joined the College of Veterinary Medicine (CVM) full-time, illustrates the toolbox he now brings to CVM researchers every day.

"People do what they know how to do," Rendahl says. "When you’re familiar with more kinds of analyses, your toolbox is bigger, and you can say, 'Well, wait a minute, we can do something more sophisticated than that.'"

Today, as CVM’s dedicated, in-house statistician, Rendahl’s time is split between teaching graduate courses, such as Essential Statistics for Life Sciences, and collaborating directly with faculty, clinicians, residents, and students. While Rendahl can jump into a project at any stage, he emphasizes that the best results happen when he is involved before the first data point is ever recorded.

"Best practice is that someone will reach out to me when they're planning their study," Rendahl says. "We'll chat about what they're planning and make an analysis plan... so that when they're collecting the data, they can collect it in a way that is going to make sense for an analysis that will help answer the questions they're going to want to answer."

Designing for discovery

One of the most frequent challenges Rendahl helps solve is how to handle different levels of data—a concept that can get "really complex, really fast."

He uses a diet study for pigs as a classic example of why the obvious math isn't always the right math. If you are feeding animals by the pen, everyone in that pen gets the same diet. "They might be more similar to each other because they're in the same pen," Rendahl says. "So even if you have data on the individuals, you need to be looking at the pen-to-pen variability to make an inference about the diet."

This type of multi-level variability is a recurring theme in his work, whether he is looking at true experiments or observational studies in the veterinary clinic. In the clinic, researchers often look back through records to compare one group of animals to another.  These types of analyses require sophisticated statistical designs to ensure the comparison is fair.

For example, if a researcher is comparing two different heart medications, they have to ensure they aren't accidentally comparing older, sicker dogs who received one drug against younger, healthier dogs who received the other. Rendahl helps build models that account for those differences, ensuring the "signal" of the treatment isn't lost in the "noise" of the patients' backgrounds.

Beyond the design of the experiment, Rendahl is often the one asking the most practical question of all: Is there enough data to even start? "We also talk about sample size and power," Rendahl explains. "Let's make sure you are collecting enough data to be able to [answer your questions]." By identifying these complexities during the planning phase, Rendahl ensures that the research remains sensible for the questions the team wants to answer—and that the results are robust enough to stand up to the peer-review process for publication.

A window into many worlds

Rendahl describes himself as a generalist, and his daily work provides a unique window into the sheer variety of research happening across the college. On any given day, he might be working with clinician-researchers at the veterinary hospital to help pets like dogs and cats, or working with livestock researchers to study calf health and well-being. He even assists basic science researchers who use animal models to answer fundamental questions about the immune system and virology.

For Rendahl, this diversity is the best part of the job. "I like being able to use mathematics to understand the world around me," he says. "One of the neatest things about being a statistician is getting to learn about all the cool research that people do, and people here do such interesting work. It's really fun to be able to be part of that process."

Whether he is learning about omics (genomics, proteomics, and metabolomics) alongside his colleagues or helping surgeons analyze ways to repair knee joints, his goal is to make sure the science is as rigorous as it is interesting. "I've always liked science and understanding scientific questions, and understanding how things work," Rendahl says. "I didn't really have any particular direction that was going to go until I started collaborating with people here."