consulting write-up

I took my first statistical consulting class this semester, and even though I had done some informal consulting through my research positions, this experience felt noticeably different. My initial emotions were fear and pressure, pressure to know all the statistics, all the methods, all the models, and all the tests. If mastery of every technique was what it took to succeed, I felt I was already positioned to fail. But fortunately, through Professor Schuler classes and the insights shared by our guest lecturers, I quickly learned that successful consulting isn’t about having an encyclopedic memory of methods. Instead, it is about cultivating the ability to think through problems carefully and calmly, to listen attentively, and to guide clients toward clarity about where they are and what they need to move forward. That realization was both liberating and empowering. It shifted my focus from trying to prove what I know to developing the skills that truly matter: asking good questions, identifying the core of a client’s challenge, and helping them see a path from confusion to progress. This is the bulk of what I have been learning and implementing throughout my consulting sessions and it has transformed not only how I approach clients, but also how I see myself as a collaborator and problem-solver. Below I share some nuggets and lessons that I have learned from my own experience.

First and foremost, I would say that preparation is key. It makes an enormous difference to walk into a consultation with at least a basic understanding of the client’s research question and the kind of support they are seeking. For me, reviewing the information provided through the sign-up form, even briefly, has been incredibly valuable. It gives me the opportunity to do a bit of homework beforehand, familiarize myself with the context, and anticipate the types of methods or explanations that might be relevant. This preparation has significantly boosted my confidence and has allowed the consultation to flow more smoothly and productively. I therefore highly recommend making space for some level of preparation whenever possible.

In this emerging age of AI, where anyone can access and apply off-the-shelf statistical or machine learning methods, what truly sets statisticians apart is our ability to think from first principles. Our value does not lie in simply running models, but in taking a step back to calmly reason through what is being done, why it is being done, and what assumptions and considerations must be accounted for. This kind of principled thinking is essential in consulting. Many of these considerations are blind spots for clients, but they should not be blind spots for us. During our consulting classes, I have consistently observed Professor Schuler model this approach. He takes the time to pause, reflect, and walk us through a mental exercise: What exactly is the research question? How else might we look at it? What concerns or pitfalls arise if we choose one method over another? Thinking carefully about the data-generating process,how each variable came to be, whether data are missing and why, how many time points exist, and the ordering of events, can be the difference between a naïve analysis and a rigorous one. This emphasis has been incredibly meaningful for me and continues to reinforce that we must understand the research question deeply before diving into any toolbox of methods. In my own consulting, I now make a deliberate effort to have clients clearly articulate what they hope to achieve with their research. This simple step has helped me work with them to identify whether they are asking a descriptive, causal, or predictive question, and in turn, determine which methods are most appropriate. This practice not only grounds the analysis, but also builds clarity, trust, and alignment between the client and consultant.

As you will quickly notice, statistical consults are as dynamic and varied as they come. One day you may be working with chemical compound measurements in environmental science, the next day analyzing biomarker data in genetics, fMRI time-series, or complex survey data. Every problem is unique, and many will fall far outside your immediate expertise. In these situations, your client becomes the subject-matter expert, and it is not only acceptable but essential to ask questions, even ones that may seem basic or obvious. Clients often appreciate these clarifying questions because they signal genuine engagement and a desire to understand the problem fully rather than making assumptions. The more clearly you grasp the scientific context, data structure, and underlying processes, the better equipped you are to guide them toward an appropriate and rigorous analytical approach. Embracing curiosity and humility is a strength in consulting. It ensures that your statistical expertise is applied thoughtfully, accurately, and in ways that truly support the client’s goals.

My next piece of advice is one that came to me as a genuine realization. You are not expected to know everything, and that is completely okay. Don’t hesitate to make use of resources to familiarize yourself with new or unfamiliar approaches. Many times, clients will come to you with a method they were told to use by their advisor or colleague; something you may not have encountered before or that they themselves may not fully understand. Look at this as a shared opportunity to explore and learn together. Remember, it is perfectly acceptable to look something up during a consultation, to think aloud with the client, or to say that you need more time and will follow up later. Many times, less is more. Early in my consulting experience, I often fell into the trap of sharing everything I knew with a client all at once. While this came from a place of enthusiasm and wanting to be helpful, it wasn’t fair to them. Most clients come in with limited statistical training, and being presented with a long list of advanced methods can feel overwhelming rather than empowering. Instead of giving clarity, it can create confusion and discourage them from engaging with the analysis. As I gained more experience, I learned the value of starting simple and only building up to more sophisticated approaches as the client demonstrated interest or as the problem required. For instance, if a client wanted to estimate a causal relationship using repeated measures data and came in thinking about GEE, I would begin right there, walking through the strengths and limitations of GEE in plain language. If they later expressed curiosity about more rigorous or flexible approaches, I would then introduce methods like longitudinal TMLE. I often shared additional explanations or helpful resources in a follow-up email, giving them time to absorb the concepts at their own pace.

It’s an important and often underestimated skill to be able to communicate statistical concepts to a non-technical audience. No matter how sophisticated the method is, clients will ultimately place greater value on your ability to explain with clarity, guiding them through the process, and ensuring they feel informed and included. Effective communication builds trust and helps your clients appreciate the relevance of your recommendations. This often means going beyond technical jargon and meeting people where they are. If necessary, make use of visual aids on the whiteboard, simple diagrams, intuitive analogies, and clear, accessible plots. These tools can bridge the gap between complex methods and practical understanding. The goal isn’t just to demonstrate your expertise, but to translate it into meaningful information that empowers your clients to make confident, informed decisions. Ultimately, I view consulting as a service to the BPH community and beyond. Sometimes the most valuable thing we offer is simply being present, listening carefully, asking thoughtful questions, and helping clients clarify their ideas. These moments of engagement are often deeply appreciated, even before any formal analysis is done. When we are able to arrive at feasible solutions, suggest methodological approaches, or guide clients toward the next step in their research, that is an added benefit. There are definitely consultations that leave me feeling less knowledgeable than I expected or momentarily aware of my own limitations. But even in those moments, I walk away having learned something new about a field, a dataset, or an emerging method. Consulting gives me a window into the exciting and diverse research happening all around me. It reminds me that being part of a collaborative academic community means continuously growing, asking questions, and supporting one another.