Contact information
Megan Higgs, megan.higgs@critical-inference.com
Unique statistical consulting
The goal of my consulting work through Critical Inference LLC is to provide a unique kind of support to researchers, practitioners, and decision makers — anyone in the position of interpreting and assessing research or arguments based on statistical inference.
I focus on critically evaluating designs and inferences, assessing trustworthiness of results and conclusions, increasing awareness of limitations of statistical methods, and asking crucial follow up questions to research conclusions or decisions — as opposed to stereotypical statistical consulting focused on the technical aspects of “doing statistics” (e.g., routine data analysis, power analysis, writing up statistical results, etc.). For example, I will not take your data and simply return statistical results to you, though I am available for aspects such as identifying and articulating potential challenges in making inferences given your data, understanding underlying assumptions of potential statistical methods, designing future studies, or assisting in reviewing relevant work of others.
What I offer
The services I offer are very individualized — tailored to your needs, goals, and the context of the problem. However, they generally fall into one of the following categories.
Statistics coaching
I provide professional coaching (or tutoring) in statistical inference and methods. The goal is typically to provide support and structure for ongoing learning for professionals who have a desire to further their statistical education for personal or professional reasons. An individualized plan is developed based on goals and time available, and typically takes the form of assigned readings, tutorials, interactive lessons, and personalized discussions with me. This can be an effective way to continue your education in Statistics in the direction you choose, with individualized content and discussions, as well as flexible schedule.
Independent and external reviews
I provide independent, external reviews of research or protocols utilizing statistical methods. Reviews can take various forms – from an informal discussion to a formal report.
Guidance evaluating specific research or claims
Related to performing more formal independent and external reviews, I also review and evaluate research plans, disseminated research, or any claims based on statistical inference — from the perspective of a statistician passionate about science and how we go about justifying inferences and subsequent decisions. For example, I can provide a statistician’s perspective on methods and conclusions disseminated in a particular research paper, or even assist practitioners in weighing evidence for or against changing a practice. This type of work may involve close collaboration with you through discussion and meetings or take the form of written report describing my evaluation or recommendations.
Assistance in scientific writing about statistical inference
I assist in writing about statistical methods, associated results, and conclusions based on them. My interest is not in formulaic presentation of quantitative results, but lies in providing adequate transparency in methods, articulating underlying assumptions, and honestly conveying limitations of inferences based on the design and analysis.
Guidance for creative raw data visualization
While I will not “do your analysis,” I firmly believe in the importance of creative visualization of raw data as an integral part of the process of data analysis. I have extensive experience developing unique and creative displays of complex data and am happy to assist in the development of ideas and in some cases, providing programming in R statistical software to implement the ideas.
Seminars
Seminars are typically about one hour and involve less interaction for participants. These are closer to what would be expected of a lecture or academic talk. The potential range of topics is broad — from introductions to statistical inference to topics surrounding scientific inference, ethics, and philosophy of science in practice.
Roundtable discussions
If you have a group of people interested in delving into a particular topic related to the use of statistical methods and inference, I provide readings, organization, and facilitation for discussions. Topics can range from discussion of a specific paper or statistical norm in your field to discussions of philosophy of science and ethics related to use of statistical methods in practice. This guided-discussion format is less formal than the workshop and expected to be very participant driven. In addition to organizing and facilitating, I offer a different perspective from within-discipline peers and provide exposure to ideas that span discipline boundaries.
Workshops
I develop custom workshops based on clients’ interests, backgrounds, and needs. The minimum time commitment for a workshop is 2 hours and the hours can be spread out over multiple days. Some examples of topics are an overview of the history of Statistics in science; limitations of and myths surrounding statistical inference; recognizing common statistical pitfalls; philosophy of science and connections to Statistics; raising awareness and understanding of assumptions underlying particular approaches; in-depth discussion of a particular method to understand its benefits and limitations for a particular type of problem; background information about such things such as the “replication crisis”, etc. Workshops typically include intellectually stimulating readings, activities, and guided discussions. The total cost is typically negotiated for each request and depends on preparation needed for the specific topic and type of structure and materials requested. Workshops can range from a minimum of a few hours for group of people to on-going work with an individual over the course of several months to take a supported deep dive into a topic. For the latter option, I request a minimum 40 hour commitment for a total cost of $5000.
My experience and expertise
I have over two decades of experience in scientific research and statistical inference, as well as teaching and collaborating. My formal education started in biological sciences and then moved to Statistics after doing my own research and recognizing the importance of understanding statistical inference given its central role in modern scientific culture. My work comes from passion for science, research, and data — combined with love of philosophy and logic. You can read more about my background and path to the founding of Critical Inference LLC on the About Me page. Here is a brief list of the degrees stemming from my formal education, along with positions held after earning my PhD in Statistics (my CV is also provided).
- PhD and MS in Statistics
- MS in Kinesiology
- BS in Biology and Biomedical Science
- Assistant and Associate Professor of Statistics (tenured) at Montana State University for 8 years
- Statistician for an environmental consulting company for 2 years
- Director of Statistical Consulting and Associate Research Professor at Montana State University
- Currently founder and sole member of Critical Inference LLC — writer and consultant
Why is this type of consulting needed?
The use of Statistics in science has become very method- and answer-focused. The crucial challenges inherent in making inferences from limited information are often ignored or swept under the rug. Science and decision making can benefit from deeper reflection on the limitations of our methods and processes used in practice to arrive at conclusions. I am not against use of statistical methods, but strongly believe we should be more mindful in how we apply the ideas and methods relative to their foundations and the assumptions and theory supporting them.
With the daily demands of our jobs and lives, it is often impossible to find the time to dig deeply enough into recommendations, papers, decisions, etc. — even if we must rely on them for our work. Statistical inference is deeply engrained in many scientific disciplines and reaches into the media on a daily basis. Few people relying on the resulting conclusions have a deep understanding of the theoretical and philosophical foundations, or the history of how it came to play such a prominent role in our society. And, it is unrealistic to expect every user of statistical work to have this knowledge. There is a lot that can be gained by collaborating with a formally trained, experienced statistician who has thought deeply about inference and use of statistical methods. There is far more to statistical inference than the technical skills of carrying out a data analysis. This is the work I can, and want to, do for you.
Statisticians are often viewed as technicians — whose main purpose is to carry out existing methods and produce new improved methods. After 20 years as a researcher and applied statistician, I am trying to fill a void I found within the possible breadth of statistical consulting — and it also happens to be what I love. I repeatedly witnessed the misunderstandings and misuses of statistical inference across many disciplines, and have chosen to no longer be part of a system that attempts to force me to carry out methods and follow workflows that I professionally disagree with. I believe it can be productive to separate the work involved in carrying out statistical inferences (the “doing statistics”) from the work involved in critically evaluating statistical inferences made by others. The name Critical Inference comes from combining statistical inference and critical thinking – with the goal of promoting the hard work of critically evaluating and justifying inferences, rather than emphasizing calculations, fitting of models using a computer, and use of default criteria.
Example topics
Here are some examples of potential workshop, seminar, or roundtable discussion topics:
- Dig into the underlying assumptions of statistical methods you often use and critically evaluate their reasonableness relative to your goals and objectives.
- Improve your understanding of assumptions underlying methods used by others that might differ from your usual approach and whether they may be a good fit for your work.
- Gain background information and current views from multiple perspectives about topics such as the “replication crisis” in some scientific disciplines
- Learn about the history of use of statistical methods in practice over the last century. This can also include earlier history about the development of the concept of probability and its unsettled foundations.
- Strategies for justifying inferences beyond simple reporting of statistical summary measures
- Strategies for coming to a middle ground of using quantitative and qualitative methods to support inferences
- Guided discussion to evaluate a survey you or others have used or intend to use. This can include critiquing the survey itself, along with assessing quantitative inferences that may be based on it relative to what can be measured.
- Guidance for incorporating creative graphical displays of data into the analysis process — as part of the critically evaluating assumptions underlying your inferences and to better understand the information in your data before aggregating and simplifying.
- Discussions of philosophy and/or ethics of science, particularly those related to our use of statistical methods and inference in science today
- Guidance through the process of study design — to think critically and deeply at the beginning of the process and justify choices and decisions and decide on information to pre-register.
- Other topics or combinations of topics that you or your team would find useful or interesting!
Rates
Rates depend on the type of consulting and will be agreed upon with each client before work is done. Up to one hour is typically offered free of charge to gather information and discuss goals, ideas, and deliverables.