FAQ

  1. What is a statistical collaborator?
  2. How can collaborative statisticians improve research and decision-making?
  3. What is LISA?
  4. What is a statistical collaboration laboratory?
  5. How does one become a statistical collaborator?
  6. What is a LISA Fellow?
  7. Who is providing funding for LISA Fellows?
  8. What criteria will be used to choose LISA Fellows?
  9. What is LISA 2020?
  10. Why the focus on developing countries?
  11. Why is LISA at Virginia Tech the appropriate starting place and hub for the LISA 2020 network?
  12. Is the LISA 2020 program scalable and sustainable?
  13. How does LISA train statisticians to become statistical collaborators?

What is a statistical collaborator?

In LISA, we use the term statistical collaborator to denote the statistician or data scientist who works with the client/researcher to answer research questions through the use of statistical thinking and statistical analysis. Rather than merely answering the client’s statistical questions and sending them on their way, the “collaborator” makes sure he or she first understands the research, business, or policy problem; that the client understands the statistical solution; and that the problem is actually solved and that the client has the tools to implement the solution.

We believe that many statistical “consultants” perform in the manner described above, but that the term “collaborator” better implies the depth of engagement necessary to be an effective statistician than does the term “consultant”. Thus, we almost always use “collaborator” and “collaboration” to describe the best practices and outcomes of a “consultant” and a “consultation”.

How can collaborative statisticians improve research and decision-making?

LISA statistical collaborators meet with researchers to discuss their research goals, the nature of the data collected or to be collected, how to best collect the data, how the data can be analyzed to answer the researcher’s specific questions, what the statistical results mean in terms of the research goals, how the researcher can explain the results to his or her intended audiences, and how the researcher can use the results to solve problems or make decisions. We at LISA believe that statistics can enable and accelerate all phases of research, as shown the figure below.

 

Statistics Enables and Accelerates All Phases of Scientific Research

 

What is LISA?

LISA is the Laboratory for Interdisciplinary Statistical Analysis at Virginia Tech in Blacksburg, Virginia, USA. It was originally created in 1948 as the Statistical Laboratory, part of the Virginia Agricultural Experiment Station, and then was reorganized and renamed in 2008. LISA’s mission is to train statisticians to become statistical collaborators and promote the value of statistical thinking and data science in all phases of scientific research by helping researchers design experiments; collect, analyze, and plot data; run statistical software; interpret results; and communicate statistical concepts to non-statisticians. LISA provides statistical advice, analysis, and education free of charge to Virginia Tech researchers by offering individual collaboration meetings, walk-in consulting, educational short courses, and support for interdisciplinary research projects. In 2012, the 61 statistical collaborators of LISA helped a total of 1314 researchers from 77 departments at Virginia Tech. LISA collaborated with researchers on 412 projects from 66 departments, provided statistical advice for 396 visitors to LISA Walk-in Consulting, and taught 506 attendees how to apply statistics in their research during 25 educational short courses. Dr. Eric Vance has been the director of LISA since 2008. He passionately believes that training statisticians to become statistical collaborators will increase the impact of statistics worldwide and thereby improve the welfare of humankind. For this reason he is the primary champion and leader of LISA 2020.

What is a statistical collaboration laboratory?

A statistical collaboration laboratory consists of statisticians and data scientists who have been trained to communicate and collaborate with non-statisticians and who want to collaborate with researchers, policy makers, and business leaders to use statistics to make better decisions and improve our understanding of the world. The purpose of a statistical collaboration laboratory is to help create new knowledge that will enable and accelerate research and lead to better, data-based decisions in the public and private sector. Another purpose of a statistical collaboration laboratory is to educate and train other statisticians and data scientists to become statistical collaborators.

One example of a statistical collaboration laboratory is LISA. LISA is a laboratory for interdisciplinary statistical analysis and collaboration. Much more than a service center for the university, LISA is a laboratory of understanding and knowledge creation that creates knowledge in at least four ways:
1. LISA helps researchers answer questions they could not have answered without expert statistical advice.
2. Because LISA collaborators are experts in data, they can suggest new questions researchers can answer with their existing data or can help design innovative experiments or studies to collect data that will answer novel questions that help the researchers achieve their overall research goals.
3. When LISA collaborators encounter new types of data for which standard statistical methods do not apply, they create new knowledge by developing novel statistical methods that enable researchers to extract useful information from their data.
4. LISA is researching the process of statistical collaboration itself, advancing knowledge in best practices for statistical collaboration and how to improve one’s statistical collaboration skills.

How does one become a statistical collaborator?

First, one must learn the theory, methods, and application of statistics. These are typically learned in university graduate programs in statistics and data science.

Then the statistician must learn how to communicate and collaborate with non-statisticians. In short, the statistician must learn how to make statistics and data science useful to non-statisticians.

Finally, one must put the learning and training into practice, meeting with people who have real problems and real data to find and implement a real solution.

A true statistical collaborator must also collect and analyze data on their own statistical collaborations (get objective feedback on their performance) to determine what they do that is successful and how they can improve their practice of statistics and data science.

What is a LISA Fellow?

A LISA Fellow is a statistician who is educated and trained in LISA at Virginia Tech to become a collaborative statistician. The LISA Fellow will gain experience and improve their practice of statistics while working in LISA as an Associate Collaborator and then as a Lead Collaborator.

Who is providing funding for LISA Fellows?

Google provided the funding for our first LISA Fellow, Olushina Olawale Awe from Nigeria. We do not currently have funding for additional fellows.

LISA Fellows will need to provide proof of funding for their visa application. We expect foreign (non-US) governments and universities to fund statisticians to be visiting scholars and LISA Fellows in the LISA 2020 program. In this way the foreign governments will demonstrate their support for the returning LISA Fellow to create a statistical collaboration laboratory at their home university or institution.

What criteria will be used to choose LISA Fellows?

The LISA Fellow will be chosen based on three criteria: 1) Personal characteristics including passion for statistical collaboration, technical expertise, and experience; 2) The demonstrated need for a statistical collaboration laboratory at their home university or institution; and 3) Potential support and resources available in their home country to create a statistical laboratory. The LISA Fellow will be meeting with clients from a wide array of disciplines. The LISA fellow will need a strong statistical background (preferably a master's or PhD in statistics or a related field) to be able to assist the clients as we help develop the fellow's communication and collaboration skills. The goal of the LISA 2020 program is to create a network of at least 20 statistical collaboration laboratories in developing countries by 2020. To accomplish that, we must choose fellows whose university/institution have a need for a statistical laboratory. However, simply needing the laboratory is not enough, we need to know that their will be support for the returning LISA fellow to create and sustain the laboratory.

What is LISA 2020?

LISA 2020 is shorthand for the plan for LISA Fellows to create a network of 20 statistical collaboration laboratories in developing countries by 2020. How many LISA Fellows? At least 20. How many new statistical collaboration laboratories? At least 20. By when? The year 2020.

Why the focus on developing countries?

Because modern research creates such large amounts of data across the fields of science, social science, and the humanities, statisticians are needed who can devise methods to guide experimental design and data analysis to make sense of all that information.

One technically sound statistician who is trained to effectively communicate and collaborate can enable 50 or more research projects per year. Those research projects can have a positive impact on thousands of people. This impact is increased by training statisticians who are local, who have a relationship with researchers, and who know what data have been collected or can realistically be collected. A statistician who lives in and around the area of research has a better understanding of what those data really mean and how they can be appropriately analyzed. This kind of local statistician has a vested interest in helping to solve the research problem for the benefit of society.

Why is LISA at Virginia Tech the appropriate starting place and hub for the LISA 2020 network?

Virginia Tech is strongly committed to its land-grant mission of transforming knowledge to practice in the service of society. Its motto is Ut Prosim, “That I May Serve,” and this motivates LISA’s efforts. Founded in 1948 as the Statistical Laboratory to assist agricultural researchers design experiments and analyze data, LISA has evolved and grown into the Laboratory for Interdisciplinary Statistical Analysis, reorganized and renamed as such in 2008. We serve our clients by helping them answer their research questions and by teaching them sound statistical practice to guide their future endeavors. We serve the field of statistics by strengthening and advancing the statistics profession through the stimulation of research into effective methods for statistical collaboration and by developing new methods to analyze data to solve real-world problems. We serve everyone who will benefit from the development, application and dissemination of statistical science by educating and training statisticians to put their expertise into practice to make statistics of service to society.

Is the LISA 2020 program scalable and sustainable?

LISA currently has 22 graduate students, one director, and one assistant director working as statistical collaborators and more research clients requesting our services than we can comfortably handle (412 collaborative projects from 66 different departments in 2012). The presence of 1-5 already technically trained LISA Fellows per year will be a welcome addition to the LISA team. Because the LISA 2020 laboratories will be run by local fellows and (ideally) funded by their home universities or governments, LISA will not be responsible for the day-to-day operations. Instead, LISA 2020 personnel will serve in an advisory capacity. Just as LISA lead collaborators mentor associate collaborators, established LISA 2020 labs will be encouraged to mentor the newest labs.

Once we make progress creating the LISA 2020 network, and as soon as we are successful in publicizing our efforts and accomplishments, we expect to partner and cooperate with other organizations in the advancement of the LISA 2020 vision. We expect that some LISA 2020 fellows will receive their training and experience from other statistical collaboration laboratories and consulting centers, perhaps from those run in the future by current LISA students or from future laboratories in the LISA 2020 network.

How does LISA train statisticians to become statistical collaborators?

LISA’s primary mission is to train statisticians to become statistical collaborators, and we will extend and adapt this training to the needs of the LISA Fellows. We believe that every statistician—whether in industry, academia, or government—will be called at some point to collaborate with non-statisticians. But not every statistician has the skills or experience to effectively collaborate. Currently, our Virginia Tech students learn the fundamental theory and methods of statistics in technical statistics courses at Virginia Tech. Then in a course titled “Communication in Statistical Collaborations,” they learn, develop, and practice the essential non-technical skills necessary to be effective statistical collaborators, including:

  • Helping scientists truly answer their research questions;
  • Using peer feedback and self-reflection to improve communication and collaboration skills;
  • Collaborating with members of a team in a contributing, productive, effective, and respectful manner;
  • Structuring and leading efficient and mutually satisfactory statistical collaboration meetings with clients;
  • Listening and asking good questions;
  • Creating plots of data that tell the story of the research;
  • Explaining statistics to non-statisticians.

LISA Fellows will also learn the material covered in this course to prepare them for collaborating in LISA.

Virginia Tech students continue their training outside of the classroom while working in LISA as associate collaborators under the supervision of a lead collaborator, who is usually a statistics faculty member or an advanced PhD student. This position allows the associate collaborators to gain experience interacting with 3-8 researchers per semester from various fields, understanding the research questions, applying their statistical knowledge to help answer the research questions, and explaining the solutions in terms the researcher will understand. Weekly technical staff meetings allow collaborators to describe current projects and ask the LISA director, assistant director, and their peers for statistical advice. To further enhance their training, biweekly video coaching and feedback sessions provide the statistical collaborators with feedback on their actual meetings with clients and allow them to reflect on which techniques worked or did not work, and why. These video sessions focus on the interpersonal and intrapersonal actions of the clients and the statistical collaborators, and on the technical aspects of the interactions. Additionally, after each semester, the director of LISA reviews feedback from clients with each collaborator. When associate collaborators have gained enough confidence and experience, they are promoted to lead collaborators and collaborate with ~16 researchers per semester. Increasingly, student lead collaborators have taken a mentoring role with associate collaborators, providing them with regular coaching and feedback.