Genevera I. Allen and Jeffrey T. Leek

It is an exciting time to be a statistician! The rise of diverse Big Data, the increasing importance of data-analytic skills, and expanded access to cheap computing yield unbounded opportunities and challenges for people just beginning their statistical careers. While changes in our field affect all members of the statistical community, here we particularly focus on how our culture shapes junior faculty in statistics and biostatistics departments. Our column is based on both our experiences and stories we have heard at meetings, in hallways, and on social media and the web.

What is statistics, and who is a statistician?

Despite the excitement surrounding our field, many junior faculty members still think their work must conform to a narrow definition of “statistics” to be successful in academia. This thinking is reinforced by hiring committees, promotion committees, and senior colleagues who frequently point to a narrow set of journals or concepts as “real statistics.”

Robert Rodriguez suggested in his 2012 ASA presidential address that we need a big tent approach to statistics. As junior faculty members, we couldn’t agree more. We believe this flexibility is important to ensure our field attracts top young talent; has a seat at the table for important national discussions such as the Big Data initiative, BRAIN initiative, and online education; and puts junior statistics faculty members in position to jockey for funding amidst national budget cuts.

We suggest repositioning statistics as the science of data by declaring that any data-related problem is a statistics problem. This means statistics can no longer be defined by that work published in a small and exclusive set of journals. It must be a broad class of ideas that span disciplines and appear in theoretical, methodological, and applied journals.

How do we evaluate young faculty members?

As our discipline changes, the academic profiles and career trajectories of junior faculty members may look different from those of faculty members 10 or 20 years ago. There are many young stars in our field who have made a name for themselves in unconventional ways. But these faculty members may struggle to justify themselves at both their institutions and in our field in general.

We are discouraged to hear many junior faculty members report that only the “top 5” statistics journal publications are considered for their evaluation. What about quality statistical work published in the top venues of related fields such as machine learning, signal processing, or bioinformatics? Or, work published in scientific venues that influence the broader scientific community?

Moreover, there are many contributions that do not appear in journals. These include examples of critical efforts that are often down-weighted or ignored in institutional evaluations, grant review panels, or award committees:

• Developing and maintaining widely used statistical software
• Leading efforts to encourage reproducible research
• Implementing substantial educational initiatives online or otherwise
• Incorporating social media and technology for education and research
• Fostering connections with industries producing Big Data
• Developing new computational strategies for working with Big Data

Pressure to ensure their work fits within the narrow scope of a few journals means junior faculty members are unable or discouraged from pursuing these types of creative statistical contributions.

We suggest expanding review criteria to reflect the impact of statistical work within not only a narrow definition of statistics, but also more broadly on science and education.

How does our publication culture affect junior faculty members?

Since publications are a primary factor in evaluations, publication culture—and specifically our field’s long publication process—particularly affects junior faculty members.

Because of university budget pressures, young faculty members are expected to obtain research funding at the start of their careers. To do so, an established publication record is necessary to demonstrate the ability to perform the proposed research. With often long publication processes, young faculty members struggle to establish a convincing publication record in a short timeframe.

As our current review times are not compatible with the pace of scientific development, some choose to not submit to statistical journals, opting for venues in related fields such as machine learning and signal processing instead. Others, working on statistical methods inspired by specific applications, choose to publish directly in applied fields. While this may have an immediate scientific impact, it may not “count” for evaluation purposes.

We propose our flagship journals adopt a similar Big Tent approach to statistics and work to speed the publication process, for example by mandating four-week reviews.

Where do we go from here?

Young faculty members in statistics and biostatistics make up a diverse group. They range from the purest theoretician to the most applied collaborative scientists. We love this diversity and think embracing the many faces of our field only makes us stronger. At the same time, the culture of our field can sometimes create hurdles for burgeoning statistical careers. We hope our discussion and proposal will be the beginning of a conversation that continues to propel statistics to the forefront of modern science.