香蕉秀

Leading with ideas

HOW DO YOU MEASURE RESEARCH PRODUCTIVITY?
(VERY CAREFULLY)

By Andy Boynton, John and Linda Powers Family Dean

As deans, each of us has the task of not only building and improving a research culture but also measuring it. Needless to say, it鈥檚 a challenge. Just how do you put a number on intellectual output and the creation of knowledge that makes a difference?

We all have our ways. In the spirit of exchanging best practices, here鈥檚 my take on an eight-year experiment we鈥檝e undertaken at the Carroll School of Management.

Interspersed with my comments will be pages from what we simply call 鈥淩esearch Reports.鈥� These are three-page summaries compiled every spring, capturing the quantity and quality of a faculty member鈥檚 research. The reports give us a transparent way of evaluating the research, one that lends to comparison across positions and departments.

We鈥檝e found these annual summaries to be very useful. They don鈥檛 explain everything about a professor鈥檚 research, and they shouldn鈥檛, but they do offer a starting point for discussion.听They're one measure鈥攁 good first approximation of research productivity and impact.听Naturally the data points come into play when considering tenure and promotion, but in my mind the real value is that they begin a conversation.

In the pages displayed here, you鈥檒l see the standard references to 鈥淚mpact Factor鈥� and 鈥淎rticle Influence.鈥� As you know, both of these refer to the quality of journals in which our professors are published. 鈥淚mpact鈥� relates to how often the average article in these journals is cited; 鈥渋nfluence鈥� gives greater weight to citations in higher quality journals.

You鈥檒l see references to 鈥淔inancial Times,鈥� specifically the top 50 journals used by the听Financial Times听in compiling its rankings of business schools from the standpoint of research output. Our measures also take account of how often a given journal鈥檚 articles are cited in Web of Science, and how often articles written specifically by our faculty are cited in Google Scholar.

In addition, metrics are influenced by whether a faculty member鈥檚 article is coauthored, and if so, how many collaborators there are (more about how we don鈥檛 necessarily penalize coauthorship in a moment).

Here is the first page of the Research Report that every faculty member receives. It shows how their statistics are calculated:

On the second page, the faculty member sees his or her research statistics and how they compare with colleagues in the department and school. I have the honor here of using as a sample the May 2016 Research Report for Jeffrey Pontiff, the James F. Cleary Chair in Finance.

As you鈥檒l read in this edition of听Carroll Capital, Jeff recently won the听Journal of Finance鈥檚听2016 Amundi Smith Breeden First Prize for his paper, 鈥淒oes Academic Research Destroy Stock Return Predictability?鈥� He collaborated with our former doctoral student R. David McLean, now at Georgetown University鈥檚 McDonough School of Business.

Research Statistics for Jeffrey Pontiff (Finance)

I alluded earlier to the issue of coauthorship. I say 鈥渋ssue鈥� because there鈥檚 more than one way to account for articles coauthored by our professors. On the one hand, we want to encourage collaboration; on the other, we want to see signs of independent authorship as well. The upshot: we include some measures that (in effect) subtract points for coauthorship and others that don鈥檛.

Finally, I return to the broader question of quantifying the value of knowledge. I strongly believe that everyone鈥檚 research is nuanced, and to understand it well, you have to get beneath the data. Still, the data can tell you a lot.

All of us have goals in this regard鈥攁t the Carroll School, we are focused heavily on producing high-quality, impactful research. We need a measuring system compatible with this goal, not necessarily the whole picture of scholarly output, but at least a snapshot. We keep tinkering every year with our measuring system to make sure we鈥檙e getting the clearest possible picture.

I would be thrilled to share more details about our approach with readers of听Carroll Capital, and hear about your efforts to measure research accomplishments. Please feel free to听send me an e-mail.

I leave you now with the third and final page of Jeff Pontiff鈥檚 Research Report. 鈥淲OS鈥� refers to Web of Science, and the 50 journals ranked by the听Financial Times听are assigned a value of 1 (all others carry a value of 0).

List of Publications for Jeffrey Pontiff (Finance)