I Collect Data

A couple of years ago I attended a meeting where Russell Belk, a professor of marketing from the University of Utah, gave a talk about “collecting,” i.e., the phenomenon in which people go to great lengths to obtain large numbers of coins, stamps, or other objects. After his presentation, I approached Belk and said (with just a touch of smugness) that I didn’t collect anything. But later that day, when during my talk I mentioned that I had “collected some data” to test a hypothesis, Belk smiled and said “Ah! You are a collector.” I sputtered some reply and not so deftly returned to my talk, but his comment stayed with me. I have indeed spent most of my professional life collecting data. Was I no different than those who put a great deal of effort into collecting Elvis memorabilia, fancy cars, or Pokémon cards?

When I breakfasted with Belk some months later, we discussed the parallels between what research has shown about collecting and how academics (not just me!) behave as data collectors. The similarities were many; below, I sketch out five of them.

1. Collectors display their collections. If you walk into the home of a collector, it rarely takes long before the nature of his or her collection becomes obvious, as it is displayed on the walls, in cases, or wherever is appropriate. Further, collectors often gather together in hotels or other sites to show others what they have acquired. Similarly, we data collectors display portions of our collections in journals created especially for the purpose of reporting data; we make posters and give talks to other collectors at conferences; we even keep records of our collections in our curriculum vitas. Indeed, consider how infrequently academics collect data that they do not display to someone; unless the data are really poor, we typically work hard to show others what we have collected.
2. Connoisseurship. Collectors are connoisseurs. Although many of us are unable to tell the difference between two bottles of chardonnay or the worth of a coin minted in one year or another, collectors of these objects recognize fine distinctions to which the rest of us are oblivious. Similarly, we data collectors can discern between data deserving publication in the top journals of our field, data which will find print in other less prestigious journals, and data doomed to the dustbin of oblivion. But were we to show data of these different levels of quality to our friends who do not collect data, they would usually be equally impressed by what we have collected or equally dumbfounded that we have put so much effort into proving a point that their grandmothers already understand.
3. Specialization. Collectors tend to specialize, collecting primarily one sort of object. Generally speaking, one does not find an individual who collects both masks from the South Seas and yardsticks. Collectors instead find everything they can that belongs to a particular category of object, as such specialization increases their success as a collector. Similarly, we data collectors tend to specialize in the types of data we collect. We choose a particular discipline in which to pursue an advanced degree, and then specialize in a particular sub-area of that discipline, and then focus on a particular topic within that sub-area. As a result, most of our curriculum vitas show a reasonably consistent set of data collected about one topic, be it cellular functioning, autonomic arousal, or a certain chemical compound. By specializing our collections, we increase our chances of successfully finding more data to add to our collection. Which bring us to:
4. Growth of collections. Collectors incessantly try to obtain more objects relevant to their collection. They attend trade shows, look in magazines, and stop in stores during family vacations on the off chance that they might find just the thing to fit an empty niche in their collection. But not just any item will do: The collection must grow through the acquisition of novel objects, not the repeated acquisition of duplicates of the same object. Similarly, we data collectors typically collect data at every turn, awaiting the next trip abroad to search out an undiscovered species or the new influx of introductory students who will fill out our surveys. Many of us have enough data stored up to churn out several papers, yet we insist on collecting still more. These new data rarely reproduce data we have already collected, but instead expand on our collection in some subtle way. For all of us collectors, growth comes from diversity.
5. The search for the perfect object. Most collectors have an idea of some “perfect” member of their collection: the vintage of a particular year, the first edition book, the unblemished doll still in its original package. Such perfect objects are usually among the most highly prized members of the collection, the ones displayed with most prominence, and the ones first pointed out to visitors. Similarly, we data collectors often have images of our “perfect” data collection, which of course varies according to our methodological proclivities. For some, it is an unconfounded experiment with all hypotheses (including the four-way interactions) supported at p values less than .001. For others, it is a well-preserved, largely intact fossil. For yet others, it is an undocumented comet spinning through the heavens. Regardless, the search for these perfect data keeps us active for many years. If we do find what we are looking for, we tell everyone about it, and are happy to reap the attention of our colleagues (as well as perhaps the occasional newspaper reporter).

CONCLUSION
Of course the fact that similarities exist between “object” collectors and “data” collectors does not imply that all aspects of the endeavors are exactly the same. Collectors of “objects” seem to do so for a variety of reasons, including curiosity, challenge, and even psychopathology. We data collectors typically have the additional purposes of collecting data so that we may ameliorate some unfortunate condition or so we may find “the truth” (or at least that truth supportive of our favorite theory). Additionally, we data collectors often have two other motivations which must not be forgotten: Our collections pay the bills and provide the strokes that bolster many of our egos (i.e., we rarely publish anonymously).

Despite these differences, I find the similarities between object collectors and data collectors provocative, for they show once again that many of the same dynamics behind our behavior as scientists and academicians are, at base, shaped by more general psychological tendencies.

We can not escape the facts that we data collectors are people first, and that ours is a culture that operates by the same types of rules that govern other cultures. Even when those other people and cultures are trying to amass large numbers of Beanie Babies.