Wednesday, September 11, 2013


Here’s another acronym for you:  Garbage In, Garbage Out.

Research is both an art and a science.  At its best, it can draw inferences from disparate facts.  But it can’t create facts out of thin air (despite how much a zealously optimistic client may want you to).

The day a dozen years ago was overhung by glorious blue skies over the eastern half of the country.  I got an email from a colleague: “Isn’t it crazy about that plane in New York?”  I summoned my usual go-to news website, USA Today; the story was just breaking, with minimal detail.  The Wall Street Journal posted a brief advisory that a plane had crashed into a skyscraper in lower Manhattan, warning that trading on the NYSE might be affected.

As the saga unfolded, it became clear that the day—indeed, the week—would be like none other.  As everyone in the office tried to press on with their routines, I got a tasking from administration: Account for any alums who worked either in the World Trade Center or the Pentagon.  This ad hoc request was reliant on a database where employment information routinely played second fiddle to home info.  I know; shocker.  The process took a long time, but it needed to—to be right.  It was not meant to be, however.  Out of the thousands of graduates, I found only two with employment addresses at the Pentagon, both evidently long out of date.  And tragically enough, two of our graduates perished at the World Trade Center—neither one with business info current in the system.

There are so many bigger things to consider in this context than whether an institution has up-to-date workplace data on its constituents.  But this, for me, typified the helplessness a research professional may encounter relying on a feckless system that purports to capture useful information.  Much of the problem is, as with the blind men and the elephant, differing perspectives and incongruent agendas with respect to constituent data.

How to fix it?  Know specifically what your data needs might be, first off.  If you foresee ever needing work info, email addresses, cell numbers—don’t wait around passively for the Data Fairy to drop that down your chimney.

No prophet is accepted in his own country.  Administrators too often believe that their own researchers are fallible, while vendors?  They are wizards.  Some are, to be fair, but c’mon—how many times has an administrator said to you, “Can you buy a list of X for me—cell numbers, email addresses, employment data?”  Nothing against vendors (some are my best friends), but again—c’mon!  What’s the best approach to making sure that the info in your database is current?  Build it directly from the source.

That’s never as easy as it sounds.  But first—you have to make sure you have a place at the table when any outreach to constituents is discussed—if not direct involvement, make sure you have a proxy there to speak for you.  If a questionnaire is going to be distributed—to alums, to parents, to the community—what questions do you need to be asked?  Be a part of that planning.

And how do you obtain “buy in” from the constituent, to get them to freely (and accurately) provide their valued information?  It goes without saying that you never do anything to betray that trust.  But you may first have to provide some sort of quid pro quo.

For instance, one of my past employers encouraged its alums to enter a giveaway for an iPod –in exchange for their email addresses.  About thirty percent of the sample responded, which was a big improvement on what we had.  Value was added on both sides.

LinkedIn pages provide incredible self-reported info, but be wary—there’s no requirement for those pages to be kept current, and I’ve seen forgetful business moguls (or those Unclear on the Concept) create as many as three separate LinkedIn pages, all with divergent data.

Also, vendors such as LexisNexis who bundle information can be of paramount help in locating people (including cell, email, and employment info), but be aware of how to interpret a lapse in a person’s most recent address, which may indicate any of the following: 1) it is a duplicate record—search again to find the one most current, 2) the person is living out of the country, 3) they are in a residential medical facility, 4) they’re incarcerated, 5) they’re serving in the armed forces, or even 6) they’re taking the challenge seriously to “live off the grid.”  Also check any available spouse’s record (or that of any other cohabitant) for possibly more up-to-date info.

The best research operation is one supported by a database with the utmost integrity, both for the blue-sky days and for those when things go cataclysmically sideways.

Tim Dempsey, Director-at-Large, APRA MidSouth

No comments:

Post a Comment