Can an Anecdote be Data?
Over at the blog, Confessions of a Community College Dean, a commenter wrote that
Academics of all people should remember that the plural of anecdote is not data.
I find this to be a humorous, and in some way, interesting quote. I also find myself “engaging” with the quote in ways that I didn’t expect. This statement challenges really two fundamental concepts–that of anecdote, and of data. When does a collection of anecdotes step out of a collection of “stories” and actually add up to real, actionable data?
Data is, at the most broad, a collection of anecdotal evidence that builds a case over time. In fact, case-study analysis is simply a rigorous approach to gathering anecdotal evidence. Yes, you are limited in the amount of pure statistical analysis that can be accomplished through the gathering of anecdotal evidence, but anecdotes bring a richness to the data that one cannot get through simple “number crunching.”
In addition, an anecdote could be considered a data point of one. And sometimes “one” is all you need. How many times should a car not start before you suspect a problem? How many space shuttles do we need to blow up before we learn about O-rings? If one is seeking to improve service, enhance quality, or improve a reputation, then every negative “anecdote” is a story of failure–failure that must be controlled, corrected, and eliminated.
On the other hand, anecdotes of success should be plentiful, and instructive, if they help one overcome the failures in the other negative anecdotes. Anecdotes about repeated success help in that they demonstrate that success can be achieved, and perhaps can be “replicated” with appropriate processes.
Let’s look at one example. In a blog written by one of my students, he struggles with poor customer service from (stand by for stereotype) “the cable guy.” At the end of the day, the cable guy dragged mud over the carpets, cut two holes in the wall (only needing one) and then had to leave so couldn’t repair the damage to the wall.
The best part about the entire story was, the installer had to be at another appointment and couldn’t fix the hole in my wall. I asked what he was going to do about the wall and he gave me some putty and a sponge and explained how to fix it. He then gave me my $80 installation bill and was on his merry way.
Is it any wonder this customer was upset?
Comcast (as Wesley points out) “wants to be perceived as the leader in cable services.” If one seeks to deliver a high quality service, error free, on-time, every time, then a single story of failure is a failure. In fact, what the Comcast blog posting points out is that a service provider, Comcast, is failing in key aspects of the delivery of that service.
The Fitzsimmons’ write, in their book Service Management that reputation is often a key dimension to the provision of a service. They write:
The uncertainty that is associated with the selection of a service provider often is resolved by talking with others about their experiences before a decision is made. Unlike a product, a poor service experience cannot be exchanged or returned for a different model.
So the “take away” here is that for an individual about to make a decision an anecdote is “data.” And as negative experiences begin to collect, the weight of these anecdotes can tarnish (perhaps irretrievably) the reputation of an organization, or a person.
So, does the simple statement “the plural of anecdote is not data” hold true? I think not. Cute, but alas, wrong.
Yes anecdotes are clearly data, but thy might not be the relevant data. I think the idea behind the statement is rather that you should be careful with anecdotal evidence and bear in mind that it will be heavily colored by your surrounding, e.g. your circle of friends, which might not be representative. So, yes for everyday decisions that only affect you, anecdotes are easily available and appropriate data. After all, your friends should be representative for you.
But for public policy issues like higher ed’s connection to health care, which happened to be the topic over there, proper statistics might be the way to go. I have to admit that I am tired of people repeating personal anecdotes instead of digging up actual numbers like life expectancy or frequency of malpractice or whatever.
BTW, I really didn’t get your idea of having no choices in a public health care system: Back in Europe, I just went to another doctor whenever I was not satisfied with mine. And yes, my friends’ opinion about a doctor was an important point. Now that I live in the US, my health care insurance would require me to fill out tons of paperwork to change my primary care physician. I had considerably more choice back home.
Thanks for joining us over here, and leaving a comment!
I certainly agree with you about anecdotes being tainted by perspectives and attitudes of the person relating the tale. And I agree that we should also consider more quantifiable data when making policy decisions.
I actually replied back at http://suburbdad.blogspot.com/2008/02/i-need-this-class-to-stay-on-my-parents.html
to your comment. My point wasn’t about not having choices, but rather the impact on choices that a single-payer system can have.
Just so you know, this initial post was not a result of the health care discussion, but a different topic.
Thanks for visiting, and feel free to look around!