What Does This Mean to Me, Laura?

Post Categories

What Does This Mean to Me, Laura?

/ 3


The plural of anecdote isn’t data

16th November 2015

Tablet device on desk showing financial dataA couple of weeks ago, I tweeted this:



The response to that, both online and off, was very interesting. It was liked and retweeted a bit, but also instigated some discussion on Twitter.

Responder A: “In my exp the squeaky wheels get the grease. Policy & design decisions have been reversed based on a single patron’s comment.”
Responder B: “in my experience many senior staff have selective hearing and only hear some squeaks.”
Responder A: “Heh. At my last POW the approach seemed to be to listen to *everyone* & then do what they asked for where poss.”
Responder A: “I suspect this was partly due to receiving patron feedback so infrequently – we really took it to ❤️ when it came ”
ME:                    “The problem, of course, is that anecotes are not the plural of data.”
Responder B: “nice approach but eventually you end up with opposite requests. That’s what libs don’t deal with well. ”
Responder C: “lots of squeaks sounds like the perfect time to go get some data”
Responder B: “otoh if you wait for ‘data’ before changing things you’ll never experiment.”
Responder C: “there’s still value in checking whether something is an issue for a small group or a large one”
Responder B: “yes but I’ve seen ‘no data’ used as a conservative roadblock to change.”
Responder D: “anecdotes are certainly used as a roadblock to change.”
Responder A: “By some weird coincidence I was faced w/ this today. Training staff in new ILS & being given anecdotes as ‘proof’ that 1/2”
Responder A: “2/2 the new system was no good/couldn’t support their workflows. It was bullshit, & just used as an excuse to reject change.”
Responder A: “Change that is coming no matter how much they drag their heels. I am still steaming.”

I’m a tech. I like having numbers (although I will be the first to tell you that I am very bad at making them do things mathematically).  I always cringe when someone tells me “But the patrons want…”  How do you know this? Did you do a survey?  Is that what your web metrics show?  Can you quote a study?  Storytelling is, without, question, a powerful tool when used in conjunction with the conveying of information. When you can tell a story with your statistics, it becomes a very clear picture.  Stories, alone, are just that…stories.  I don’t think they should be discarded entirely; rather, they need to serve as a jumping-off point for further research.  Consider the story to be the hypothesis behind your next survey or deep dive into metrics. Can you prove it true or false?  The story itself cannot reasonably be the data.

What do you think? Agree or disagree? What are your experiences with “But the patrons want…?”


  • Tara Li
    Posted at 1:26 pm November 16, 2015
    Tara Li

    Sorry, but I have to disagree somewhat here.

    Anecdotes *are* data, just data of extremely low quality. Enough anecdotes, and you end up with anecdata, an indicator that something needs to be investigated more rigorously. Listen to the anecdotes, but don’t presume the conclusion you should reach is necessarily the obvious one presented. “We can’t figure out the new ILL system, it’s crap!” Does this mean there are actually problems with the ILL system, or does it mean that the training materials for the ILL system needs to be revised, or is it that the librarians are so comfortable with the old system, and don’t want to change? In the latter case, would it be possible to put an interface that resembles the old system *over* the new system, to ease the transition?

    To me, the canonical example of this, and it *is* an anecdote itself, is the refusal of European scientific academies to admit to the possibility of meteorites actually falling from the sky, despite numerous reports of exactly that from non-scientist observers, until they actually observed it themselves in a rather massive meteor shower.

    Your patrons are telling you *something* – it’s up to you to figure out what they *really* want, rather than what they think they want, or what they say they want. (But going by what they *say* they want is often a good first step!)

  • Laura
    Posted at 1:31 pm November 16, 2015

    I don’t think that we actually disagree, except possibly in the semantics department.

    “Listen to the anecdotes, but don’t presume the conclusion you should reach is necessarily the obvious one presented.”–Tara Li

    “I don’t think they should be discarded entirely; rather, they need to serve as a jumping-off point for further research. Consider the story to be the hypothesis behind your next survey or deep dive into metrics.”–Me

    Anecdotes *can* be an indicator of an issue; I can’t and won’t argue with that. My problem (and that of others involved in the Twitter chat) is when that’s as far as things go. They’re often presented as proof positive that a problem or fact exists for a large group of people, with no further foundation to stand on.

  • Freya Anderson
    Posted at 6:54 pm November 16, 2015
    Freya Anderson

    I don’t disagree with anything here, but just wanted to point out that time is a factor and in today’s world we need to stay as nimble as we can. When considering major changes, we research our options and our users, usually with focus groups and/or surveys, although we’re always interested in other options that are relatively inexpensive in time and money. Often, though, I think a willingness to change course is more important. When we’ve made changes based on anecdotal information, we need to remember to be especially vigilant in observing how those changes are received. In those cases where they turned out to be missteps, we need to be willing to reverse or modify the changes, rather than being wed to our cool ideas.

Comments are closed.