The other day my wife cooking a new dish for dinner. As I was taking my first bites of it, she asked, "What do you think? I think it came out well." I laughed, and responded, "Thank you for letting me know the right answer to your question."
In that situation, I found it funny. But a few days later, I had another experience which isn't such a laughing matter. I was on a college tour with my son. At the end of tour, the tour guide (whose name has been changed to protect the guilty) said "I hope you've enjoyed the tour. If so, there's a survey you can fill out to provide your feedback, and remember my name's Greta. If you didn't like my tour, there's no such thing as a survey, and my name's Megan."
As a parent I could chuckle, but as a researcher I cringed. There's a reason the admissions office of that university created this survey. Someone went to considerable effort to come up with a good set of questions, and someone is also going to go to considerable effort to analyze the data. But what the tour guide said is going to undermine those efforts and possibly render the data meaningless.
At one level, this is just an extension of my earlier post about the difference between asking for evaluation and asking for guidance. Presumably the reason the tour guide said what she said is that the university uses the surveys in part to evaluate the performance of the tour guides. If so, then the guide shouldn't be the one asking people to fill out the surveys, because she has a clear conflict of interest. Or better yet, the university could rethink how they use the surveys, to shift the focus from evaluation - which no one enjoys - to professional development, or improving how they train guides, or something else that isn't so ego-bruising to the guides. That way the guides will be more comfortable asking for honest feedback rather than positive feedback.
But this story also illustrates a broader psychological principle that I think anyone trying to run a customer-centric business should be aware of. That principle is called social desirability. Sounds like a good thing, yes? For making our social interactions go smoother, it might be. For getting accurate information from people, it's a big headache.
To understand how social desirability works, suppose you are an employer where I am in Washington state, where recreational marijuana is legal. In your business employees have to operate potentially dangerous machinery like forklifts. You suspect that some of your employees are showing up high, creating a workplace safety issue. But you're not sure how big the problem is, so you decide to do a little research. You know that if you ask people directly, they're likely to lie. So you decide you'll do an anonymous on-line survey - even if someone says they smoke a joint every morning before work, you'll have no way of knowing which employee that is. And you take great pains to explain that to your employees before sending out the survey. So in that case they'll be honest, right? Perhaps not.
To understand why, it helps to consider a less loaded example. At its core, social desirability is about managing our image, and presenting ourselves as following the norms of a group we identify with. There are two parts to this. First, we want people to think of us as "good" group members - people who behave in the way that is expected within the group. Second, we want to think of ourselves as a good group member. Of course, the standards of behavior within a group are typically aspirational, not a description of what everyone in the group actually does. So you often end up with a discrepancy between the standard, and what individual people actually do.
For example, in a workplace setting, there might be a standard that people show up to work on time. Miguel tries hard to do this, but he can't drop his daughter off at daycare until 15 minutes before he's supposed to be at work. So if anything goes wrong, like a traffic accident on his route, he ends up late. Suppose Miguel's supervisor notices this, and asks him to honestly say how many days a week he is late. He can't say "never," because the supervisor has seen specific instances where it happened. But what he can do is shift the number downward, to reduce the discrepancy between what he says and the group norms. So if he's actually late three times a week, he might say "once or twice a week."
Now suppose his supervisor is both an enlightened manager and savvy about these sorts of things. So she says that she understands Miguel is in a difficult situation, but that when he arrives late it negatively impacts the business. So what she wants him to do is honestly think about how often he's late per week, and write that number down on a note card. He doesn't need to show that number to anyone, but she asks him to take the next month and work on figuring out how he can be late less often. At the end of the month, she wants him to use the other side of the note card and again write down how often he's late. She doesn't want him to show her the card or share the numbers with her, but she would like to have a brief conversation about how he thinks his efforts have worked.
So will Miguel be truly honest now? He still may not. Remember, it's not just about managing how other people see you, it's about how you see yourself. Miguel might know that in fact he's late about three times a week, but it's one thing to know that and another thing to write it down and have the number stare back at you. So Miguel might find a way to sugar coat the situation. If he can think of a week he was only late once, he might write "1-3 times a week."
Social desirability is one of the reasons why it requires a lot of training and experience to write good survey questions. It can come into play in any situation where how a person answers a question might impact how you think about them, or how they think about themselves - in other words, anything having to do with the customer relationship. For this reason, I don't believe it's a good idea to take a DIY approach to surveying your customers. We are happy to do surveying for you, or to help you create a good survey. I would also say that of all the ways to collect data from customers, surveys are probably the most vulnerable to social desirability effects, so we're also happy to help you figure out other ways to collect data where it isn't as big of an issue.