3a665dfd4dae060bec7a0638768eab12_400x400We are sometimes asked why we don’t have the ‘best friend at work’ question in our engagement and culture surveys. Apart from it being a Gallup copyrighted question, we have a few good reasons.

1. Even on basic grammar considerations it isn’t that great a question. The question consists of the following statement which is accompanied by an agreement scale:

‘I have a best friend at work’

Does it mean my best friend works with me? Or does it mean that I have a colleague who I like more than other colleagues who I would consider a friend? Or does it mean I can identify anyone as being more friendly than my other colleagues?

Many people tell us they have a ‘good’, ‘close’ or even ‘totally awesome’ friend at work, but ‘best friend’ is often described as confusing.

Additionally, considering Gallup’s stated ratonale behind the question, it isn’t clear the wording best matches this. Here is the rationale as explained in Gallup’s recent meta-analysis.

“Managers vary in the extent to which they create opportunities for people at work to get to know one another and in how much they value close, trusting relationships at work. The best managers do not subscribe to the idea that there should be no close friendships at work; instead, they free people to get to know one another, which is a basic human need. This, then, can influence communication, trust, and other outcomes.”

If we want a question to assess how well managers facilitate close friendships at work it seems odd to frame it around having a ‘best friend’; ‘best friend’ is usually a term people reserve for a very singular and special relationship. Even in the narrower context of someone having a ‘best’ friend at work it seems odd to confuse people who might value a few colleagues as close friends but not have a singular ‘best’ friend at work. It’s not great practice to use absolute type words in survey questions (like best, or always or never) unless you have a very good reason for doing so, and it can have substantial impacts on statistical properties and response biases.

2. More importantly, we don’t see any evidence that this sort of question is related to engagement or retention measures. When a client wishes to ask this question we have instead asked if people have a ‘close’ friend at work. This seems to match the intent of the question better anyway. However, when we look at the relationship between this question and motivation, intent to stay, or whether people would recommend their workplaces, we find no relationship. That’s right – zip (or as statisticians like to say: it’s so close to zero that it’s statistically indistinguishable from zero).

We don’t make any claims about Gallup’s own research. It’s difficult to independently test or validate because you can’t use the questions due to copyright. However, we note that in the latest meta-analysis there is no reference to relationships that the individual questions might have with outcomes – these are only shown for the aggregate of all Q12 questions together. As this recent and compelling article from LeadershipIQ makes clear – it is our job to listen to what we hear employees saying, not what Gallup says they are saying.

If you really want to ask whether your workplace or a manager or your culture fosters close relationships, we prefer to ask these sorts of questions (see footnote):

“There is at least one person I can speak to openly and honestly here”

“I truly value the relationships I’ve developed here (or in my team)”

“I feel close enough to people here to feel comfortable”

Do you really need a ‘best friend’ at work?

Read more at the CultureAmp Blog

Footnote: Questions in employee surveys and many psychological assessment instruments are often referred to as items. Items may not take the typical form of a question but instead may be comprised of a statement which people then indicate a level of agreement or disagreement with. These questions/items take this same approach.

mcpherson 2

mcphersonAbout the Author: Jason McPherson is the Chief Scientist forCultureAmp, developers of Murmur, the first People Intelligence Platform built specifically for metrics-driven organizations. Jason is a research psychologist interested in technology amplified research.

He was previously a founding team member with Kenexa in Australia, and Asia Pacific Research Associate at Towers Watson.  Follow Jason on Twitter@DrPeopleGeek or connect with him on LinkedIn.