Language Log

We eventually wound up walking into this complete other study

After trying for a couple of weeks to ignore the news stories about that workplace-swearing study, I've given up -- the flood of "check it out" emails from readers finally led me to look into the research behind the media coverage. This confirmed, once again, that things that seem "too good to check" are probably exactly that.

In this case, you can check it out yourself by reading Yehuda Baruch and Stuart Jenkins, "Swearing at work and permissive leadership culture: When anti-social becomes social and incivility is acceptable", Leadership & Organization Development Journal 28(6): 492-507, 11/6/2007. The authors review (some of) the literature on swearing, and propose an "An emergent model for the consequences of swearing at workplace":

This is not a model in the statistical sense, which might be tested in relation to counts or measurements of various sorts of relevant behaviors or characteristics of individuals and groups. Rather, it's a "boxology" that expresses graphically a number of qualitative hypotheses. In particular, the authors adopt an established distinction between two types of swearing, "social swearing" and "annoyance swearing", and hypothesize that social swearing has positive effects on stress release and social cohesion (and therefore on individual and group well being), whereas annoyance swearing has negative effects.

This could all be true, though it's not obvious (for example) that annoyance swearing has negative effects on stress release. And someone might suggest that it's also important to distinguish aggressive swearing, which expresses hostility towards other people, and also to distinguish whether the object of hostility is within the group or outside it. Having had those thoughts, you could change the boxology to symbolize them.

But whatever the boxology, how do the authors of this study test it? They decided that measuring and counting things would be too hard:

While the model presented above fits well with a positivist approach of listing and testing an explicit set of hypotheses, the nature of the subject would not allow for using a conventional quantitative method (e.g. questionnaires). Collecting first-hand data on swearing is a challenging task, and needs to follow guidelines for conducting research on sensitive topics.. Firms tend to be reluctant to admit that swearing characterizes part of their operation. A direct, "front door" approach is troublesome for a number of reasons:

* The study will potentially be intrusive, time-consuming, and disruptive to the individuals or organizations involved. The firm may expect some consideration in return for their investment of staff time.
* Obtaining accurate data requires a long time frame. Participants may initially be shy about their conversations being recorded, and will tend to modify their behaviour in order to create a good impression.
* Research of this type is an ethical minefield. The firm will be concerned about its reputation and image, whereas the employees will be concerned about control of the data in order to maintain good relationships with their colleagues and employer.

A more relevant and feasible way to collect sensitive data is action research, where the researcher is engaged in the organization as a participant. This helps to overcome the above mentioned issues, and enables discussion of real life in organizations. The data collection for this study was conducted while the second author was employed in a temporary position in a mail-order warehouse.

In other words, Jenkins had a part-time job, and made some observations about cussing in that context.

That's not quite all they did: they also gathered a set of "cases, vignettes, and examples" in "six focus group discussions (four in a southern USA state, two in England)".

Full time and part time employees, mostly students in class discussions in groups of 10-20, were asked to reflect on both positive and negative use and "application" of swearing in the workplace. They were working in a variety of sectors, and as expected, the responses reflected certain variance amongst sectors and occupational groups.

It's hard to be sure, but this is consistent with the view that the second author attended a university in the U.S., and discussed cussing in some recitation sections where he was a teaching assistant (or perhaps a student); and then repeated the experience at UEA in Britain.

Now, I've got no problems with ethnography; and discussions with students are a great source of ideas and anecdotes. But let's be clear: this "study" consisted of one person's observations about cussing at a part-time job he once had, along with notes from six classroom discussions with students about their own experiences with cussing on the job.

The ironic thing is that pretty much all the reporters writing stories about this research, and most of the people reading those stories, have had at least this wide a range of experience of cussing in the workplace, and probably almost as much experience of swapping stories about cussing among different groups.

Thus they have roughly as much empirical basis as Baruch and Jenkins for evaluating hypotheses about how workplace swearing works and what its positive and negative impacts might be. But because this is a "study" presenting the results of "research", the journalists and their readers are all solemnly considering it as testimony of a qualitatively different character from their own life experience.

Something strikes me as really funny about this. In particular, it reminds me of one of my favorite pieces of science-news satire, "New Study Finds College Binge Drinking To Be A Blast", The Onion, 3/24/1999:

AMHERST, MA—Researchers at the University of Massachusetts released a surprising new study Monday indicating that, contrary to long-held beliefs about its destructive effects, collegiate binge drinking is a fucking blast.

"Data collected at bars and fraternity parties on the UMass campus has yielded unexpected conclusions with regard to the practice of binge drinking," study head Dr. Albert Greaves said. "Over the course of our research, a consistent pattern emerged demonstrating that binge drinking seriously kicks ass."

What I like best about this article is the blending of rhetorical styles -- for example:

According to Greaves, much of the UMass team's research was conducted at a party at this one guy Matt's place. "My colleagues and I were doing beer bongs, keg-stands, Jell-O shots, Jager shots—you name it," Greaves said. "We were totally binge drinking and just having a great fucking time. The best part was the crowd—the study was packed, and there was this amazing random sampling of hot chicks. I was so drunk, I couldn't figure out what the source of the unusually large hot-chick sample was, but by that point, I really didn't care."

When the keg was tapped, Greaves and his team went looking for a place to gather more data. "We heard there was this awesome study on Church Street, but we didn't have the address, so we just went wandering around," Greaves said. "We eventually wound up walking into this complete other study where we didn't know anyone. Unfortunately, it turned out to be totally lame—most of the people there were in the non-drinking control group. We had fun for a little while busting on them, but pretty soon we split."

Another example:

"Dr. Schmid is what we scientists term a fucking booze monster," team member Dr. James Podriewski said. "This one time, we needed a whole bunch of Wild Turkey and tonic water for a study that was just getting going at midnight, so we sent him out to this store that's open until 2 a.m., and we're waiting for, like, hours until he finally comes back, and he doesn't have any of the stuff, but he's carrying this big fucking railroad-crossing sign, and he's all like, 'Guys, check out the sign I found.' It was funny as shit. I swear, I was laughing so hard, I almost left a urine sample all over my pants."

The use of this is especially nice.

For more on cussing from a cross-cultural and comic-strip perspective, check out this Language Log post and the links at the bottom of it, especially this one.

Posted by Mark Liberman at October 22, 2007 07:03 AM