Australian Linguistic Society

Researching language in the workplace: a participatory model

Maria Stubbe, School of Linguistics and Applied Language Studies, Victoria University of Wellington,


The Language in the Workplace Project1 was begun in 1996 with two broad goals in mind: (1) to analyse the features of effective interpersonal communication in a variety of workplaces from a sociolinguistic perspective; and (2) to explore the practical implications of these findings for New Zealand organisations. The initial motivation for this project was an obvious gap in the available research evidence on workplace communication, especially research specifically relevant to the New Zealand context. The workplace is a setting where many people spend a great deal of their time, and one which is socially very complex. From an academic perspective, therefore, it is a potentially fruitful ground for looking explicitly at areas of interest to sociolinguists such as language and power, cross-cultural communication and gender issues, as well as more general questions relating to the use of language in this setting (see also Holmes 1998). From an applied perspective, effective communication with clients and colleagues is universally recognised as being crucial to the smooth and productive running of an organisation, and as a vitally important area for human resource development. Quality research is a prerequisite for developing soundly based communication skills evaluation and training programmes.

However, there is remarkably little research which examines in detail how people actually communicate verbally with their colleagues at work on a daily basis, and how they use language to manage the inevitable tensions between their various social and professional roles. Although there is a considerable body of linguistic research on institutional discourse, with some exceptions (e.g. Clyne 1994), this has tended to concentrate on specialised contexts such as classrooms, courtrooms and doctor-patient interactions. In the management field, most reported studies on organisational communication use material derived from indirect sources such as self-report data, interviews with significant personnel, and anecdotal observations. There is also a need for applied research which recognises diversity and leads to open-ended observations and guidelines. Many management textbooks and courses fail to acknowledge the complexities of real language use, and are inappropriately rigid and prescriptive in their recommendations for improving workplace communication. This is perhaps unsurprising, given the dearth of research based on genuine, "real life" data drawn from actual interactions in workplace contexts, which is where people actually conduct the day-to-day business of an organisation.

The challenge for the Language in the Workplace Project team was to create a methodologically sound research design for collecting and analysing workplace data which also provided a credible basis for developing practical applications. This raised a number of interesting methodological and ethical issues of a sort which do not usually face researchers engaged in more traditional types of sociolinguistic research. The next section outlines the various factors which had to be built in to the project design, and describes the participatory research model developed by the project team as a way of addressing these. The final part of the paper illustrates how this innovative approach has considerably enriched the linguistic analysis of the data, while at the same time producing a range of positive and practical outcomes for the individuals and organisations involved in the project.

Database and methodology

During 1996 and 1997, the Language in the Workplace team collected a corpus of 330 interactions in four New Zealand government agencies. These workplaces include one where there is a high proportion of women and one with a high proportion of Maori2 workers, in addition to two workplaces with an ethnic and gender balance more closely reflecting the New Zealand norm. Altogether, 251 people (152 women and 99 men) from a range of ages and levels within each organisation were recorded. In terms of ethnicity, 111 of the participants are New Zealand Pakeha, 114 are Maori, and 26 are from other ethnic groups, such as Samoan or Chinese. The bulk of the data consists of small, relatively informal work-related meetings and discussions ranging in time between twenty seconds and two hours. Such meetings fulfil a wide variety of purposes in these workplaces: to plan, to convey instructions, to seek advice, to check reports, to solve a problem or do a task, to provide feedback, to evaluate proposals, and so on. The database also includes some other types of interaction, such as social talk and telephone calls, in addition to a number of larger and generally longer meetings which were videotaped. This data comes mainly from policy and advisory units, an environment where talk is integral to the core business of the workplace. It thus provides an especially rich source of data for investigating how language functions in the ongoing construction of workplace relationships.

In designing the project methodology, the team had to consider and weigh up a number of factors. In the first instance we needed a practical method of collecting a reasonably large and representative database comprising high quality, natural interaction data from everyday workplace contexts. This data had to provide a suitable basis for detailed discourse and pragmatic analysis, and allow us to take explicit account of socio-cultural factors such as gender, culture, and relative status. The data collection process could not be too intrusive (in order to avoid "tape shyness"). It also had to meet certain logistical and technical requirements- for instance it could not be overly time-consuming or disruptive for either the individuals or organisations involved, it had to be achievable within a limited time frame, the technical quality of the recordings had to meet a minimum standard, and each recording had to be accompanied by a certain amount of demographic and contextual information.

Secondly, because we wanted our analysis to be as accurate as possible, and eventually to feed the results into practical applications, we needed to establish an ongoing relationship with our participants. Because it was difficult to predict at the beginning exactly what form this collaboration might take in a given organisation, the methodology had to be sufficiently flexible and adaptable to evolve with the project. The process of data collection and analysis also had to be managed in such a way that some issues of concern to the participating organisations and other "end-users" of the research could start to be addressed within a reasonably short time frame. In other words, we had to build in some short-term outcomes as well as the longer-term results more typical of the research process. Even though the participating organisations were convinced that the research had the potential to be useful in the longer term, they still expected a more immediate, concrete benefit in return for their investment of staff time and goodwill.

Thirdly, the research design had to meet a number of stringent ethical requirements, the most immediate of which were ensuring that genuine informed consent was obtained from everyone who was recorded, and guaranteeing confidentiality to the individual informants and organisations involved in the project. These are of course quite usual considerations in any social science research, but they acquired an extra edge in the workplace context, where people are very aware of the need to protect sensitive information and to look after their relationships with their clients and colleagues. Even though they knew that the content of their interactions would not be the focus of the research, our participants were initially very wary of losing control over any data that could potentially identify and compromise individuals or the organisation concerned. Finally, we wished the research process to be as open and empowering as possible, and to avoid any exploitation or misrepresentation of our informants. We thus based our design as far as possible on the action research principle of "research on, for and with" our participants (Cameron et al 1992: 22)

The project team devised an innovative research methodology in order to meet the project's objectives and accommodate the various design constraints outlined above. Essentially, this involved establishing and maintaining an ongoing dialogue with the organisations involved, and giving participants maximum control over the collection and subsequent uses of the data. The research team and participants collaborated in setting the research agenda and exchanging relevant information, thus allowing the particular needs identified by each party to be addressed, and meeting the overall goals of the project.

In outline, the participatory model worked as follows:

1 Once a candidate organisation was identified, we held an initial meeting with senior management to discuss our research proposal, seek their input and gain their agreement in principle to being involved. This was followed by an open presentation to all interested parties in the organisation. Here we provided background information on related research, the aims, methods and expected outcomes of the project, and gave opportunities for questions and comments.

2 A group of volunteers representing a range of roles and levels within the organisation then agreed to record a representative range of their everyday interactions at work over a period of two to three weeks. Some kept a recorder and microphone on their desks, others carried the equipment round with them. All those involved agreed to be recorded in advance of any data collection. Throughout the data collection process participants were free to edit and delete material as they wished. Even after they had completed recording and handed over the tapes, they could ask us to edit out material which they felt in retrospect they did not wish us to analyse or use in any published material.

Every person recorded was asked to fill in a sheet providing demographic information, as well as explicitly agreeing that the data could be used for linguistic analysis by the project team. Ethnographic data was also collected by means of workplace observations, contextual notes provided by the participants at the time of recording, briefing and debriefing sessions with volunteers, and follow-up interviews with selected informants involving the reflexive analysis of data extracts and summaries.

3 At the end of the first year, once the initial data collection and some preliminary analysis were complete, formal feedback sessions were held at each workplace. These took the form of open forums, including a report on progress and findings to date, discussion of any issues arising and possible future directions and applications for the research. Liaison with each workplace has continued since via Research Associates attached to the project in two of the workplaces, and through several other individuals with a special interest in the project. In particular, we consulted regularly as we developed the next stage of the research, seeking input on what directions and practical outcomes would be most appropriate and useful to their organisations.

4 During the intensive data analysis phase of the project, we selectively followed up certain interactions. This follow-up functioned as a way of (1) checking that our interpretations were on track, (2) trialling some of ways the project data and our analysis of it might be applied, and (3) providing another opportunity for feedback in both directions. Once a particular aspect of the data analysis was complete, we sought feedback on the output from our participating workplaces. To date, this has taken the form of reviewing academic papers, piloting training materials and workshops, and collaborating in the development of a workplace communication evaluation model.


To what extent did this methodology deliver the desired outcomes? Overall, it has greatly exceeded our expectations, and is now being successfully applied, with some minor modifications, in a range of other workplaces linked to the project. Sacrificing control over exactly what data would be collected and from whom might appear to be a major disadvantage. This did not in fact turn out to be so, as the "dripfeed" method of collection allowed us to monitor the data as it came in. We thus still achieved a representative mix of informants and data types, as summarised above. There were some other minor limitations on the types of data we were able to collect. For example, very brief transactions were generally not included by our informants for logistical reasons- it was generally too disruptive of the interaction itself or too time-consuming to obtain consent and provide the necessary contextual information. We also found that meetings involving more than three or four people were of limited use for analysis unless they were recorded on video, as it becomes well nigh impossible to distinguish all the voices and transcribe the interaction from an ausio recording alone. We did videotape some larger meetings, but on the whole these are less well represented in the initial database, a gap that has since been filled by meeting data from three other workplaces.

On the other hand, by leaving control of the recorded material in the hands of our informants, we developed an excellent research relationship with them. As a result, they trusted us with a wide range of fascinating material in return for guarantees of anonymity and confidentiality. This collaboration has generated a large and very rich corpus of naturally occurring spontaneous workplace talk for analysis by the project team - the only extant one of this size that we know of. This is first and foremost natural data: after an initial period of self-consciousness, people usually relaxed and reverted to their normal behaviour- they generally reported that the tapes were representative of their routine interactions, and they certainly sounded so to us. The following excerpt is a typical example. (All names are pseudonyms).

(1) Context: Bruce comes over to Brenda's desk with a query

Brenda:	hello Bruce [laughs] oh no don't ask me about that
	funding I don't know anything about it [laughs]
Bruce:	[laughs] well I-
Brenda:	yes I know [laughs]
Bruce:	Yvette gave me these and said um our section's 
	supposed to be doing something with these-
Brenda:	right
Bruce:	but we haven't been
Brenda:	yeah
Bruce:	and er she said you might know something about 
	them so to come and have a chat to you
Brenda:	well no this one here our section have had nothing to do
	with it I always understood your lot were doing it all along
Bruce:	well that's interesting 'cause we haven't been
Brenda:	I wonder if anyone in X might know something about it? 
	you might have a quick chat to Liam there but we've 
	definitely done nothing on it
Bruce:	right well I'll go and have a chat to Liam about it 

As people became more accustomed to the recording process, the amount of material they edited, or which they asked us to edit, decreased dramatically. One striking example of this occurred when one informant with an employment-related grievance recorded a full and frank discussion about it with a friend over lunch, and also a subsequent lengthy meeting with a senior manager about the same issue. Moreover, because people were recording a large number of their interactions within a short space of time, we regularly obtained a series of interactions which were similarly linked in some way.

For many informants, taping often became a routine part of the working day. One interesting consequence of this was that over time the topic of the recording process developed into a legitimate topic of small talk. In other words, the recording process became so normal for some participants that it was relegated to the periphery of their attention and reference to it became routinised, as in this example:

(2) Context: Barbara, a policy adviser talking to Ruth, her manager

Barb:	hey Ruth, I've got a little problem
Ruth:	yeah? what with?
Barb:	well I was just sending you a complicated email
        but if you've got ten minutes- 
Ruth:	TEN minutes? I'd just like to say I'm up to tape 2, side 2
Barb:	Well done
Carol:	(in background) you talk a lot
Ruth:	no no what I did was taped the meeting which
        was Hinemoa talking to us about	a lot of
        interactions. So that was an hour and ten minutes
        so that's why I'm ahead of the game
Barb:	oh well done um I've finally just had a look at 
        these questions and ... 

In another interaction, an ironic reference to whether or not swearing is suitable material for recording provides a humorous way to close what has been a somewhat contentious discussion:

(3) Context:Two analysts debate which one of them should take on an unwanted task

Ann:	I can't take that on because with what I'm doing here
	and my manager is quite adamant about sticking to it
Brian:	mm see and it's nothing to do with us
Ann:	yeah and-
Brian:	so the only generalist at the moment is Graeme
Ann:	yeah
Brian:	there's no other analysts working on it
Ann:	no- no other analyst working on this area
Brian:	okay [laughs] okay fuck up
Ann:	[laughs] you want to swear [laughs] you're allowed 
	to swear because they want to hear swear words
Brian:	[laughs] oh well I'm going fuck
Ann:	okay thanks for being recorded
Brian:	that's all right  

We also received a great deal of positive feedback from informants about the benefits of the data collection process to them personally and professionally. They reported that they had gained useful insights into actual patterns of workplace communication, as compared to their perceptions (e.g. the amount of gossip, small talk, or swearing which occurred; who they tended to interact with, for what purposes, and where.) The participants found it particularly illuminating to have the opportunity to look in detail at some of their own interactions during the feedback and follow-up sessions, and to have the opportunity to provide the research team with suggestions to guide the next stage of analysis.

Another fascinating aspect of using this participatory model was the degree to which going through the data collection process with our participants was in itself a vehicle for finding out a great deal about the organisational culture of each workplace, including power and role relationships, typical patterns of interaction, and general communication issues. We were thus able to gather a great deal of detailed ethnographic data indirectly. For example, the managers among our volunteers generally filled their quota of tapes relatively quickly. This was in part because their jobs involved them in a lot of talking, but also because it was often easier for them to gain the cooperation of other participants in an interaction than it was for more junior staff members. Junior staff tended to be more conservative initially about what constituted material that was too sensitive to record, and often had rather different perceptions of the prevailing organisational culture to those of their managers.

The research process allowed us to develop a relationship of mutual trust with our informants, so that we were able to come back later to ask more questions, get feedback on our interpretations, and obtain permission to use specific transcripts for presentations, workshops, and publications. The methodology adopted for this project has thus made it possible to adopt a qualitative, grounded approach to the analysis of data, and to start on detailed analyses at a relatively early stage in the life of the project. Aspects of the data analysed to date include the construction of professional identity, collaborative problem-solving, meeting structures and processes, directives, and the functions of humour and small talk (see for example Holmes, Stubbe and Vine in press; Holmes 1998; Holmes in press; Stubbe, in press). This information is now starting to be used in the development of a range of practical training workshops and materials.


The unique participatory research model described here initially evolved in response to the particular challenges inherent in collecting a large amount of natural interaction data in a workplace setting. However, this methodology has turned out to have a number of significant advantages over more traditional sociolinguistic methods, and as such has the potential to apply well beyond this one project. We have obtained a valuable database, comprising both recorded interactions and a large amount of ethnographic information which has considerably enriched the analysis undertaken so far. This model has also produced ongoing benefits for the participating workplaces in terms of a range of practical outcomes. As such, it has proved to be an eminently practical means of marrying theoretical and applied research goals.


1 The methodology described in this paper was developed and implemented by Janet Holmes, Project Director and Maria Stubbe, Research Fellow, supported by Bernadette Vine and several research assistants who helped collect and transcribe the data. I would like to acknowledge here the invaluable contribution of all those individuals who allowed their workplace interactions to be recorded and analysed, as well as the assistance provided by Research Associates Harima Fraser (Te Puni Kokiri) and Frances Austin (Ministry of Women's Affairs). The Language in the Workplace Project is funded by the New Zealand Foundation for Research, Science and Technology.

2 The Maori people are New Zealand's original Polynesian inhabitants, now comprising approximately 15% of the population. Pakeha is the Maori word commonly used in New Zealand to refer to European people, primarily those of British descent.


  • Cameron, Deborah, Elizabeth Frazer, Penelope Harvey, M.B.H. Rampton, and Kay Richardson 1992. Researching Language. Issues of Power and Method. London: Routledge.
  • Clyne, Michael 1994. Inter-cultural communication at work. Cultural values in discourse. Cambridge: Cambridge University Press.
  • Holmes, Janet, Maria Stubbe and Bernadette Vine (in press). Constructing professional identity: "Doing power" in policy units. In Sarangi, Srikant and Celia Roberts (eds). Discourse in the Workplace: Communication in Institutional and Professional Settings. The Hague: Mouton.
  • Holmes, Janet 1998. No joking matter! The functions of humour in the workplace. Proceedings of the Australian Linguistics Society Conference July 1998. University of Queensland: Australian Linguistic Society.
  • Holmes (in press). Small talk in government departments. To appear in Justine Coupland (ed) Small Talk. London: Longman.
  • Stubbe, Maria (in press). Striking a balance: Language, gender and professional identity. To appear in Proceedings of the Fifth Berkeley Women and Language Conference April 1998. Berkeley, University of California: Berkeley Women and Language Group.