An exploratory study of facilitation in distributed requirements engineering

Abstract

There is an increasing need to build an understanding of group facilitation in distributed requirements engineering. Group facilitation is an important element of group approaches to requirements engineering, which is being challenged by the emergence of distributed practices in software specification and development. Thus thorough empirical investigation of human facilitation in computer-mediated requirements meetings is needed. This paper presents an exploratory study of facilitation in distributed requirements teams. Three professional facilitators mediate 15 three-person groups in an experimental simulation of software requirements negotiation. Facilitation in face-to-face meetings is contrasted with four group settings in which the facilitator is physically separated from the group or co-located with key stakeholders. Rich qualitative and behavioural data enables an understanding of differences and similarities in the facilitation of face-to-face and distributed groups and of aspects that were detrimental or beneficial to their facilitation. The empirical evidence indicates that a reduced richness of social behaviours in computer-mediated group settings (1) made the group facilitation problematic but also (2) enabled certain facilitation support in the medium itself. The findings of the study are discussed in the light of existing models of the role of the facilitator. Advice from the expert facilitators is presented in the form of recommendations for the facilitation of distributed requirements teams, and worthwhile directions for further research are outlined.

Appendices

Appendices

1.1 Appendix 1

The role description for system users is shown in Table 4 and the requirements to be negotiated is shown in Table 5.

Table 4. Role description for system users

Table 5. Requirements to be negotiated

1.2 Appendix 2: SYMLOG analysis

SYMLOG [1] is a methodology for the observation, coding, and analysis of group dynamics. It assumes that group interaction can be measured in terms of three dimensions. Each dimension represents a continuum on which people can be positioned according to their behaviour as group members. Dominant or submissive behaviours associated with perceived influence and power constitute an upward–downward (U–D) continuum. The socio-emotional oriented behaviour is assessed along a positive–negative (P–N) continuum. It illustrates a person's concern for establishing and maintaining interpersonal bonds with others. The task-oriented behaviour is assessed on the forward–backward (F–B) continuum; it illustrates an individual's concern for performance.

These dimensions are the three orthogonal dimensions in a cube (see Fig. 9), where the one-, two-, and three-letter codes (e.g., PB, UNF) represent behavioural codes [29]. Using the standard SYMLOG coding schema, a field diagram of the interaction is generated (such as those in Figs 7 and 8). The field diagram is a visual representation of images of group members in a two-dimensional plane whose vertical axis is the F–B dimension and the horizontal axis is the P–N dimension. The image of each group member is represented in this field as a circle whose radius conveys the level of dominance. The larger the circle, the more dominant the person.

Fig. 9

SYMLOG three-dimensional space

These field diagrams are obtained as follows. For each interaction act, a SYMLOG coder records the following information: (1) the time of the act, (2) who is the sender, (3) who is the receiver, (4) a specification of whether the observed behaviour was overt or non-verbal, (5) the behavioural code, and (6) a comment describing the behaviour topic. For instance, a message sent by John to Mary showing dominant and unfriendly behaviour, at minute 23 of the interaction, would be coded "23 John Mary ACT UN 'Overrides others in conversation'."

After all interaction acts have been coded, raw interaction scores are calculated for each group member (initiator of messages) with respect to U, D, P, N, F, and B directions (see Table 6). For example, if a message sent by John was rated UN then 1 U is added to u _i and 1 N to n _i counters respectively. These counters are defined for each group member and are used in defining plotting data for the SYMLOG field diagram. These counters are shown for the P–N dimension, but it applies to the other two dimensions as well. Once the raw interaction scores are computed for all members, they are transformed into aggregated data for plotting SYMLOG field diagrams using the formula for Location of each Participant (Table 6). The field diagram summarises the average group behaviour for each member as a circle in a two-dimensional plane and whose radius conveys the level of dominance (the value of U–D dimension).

Table 6. Scoring procedure for the location of each participant on the field diagram