Variables, Operationalisation and Measurement

Both technical uncertainty and technical complexity are multi-dimensional constructs. Components of technical uncertainty and complexity are identified from the literature. Here, these components are treated as variables to indicate the levels of technical uncertainty and complexity. Operationalisation and measurement scales for each of the variables are further described below.

3.2.1 Technical uncertainty

Five components of technical uncertainty were identified from literature. It considers the lack of knowledge or absence of information of the state of the technical arrangements to carry out project work. Such absence of information may occur throughout a project at the input, process and output. We adopt Perrow’s (1970) perspective of considering technical uncertainty as the knowledge required to carry out organisation tasks, whose framework has been applied by Gidado (1996) and Shirazi, Langford and Rowlinson (1996) in the construction context. Perrow’s framework was found to have high validity (Withey, Daft and Cooper, 1983). In line with the scale developed by Withey et al for Perrow’s framework, we operationalise input, process and outcome uncertainty in terms of the degree to which project members has clear

understanding of what is needed to execute the project tasks. A low level of understanding would be associated with high level of uncertainty (i.e. a reverse scale).

Indicators of uncertainty are listed below.

(a) Input: The degree to which project member clearly understands the design parameters, interfaces, resources required and techniques used.

(b) Process: The degree to which project member clearly understands the major steps in the process, properties of the material and techniques.

(c) Outcome: The degree to which project member clearly understands the scope of work, employer’s project goals and quality requirements.

Shenhar (2001) obtained objective measures of uncertainty by considering the number of design cycles and/or prototype testing. It was established from Shenhar’s studies that engineering activities increased with the level of technological uncertainty.

In line with Shenhar’s scale, we further operationalise process uncertainty as the number of iterations of design development and prototype testing before approval or mass production.

Conceptualising technical uncertainty as newness of technology was based on the work of Shenhar (1997, 1998, and 2001). He measured technical uncertainty by the percentage of new technology employed in the project, which has been criticised by Williams (1999) as a variable difficult to quantify. In contrast with Shenhar’s measurement of newness of technology as percentages, we adopted the scale developed by Swink and Calantone (2004) in the manufacturing context, which considers the degree to which technical means are new to the project member; i.e. a scale that is relative to the project. This is based on the rationale that the more that technical means are novel to the project members, the less the project members are familiar with the means and hence the higher the level of uncertainty. The operational measure is:

(a) Technological novelty: The degree to which design methodology, construction techniques and materials are new to the project member.

Finally, technical uncertainty arising from variations to the project is operationalised by considering the frequency to which changes occur. A high frequency would be associated with high degree of uncertainty. This variable considers the following aspects of change. The operational measures of each of the variable for technical uncertainty are summarised in Table 3.1.

(a) Frequency of changes arising from changes in employer’s requirements and project goals.

(b) Frequency of changes arising from variations to project interface that necessitated design adjustment.

(c) Frequency of changes arising from variations to project interface that necessitated adjustment of construction activities.

Table3.1 – Operational measures of Technical Uncertainty

Variables Operationalisation Reference

Input Knowledge of the inputs Withey, Daft & Cooper (1983)

Process 1. Knowledge of the technical means

2. Number of engineering iterations Morris & Hough (1987) Withey, Daft& Cooper (1983)

Shenhar (2001) Outcome Knowledge of the scope of work and

employer’s project goals Tuner & Müller (2003) Withey, Daft & Cooper (1983)

Technological

Novelty Degree to which project members are familiar

of the means Swink & Calantone (2004)

Variation Frequency of changes De Meyer, Loch & Pich (2002)

3.2.2 Technical complexity

From the literature review, technical complexity comprises of 4 variables: scope, differentiation, interdependence and concurrency.

In selecting an appropriate scale to measure scope, several operational scales were considered. The number of project activities identified from the project‘s programme schedule and work breakdown structure was commonly used as an indicator (e.g. Shenhar, 2001). From our experience, employers may likely dictate the level of detail to be reported in the works programme. The detailing of programme schedule is also subject to the experience and personal preferences of project members.

Hence, it may be difficult to ensure consistent representation of project scope if we consider the number of project activities as scope. Shenhar (2001) offered an alternate measure of complexity by considering the number of subsystems involved in the project.

Baccarini (1996), Ivory and Alderman (2005) and Williams (1999) had also associated complexity with the number of project elements.

To the civil and architectural discipline, “system” is not a common language to the practitioners. Moreover the number of project elements presents high variability in its measure. Hence we adopted Shenhar’s alternate measure of project scope – dollar value. This indicator of project scope is popular among questionnaire surveys in the construction industry (e.g. Chen, 陳宗儀, 2003; Chiu, 邱文杰, 2002) and could also be used to categorise groups of projects.

Technical differentiation considers the diversity and variety of the technical means necessary to deliver the project outcome (Baccarini, 1996). Complex engineering

projects often require inputs from different engineering disciplines. Complex projects are also characterised by multiple combinations and assortments of materials, construction machineries, techniques and trade skills (Nam and Tatum, 1988). The wider the variety of engineering disciplines, materials, machineries and techniques are involved in the project, the more is the project differentiated technically (Baccarini, 1996; Morris and Hough, 1987), and hence the higher the level of complexity. We consider the following as indicators of technical differentiation.

(a) The variety of engineering professions involved in the project.

(b) The variety of materials and equipments used in the project.

(c) The variety of construction machineries employed in the project.

(d) The variety of construction techniques and technologies employed in the project.

The third variable considered under technical complexity is technical interdependence, which is a concept applicable to tasks in execution. Thompson’s (1967) scale of interdependence has been applied extensively in the literature. Morris (1988) and Williams (1999) considered that construction activities are inter-related reciprocally. They further suggested that the predominant type of interdependence in the construction industry is reciprocal interdependence. If one is to apply the Thompson scale in the survey, we are of the opinion that the Chinese textbook translations¹¹ of sequential, pooled and reciprocal interdependence is not easily understood by practitioners of the construction industry. Hence we focused on the semantic meaning of

11 See 李茂興、李慕華、林宗鴻，1994；彭文賢，1996；龔平邦，1984；張苙雲，1990. “Sequential

interdependence” 中譯「連續性互動、互賴」；”pooled interdependence” 中譯「波及性互動、互賴」；”reciprocal interdependence” 中譯「互惠性互動、互賴」.

the word ‘interdependence’ and choose to operationalise it by the extent to which activities are dependent on each other during its execution. The following list of activities is chosen to represent interdependence.

(a) Interdependence between different engineering professions.

(b) Interdependence between design and construction activities.

(c) Interdependence between different construction activities.

(d) Interdependence between functional departments within the project organisation.

(e) Interdependence between different subcontractors and sub-suppliers.

(f) Interdependence between the project and other employer’s contractors.

The last variable concurrency considers the overlap of activities in their execution.

We operationalised concurrency by the degree to which activities overlap in execution in the same given frame of time. Following Williams (1999), concurrency increases technical complexity. The more that different activity is overlapped in their execution, the higher the level of technical complexity. The measures are as follows.

(a) In the same time frame, the degree of concurrent execution of different design units.

(b) In the same time frame, the degree of concurrent execution of design and construction activities.

(c) In the same time frame, the degree of concurrent execution of different construction activities.

The operational measures of technical complexity are summarised in Table 3.2.

Table3.2 – Operational measures of Technical Complexity

Variables Operationalisation Reference

Scope Dollar value of project Shenhar (2001)

Technical

Differentiation Number of technical disciplines, materials, machinery and techniques necessary to delivery project result.

Baccarini (1996)

Interdependence Degree of interdependence between entities, activities, internal and external interfaces. Concurrency Degree of overlapping execution in any

given time frame.

3.2.3 Project organisation structure

In project organisation structure we consider some of the more important structural dimensions arising from the literature review, namely: size, organisational differentiation, integration and information process.

Size is operationalised by the head count of project staff. Having considered the temporary nature of projects, we dichotomised staff number into full time and part time staff (Robbins, 1990).

Organisational differentiation refers to vertical hierarchy and horizontal departmentalisation (Robbins, 1990). Vertical hierarchy was measured by office

hierarchy (Eccles, 1981) and considers the number of levels in the vertical structure.

Since subcontracting is a common practice in the construction industry, we further operationalised organisational differentiation by the number of subcontractors and suppliers involved in the project. Although several tiers of subcontract relationship are common in practice, we consider only the main subcontractors, in line with the project level of analysis, who have direct contractual relationship with the main contractor.

We operationalised integration by adopting the Galbraith and Kazanjian (1986) scale¹², which has wide application in both organisation and project management literature. The level of integrative effort¹³ is measured by an index calculated using the equation given below. This equation assigns different weights to different integrative mechanisms in the rank order given in Table 2.10, and is consistent with Daft (2004) given in Fig. 2.6. Separate indices for internal and external integrative efforts are calculated. This method is consistent with Galbraith and Kazanjian (1986) that the integrative mechanisms are additive and no mechanism is substitute for another. This is also in line with Lawrence and Lorsch (1967a) that effective organisations employ more integrative devices than less effective organisations.

Integrative effort = (rules x 1) + (regular report x 2) + (planning x 3) + (direct contact x 4) + (integrator x 5)

Finally, information processing concerns the manner by which information is

12 The list of integrative mechanisms in Table 2.10 consists of 9 items. Respondents to the pre-test survey commented that engineering project is itself a “temporary task” and/or a “team”. These 2 items were dropped from our questionnaire. Furthermore, the respondents were not able to distinguish the conceptual difference between “liaison role”, “integrating role” and “integrating department”, these 3 items were combined as one in our questionnaire. Finally, we have targeted our respondents as the highest ranking officer of the project organisation, and therefore “hierarchy” was also dropped.

13 The term “integrative effort” was adapted from Peng (1996) (彭文賢(民 85), 第七章, “互動能量”).

collected, processed and disseminated within the project organisation and among the subcontractors (Kmetz, 1984). We follow the procedure advocated by Daft and Lengel (1986) by measuring the frequency of use for different types of media, separately for within and outside of the project organisation. An index representing the intensity of information processing was also created by assigning weights according to the

“richness” of the media (Daft and Lengel, 1986; Daft, 2004), expressed in the following equation.

Intensity of information processing = [freq. of using report] + [freq. of using letter] + [freq. of using e-mail] + [freq. of using telephone] + [freq. of using face to face direct contact] + [freq. of using meeting]