LITERATURE REVIEW
A study in 1998 found that three formal variables (i.e. co-department membership, supervisor-subordinate relationship, and gender) and two emergent network variables (i.e. task communication and acquaintanceship) were predictors of coworkers' perceptual congruence on their organisation's social structure. Perceptual congruence was defined as the extent members agreed on their perceptions of the organisation's social structure. Coworkers from the same department tended to perceive their organisation's social structure more similarly than coworkers from different departments. Coworkers who were in a supervisor-subordinate relationship had greater perceptual congruence on their organisation's social structure than those who were not. Coworkers from the same gender group also tended to have greater perceptual congruence on their organisation's social structure than those of different genders. In the task communication network and the acquaintanceship network, coworkers who reported mutual ties had greater perceptual congruence of their organisation's social structure than those who reported asymmetrical or null communication or acquaintanceship ties (Heald, Koehly, & Wasserman, 1998).
The data were collected through a series of three structured interviews with individuals in an organisation. The topics included individuals' task-related communication network, their acquaintance network, and their participation in the organisation's work flow. Individuals were also given a roster of all individuals in the organisation and asked to identify those with whom they communicated about work-related issues during the past two months and those whom they personally knew (Heald, Koehly, & Wasserman, 1998).
In 1995, a federally-sponsored agency in Tampa, Florida conducted an applied social network analysis study to assess the dozens of economic development organisations in the Tampa Bay area. Social network analysis was selected for its strength in assessing communication, relations, cooperation, and inter-organisational environments. The study sought to identify levels of communication, and assess overall levels of inter-agency coordination amongst several other explicit goals (Hagen, Killinger, & Streeter, 1997).
In this study, a written survey was designed to obtain information from 37 organisations. A matrix of inter-orgnanisational relations and frequency of formal communication could then be constructed. 31 out of the 37 organisations completed the questionnaire, with respondents indicating that they communicated formally with each other (those on the list) daily, weekly, monthly, yearly, or never. The survey was administered by mail, with phone and facsimile follow-up. Social network analysis was applied to the matrix to identify levels of inter-organisational communication, clusters of development activity, and leading organisations. Four network measures were selected: centrality, equivalence, centralisation, and density. Centrality was used to identify network leaders; equivalence was used to identify sets of network actors with similar roles in the network; centralisation measured the degree to which an entire network was focused around a few central nodes; and density was used to measure levels of uncentralised inter-organisational cooperation. The four network measures were individually analysed (Hagen, Killinger, & Streeter, 1997).
In another study, an analysis was conducted on the Mexican power network. Its objective was to examine the network's topology, function, and centrality values. The focus was the network's 37 core actors, who have played a central role in Mexican politics after the 1910 revolution. The study aimed to explain which of these actors influence political events - mainly the control of presidency - and defined the cliques and their changing values. It was discovered that the core network was actually a superimposition of structures, where an actor often belonged to various networks simultaneously (Mendieta, Schmidt, Castro, & Ruiz, 1997).
The study claimed that a network was formed by dyads interconnected by links, representing formal, informal, and organisational relationships. A network was the logical sum of networks or sub-networks. Each group or sub-network represented the points of coincidence of a set of groups based on their common belonging, or a common interest (political power). The group's bonding created a distinctive sense of belonging which impacted upon the nature of the network (Mendieta, Schmidt, Castro, & Ruiz, 1997).
In a 1994 study, Karathanos reviewed the theory on coalition formation and what it might mean to a manager in an organisation. She proposed that communication network analysis would be a useful tool for discerning particular types of organisational coalitions and subsequently dealing with them. She defined coalition as a means-oriented alliance among groups or individuals who differed in goals (Karathanos, 1994).
Coalitions were seen to predominate where there was a lack of security, especially in large, anonymous organisations. For psychological protection, people formed coalitions to counterbalance the formal demands of the organisation and make life within it more acceptable and meaningful (Karathanos, 1994).
In another study on the macro and micro level human resource implications of network analysis, the researchers claimed that the nature of work relationships was constrained by network and hierarchical forms of organisation. An understanding of and managing the networks in organisations might be used to improve existing programmes in achieving non-discriminatory employment practices (Stephenson, & Lewin, 1996).
The hierarchical model of organisation proposed in the study emphasised formal reporting relationships of subordinates to superiors, functional specialisation, and the division of labour as the leading principle of job design. Nevertheless, attention was brought to the importance and power of informal groups, collectives, and cliques within formal organisations. Informal collectives of employees within formal organisations were claimed to be of critical importance because they reflected the actual behaviour of real-world organisations and could strongly influence the performance of organisations. The researchers also identified the behaviour of "organisational networks", which were defined as collectives and groupings of employees formed in response to certain orgnaisational characteristics (Stephenson, & Lewin, 1996).
Network data were collected using network survey data, personnel records, in-depth interviews, and selected public documents. Respondents were asked to name the people in their respective divisions or work-group with whom they had the most interaction on a variety of dimensions. They were also given a form to identify their frequency of interactions with other members on the list (Stephenson, & Lewin, 1996).
What was interesting in the findings was that the promotion of a key player in a network could "effectively catapult" the person out of the network into a very different world of hierarchical authority, where "old friends become suspicious and remote" (Stephenson & Lewin, 1996). Hence, management could take that into account in selecting its candidates.
In another study, network analysis was used to describe purchasing workflow patterns within an industrial firm. Prescribed networks (i.e. hierarchical-level departmental membership and product-purchasing assignment) and emergent networks (i.e. position on the organisational boundary and centrality links within the firm's buying system) were investigated. This study was part of a larger project to conduct an in-depth investigation of the complexities of organisational buying, and the effects of one's position in the prescribed and emergent networks on the influence of purchasing-related decisions (Buckles, & Ronchetto, 1996).
In this study, network analysis was employed using an in-depth case approach, which examined the purchasing networks within the organisation. Data collection was conducted through the use of a snowballing technique, beginning with personal interviews to identify all those within the emergent purchasing workflow network, and these same participants identified others who fitted the same qualifications of the study. Assessments of specific variables were also done through a written questionnaire (Buckles, & Ronchetto, 1996).
The study of communication and social networks is seen to be applicable to all types of organisations. The increasing importance of understanding and managing formal and informal systems among members is crucial in improving the overall effectiveness of an organisation.

B-LEVEL MODULE

LITERATURE REVIEW
Network analysis is a systematic means of examining the general configurations of communication relationships, both formal and informal, within an organisation (Johnson, 1993). The resulting "communication maps" can assist in the identification of opportunities and problems (Hamilton, 1987), such as where information flow is blocked or overloading the communication network, who is blocking or overloading the information flow, and thereby construct new structures to reduce information blocks or overloads (Downs, 1988). It can also lead to the restructuring of formal and informal channels so that they complement each other (Hamilton, 1987).
Network analysis is concerned with flow rather than content, and finds out who is talking to whom (Hamilton, 1987). There are three sets of properties of networks:
1) transactional content - refers to what is exchanged by the social subjects, such as expression of affect, influence attempt, goods or services, and information;
2) nature of the links - this refers to the strength and qualitative nature of the relation between two social subjects, such as intensity, reciprocity, clarity of expectations, and multiplexity; and
3) structural characteristics - the overall pattern of relationships between the system's actors, such as size, density or connectedness, clustering, openness, stability, reachability, centrality, liaison, bridge, gatekeeper, and isolate (Tichy, Tushman, & Fombrun, 1992).
Network analysis can be used to identify several network roles, such as isolates, opinion leaders, gatekeepers, cosmopolites, bridges, and liaisons. Isolates are members who have minimal contact with others in the organisation. They are either "hiding out" or are being avoided in the organisation. Opinion leaders are informal leaders who guide organisation members' behaviours and influence their decisions, although they do not necessarily hold formal authority that is prescribed by the organsation. Gatekeepers are in the middle of a network and they control information flow among organisation members. They have the power to decide what information is or is not important and can help members avoid information overload by channeling only important messages to them. Cosmopolites are individuals who connect the organisation to its environment. Bridges are organisation members who connect a clique to which they belong with members of another clique. These individuals help share information between cliques and facilitate inter-group coordination. Liaisons are similar to bridges, but they connect two cliques without themselves belonging to either one (Kreps, 1990).
Internal organisational communication consists of formal and informal communication channels. Formal communication channels are dictated by the planned structure established for the organisation and rarely satisfy completely the information needs of its members, so members develop an informal system or grapevine to gather information which they cannot get from formal channels. Informal communication channels are not planned and generally do not follow the formal structure. Instead, informal channels emerge out of natural social interactions among organisation members (Kreps, 1990).
Communication flows can be downward, upward, or horizontal. Downward communication flows from upper management to lower levels in the hierarchy; upward communication flows from lower-level employees to higher-level personnel; and horizontal communication flows among members who are of the same hierarchical level (Kreps, 1990).
The central unit of analysis in a network is the communication relationship between any two organisation members. Networks differ in size and structure. There are three types of communication networks: the total system networks map the patterns of communication throughout the entire organisation; clique networks identify groups of individuals within the organisation who communicate more exclusively with one another than with other members; and personal networks are the individuals who often interact with an organisation member (Kreps, 1990).
Network analysis can be done by examining lists of telephone calls or non-intrusively observing the members by standing in a corner in the actual organisation (Hamilton, 1987). This is also known as residential analysis, where the auditor conducts observations of communication behaviour within the organisation (Hellweg, 1997).
In any method of network analysis, the auditor should aim for a complete record over a given time of who talks to whom, and of all connections and instances. Completeness is vital - including all contacts, whether job-related or not. However, network analysis should not be used independently, but triangulated with other instruments as it does not have the detailed accuracy necessary in a stand-alone methodology (Hamilton, 1987).
There are four basic network-data collection methods:
1) positional analysis which entails the use of formal organisational data;
2) reputational or attributional method which attempts to get behind the scenes by using judgments of selected community members who are the "first-rate leaders";
3) decisional analysis which selects a number of key issues, identifies participants in decision making, determines outcomes and then appraise relative influence; and
4) interactional method which asks individuals to report their interactions or influence attempts over a period of time and for particular content areas (Tichy, Tushman, & Fombrun, 1992). Examples of this method include asking employees to maintain a communication diary by noting down communication episodes that have occurred for a specified period of time, the initiator of the communication, the channel involved, the purpose of the communication, its length, and an evaluation of its usefulness, timeliness, importance, and accuracy, or asking employees to fill in questionnaires about who they communicate with, how frequently, and for what purpose through the process of recall (Hellweg, 1997).
To conduct a successful network analysis, there are four key principles to keep in mind:
1) data should be collected in a logistically efficient manner - identifying who should be studied, what message topics are important, and what work groups to assess;
2) data should be analysed so that conclusions are scientifically justified;
3) data should be interpreted so that results are intelligible; and
4) conclusions should be integrated into the overall strategic plan for organisational development (Downs, 1988).
Some general steps are further recommended:
a) give each person a form and a list of all the people in the organisation and ask them to check off each interaction on the form and the time it takes place;
b) instructions should be clear and simple;
c) if possible, the auditor should pre-test forms and instructions with a small random group. The auditor should also hand-deliver the forms, instruction sheets, and return envelopes to each person the day before; and
d) give people who interact with many people in a day, e.g. clerks and receptionists, special help in filling in their forms so that they can handle them more efficiently (Hamilton, 1987).
The auditor should also find out how typical the day was for each person. The aim is not to find out if things "always" or "never" happen the way they appear in the network analysis data, only that they are typical. If the day in question is 80 per cent typical, it is adequate for audit purposes (Hamilton, 1987).
There are two approaches to study the data - the macro and micro approaches. The macro approach is to create a simplified master chart that records the total number of interactions that each person participates in. Identify the frequent communicators, isolated members, those who carry an average communication load, and groups of people who set a norm for their type of work. The micro approach is to group the forms together by department, section, and group of people whose forms indicate that they communicate regularly. Using the matrix proposed by Hamilton (1987), the auditor can find out how the people within these units interact. The auditor can also draw a network diagram which identifies frequent communicators easily. It is also possible to find links or people who hold clusters together and isolates (Hamilton, 1987). The optimal approach would be to conduct both the macro and micro approaches to obtain a more complete picture of the organisation networks.
Johnson (1993) suggests there are three primary means of depicting network configurations: communigrams, individual patterns of relationships, and network indices. A communigram is a graphic representation of the network of relationships in the organisation. Individual patterns of relationships incorporate network roles, where an individual's communication role is determined by the overall pattern of his or her communication linkages with others, e.g. roles of a liaison or linking pins. Network indices refer to the use of various mathematical formulae or indices to reflect particular patterns of organisational communication relationships (Johnson, 1993).
From the network diagrams of the departments, the auditor can determine the most appropriate communication network from the configurations. Possibilities include the chain, circle, wheel, and Y-fork. The auditor can also compare the conceptual maps with the physical locations of the people, and find out how physical location affects communication. Following that, the auditor can compare the conceptual and spatial maps with the organisation chart to learn how much the formal lines of responsibility reinforce or counter actual communication. If possible, the auditor should compare two "runs" of network analysis to see whether time affects the flow of communication (Hamilton, 1987), i.e. conduct the same network analysis on the same organisation twice with a time lapse in between.
There are a number of computer algorithms available for analysing sociometric data (Hutchinson, 1992). Computer programs that are capable of drawing scattergrams, networks, and matrices will be useful to the analysis (Hamilton, 1987).
A useful framework for discussing link properties is the model of social interaction developed by Johnson, which posits that elements at deeper phenomenal levels, such as emotions and relationships, determine those elements that are at more surface levels, such as content. A fundamental premise of the model is that social interaction is characterised by the manifest acts usually observed during the course of an interaction, as well as the elements that underlie and determine these acts. Relationships and emotions are the two underlying elements in Johnson's model of social interaction (Johnson, 1993).

One of the limitations of network analysis is that there exists the assumption of a closed system, which imposes imaginary boundaries which may not exist in reality. Other limitations include the necessity for a high response rate, and the increased interviewing or observational expenses if more detailed information is needed (Tichy, Tushman, & Fombrun, 1992).
One of the most pressing problem areas of network analysis is the nature of overlap, or correspondence, between differing networks such as friendship as opposed to work. Organisations are composed of a variety of overlapping and interrelated networks of differing functions (Johnson, 1993). Most individuals belong to more than one social network in an organisation and that complicates the analysis of network data.
There is also a problem associated with the context of networks as to where to draw the boundaries around networks. Moreover, organisations are constantly changing, and the linkages between organisation members are increasingly becoming less permanent and stable. This raises questions of how appropriate network analysis is an indicator of communication structure (Johnson, 1993).
Network studies are also claimed to be too descriptive, as there is no explicit theoretical guidance at present (Johnson, 1993).
Nevertheless, network analysis has inherent advantages. Network analysis is a very practicable method for examining the overall configurations of communication contacts in a large social system. It also provides very specific and direct information on the pattern of an individual's linkages, since networks are based fundamentally on dyadic linkages. Moreover, network analysis permits the derivation of other measures from the aggregation of these individual linkages, such as clique identification and roles. The relatively value- and content- free characteristics of network analysis may also account for the popularity of the use of network analysis (Johnson, 1993).

OPERATIONALISATION OF NETWORK ANALYSIS
In order to conduct a network analysis on an actual organisation, the following network analysis form is an example of the actual form to be used. It is adapted from Hamilton (1987), in which she provided an example of a detailed network analysis form. However, the format of the form is altered to make it more user-friendly for the respondents. Some of the changes made are:
a) putting the instructions in boxes for emphasis and to allow more ease in reading;
b) using wider line spacing to allow more ease in reading and provide a neater presentation;
c) using different font sizes to create different emphasis on the various instructions;
d) providing examples of how the particulars should be filled in to provide the respondents with an idea of how to fill up the form, so as to reduce the chance of error in the information;
e) providing lines to enable the respondents to write their responses neatly;
f) adding an additional category of what the communication is about, i.e. the purpose of the interaction, to gain more insight whether the information exchanged is work-related, personal, or social;
g) the actual network analysis form is printed horizontally (landscape style) instead of vertically (portrait style) to enable the respondents to read and fill in the all the categories at one glance and to facilitate better presentation;
h) the margins on all four sides have been adjusted to fit the instructions, categories, and lines on exactly four pages to enable a neater presentation; and
i) the actual network analysis form to be used is page-numbered from one to four, but for the purpose of submitting it together with this paper, it has been page-numbered from 12 to 15.

Survey page 1

Survey page 2

Survey page 3

Survey page 4

A-LEVEL MODULE

IMPLEMENTATION OF NETWORK ANALYSIS
The organisation selected is part of a division in the Productivity and Standards Board of Singapore (PSB). There are a total of five departments under the CEO, each headed by a general manager. The department of interest here is the Workforce and Local Enterprise Department, which is further subdivided into two divisions. The division selected is the Local Enterprise Division, which is further subdivided into several sectors. The network analysis was initially conducted on three sectors - manufacturing, business services, and upgrading services, but due to the poor response rate, only the responses from the business services sector would be analysed.
There are a total of 15 staff in the business services sector - one deputy director and six senior officers, including an attached administrative department headed by two persons. The unique characteristic of the administrative department is that they act as "support staff" to the sector, but it does not belong to the sector exclusively. Any organisational member from other sector who require additional help can approach these administrative staff.
The network analysis form was delivered personally to the offices and the respondents were asked to fill in the forms starting from the next day of work. The completed forms were then collected personally at the end of the day. The duration of the data collection was one working day.
Out of 40 forms delivered to the three sectors, five responded from the business services sector, three from the manufacturing sector, and two from the upgrading services sector. Hence, the responses from the business services sector were selected for analysis due to the highest number of responses collected.
Although the analysis was carried out on only five network analysis forms, the number of communication contacts was overwhelming and the master chart looked like a spider's web (please refer to master chart on page 18). The macro approach was adopted from Hamilton (1987) to produce the master chart.
PART OF THE FORMAL ORGANISATIONAL CHART OF PSB

CEOLEE SUAN HIANG

INTERNAL AUDITDEPARTMENTDEPUTY DIRECTORJANET YAP

WORKFORCE & LOCAL ENTERPRISEGENERAL MANAGERHENRY HENG

LOCAL ENTERPRISE DIVISIONDIVISIONAL DIRECTOR LOH KOK CHOYDEPUTY DIVISIONAL DIRECTOR (1) DANNY LAMDEPUTY DIVISIONAL DIRECTOR (2) TAN PHENG HUAT

BUSINESS SERVICESSECTORDEPUTY DIRECTORGILBERT LOO

ADMINISTRATIVE DEPARTMENT2 HEADS OF DEPARTMENT1 SENIOR OFFICER4 ADMINISTRATIVE EXECUTIVES

6 SENIOR OFFICERS

MASTER CHART OF COMMUNICATION LINKS

EXPLANANTION OF MASTER CHART

The red circle nodes refer to the five respondents who completed the network analysis form; the rest of the nodes refer to those who did not complete the form.
Circle 1 : Chua Guat Leng - Senior Officer (SO) from Business Services Sector (BSS)
Circle 2 : Joanna Cheong - SO from BSS
Circle 3 : Jason Ng - SO from BSS
Circle 4 : Helen Pious - Administrative Executive (AE) from the administrative department attached to BSS
Circle 5 : Yeo Huang Muay - AE from the administrative department attached to BSS

The green triangular nodes refer to other organisational members from BSS or other sectors, e.g. manufacturing, upgrading services, etc.
Triangle 1 : Roy Foo - SO
Triangle 2 : Yeo Kok Wee - SO
Triangle 3 : Tan-Chan L.H. - SO
Triangle 4 : Neo Keng Chuan - SO
Triangle 5 : Lam Chung Soon - SO
Triangle 6 : Lee Kok Seong - Deputy Director
Triangle 7 : Teng Chih Cheong - SO
Triangle 8 : Marie Wong - AE (from another sector)
Triangle 9 : Yvonne Ng - AE (from another sector)
Triangle 10 : Sim Gek Hung - SO
Triangle 11 : Wong Nyuk Min - SO
Triangle 12 : Christina Tan - SO
Triangle 13 : Derrick Ong - SO
Triangle 14 : Tan Geok Lan - SO
Triangle 15 : Mike Sim - SO
Triangle 16 : Tan Lay Ling - PSB staff (no indication of designation)
Triangle 17 : Gilbert Loo - Deputy Director of BSS
Triangle 18 : Janet Ng - AE (from another sector)
Triangle 19 : Weng Hung Min - SO
Triangle 20 : Wong Kok Seng - Director of Land/Labour Department
Triangle 21 : Steven Lam - SO
Triangle 22 : Walter Teo - SO
Triangle 23 : Jolyn Tan - SO
Triangle 24 : Sherman Loo - SO
Triangle 25 : Mr. Seow - Administrator (no indication of specific designation)
Triangle 26 : Zahri - Manager (no indication of specific designation)
Triangle 27 : Chan Yoke Kwan - SO
Triangle 28 : Anthony Ong - SO
Triangle 29 : Wong Wai Nam - Director of Upgrading Services Sector
Triangle 30 : Anne Cheong - AE (from another sector)
Triangle 31 : Chai Li Chin - AE (from another sector)
Triangle 32 : Yeo Cheng Gek - personal assistant (no indication of specific designation)

The blue rectangular nodes refer to individuals who are not part of the organisation at all, e.g. visitors, suppliers, etc.
Rectangle 1 : Chan Kah Guan - Head of Economic and Development Board
Rectangle 2 : Daniel Tan - Lucky Industrial Trading
Rectangle 3 : Merlin Serano Sardoma - Yeo Huang Muay's maid
Rectangle 4 : Chi Pang Chee - consultant
Rectangle 5 : Tan Keng Suan - consultant
Rectangle 6 : Shime - Chua Guat Leng's Japanese friend
Rectangle 7 : May Lee - bank officer
Rectangle 8 : Bernard Lim - bank officer
Rectangle 9 : "Ah Tee" - "printer man"
Rectangle 10 : Leong Mun Poh - visitor

The purple hexagonal nodes refer to the other organisational members from the administrative department attached to BSS.
Hexagon 1 : Chew Mok Lee - Head
Hexagon 2 : Mohinder Kaur - Head
Hexagon 3 : Pushparni - AE
Hexagon 4 : Cindy Tay - AE
Hexagon 5 : Michael Yip - AE
Hexagon 6 : Shirley Wee - SO


SUMMARY OF ANALYSIS
Chua Guat Leng (SO) Joanne Cheong (SO) Jason Ng(SO) Helen Pious(AE) Yeo Huang Muay (AE)
Total no. of links 30 18 18 18 31
Medium used most often Face to face Face to face Face to face Face to face Face to face
Medium used least often Telephone Written communication Telephone Telephone Written communication
Most no. of links with whom Christina Tan(SO) Yeo Huang Muay & Chua Guat Leng Chua Guat Leng Shirley Wee(SO) Chua Guat Leng
Least no. of links with whom None* Helen Pious Helen Pious Jason Ng & Joanne Cheong None*
Communicated most with whom about work Christina Tan (SO) Chua Guat Leng Chua Guat Leng Shirley Wee Jason Ng, Michael Yip & Sim Gek Hung (SO)
Communicated least with whom about work Yeo Huang Muay & Helen Pious+ Helen Pious & Yeo Huang Muay & Jason Ng# Joanne Cheong & Helen Pious# Joanne Cheong & Jason Ng Joanne Cheong#
Communicated most with whom socially Yeo Huang Muay Yeo Huang Muay Chua Guat Leng & Joanne Cheong Pushparni (AE) & Yeo Cheng Gek (PA) Chua Guat Leng & Joanne Cheong
Communicated least with whom socially Joanne Cheong & Helen Pious# Chua Guat Leng & Helen Pious Helen Pious & Yeo Huang Muay All four Jason Ng & Helen Pious
*In this category, there are too many individuals with whom the five members had no links at all (since the lowest possible value of the least number of links is zero), hence, the comparison is limited to only within the five members themselves.
#In these categories, there are again too many individuals with whom the five members had no work/social links at all, hence, the comparison was limited to only the five members themselves. Although the person might have links with the other four, if the links are not work-related/social-related, they will be considered null.
+In this case, since Chua Guat Leng had work links with all four members, those of the lowest number of work links, i.e. one, are selected.

Administrative Executive Yeo Huang Muay is the most frequent communicator among the five members. This could be due to the nature of her work, which allows her many opportunities to interact with others. As the administrative staff are also known as "support-staff", and the administrative executives can offer their help to members from other departments if the latter are short of help. Hence, this increases the administrative executives' chances of meeting other people. However, since the other administrative executive, Helen Pious, has fewer links than Yeo Huang Muay, it could be that the latter is more motivated to offer her help to other departments, or the latter indulges in more interactions than Helen Pious. On the other hand, it could be members from other sectors approach Yeo Huang Muay for help because she is more approachable or readily available than Helen Pious. This is evident from the fact that two of the senior officers, Jason Ng and Joanne Cheong, do not communicated with Helen Pious at all. Moreover, in comparing the number of communication links between Helen Pious and Yeo Huang Muay, the latter appears to be communicating more with members inside and outside the sector.
The medium used most by the members for communication is personal interaction, i.e. speaking face to face. This could be due to the office layout, whereby the organisational members are located in close proximity and it is easier to speak to the person by moving over to the person's table than by using the telephone or e-mailing. It is interesting to note that the telephone and written communication are used the least in communication. Written communication, which includes e-mailing, sending memoranda, or faxing, is not a popular choice of communication medium in this department. Even when communicating with members from other departments, the members prefer to walk over to the other departments than using the telephone or e-mailing. Either the departments are located in such close proximity that it is much more convenient to walk over, and the opportunity to meet the person face to face outweighs the convenience of not having to leave one's seat. Perhaps the organisational members prefer face-to-face communication due to the sincerity and intimacy involved, especially when the interaction is for social purposes, e.g. asking someone out for lunch, asking if someone enjoyed his or her holiday, etc.
Three out of the five members indicated that the person they communicated most with is within their department, i.e. Chua Guat Leng, whereas the other two communicated most with someone from another department. For the three senior officers, it is also the same three, whom they had the most links with, whom they communicated most about work. It is interesting that Chua Guat Leng communicates with someone outside of the department regarding work issues, but little light can be shed on that as the respondents only indicated the purpose of the communication was work- or social-related. It is even more interesting to note that the other two senior officers, Joanne Cheong and Jason Ng, communicate most with Chua Guat Leng about work. It is likely that Chua Guat Leng is the opinion leader in this case, since she does not communicate with them about work but relate to someone else out of the department. Yeo Huang Muay also communicates most often about work with Chua Guat Leng, but it is probably because the former attends to most of the latter's administrative tasks. However, Helen Pious communicates most about work with someone outside of the department, who is also of higher authority than Helen Pious. It is puzzling that Helen Pious should consult with a senior officer outside of the department about work instead of the two senior officers in the same department as her. Perhaps she is on more intimate terms with Shirley Wee, hence the higher communication linkages.
The people whom most of the five members communicate with the most about social issues are also the very same persons they communicated the least about work. This provides us with more insight about the content of the exchanges between two individuals. For instance, Joanne Cheong shares the highest number of links with Yeo Huang Muay, but those links are actually about social content. Similarly for Yeo Huang Muay, her highest number of links is with Chua Guat Leng, but they are of social content rather than work-related.
It is surprising that the five members communicated much more with members outside of their department than among themselves. The senior officers, for example, seem to interact more with other senior officers than with their counterparts in the same department. This could indicate a problem in communication among the five members. For instance, the three senior officers seem to be concentrating their links with Yeo Huang Muay than with Helen Pious, although they are both administrative executives. Whether a bottleneck exists in Yeo Huang Muay's case, it is difficult to infer, as she participates in a mix of social- and work-related tasks with the members.
There also seems to be few interactions between each senior officer and the director, Gilbert Loo (triangular node 17). There is virtually no communication at all between the director and the administrative executives. Perhaps the flow of communication is downward, whereby the director hands the duties down to the senior officers, who further direct the administrative executives. It is regretful that no insight can be gained from the director's point of view as he did not respond to the network analysis.
Similarly for the administrative executives, the communication links with their heads of administrative department are few, suggesting a downward flow of communication. There seems to be more communication between one level and above or below on the hierarchy, but none two levels above or below, i.e. senior officers communicate more with the directors than administrative executives with directors. Perhaps there exists a formal protocol for communication across levels in the hierarchy.
It may be inferred that there is more communication on the same hierarchical level, i.e. horizontal communication among senior officers. They communicate a great deal among those of the same rank, and relatively less with the higher or lower levels. However, the administrative executives communicate more with the senior officers than with their own counterparts. It is difficult to understand why due to the dearth of information, but the reason why Helen Pious is an isolate among the five members could be due to her race (she is an Indian). However, this is merely a conjecture, as it could be due to other factors unknown in this analysis.

EVALUATION
Although the lower-than-expected response rate was initially disheartening, it might have proved to be a blessing in disguise as the number of networks was overwhelming. Analysis was tedious as the organisational members did not restrict their interactions within their department, but extended them to other members from various departments. There was also a high number of interactions with outsiders, i.e. individuals who were not part of the organisation at all, and that further complicated the analysis.
It was inherently difficult to trace the formal or informal communication networks of the business services sector as the director did not participate in the network analysis; most of the communication links with individuals of higher authority were from different departments; the different flows - downward, upward, or horizontal, were impossible to trace as none of the respondents indicated who initiated the interaction; and the social interactions appear to be haphazard and concentrated among senior officers. Moreover, the sample was too small to draw any definite conclusions. Hence, the usefulness and effectiveness of the formal and informal communication systems could not be evaluated upon.

MODIFICATIONS FOR A NETWORK ANALYSIS ON A LARGER SAMPLE
From this exploratory network analysis, a larger sample would increase the number of links exponentially, thus overwhelming the auditor when it comes to analysing the networks. If the organisation consists of several other departments, the potential of communication linkages will further increase due to inter-departmental communication. Limiting the links within the department is not advisable as it would limit the insight the auditor can draw from each member's interactions.
To prevent the likelihood of a poor response rate, the auditor could provide an incentive for the organisational members to complete the network analysis form, e.g. a token of appreciation.
Besides distributing the forms, the auditor can obtain permission to conduct participant-observation on the organisation. In this way, the auditor can gain more insight in the physical structure of the office layout, the non-verbal communication, and the overall atmosphere of the working organisation.
The auditor should provide a clear and simple layout in the network analysis form and simplify the process of entering the information because having too much detail to enter might exasperate the members, or they might regard it as an invasion of their privacy. The information collected might be less insightful, but could prevent less members from being turned off.
Personal interviews should be conducted to find out more about the members' responses, as well as to clarify information that is unclear to the auditor. This is also a good opportunity for the auditor to find out other communication innuendoes that are difficult to write down and understand more about the grapevine or how the communication network flows in the organisation, e.g. upward, downward, or horizontal.
If the sample appears to be more than the auditor can handle, he or she might have to employ additional help in preparing, distributing, collecting, and analysing the network analysis form. It would be beneficial if the auditor could conduct a pre-test of the network analysis to gain some idea of the potential problems he or she might faced in the actual implementation of the network analysis.

BIBLIOGRAPHY

Buckles, T.A., & Ronchetto, J.R. (1996). Examining an industrial buyer's purchasing linkages: a network model and analysis of organisational buying workflow. Journal of Business & Industrial Marketing, 11(6), 74-92.
In this article, network analysis is used to describe purchasing workflow patterns within a single industrial firm. Both prescribed networks (i.e. hierarchical level departmental membership and product purchasing assignment) and emergent networks (i.e. position on the organisational boundary and centrality links within the firm's buying system) are investigated. Results provide implications for purchasing managers and organisational buying researchers.
Downs, C.W. (1988). Communication audits. United States of America: Scott, Foresman and Company. [HD30.3D751]
The chapter on communication networks provides a detailed summary of the purposes of network analysis and illustrations of a data collection form which includes questions on specific message topics. The reliability of network data can be calculated using a specific formula, and a statistical technique known as smallest space analysis is used for communication network analyses.
Hagen, G., Killinger, D.K., & Streeter, R.B. (1997). An analysis of communication networks among Tampa Bay economic development organisations. Connections, 20(2), 13-22.
[http://www.analytictech.com/connections/v20(2)/tampabay.htm]
This study aims to identify the levels of communication, assess overall levels of inter-agency coordination, identify leading economic development agencies and identify sectors of development or economic activity in economic development organisations in Tampa, Florida. Social network analysis is selected in this study for its strength in assessing communication, relations, cooperation, and inter-organisational environments.
Hamilton, S. (1987). A communication audit handbook: Helping organisations communicate. Great Britain: Pitman Publishing. [HD30.3H221]
The chapter on network analysis provides a simple introduction of what network analysis is about and what are some of the ways to conduct a network analysis. General steps are provided as a guide to conduct a network analysis. One useful illustration is the network analysis form which lists some of the questions used in a network analysis. Tips on how to analyse and interpret the data are provided as well.
Heald, M.R., Koehly, L.M., & Wasserman, S. (1998). Formal and emergent predictors of coworkers'perceptual congruence on an organisation's social structure. Human Communication Research, 24(4). Thousand Oaks: Sage Publications. [pP91.3H918]
This study investigates several factors that are hypothesized to influence the perceptual congruence among organisational members. It proposes that members' perceptual congruence on the organisation's social structure is influenced by similarities in formal structure, demographic attributes, and emergent structure. Analysis on social network data is conducted to validate the hypotheses.
Hellweg, S. A. (1997). Formal and informal communication networks. In P. Y. Byers, Organisational communication: Theory and behaviour. United States of America: Allyn & Bacon. [HD30.3R63]
This chapter elaborates on the formal and informal network systems in an organisation, as well as network structures such as the chain, Y-network, wheel, and circle, network roles, network descriptors such as dominance, centrality, flexibility, reachability, network strength, reciprocity, symmetry, and openness. The grapevine is taken to be synonymous with an informal network and tips to control the grapevine are presented.
Johnson, J.D. (1993). Organisational communication structure. United States of America: Ablex Publishing Corporation. [HD30.3J67]
The chapter on network analysis elaborates on the relationships or links between two nodes, using Johnson's model of social interaction. The model contains six elements which constitute the domain of social interaction and posits that emotions and relationships which are at a deeper phenomenal level than content or transference determine these surface acts. Some configurations of relationships among entities are presented and certain strengths and weaknesses of the network analysis as a social science technique are highlighted.
Karathanos, P.H. (1994). Communication network analysis and dysfunctional organisational coalition. Management Decision, 32(9), 15-19.
This study reviews theories on coalition formation and what it might mean to a manager in an organisation. It further proposes that a theoretic understanding of coalitions, together with a communication network analysis, would be a useful tool for discerning particular types of organisational coalitions and subsequently dealing with them.
Kreps, G.L. (1990). Organisational communication: Theory and practice (2nd ed.). New York: Longman. [HD30.3.K92]
The chapter on internal organisational communication systems explains what is meant by formal and informal communication systems. It elaborates on the management information systems, and centralised and decentralised communication systems which are supplementary to this topic. This chapter also highlights the role of the management in the grapevine or informal system and explains the communication network structures.
Mendieta, J.G., Schmidt, S., Castro, J., & Ruiz, A. (1997). A dynamic analysis of the Mexican power network. Connections, 20(2), 34-55.
[http://www.analytictech.com/connections/v20(2)/mexican.htm]
This study examines the configuration, function, and role of the Mexican network of power. Its objective is to study the network's topology, function, and centrality values. The dynamic change of the network is addressed, and the concentration of power is claimed to change because the disappearance of actors eliminates connections and the inclusion of actors creates different cliques, redefining centrality within and among cliques.
Stephenson, K., & Lewin, D. (1996). Managing workforce diversity: macro and micro level human resource implications of network analysis. International Journal of Manpower, 17(4/5), 168-196.
This study presents a new method of identifying and analysing informal networks in business enterprises. The analysis is intended to assist public policy-makers in focusing their macro-level anti-discrimination efforts on "real world" organisational behaviour. The network analysis also serves to enhance our understanding of organisational outcomes and how such outcomes can be influenced and changed through conscious changes in the management of organisational networks.
Tichy, N.M., Tushman, M.L., & Fombrun, C. (1992). Social network analysis for organisations. In K.L. Hutchinson, Readings in organisational communication. United States of America: WM.C. Brown Publishers. [HD30.3H976]
This book is a compilation of various studies on organisations and several which are of some relevance here are the studies on organisational grapevines, rumour as communication, and social network analysis for organisations. The social network analysis for organisations is the most relevant to this paper and it provides the basic concepts of social network analysis, as well as the strengths and weaknesses of various data collection and analysis techniques used to study social networks.