POPULATION CENSUS PUBLIC COMMUNICATIONS:
A MARKETING APPROACH

John Paice
Demography
October 1993


CONTENTS

Introduction

Previous Australian Census experience

Strategies to achieve response

Social Marketing
Definition
Exchange
Target groups
Competition
Product
Price
Place (Distribution)
Promotion

Evaluation
Product
Price
Place (Distribution)
Promotion

Conclusion

INTRODUCTION

While in Australia it is now standard practice to apply marketing concepts to the output of official statistics, they have not been applied up to now to data collection. The collection phase of the 1991 Australian Census of Population and Housing is used as a case study to give an insight into the applicability of marketing concepts to an official statistical agency's public communications.

The Australian Census of population and housing is conducted by the Australian Bureau of Statistics (ABS) as part of its mission to assist and encourage informed decision-making, research and discussion within governments and the community.

The objective of the Census is to measure the number and key characteristics of people in Australia on Census night. It is held every five years, the last one being on 6 August 1991.

While the 1991 Census was a very complex and large operation, the basic concept of the collection phase of the Australian Census was simple. A self-enumeration method was used to obtain information where households were asked to complete Census forms with relatively little assistance from Census collectors. Census collectors dropped off Census forms in a one and half week period before Census night and picked up completed Census forms in a similar period after Census night.
The collection phase was a significant communication exercise. Communication was required with the 42,000 temporary staff involved and to the 6.5 million households who needed to be made aware of the Census as well as to understand, complete and return the Census form, a major communication product in its own right. Logistical and communication considerations required most of the products affecting public communication to be in place well before Census night leaving public relations as the major avenue for flexible response. The communication products apart from the form included the information booklet delivered with the Census form, collector instructions and television advertisements.

PREVIOUS AUSTRALIAN CENSUS EXPERIENCE

Of the Australian Censuses conducted by ABS and its predecessor from 1911, the first about which concerns were publicly raised to any significant extent was the 1971 Census. In particular the newly formed Australia Party attacked the requirement for names and addresses (Cole 1976). In response to these concerns the Treasurer ordered the destruction of all Census forms currently held and directed that the Census forms for the forthcoming Census be destroyed as soon as the data, excluding names and addresses, had been transferred to magnetic tape. In Australian Censuses from 1911 to 1971, forms for the two previous Censuses were retained at any time (ALRC 1979).

For the 1976 Census, apart from seeking (unsuccessfully) substantially more funds for public relations, there was no significant effort made to address privacy concerns (Paice 1988). Considerable public criticism of the 1976 Census occurred. While much of the criticism was directed at the nature and number of questions asked, objections were also made about Census procedures. In May 1976, just before the 30 June 1976 Census, the Attorney-General requested the Law Reform Commission to take account of the implications of the Census for individual privacy in the preparation of its report on a wide-ranging reference on privacy (ALRC 1979).

The Commission's 1979 report on Privacy and the Census signalled a significant change for the ABS in the taking of the Census. Many of its recommendations were adopted, including one that there should be an intensive advertising and publicity campaign to explain the Census and the measures taken to protect confidentiality. Perhaps of even more importance was a change by ABS to a more open and public approach in developing each Census. Compared to the 1976 Census, relatively little controversy has occurred with Censuses conducted since then.

STRATEGIES TO ACHIEVE RESPONSE

To achieve a high level of response in a Census, strategies are adopted with an intention of affecting the beliefs and behaviour of society. Paisley (1989) records that USA's Forest Service has developed a paradigm of the 'three Es' to protect forests from public misuse: engineering, enforcement and education.

In the case of the Australian Census, what is engineered and directly affects the public is the Census form and associated documents and procedures. These documents and procedures were subject to careful design that included extensive testing. A series of 8 field tests and 1 dress rehearsal occurred.

Apart from the tests, form design was also assisted by focus group comment and external to ABS expert comment. In procedures, the basic drop off and pick up design incorporated in the collection phase is advantageous. It fulfils one of the conditions specified by Lazarsfeld and Merton (1949) for successful public communications campaigns in that it enables face-to-face communication between collectors and the public. To the extent the public discusses ABS public communications on the Census with collectors, it means that 'they will process (the relevant) information better' (Kotler and Roberto 1989, p.11).

Enforcement is an option used in the Census but only as a last resort. While the public generally assume that the form has to be completed (Elliot and Shanahan 1989) in a legal sense this only occurs after a formal written direction. Householders are initially requested to complete Census forms. For those householders who do not cooperate the power exists to formally direct a person in writing to complete and return a form. This power was invoked in 582 cases in the 1991 Census, which is minute when considering the 6.5 million households involved. There were 157 prosecutions subsequently approved for cases where notices of direction were not complied with.

Education is associated with the process of communicating and one approach, social marketing, is the focus of the remainder of this paper.

SOCIAL MARKETING

Definition

Social marketing involves the design, implementation and control of programs aimed at increasing the acceptability of a social idea or practice in one or more groups of target adopters. (Kotler and Roberto 1989, p.24).

This translation of marketing concepts to social issues misses out on the description of the process itself contained in the American Marketing Association (AMA) definition of marketing the process of planning and executing the conception, pricing, promotion and distribution of ideas, goods and services to create exchanges that satisfy individual and organisational objectives (McColl-Kennedy et al 1992, p.5)

This definition incorporates the so-called 4 Ps of marketing - product (the conception of ideas, goods and services), price, promotion and place (distribution). These represent the four main interrelated factors, the marketing mix, that an organisation can vary to achieve maximum response from a target market/group. Other factors exist which an organisation is not so able to influence including the legal, political, economic and social external environments.

Common to both definitions is a focus on the customer satisfaction - the satisfaction of individual objectives in the AMA definition and acceptability in a target group(s) in Kotler and Roberto's definition.

Exchange

The AMA definition also refers to exchanges. In the case of the Census no direct benefit is offered in return for the provision of statistical data. While the data is transformed into statistical information of value to the community it is not necessarily of value to the individual providing the data.

This suggests that community benefits, indirect as they may be to the individual, do help to motivate statistical respondents to provide accurate data. Research in the USA indicates that individuals do perform an internal benefit-cost analysis before deciding whether to provide personal information, based on such factors as previous experience with invasions of privacy, probable benefits and anticipated costs (Cox et al 1986). While Cox et al observed that social science literature argues that most individuals do not consistently or routinely employ such formalised thought processes, they suggest that the underlying notion of balancing benefits and costs may be relevant in those instances in which the individual's attention becomes focused on one aspect (benefit or cost) of an issue. In the USA's 1990 Census it was found that public understanding and confidence in the use of Census information contribute significantly to participation in the Census (Kulka et al 1991).

There is the implication though that those disillusioned with the community will discount the benefits the community receive and seek to avoid being counted. This hypothesis is consistent with findings from a March 1986 Census test held in Los Angeles where public information programs were not very effective for the hard-to-enumerate population (Moore and McDonald 1987). In USA's 1990 Census a 'general sense of alienation and a lack of connectedness with society do appear to inhibit participation' (Kulka et al 1991, p.53). The problems of Australia's 1976 Census could also partly be attributed to such disillusionment. The 1976 Census was held 8 months after the Government had been dismissed in controversial circumstances.

Research before the 1991 Census using focus groups indicated that there was a concern that the benefits of the Census for the community did not exist. In a November 1990 report ABS was advised that although the community overall was favourably predisposed towards the Census, perceived shortcomings existed.

- As has been noted in the past, there are some doubts within the community as to the effective use of Census statistics. Cynicism at what is seen as the past failure of politicians and governments to effectively use Census data was on this occasion compared to the failure of some commercial or business enterprises not to accurately plan ahead.

- Many were critical of what they saw as the absence of examples of how past Census data has been put to effective use. Tangible results of the successful use of Census statistics was frequently requested.

- There was some evidence to suggest that community confidence is at a particularly low point at present and the general feeling of 'doom and gloom' which permeated throughout the study has led some to consider 'why bother' filling out the Census (form). This suggests a need to constantly reassure and reaffirm the community of the benefits of the Census.

- There were some concerns about the aspect of privacy but they were not as strong as has been observed previously. (Elliot and Shanahan Research 1990b, pp 8,9).

It was because of this and prior research with similar findings that the ABS's public communication program for the 1991 Census had as its main focus the benefits of statistics from the Census in helping to plan services provided to the community by government, business and private non-profit organisations. While the link between statistics and community services is obvious to people involved in the process, it is generally not apparent or appears to be tenuous to those not involved.

Such an approach is consistent with the classical input/output model of communication that assumes persuasion involves communicating new information from an outside source. It is also consistent with the self-persuasion theory that holds persuasion involves activating information already held by a person. This is achieved by increasing the salience of supportive information that the target person already accepts (McGuire 1989).

Consistent with the concept of individuals balancing benefits and costs, there was a concerted attempt to focus the media's attention on positive, from an ABS viewpoint, aspects of the Census. Also as a result of this reasoning, the information booklet delivered to householders primarily concentrated on the benefits of Census statistics and gave no prominence to the section on confidentiality entitled in the reassuring manner 'How we protect your privacy'.

Statistics Canada has the view that 'more effort needs to be applied in personalising the Census to individual needs and interests rather than stressing the completion of the questionnaire as a civic duty' (Statistics Canada 1992, p.5). The theme for Canada's 1991 Census public communications was 'What's in it for me'. The ABS has placed the emphasis on community rather than individual benefits as individuals can often be disadvantaged by decisions taken in the community interest. The ABS approach is consistent with findings from a June 1990 telephone survey of Census participation in the USA that found the two most common reasons reported for returning Census forms were 'patriotic duty or civic responsibility' and 'importance to community' (Dingbaum and Thomas 1991).

Never-the-less it is reasonable to assume that a direct benefit is likely to be more a motivating force that an indirect benefit and ABS should be alert to providing direct benefits where it can generally apply. One such opportunity will arise with the release of community profile products from the Census. There is significant public interest generated by stories based on Census data. This interest in part could be catered for by promotion of community profile products to the general community and their availability either through local libraries or direct from ABS.

Target groups

Kotler and Roberto's definition refers to target groups and in marketing theory the concept of target markets exists. The idea is that a heterogeneous group of consumers is divided into homogeneous groups with relatively similar needs.

An approach to identify target markets is a competitive necessity as to otherwise would mean that an organisation would fall for the majority fallacy - producing products to serve most consumers. Such products are not generally able to compete with those marketed to accommodate the needs and wants of particular target markets.

How does this theory relate to the Census? Even though the Census is one of those increasingly rare mass market operations, and it is critical to maintain this focus, the concept of target markets is still applicable. For the 1991 Census collection phase the most important external groups were

- Politicians, both federal and State,
- Aboriginal communities,
- Ethnic communities,
- Privacy groups, and
- Media.

The role of politicians is particularly important in the running of a successful Census even though cynicism exists in the community of the use made by politicians of Census data (Elliott and Shanahan Research 1990b). First, although a Census is required by legislation to be conducted every 5 years, it is the Federal Government that decides the content of and allocates resources for the Census. The content is, unique among statistical collections conducted by ABS, subject to federal parliamentary approval in that regulations specifying the content of the Census can be disallowed.

Second, State Governments are also involved. Under agreements between the Federal and State Governments, the ABS provides statistical services for State Governments. Given the importance of the Census to State Governments as well as the importance of their support for the Census, correspondence between the heads of governments occurred before decisions were taken on the Census.

Third, politicians and their staff also have extensive direct contact with the community and this was particularly relevant for promotional purposes.

The variations that occurred in the marketing mix for each target group is referred to in relevant sections which follow.

Competition

Marketing theory requires attention to be paid to competition. Direct competition in the form of other statistical collections conducted by ABS and others does exist although at the time of the Census only a small proportion of households is affected. What could be more significant is competition for time to enable the Census form to be completed and competition for attention. One study indicates that this was a factor in non-response to USA's 1990 Census (Dingbaum and Thomas 1991) although Kulka et al (1991) have indicated that it was not a major one.

Time competition can be taken into account by minimising the time required to complete the Census form, providing benefits for its completion and early personal follow-up for the completed form that is hard to avoid.

The competition for attention is large. Taking advertising alone, Australians are exposed to over 2000 messages every day (McColl-Kennedy et al 1992, p.317). Attention was achieved in TV advertising by using a creative concept that had not been used before and by 'positioning'.

Positioning helps to cut through the competition for attention. Positioning involves a unique and meaningful benefit being presented with an aim of influencing the perception of a target market (Rothschild 1987).

For the Census the positioning can be seen as

- the Census provides statistical information of great importance for the community, and
- the Census involves everyone.

The second position was reflected in the 1991 Census 'tag line' or promotion unifying message of 'we can't do it without you'. The first position was reflected in the 1986 Census tag line of 'helping to make a better Australia'.

Product

Because the Census aims to obtain comparable data for everyone at a specified time, and that compatibility and efficiency are enhanced by limiting form design options, forms are not designed for every target group. Exceptions exist in the cases of non-private dwellings such as hotels and hospitals and remote Aboriginal communities. Information asked of private dwellings such as rent and mortgage payments and number of bedrooms are not asked for non-private dwellings. In the case of Aboriginal communities in remote and some other areas where problems of literacy mean that the self-enumeration method is inappropriate, an interview form is used to obtain data.

For three target groups supplementary support was provided beyond that available for all Australians. In the case of Aboriginal communities, Aboriginal people were employed where appropriate to facilitate communication. For ethnic communities a telephone interpreter service was established, multilingual collectors and interpreters employed and walk-in inquiry centres established. For disabled people with hearing or speech impairments a telephone typewriter inquiry service was established.

Price

Price is what has to be given up to acquire a product. From the respondent viewpoint this can be viewed in terms of intrusion into privacy, second as the time taken to fill out a Census form and third as the financial cost to the community.

Attitudes to privacy change over time (Bulmer 1979) and judgements, based on testing, are made so that only non-controversial topics are included in the Census. Self-enumerated Censuses are subject to much greater restriction on the nature of questions included than where personal interviewing is involved (Paice 1988).

On the second aspect, the time spent by a respondent in completing a Census form in the 1991 Census averaged 22 minutes (FSA 1991) and there was no significant concern expressed that this was excessive.

Third, to conduct, process and disseminate the results of the 1991 Census required ABS to undertake additional expenditure of approximately $113 million, or around $6.50 per person (ABS 1991a). (One householder offered to pay $6.50 for taking part in the Census!)

Australia often compares its Census operations with Canada, New Zealand, UK and USA because of cultural, language and historical commonalities. It appears that of these countries, the only one to run a cheaper Census than Australia per head of population is UK. The UK does not have the travel costs that Australia has and, in addition, uses sample processing which was not used in Australia's 1991 Census (ABS 1990).

Place (Distribution)

The personal contact by collectors in Australia means that communication and response can be tailored to each circumstance within the guide-lines established. This includes the provision of assistance from the collector. It is also more difficult for respondents to delay response than where forms are mailed back as happens in Canada and USA.
Provision exists for the mailing back of Census forms on an exception basis in Australia.

Promotion

Promotion is communication and can include public relations, advertising, personal selling and sales promotion activities. It has been described as

- actively reaching out to the right people with the right message at the right time in order to obtain the right effect (Solomon 1989, p.93).

There were three main phases to promotions in the 1991 Census. The first phase set the scene, the second to help recruit collectors and their immediate supervisors as well as to provide some awareness that 1991 was a Census year and the third to achieve community awareness and support.

The scene setting phase sought to ensure that key public concerns were taken into account in Census planning. This phase saw the release of the following key public reference documents which supported promotion activity throughout the Census

- 1991 Census of Population and Housing: Preliminary ABS Views on Content and Procedures (2101.0) issued in January 1988 for comment by users of Census data and the general public.

- 1991 Census of Population and Housing: Content and Procedures (2613.0) which was tabled in the Parliament in May 1989 when the Government announced its decisions on the 1991 Census. The document was also generally circulated. The proposed question design was included for each topic included in the Census. For the first time detailed reference was also made to topics proposed by users of Census data for inclusion in the Census but rejected.

- Aboriginal Enumeration Strategy circulated to key organisations involved in Aboriginal issues, initially in draft form for comment.

- Ethnic Enumeration Strategy circulated to key organisations involved in ethnic issues, again initially in draft form for comment.

- How Australia Takes a Census (2903.0) issued in April 1991 that contains a copy of the Census form. For the first time this document was published before the Census. In the 1976 Census the form was not made public until just before the Census. In the 1981 and 1986 Censuses forms were released well before the Census but not in a published form.

The second phase included public relations, newspaper advertisements and radio advertisements in problem areas.

The third phase included public relations, newspaper and television advertising. These activities were apart from the personal selling of the Census by collectors.

Collectors have an important role in eliciting response. A common description of the behavioural or psychological process that occurs before a person acts is given by the acronym AIDA standing for awareness, interest, desire and action (McColl-Kennedy et al 1992). There is a common assumption in communications literature that mass media may be better in achieving awareness of a campaign's message and knowledge, while interpersonal communication may be more effective in motivating actual behaviour change (Solomon 1989). While collectors can contribute to all four stages of AIDA their prime contribution can be seen in the desire and action stages where their very existence, besides any thing they orally communicate, can be a factor. While Australia has this advantage, in the USA it has been recognised that it would be useful to search for ways of transforming knowledge and attitudes held by people into action by looking at people to people communication (Dillman 1991).

Promotion was tailored to target groups. Federal and State politicians were sent the May 1989 contents and procedure document. Then just before the Census each politician's electorate office received an information kit that included an inquiry guide as well as the document How Australia Takes a Census.

For Aboriginal communities, 20 field officers were employed for periods averaging 4-5 months with a key focus being communication. This was in recognition that communication through the normal channels is ineffective in Aboriginal communities where credence is often only given to word of mouth contact. It was also in recognition that research for the 1986 Census had uncovered resentment in the general community against Aboriginal (and ethnic) people and wider canvassing of messages for Aboriginal people could be counter-productive. A special video was produced along with posters and leaflets for Aboriginal people.

For ethnic communities the promotion strategy included the production of the main television advertisement in 16 languages, leaflets in 20 languages as well as a separate ethnic media launch.

Privacy groups apart from the Human Rights and Equal Opportunities Commission were approached with the key materials but only a limited response occurred. Extensive discussions took place with the Commission to ensure that the Census was conducted consistent with legislation administered by it.

With the media a significant effort occurred which included the production of 65 media releases as well as the placement of media liaison staff in the major cities at the peak of collection operations.

EVALUATION

Product

The 1991 Census was a success in that it has delivered a high quality product. The initial estimate of the undercount of the population indicates that it was slightly less than the 1981 and 1986 undercounts of 1.9% that were well below the 1976 undercount of 2.6%.

85% of householders found 'the Census form was straightforward and easy to complete' with 7% disagreeing and 8% not taking either view. 76% of householders were satisfied with the service provided by the Census collector (33% very satisfied) with 7% not satisfied. Most of the dissatisfaction arose out of system induced rather than necessarily personal problems - collectors not keeping appointments and not providing privacy envelopes at drop off (FSA 1991).

The overall picture emerges that while there remains a need to seek improvement, there is a high standard of product being delivered.

Price

The general acceptance of the form and the nature and extent of complaints received by ABS indicates the price paid by respondents was generally accepted as reasonable.

Place (Distribution)

The effectiveness and efficiency of using a drop-off and pick-up method is apparent in an international context. The effectiveness can partly be attributed to the action facilitating effect of personal pick-up. It also avoids the problem USA has with mailed out Census forms getting mixed up with other unsolicited mail, a factor that exists even though it does not appear to explain the decline in mail response rates between the 1980 and 1990 Censuses (Fay et al 1991).

Only 3% of households mailed back their Census form in Australia which for the most part reflect the people involved having more confidence in the mail system than in their collector. Given that this 3% was a significant increase on the 1986 experience, arrangements need to be made to facilitate return of a larger proportion of forms by mail. Separately promotion activity needs to occur aimed at boosting the credibility of collectors and hence the existing system.

Promotion

Bipartisan support was achieved from Federal and State politicians. Aboriginal communities were generally supportive. A problem arose in ethnic communities with a concern that non-Christian religions were not being recognised but generally ethnic enumeration ran very smoothly. One privacy group, the Victorian Council for Civil Liberties, went public on Census Day with a statement based on misconceptions. This may reflect a need for ABS to be more assertive in seeking to discuss the Census with such groups at an early stage. Generally the media gave favourable coverage to the Census with 1,400 articles appearing in print. While they were generally supportive, one editorial was based on wrong information. To prevent misunderstandings in editorials, editorial staffs need to be directly targeted.

The scene setting phase was a success in that all the major issues raised while the public was involved in completing Census forms had been anticipated.

Recruitment was the second major promotion phase. In the recruitment of collectors and their immediate supervisors, problems existed in some areas in getting the message of job availability through to the public even though ABS received 75,000 applications in 1991 compared to 60,000 in 1986. Experience indicates that if the message of job availability was able to get through to the target groups involved, in most areas there are people interested in the home based, part time jobs. This job design is much easier to promote than the short term full time jobs that USA has. The main target groups were the increasing number of early retirees and the declining number of full time home carers.

In the next Census it is planned to advertise for collectors after supervisors have been recruited. This will enable early involvement of supervisors in word of mouth promotion activity where necessary. Greater use of local papers for advertising as well as public relations activities using local radio are also planned.

Reflecting the third major promotion phase, community awareness and support, 11% of collectors reported that more than 25 householders out of an average of about 200 households were not aware of the Census when contact was made. In August 1991 after the Census 7% of householders didn't know what the Census was used for, 52% believed that the Census was completely confidential and 35% were aware that ABS was responsible for the conduct of the Census (FSA 1991).

Overall a picture emerges of a successful community awareness and support phase. It would be surprising if the 7% of householders who didn't know what the Census was used for could be significantly reduced or for a great increase in the number who accept that the Census is confidential. The 52% of householders who believed that the Census was completely confidential was equivalent to the situation 10 years earlier after the 1981 Census (Paice 1988). While awareness that ABS was responsible for conducting the Census was only a secondary objective of the communication campaign, it did achieve a recognition rate of 35%, well above that for unemployment (24%), consumer price index (14%) and balance of payments (10%) statistics (FSA 1991).

While the number of households who were unaware of the Census when the collector called was relatively small, this could probably be made smaller by launching TV advertising prior to collectors commencing their work. Because funds do not permit TV advertising for an extended period, public relations, the information booklet (51% of householders claimed they read or browsed through it - FSA 1991) and collectors would then provide the carry through promotion activity. Such a plan would be facilitated by a TV advertisement with the major impact of the 1991 Census advertisement.

80% of householders in October 1991, 2 months after the Census, claimed they could remember the TV advertisement with 68% claiming it a source of information (FSA 1991). It could well be used again in the 1996 Census. As noted in section 5.3, the credibility of collectors needs to be boosted .

CONCLUSION

The application of marketing concepts to public communications associated with data collection can at times appear to be forced as in, for instance, the suggestion that competition can be interpreted as being competition for time. On the other hand, it facilitates a focus on respondents and a wide view of communications which helps place into perspective promotional issues which are the main focus of public communication specialists. As such a marketing approach provides a useful insight into public communication issues associated with data collection and forms an, as yet, under used tool for public statisticians.

REFERENCES

ABS (1988). Provision of Information to the Australian Bureau of Statistics. Internal paper, 26 May 1988.

ABS (1990). International Cost Comparisons of Population Censuses. Internal ABS Census working paper 90/2, March 1990.

ABS (1991a). How Australia Takes a Census (2903.0).

ABS (1991b). Communication Strategy for 1991 Census. Internal paper, July 1991.

ABS (1991c). Review of Communication Aspects. Internal paper for 1991 Census Collection Phase Evaluation Conference.

ABS (1991d). Census of Population and Housing: Collection Phase - Strategic Directions for the Next Census. Internal paper for 1991 Census Collection Phase Evaluation Conference.

ABS (1992). The Value and Acceptance of the Australian Census of Population and Housing. Conference of European Statisticians op cit, paper GE.92-30833.

ALRC (1979). The Law Reform Commission. Privacy and the Census. Australian Government Publishing Service.

Bates, N and Whitford, D (1991). Reaching Everyone: Encouraging Participation in the 1990 Census. Paper presented at

American Statistical Association's August 1991 Annual Meeting, Social Statistics Section.

Bulmer, M ed (1979). Censuses, Surveys and Privacy. The Macmillian Press.

Bureau of the Census (1991). Annual Research Conference March 1991: Proceedings.

Cole, R (1976). The Census of Population and Housing, 30 June 1976. Australian Journal of Statistics, 18(3).

Conference of European Statisticians (1992). Papers prepared for 4-6 May 1992 work session on population and housing

Censuses, Conference of European Statisticians, Statistical Commission and Economic Commission for Europe.

Cox, L, McDonald, S-K and Nelson, D (1986). Confidentiality Issues at the United States Bureau of the Census. Journal of
Official Statistics, Vol 2, No 2.

Dillman, D (1991). Discussion in Bureau of the Census op cit, 55.

Dingbaum, T and Thomas, K (1991). Telephone Survey of Census Participation. Paper presented at American Statistical Association's August 1991 Annual Meeting, Social Statistics Section.

Elliott and Shanahan Research (1989). An Exploratory Study, Community Attitudes, The 1991 Census. Report prepared for ABS, November 1989.

Elliott and Shanahan Research (1990). An Exploratory Study, Reaction to Census 91 Brochure. Report prepared for ABS, April 1990.

Elliott and Shanahan Research (1990). An Exploratory Study, Further Communication Research, Census 91. Report prepared for ABS, November 1990.

Fay, R, Bates, N and Moore, J (1991). Lower Mail Response in the 1990 Census: A Preliminary Interpretation. Bureau of the Census op cit.

Fein, D and West, K (1988). The Sources of Census Undercount: Findings from the 1986 Los Angles Test Census. Survey Methodology, Vol 14, No 2.

FSA (1991). Frank Small and Associates. Towards an Understanding of Designing Future Population Census Forms. Report prepared for ABS, November 1991.

Kotler, P and Roberto, E (1989). Social Marketing: Strategies for Changing Public Behaviour. The Free Press.

Kulka, R, Holt, N, Carter, W and Dowd, K (1991). Self-Reports of Time Pressures, Concerns for Privacy and Participation in the 1990 Mail Census. Bureau of the Census op cit.

Lazarsfeld, P and Merton, R (1949) as referred to in Kotler and Roberto op cit.

McColl-Kennedy, J, Kiel, G, Lusch, R and Lusch, V (1992). Marketing Concepts and Strategies. Thomas Nelson Australia.

McGuire, W (1989). Theoretical Foundations of Campaigns. In Rice and Atkin op cit.

Moore, J and McDonald, S (1987). The Census community awareness program: an evaluation of the potential and actual effectiveness of CCAP based on evidence from the 1986 Los Angeles Census Test. Unpublished Bureau of the Census report, referred to in Fein and West, op cit.

Netherlands Central Bureau of Statistics (1992). An Alternative for the Census? The Case of The Netherlands. Conference of European Statisticians op cit, paper GE.92-30662.

Office of Population Censuses and Surveys (1991). Report on Publicity for the 1991 Census: Management Summary. Internal paper, December 1991.

Paice, J (1988). Australian Census of Population and Housing: Confidentiality Issues. Paper presented at the International
Symposium on Low Cost Information Systems - State of the Art and Future Implications: September 1988. Internal ABS Census working paper 88/1.

Paisley, W (1989). Public Communication Campaigns: The American Experience. In Rice and Atkin op cit.

Rice, R and Atkin, C eds (1989). Public Communication Campaigns, second edition. Sage Publications.

Rothschild, M (1987). Advertising, From Fundamentals to Strategies. D C Heath and Company.

Solomon, D (1989). A Social Marketing Perspective on Communication Campaigns. In Rice and Atkin op cit.

Statistics Canada (1992). Respondent Relations: What the Future Holds for Census-Taking. Conference of European Statisticians op cit, paper GE.92-30657.