Legislative Demand And Risk Perception

Legislative Demand and Risk Perception

Examining a link between risk perception and demand for government intervention in the marketing of consumer products with perceived health risks.

Stephen Dann Bachelor of Arts

(Government and Law)

A dissertation submitted in fulfilment of the requirements for the degree of

Bachelor of Commerce with Honours Faculty of Commerce and Administration

GRIFFITH UNIVERSITY

1 November 1995

Synopsis

This research examines the relationship between perceived risk and legislative demand for a series of common consumer products. Legislative demand, the consumer's desire for government regulation of products or events with perceived risks, has been established from the combined research of the public policy, risk perception, risk management, marketing and consumer behaviour fields. Review of theses literatures revealed four main areas of research: the relationship between risk and desire for regulation; the prevalence of advertising regulation; and the relationship between age and gender as factors influencing demand for legislation.

To examine these research areas, a questionnaire survey was pilot tested on a sample of University of Queensland student. The main survey was conducted using Griffith University undergraduate and post graduate students.

Major findings of the thesis have been the existence of a relationship between perceived risk and legislative demand for two of the nine products tested. The research also uncovered an unique finding concerning gender and demand for legislation of tobacco products. Further, two conceptual framework have been established: the risk processing chain, and the risk response model. The risk processing chain examines the progression from a risk event to the consumer's eventual reaction, either acceptance of the risk or legislative demand. The risk response model specifically examines the relationship between levels of perceived risk, personal risk management and legislative demand.

Finally, the research examines the implications for public policy, risk management, and marketing arising from the research findings.

TABLE OF CONTENTS

Ttitle Page Synopsis

...

Table of contents

...

List of figures

...

List of tables

...

Declaration

...

Acknowledgements

...

Formatting

...

Chapter 1: Outline

1

1.1. Introduction

1

1.2. Research problem

1

i

Justification of the Research problem

1

ii

Hypotheses

2

1.3. Methodology

3

1.4. Chapter outlines

3

1.5. Definitions

5

1.6. Limitations and key assumptions

6

1.7. Conclusions

7

Chapter 2: Literature review

8

2.1. Introduction.

8

2.2. Risk Perception.

8

i

Risk, Actual Risk and Perceived Risk (Behavioural decision making theory)

8

ii

Consumer perceived risk (Consumer Behaviour).

9

iii

Types of Risk.

10

iv

v

Risk Perception: Heuristics, Biases and Information Processing.

10

Risk Response

12

2.3. Government Intervention.

14

i

Legislative Demand.

14

ii

Slovic Fischoff and Lichenstein's (1980) dread risk.

14

iii

Kasperson's (1992) Model of the Social Amplification of Risk.

15

iv

The Garibaldi Meats Case Study.

16

v

Kasperson's Legislative Demand

17

H1:

The consumer's perception of a product's risk will influence their desire for government intervention to reduce the level of risk associated with that product .

H1A The higher the perceived risk, the greater the legislative demand. v H2:

Types of Legislative Controls.

17 18 18

There will be a demand for greater legislative control of the advertising and other promotional activities of perceived

2.4. H3:

H4: 2.5.

high risk products.

19

Consumer Behaviour..

19

There will be no significant differences in legislative demand by gender.

20

There will be significant differences in legislative demand by age.

20

Conclusion

20

Chapter 3: Methodology.

21

3.1.

Introduction

21

3.2.

Quantitative v Qualitative.

21

i

Qualitative

21

ii

Quantitative

22

3.3.

Methodology

22

3.4.

Sample

23

i

Sampling Technique

23

ii

Determining sample size

25

3.5.

Process

26

3.6.

Questionnaire Details.

27

i

Product Selection.

27

ii

Section 1: Product Use Questions.

27

iii

Section 2: Product Risk Perception.

28

a) Risk Perception: General Risk of Product Use.

28

b) Specific Risk Perception: Perception of the Heart Disease Risk. 29 iv

Section 3: Product Regulation - General.

29

a) Role of Government.

29

b) Role of Government: Generic Legislative Demand.

30

c) Regulating Price, Product, Promotion and Place

30

d) Negative Response

30

e) Replication Questions.

31

a) Calfee and Ringold (1994): Consumer Beliefs About Advertising .

31

b) Barksdale and Darden (1972): Consumer Attitudes Towards Marketing and Consumerism..

31

v

Section 4: Methods of Control: Generic.

31

vi

Section 5: Specific Controls over certain products.

32

vii

Section 6: Demographics.

32

The Pilot Study.

33

Changes

22

a) Section 5: Levels of government control (Hypotheses 1 & 1A)

33

b) Section 1: Product usage levels

34

c) Section 3: Product regulation - general

34

3.7. i

3.8. Summary and Conclusions

Chapter 4: Research results.

35

36

4.1. Introduction

36

4.2. Demographics

36

4.3. Product usage

36

4.4. Heart risk

37

4.5. General risk

37

4.6. Legislative demand

37

i

Levels of Government Control

37

ii

Government Controls: General

38

iii

General Opinions Concerning Product Regulation

38

A) Heart Disease

39

B) General Health Risk

39

4.7. Hypothesis 1A : Perceived risk and legislative demand

40

4.8. Hypothesis 2: Demand for advertising controls on high risk products

41

4.9. Hypothesis 3: Difference in legislative demand by gender

42

4.10. Hypothesis 4: Differences in legislative demand by age

43

4.11. Replication studies

43

i

Mazis, Morris and Swasy (1991): General Perceived Risk

43

ii

Calfee and Ringold (1994): Consumer Beliefs About Advertising.

44

iii

Barksdale and Darden (1972) and Barksdale, Perreault et al. (1982) Consumer Attitudes Towards Marketing and Consumerism. 44

4.12. Conclusions

Chapter 5: Conclusion and Implications

45

46

5.1. Introduction

46

5.2

46

Conclusions i

Research Problem: Legislative Demand

46

ii

Hypothesis 1: Perceived risk and legislative demand

47

iii Hypothesis 2: Demand for advertising controls on high risk products 48 iv

v

Hypothesis 3: Difference in legislative demand by gender

50

A) Gender Difference in Demand for Tobacco Regulation

50

- Representation of Smokers v Nonsmokers

51

- Cognitive Processing

51

Hypothesis 4: Differences in legislative demand by age

54

5.3. Replication Studies

55

i

Replication study 1: General health risk perceptions

55

ii

Replication study 2: Consumer beliefs about advertising

56

iii

Replication study 3: Consumer attitudes towards government regulation of advertising

56

5.4. Implications for Theory and Practice

58

i

Public Policy

58

ii

Risk Management and Risk Perception

59

iii

Marketing

60

5.5. Limitations

61

5.6. Further research

62

References

63

Appendices.

75

Appendix 2.1.

Expanded risk processing chain

...

Appendix 2.2.

Kasperson's (1992) Model of the social amplification of risk and the potential impacts on a corporation

...

Appendix 3.1.

Calfee and Ringold (1994) Frequencies

...

Appendix 3.2.

Heart Foundation Food Pyramid

...

Appendix 3.3.

Pilot study questionnaire

...

Appendix 3.4.

Questionnaire

... References.

...

LIST OF FIGURES

Figure 2.1.

Risk Processing Chain: From Risk to Risk Response.

12

Figure 2.2.

Risk Response.

13

Figure 2.3.

Social Amplification of Risk and the Potential Impacts on a Corporation.

Figure 2.4.

15

Redefinition of Kasperson's (1992) Model of the Social Amplification of Risk

17

Figure 3.1.

Formula for determining sample size 'n'

25

Figure 3.2.

Determing sample size 'n'

26

Figure 5.1.

Risk Procession Model

52

Figure 5.2.

Male Tobacco Risk Procession

53

Figure 5.3.

Female Tobacco Risk Procession

53

LIST OF TABLES

Table 4.1.

Product Usage.

37

Table 4.2.

Frequencies for Methods of Control.

38

Table 4.3.

General Opinions concerning regulation of products with Heart Disease Risk.

Table 4.4

39

General Opinions concerning regulation of products with known health risks.

Table 4.5.

40

Relationship between perceived risk and legislative demand (restrictions on advertising).

Table 4.6.

41

Relationship between perceived risk and legislative demand (controls on product promotion).

42

Table 4.7.

Difference in demand for tobacco controls by gender.

42

Table 4.8.

Percentage of responses rating the general risk of alcohol, tobacco and dairy products as very harmful

Table 4.9

43

Combined Percentage of responses rating the general risk of alcohol and tobacco as 'very harmful' and 'somewhat harmful'.

Table 4.10.

Table 4.11.

44

Desire for greater government regulation of advertising sales and marketing.

44

Attitudes towards government regulation.

45

Table 5.1.

Distribution of Smokers v Nonsmokers

51

Declaration

The work presented in this dissertation has never previously been submitted for a degree or diploma in any university. To the best of my knowledge, no material previously published or written by another person has been included, excepted where due reference is made in the dissertation itself.

......................... Stephen Dann

Acknowledgements Cogito cogito ergo cogito sum: I think that I think therefore I think that I am

Special thanks to my supervisors, Dr Peter Graham and June Dunleavy, for guiding my research with their invaluable knowledge and skills. I greatly appreciate the time you have devoted to editing, discussion, reading and revising this work. I also wish to thank Professor CP Rao of Old Dominion University, Virginia, USA for his suggestions and comments.

For their continued patience, support, assistance, and hard work, many thanks go to my parents Mike and Jean, my brother Peter and to Dr Susan Dann, who had the dubious honour of being my Services Marketing Lecturer in second semester. I also wish to thank my extended clan of friends and colleagues, who have been subjected to many various ideas, thoughts and hypotheses during the last year months, some of which were related to this thesis. I would also thank Claris Corporation and Apple, for creating the tools with which this thesis was crafted.

Finally, I would like to express gratitude to everyone who filled out my questionnaire and helped turned the theory into reality, the ideas into numbers and concepts into a thesis. Without them, none of this would have been possible.

.

Formatting

This dissertation has been formatted in accordance with the guidelines contained within the Publication Manual of the American Psychological Association (Fourth Edition).

CHAPTER 1: OUTLINE

This chapter establishes the overall framework for the thesis, stating and justifying the research problem and overviewing the hypotheses, definitions and key assumptions of the research. Further, it outlines the chapter structures, methodological framework and the limitations of the research.

Research Problem

Is there a relationship between a consumer's risk perceptions and their demand for government intervention in the marketing of consumer products with perceived health risks?

Justification of Research Problem

Legislative demand has become a significant aspect of marketing practice since the rise of consumer activism, the increasing awareness of consumer rights, and the emergence of consumer lobby groups. This research specifically reviews the relationship between a perceived risk and the consumer's demand for government intervention into the marketing of 'risky' products.

The concept of legislative demand has been drawn from the combined fields of consumer behaviour, risk perception, and public policy. At present, there have been no studies specifically examining the concept. However parallels have been found in Slovic, Fischoff and Lichtenstein's (1980) 'dread risk' and Kasperson's (1992) Social Amplification of Risk. 'Dread risk' is defined by Slovic et al. (1980) as the type of risk the individual no longer thinks rationally about, instead preferring to rely on 'gut' feelings or instincts. Slovic et al. (1980) assume a 'legislative demand' in their 'dread risk' concept, stating that a risk with a high 'dread

risk' score will lead to demands for government intervention to control that 'dreaded' risk. Kasperson (1992) implies legislative demand in the Social Amplification of Risk model (See Chapter 2) by having 'government regulation' as an outcome of the model. Neither author expressly state or test the concept. Further no literature has specifically examined the relationship between risk perception of a product and demand for regulation of that product.

The major practical implication of this research is a greater understanding of the processes by which legislative controls are implemented in the marketing of legal products. Specifically, the research examines the degree to which consumer perceptions of a product influence the consumer's belief in the role of government controls over that product. Further, the research is an exploratory study into the relationship between risk and legislative demand over a narrow band of food products, and has potential for much wider industry application in the emerging technology based fields such as genetically engineered food products.

In the public policy field, the existence of a quantifiable risk perception-legislative demand concept can assist policy making in product and marketing controls. Products with high perceived risk would be expected to have more government regulations controlling their marketing practices than products with lower risk perceptions. In Australia, the tobacco industry, which has a high perceived risk product, has legislation which affects its distribution, pricing and promotional strategies. The alcohol industry, whilst having antialcohol lobby groups comparable to those of the anti-tobacco movement, does not have similar levels of government control. Part of the reason for the discrepancy in the levels of control may arise from alcohol having a lower perceived risk than tobacco (Mazis, Morris and Swasy, 1991). In addition, the legislative demand concept, coupled with Slovic et al.'s (1980) 'dread risk' would enable Government to identify areas of high risk related activities, such as nuclear power, that will have expectations of legislative controls, and preempt social activism.

Hypotheses

The following hypotheses have been developed from the literature outlined above, and expanded upon in Chapter 2. H1: The consumer's perception of a product's risk will influence their desire for government intervention to reduce the level of risk associated with that product. H1A: The higher the perceived risk, the greater the legislative demand.

Hypothesis 1 is derived from Slovic et al.'s (1980) 'dread risk' theory, Kasperson's (1992) model and Patterson's (1992) assertion of societal demand for government based risk control. Hypothesis 1A is based on Slovic et al.'s (1980) assumption of the relationship between 'dread risk' and legislative demand. H2: There will be a demand for greater legislative control over the advertising, and other promotional activities associated with perceived high risk products.

Hypothesis 2, demand for advertising controls, has been derived from the research of Jeffrey (1989), Beales (1991) and Calfee and Ringold (1990) which indicate that advertising controls are both the most common form of legislative intervention, and the type most consistently demanded by the consumer. H3: There will not be significant differences in legislative demand by gender.

Hypothesis 3, that legislative demand is not affected by gender, has been constructed to test the findings of Barksdale and Darden (1972), Slovic, Kraushappe et al. (1989) and Pidgeon et al. (1992). H4: There will be significant differences in legislative demand by age. Hypothesis 4 tests a statement by Barksdale and Darden (1972) that age is a determining factor in legislative demand

Methodology

The research will use a quantitative survey technique, sampling over 100 university students selected from both undergraduate and postgraduate levels. Although the sample will not be representative of the Australian population, the research is an exploratory work to determine if such a link between risk perception and legislative demand does exist. Consequently, should the research's findings support such a hypothesis further studies, using more representative samples can be conducted.

The questionnaire format and questions have been drawn from the literature using mostly Likert scaling. This is consistent with other research in the risk perception field. Further, many of the questions have been drawn from prior questionnaires, although no specific surveys are replicated in their entirety. The survey consists of five parts examining: usage levels of common food products, specific product risk perceptions, general attitudes towards legislative intervention, attitudes towards types of intervention appropriate for specific products, and the types of interventions appropriate for these products. The demographics collected will be age and gender. Chapter Outlines Chapter 1: Introduction Chapter 1 gives an overview of the research, outlining the research problem, its background, and the justification for the research.

The chapter also introduces the

hypotheses, research design, chapter outlines and the key limitations and assumptions which guide the research.

Chapter 2: Literature Review

Chapter 2 is the literature review which overviews research from the fields of risk perception, risk management, public policy and consumer behaviour. This chapter also

outlines the hypotheses drawn from the literature. Risk perception literature covers the definition of risk, actual risk and perceived risk, as well as the types of risk, and the factors affecting the individual's perception of risk. The risk management and public policy literature examine government intervention in risk reduction. As part of this review, the legislative demand concept, derived from Slovic et al.'s (1980) 'dread risk', and Kasperson's (1992) mode of the Social Amplification of Risk, is defined. The first hypothesis is also drawn from these two models.

Following 'legislative demand', the review examines specific types of

government controls, particularly demand for advertising which establishes the second hypothesis. Finally the chapter examines consumer behaviour in terms of consumer attitudes towards government regulation by demographic characteristics which gives the framework for the final two hypotheses. Chapter 3: Methodology Chapter 3 operationalises the hypotheses of Chapter 2. This chapter defines the sampling technique and sample frame. Further, it establishes the methodological framework for the data collection which will use a purpose built questionnaire. Each section of the questionnaire is examined and justified, including the three sections which replicate prior studies. Further, the pilot study is conducted on the initial survey design. Finally, the chapter highlights changes made to the main survey questionnaire from the results of the pilot study. Chapter 4: Research Results Chapter 4 contains the results of the survey conducted with the survey instrument from Chapter 3. This covers the demographics of the survey, and the frequencies for the major sections of the survey. Further, it describes the statistical relationships between the relevant variables for each of the hypotheses. Finally, the results which replicate parts of studies identified in the literature are compared with the statistics from the reported results.

Chapter 5: Conclusions and Analysis

The final chapter of the thesis contains the conclusions drawn from the previous chapters. It examines the implications arising from the findings of the hypotheses and replication studies for: theory, policy and practice, the private sector, and the public sector. Further, it will examine the limitations of the research, and the avenues of further research that have arisen from this study.

Definitions

The major definition to arise from this research is legislative demand, which has been specifically created for this work. Legislative Demand is defined as the consumer's belief in the level and type of control that the government should operate in any given circumstances. This definition has been drawn from three areas. First, Slovic et al.'s (1980) 'dread risk' concept proposes a relationship between 'dreaded' risks, and the consumer's desire to see regulation used to reduce these risks. Second, the legislative demand concept has been established by implication as a linkage in Kasperson's (1992) model of the social amplification of risk. Finally, Patterson, Hunnicutt and Stutts (1992) propose that when a product is identified as having a potential health risk, there are consumers who want the price, product, promotion, and / or distribution of that product regulated.

In summary, legislative demand is applicable to any situation where the consumer believes that the government should intervene to reduce the risks associated with an event, product or behaviour. Examples range from minimum health standards for food products to regulation of access to pornographic materials on the Internet. Specifically, for this research, the legislative demand definition has been limited to apply to demand for controls over the marketing of products with known health risks.

Limitations and Key Assumptions

There are four key limitations: robustness of the survey instrument, narrow subject focus, sample bias and level of generalisation available from the study. First, as an original study, not a replication, there may be problems with the robustness, reliability and validity of the survey tool. To compensate for this, the instrument has been constructed in part by replication of prior research.

Second, the research will not attempt to analyse the reasons for the consumer's levels of acceptance and expectations of legislative intervention, rather it will concentrate on quantifying the levels of support for legislative controls on the marketing of products with perceived risks.

Further, the research is concerned with the consumer's responses to

legislative controls on the core concepts of marketing: price, product, promotion, and place. It does not involve an analysis of the forces within the marketing environments which have brought about these controls, for example lobby groups and opinion leaders.

The third limitation relates to the use of university students as the sample. It is expected that the university students will not be reflective of general community standards, due to higher levels of education, potential income and social background. From these differences it is expected that the sample will have a greater predisposition towards consumer activism and government intervention (Rao, 1995). This limitation has two potential effects. First, the student sample may more strongly support the hypotheses concerning government intervention in advertising and promotion. Second, this potential bias will limit the generalisability of the research findings.

The overall relationship between perceived risk and legislative demand is not expected to be affected by any biases inherent in the student population. From the risk response diagram, which is drawn from the literature, legislative demand is expected to arise from two factors, immediate unacceptable risk, and risk that has remained unacceptable despite personal risk management. Consequently, the demand for government intervention in this study is

examined separately from other aspects of consumer activism, such as demand for business self regulation. Finally, the major aspect of legislative demand being tested is demand towards restrictions of marketing techniques, such as advertising. Calfee and Ringold's (1994) review of 60 years of American consumer opinion research, combined with Barksdale, Perreault, Ardnt, Barnhill, French, Halliday, and Zif's (1982) cross national survey (including Australia) of consumer attitudes towards marketing has shown a consistent demand for greater government regulation of advertising. As a result, the student bias is not expected to vary too greatly from the community standards towards certain aspects of legislative demand. Further, the research is concerned with the patterns of demand, rather than individual levels

The fourth and final limitation of the research is applicability of the specific results of the study to the wider community. As the study is an exploratory examination of the potential link between risk perception, and legislative demand it is not intended to be generalised to the wider community. Rather, it is expected that should the link be demonstrated, further research beyond the thesis will be conducted to test the findings on a larger, more representative sample.

Summary

This chapter provides an overview of the thesis structure. It has introduced the research problem, research question and hypotheses to be examined. The research problem has been justified and its limitations and assumptions acknowledged and methodology outlined. The following chapter will set out the review of the literature, and establish the hypotheses of the research.

CHAPTER 2: LITERATURE REVIEW

'When society identifies a potential health and/or safety problem with a product, there are those who want to regulate its manufacture, distribution, promotion and sale' (Patterson, Hunnicutt and Stutts, 1992:96).

This chapter will review the literature in the fields of risk, risk perception and risk management along with an examination of the role of government in risk management. At the present time there are two distinct strands of 'risk perception' research, consumer behaviour and behavioural decision theory / risk management. For the purposes of examining legislative demand, the emphasis has been placed upon the risk management field, rather than consumer behaviour. The review also examines Slovic's (1980) 'dread risk' and Kasperson's (1992) Model of the Social Amplification of Risk to test the 'legislative demand' concept.

Risk Perception

Risk, Actual Risk and Perceived Risk (Behavioural Decision Making Theory)

'Risk' means different things to different people (Slovic, Fischoff, Lichtenstein, 1985). As the term 'risk' has multiple meanings within the fields of consumer behaviour, risk perception, and in the 'real' world, no standard definition of 'risk' has been adopted in the literature. The lack of a single definition results from the complex spectrum of ideas encompassed by the term 'risk'.

For the purposes of this thesis 'risk' will be defined as "the objective likelihood of an adverse effect occurring within a given period of time" which has been derived from Kasperson (1992), Jeffrey (1989), and Borcherding et al. (1985). Subsequently, two further terms arise from the literature: 'objective' or 'actual' risk; and 'subjective' or 'perceived' risk.

'Actual' risk is the scientifically assessed probability of a hazard occurring with a specific set of circumstances or events within a given period of time (Borcherding et al. 1985). 'Perceived risk' is essentially an individual's subjective belief as to the likelihood, and consequences, of a hazard occurring to that person (Warner, 1992). Significantly, 'actual' risk and 'perceived' risk are not necessarily related to each other as heuristics, biases and social constructions affect the perception of risk not the actual risk itself.

Consumer Perceived Risk (Consumer Behaviour)

Bauer (1960, in Ogelthorpe & Monroe, 1994) propose that all consumer behaviour involves risk based on the assumptions that all activites include a degree of uncertainty of outcome, and that these uncertain outcomes can have negative consequences. This concept has been further expanded into five major categories:

(1) nature of perceived risk, (2) relevant types of perceived consumer risk, (3) relationship between perceived consumer risk and consumer risk, (4) effect of individual differences on risk perception, and (5) measurement of perceived consumer risk (Havlena and DeSarbo, 1991, in Oglethorpe & Monroe, 1994: 327)

Oglethorpe et al. (1994) examined the consumer perceived risk, expanding the list of outcomes to include: functional or performance, social, psychological, health and safety, and financial. Functional or performance is the risk of failure when adopting a new product. Social risk involves loss of prestige or reputation from use of a product. Psychological risk relates to cognitive dissonance, which may have longer term psychological effects. Health and safety involves risks inherent in the product or the use of the product, that is, faulty goods or health damaging goods such as tobacco. Finally, financial is the risk inherent in the purchase of durable or expensive goods that require relatively large capital investment (Oglethorpe et al. 1994). Of these consumer perceived risks, this research focuses on 'health and safety risk' as part of the legislative demand concept, which is expanded below.

The definitions and concepts of consumer behaviour risk will not be adopted in this review for two reasons. First, the legislative demand concept is reliant on behavioural decisionmaking theory and risk response, particularly Slovic et al.'s (1980) 'dread risk', Kasperson's (1992) Social Amplification Model and Patterson's (1992) societal demand statement. Second, the consumer behaviour perceived risk definitions cover a broader range of perceived risks, beyond those associated with personal health and safety. Consequently, the narrower focus of the behavioural decision theory literature is better suited to the research topic.

Types of Risk

Jeffrey (1989) created a model of three distinct types of 'actual risk': absolute, relative, and population attributable risk. 'Absolute risk' is "simply the risk that an individual will get a certain disease over a defined time period" (Jeffrey, 1989: 1194). This is normally expressed in terms such as 'the risk of death by auto accident in one year is 3 in 10,000.' 'Relative risk' is the 'ratio of chance' of suffering harm in individuals who have been exposed to a risk factor compared to those who have had no exposure to the risk (Jeffrey, 1989). For example, 'relative risk' is commonly applied to the increased risk of lung cancer in smokers compared to the lung cancer risk of non smokers. 'Population attributable risk' is the number of excess cases of disease in a population that can be attributed to a particular risk factor (Jeffrey, 1989). It is a joint function of the prevalence of a risk factor in the population (smoking), the relative risk due to exposure to this factor (increased likelihood of lung disease), and the absolute risk of the disease (likelihood of dying of lung disease for a given group in a given period) (Jeffrey, 1989).

Risk Perception: Heuristics, Biases and Information Processing

Risk perception relies on consumers being aware of risk and processing this awareness so that it has an influence on their behaviour. However, because risk perception and risk processing are based on subjective, heuristics rules, and not dependent on rational logic or

reasoning, perceived levels of risk are often inaccurate. The risk literature states that the consumer's perception of risk is influenced by many factors other than the actual risk of the hazard occurring. Risk perception and risk awareness are dependent upon individual factors within the consumer, such as the consumer's personal interest in the risk, the time factor between the action and the consequence of the risk, and the tendency of the human decision making process to be irrational with regard to the individual's own health risks (Jeffrey, 1989). Further, Jeffrey (1989) and Bettman, Payne and Staelin (1986) state that the type of risks that are most easily assessed and understood by the individual are those risks most likely to be overestimated.

Slovic, Fischoff and Lichtenstein (1980) have identified a series of heuristics and biases which affect risk perception. Heuristics are 'judgment rules' or 'rule of thumb' concepts used by most people when facing a risk perception processing event, such as deciding between the use of two cleaning agents with risk potentials (Slovic, Fischoff & Lichtenstein 1980; Bettman, Payne & Staelin, 1986). Slovic et al.'s (1980: 183-190) study classified four common heuristics: "availability, denial, personal interest, and overconfidence." Availability is the result of how easily a person can recall or imagine the consequences of the risk behaviour (Appendix 2.1.). Consequently 'availability' is affected by factors such as frequency and recency of use, and portrayal of the product in the media.1 'Denial' is the most common method of heuristic judgment, where the consumer either regards low risk events as insignificant and of little or no consequence, or regards themselves as being in control of a controllable risk (Slovic et al., 1980; Bettman et al., 1986). Further Bettman et al. (1986) raise the notion of 'self misperception' in risk processing where ordinary users regard themselves as above the level of the average user. Consequently they believe that they are either immune to common hazards2 or capable of exercising some form of control over the risk.3 'Personal 1

Slovic refers to the media as extending from new reports to contemporary cinema citing Steven Spielberg's "Jaws" as an example of how 'the media' can influence risk perception, in this case of the risk of shark attack. 2 particularly in regard to risks associated with domestic cleaning products (Bettman et al., 1986). 3 Bettman et al. (1986) raise the idea of 'incorrect self perception' with risks inherent in driving automobiles. Bettman contends that most people see themselves as above average drivers

interest' is the level to which people pay attention to information concerning a particular risk. Tonn, Travis, Goeltz, and Phillippi (1990) contest that a person's 'interest' in a risk leads to their depth of knowledge, and perception of that risk, particularly in understanding complex relationships between multiple factors in a risk situation. For example, risk knowledge of the term 'cancer' is less finely grained than 'lung cancer'. Awareness of the type and nature of a subset of the 'cancer' risk, 'lung cancer', indicates a greater knowledge of the risk, its consequences and relationships with other factors (Tonn et al., 1990). The 'overconfidence' heuristic, (Appendix 2.1.) is derived from the combination of reliance on heuristic judgments; failure to understand the operation of a system of risk elements; and a failure to perceive cumulative effects or recognise human error (Slovic et al., 1980). Jeffrey (1989) and Bettman (1986) isolated one aspect of Slovic et al.'s (1980) 'overconfidence' heuristic as 'misperception of a delayed effect' where there is a considerable delay between the risk event and the risk effect. The effect of the 'time lag' is to cause the consumer to consider themselves relatively immune to common hazards, having received no negative reaction in close proximity to the risk action (Bettman et al., 1986).

As a result of an analysis of the aforementioned heuristics and biases, the process of risk recognition, perception and response have been incorporated to develop the model in Figure 2.1. Figure 2.1.

Risk Processing Chain: From Risk to Risk Response

Risk Event

Risk Information and Awareness

Risk Information Processing [Heuristics and Biases]

4

Risk Perception

Risk Response

(Source: Developed from Slovic et al., 1980, Bettman et al., 1986; Tonn et al., 1990 and Jeffrey, 1989)

4

An expanded risk processing chain, with full heuristics is in Appendix 2.1.

This model has been constructed from the research of Slovic et al. (1980), Bettman et al. (1986), Tonn et al. (1990) and Jeffrey (1989). Figure 2.1 illustrates simplified version of the full risk processing contained in Appendix 2.1.

Risk Response

Risk response is comprised of four alternatives: personal risk management, legislative demand, consumer activism and acceptance. Personal risk management is the use of personal precautionary measures to reduce risk encounters. Legislative demand is where the individual seeks government intervention to reduce the level of risk. Consumer activism is an individual or group of consumers taking personal action against a business (Oglethorpe & Monroe, 1994). This 'action' can take the form of consumer boycotts, written complaints or private litigation. Specifically, this definition of consumer activism excludes government involvement, or demands for government involvement. Acceptance of risk occurs when the individual accepts the dangers inherent in the risk behaviour, but does not modify their activity to reduce these risks.5

For the purposes of this thesis, consumer activism is not being examined, rather the research will focus on the relationship between risk perception, levels of perceived risk and the three defined levels of risk response (personal risk management, risk acceptance and legislative demand) as illustrated by Figure 2.2.

5

Starr (1969) in Slovic (1992) comments that in voluntary risks, such as skiing, the individual is prepared to accept risk levels of up to 1000 times greater than those of 'non-voluntary' risks such as nuclear power plants.

Figure 2.2.

Risk Response

moderate to high risk

Risk Perception

moderate risk

Unacceptable Risk Level (Slovic's dread risk)

Unacceptable without Personal Risk Management [PRM]

LEGISLATIVE DEMAND Risk remains unacceptable despite PRM PRM reduces risk to an acceptable level

moderate to low risk

Acceptable Risk level

NO DEMAND

Perceived Risk Level of Risk Risk Response (Source: Developed from Slovic et al., 1980 and Starr, 1969 in Slovic et al., 1980)

The risk response model has been constructed from Slovic et al.'s (1980) research into 'dread risk' and the risk perception literature. Specifically, the model is designed to demonstrate the two paths for legislative demand, either derived from 'dreaded' risk, or where a risk remains unacceptable despite personal risk management.

Personal risk management involves the use of precautionary measures by the individual to minimise their risk from an event (Starr, 1969, in Slovic et al., 1980). If, however, the individual does not believe the risk to have been sufficiently reduced to be acceptable by their own actions, they may invoke legislative demand to assist their personal risk management. For example, taking additional care at a dangerous intersection when driving may reduce the personal risk to a level that is still unacceptable to the individual. The individual is then expected to seek government intervention to improve road conditions to reduce this risk, in conjunction with the individual's risk management. Legislative demand can arise through both initial reactions, and as a consequence of an individual's inability to sufficiently reduce their risk to an acceptable level through personal risk management.

Government Intervention

Legislative Demand

Legislative demand is the consumer's expectation as to the level and type of control that the government should operate in any given circumstances. For this research legislative demand will be limited in its application to the government control of the marketing of the products with perceived risks. The underlying assumption of the thesis is that consumers base their demand for regulation on their perception of a product's risk, rather than on the actual risk involved in using the product. Consequently, using Slovic et al.'s (1980) risk perception model where risk perception fails if a consumer can not 'imagine or recall instances of risk,' it is expected that for a consumer to demand legislation they would have to believe that the product posed a health risk (Bettman, Payne and Staelin, 1986). In particular, Slovic et al.'s (1980) 'dread risk' concept and Kasperson's (1992) model of the Social Amplification of Risk indicate a possible relationship between risk perception and legislative demand.

Slovic et al.'s (1980) 'dread risk'

'Dread risk' is defined by Slovic et al. (1980) as the type of risk about which people do not think rationally and calmly. Rather, the individual prefers to rely on instinct or 'gut' reactions due to the 'great dread' they have of the risk. This 'dread' factor is the result of a combination of subjective beliefs concerning the nature and type of risk. Slovic et al. (1980) identified ten elements comprising the 'dread risk' as: the severity of the consequences, the risk is uncontrollable, dread, the consequences are catastrophic, risk is hard to prevent, consequences are fatal, risk is distributed inequitably, the consequences threaten future generations, the risk is not easily reduced, the risk is increasing, the risk is involuntary and personally threatening. The most influential factor for legislative demand is the individual's belief that they have no form of control over the risk, either through personal risk management or avoidance. This is particularly prevalent in the 'dread risk' associated with nuclear power plants. The average person believes that their individual actions can do little to prevent, or avoid, the

consequences of an accident at a nuclear reactor. Legislative demand occurs when a hazard scores highly on this 'dread' factor where "people want to see its current risks reduced, and the more they want to see strict regulations employed to achieve the desired reduction in risk" (Slovic, 1992: 121).

Kasperson's (1992) Model of the Social Amplification of Risk

Kasperson's (1992) model of the social amplification of risk (illustrated in Figure 2.3.) contends that the effect of certain risk related events can go far beyond the initial impact of the event. This occurs through 'social amplification stations', such as individuals, groups or institutions which magnify, and focus public attention on the risk event, and the characteristics of the risk event (Kasperson 1992, Pidgeon et al., 1992).

Figure 2.3.

Social Amplification of Risk and the Potential Impacts on a Corporation.

CULTURE

Other technologies EC2 EC3

E

EC4 . . . ECn

RISK EVENT EVENT CHARACTERISTICS

Portrayal of an event ... •Symbols

Risk-Related behaviour .... •Institutions

industry Company Victims

Loss of sales Regulatory Constraints

•Signals

• Groups

Litigation

•Imagery

• Individuals

Community opposition

INFORMATION FLOW

(Source: Kasperson, 1992: 158)

INTERPRETATION AND RESPONSE

SPREAD OF IMPACT (RIPPLING).

TYPE OF IMPACT (COMPANY LEVEL).

The 'amplification' effect causes the impact of the hazard to radiate in a manner analogous to a stone dropped into a pond (Kasperson, 1992). The effect of the rippling motion is to trigger other risk activities, either in the amplification or suppression of risk reduction, which in turn causes further ripples (Kasperson, 1992). The rippling effect also moves the risk impact from being only at the level of the 'victim' to the higher levels of social agencies, reaching and affecting other areas with related characteristics (Kasperson, 1992). A practical example of Kasperson's social amplification model can be demonstrated in the Garabaldi's processed meats incident in 1995.

The Garibaldi Meats Case Study

In early 1995, a major food poisoning scare which resulted in 1 death and 21 cases of serious illness was related to the consumption of tainted processed meat products from the Garabaldi Smallgoods Company.6 When the event was receiving national publicity there was uncertainty as to whether a component product, or the processing technique was responsible for the contamination. The 'risk event' was the possibility of contracting food poisoning from the consumption of processed meat products, particularly Garabaldi Smallgoods, with its event characteristics being: risk of death, risk of serious illness and the use of processed meats. These characteristics were amplified by the media coverage of the incident, following the death of a four year old child from the food poisoning.7 Kasperson's 'rippling effect' occurred through the spread of risk information and risk association from the initial impact (food poisoning) within those who consumed the meat products, to the Garibaldi Smallgoods Company which ceased trading.8 The impact then spread from this specific company to the processed meats industry in general.9 The types of impact have ranged from loss of sales 6 7 8 9

Gibson, R (02 February, 1995). Child's Death Sparks Sausage Alert, Melbourne Age [Reuters New Archive] Food Poisoning Kills One and Infects 21. (03 February 1995). Reuters News Services. Garabaldi Firesale. (14 March 1995). Melbourne Age, [Reuters New Archive]

Bolt, C (22 February, 1995). Meat Survey to Measure Adverse Impact on Sales. Australian Financial Review, [Reuters New Archive]; Gibson, R. (08 February 1995). Smallgoods Makers Fear Sales Backlash. Melbourne Age. [Reuters New Archive].

(industry wide), community opposition (adverse perceptions of processed meats), litigation (company specific), and judicial procedures (coroner's inquest), which are major outcomes specified by the Kasperson (1992) model.

Kasperson's (1992) Legislative Demand

Kasperson's model implicitly suggests the existence of legislative demand. The model indicates a relationship between the 'spread of impact', that is, from victims to the general community and outcomes of the impact. Assuming the existence of Slovic's 'dread risk' in the 'ripple effect' this research intends to examine whether a relationship exists between consumer perception of risk and Kasperson's 'impact' of regulatory constraints. To provide a framework to test this, Kasperson's model has been redefined to exclude types of impact not applicable to the research such as loss of sales, litigation and investor flight. Further, legislative demand is indicated in the model below (Figure 2.4.) as the link between the ripple effect and the regulatory constraints.

Figure 2.4.

Redefinition of Kasperson's (1992) Model of the Social Amplification of Risk

CULTURE

Other technologies EC1 EC2 E

EC3 . . ECn

Portrayal of an event .... •Symbols •Signals •Imagery

Risk-Related behaviour ...... •Institutions • Groups

industry Company Victims

Regulatory Constraints

• Individuals Legislative Demand

Hypothesis 1 is derived from Slovic et al.'s (1980) 'dread risk' theory, Kasperson's (1992) model and Patterson's (1992) assertion of societal demand for government based risk control. H1:

The consumer's perception of a product's risk will influence their desire for government intervention to reduce the level of risk associated with that product

Hypothesis 1A identifies the relationship between perceived risk and legislative demand which is predicated on the assumptions of Slovic et al.'s (1980) 'dread risk' theory. H1A The higher the perceived risk, the greater the legislative demand.

Types of Legislative Controls

Legislative demand refers to a broad generic concept of 'increased government control'. This research acknowledges that there are many aspects of 'government controls', and consequently reviews a range of intervention strategies, both at macro and micro levels. Jeffrey (1989) defines two types of intervention strategy, individual level controls which affect consumers as individuals, and population level controls, which take effect across the whole society. Public health and information campaigns designed to better inform the consumer of product risks are listed as methods of 'individual level' controls despite the behavioural change aspects of those programs being more suited to population level intervention strategies (Jeffrey, 1989). At the population level, there are three main intervention strategies: economic incentives and sanctions, passive protection from environmental hazards, and control of promotional practices (Jeffrey, 1989; Viscusi, 1991). The economic incentive appropriate to product risk control in consumer behaviour terms is the use of taxation. This does not deny right of use of the product but instead attempts to restrict the level of use to a socially acceptable level by making the product less attractive to the consumer (Jeffrey, 1989). Control of the physical environment manifests itself in two ways, regulation of product design, such as minimum quality standards, and limitations on the

conditions of sale to reduce the availability of the risk factor to the public (Jeffrey, 1989; Viscusi, 1991).

The third method, control of promotional practices, is the most commonly applied approach (Jeffrey, 1989). The literature has concentrated primarily on advertising control as the main medium of government intervention for two reasons. First 'regulation of advertising is one of the most visible aspects of consumer protection' (Beales, 1991:101). Second, there has been a consistent demand for increased levels of government control of advertising (Calfee and Ringold, 1994). Consequently Hypothesis 2 focuses on the promotional controls rather than incorporating all elements of the marketing mix. Hypothesis 2, based on the assumption of demand for advertising controls, states: H2:

There will be a demand for greater legislative control over the advertising and other promotional activities of high perceived risk products.

Consequently, it is expected that legislative demand will exist for controls over promotion and advertising of products with perceived health risks. The literature also includes the 'final' option of an outright product ban. As Viscusi (1991: 131) states if 'declines in product use is the social objective then a total product ban is the most effective measure'.

Although not reviewed extensively in the literature, minimum age requirements, and government warnings are also government intervention strategies. Both methods are currently applied by the Australian government with products such as alcohol and tobacco where the minimum legal purchasing and consumption age is 18 years. Compulsory government warning labels have also been in force for tobacco products since the 1980s and most potential harmful products, ranging from detergents to microwave ovens carry voluntary warning labels or instructions (Viscusi, 1991).

Consumer Behaviour The final aspect of the literature review examines consumer behaviour and attitudes towards marketing and government regulation of advertising. Barksdale and Darden's (1972) research into consumer attitudes towards marketing and consumerism specifically examines the area of government regulation. They analysed responses by a series of classification variables including sex and age.

The major findings revealed little, if any, significant

differences between the responses by classification according to gender

(Barksdale &

Darden, 1972). The lack of difference between gender in support for government regulation is affirmed by the later research of Slovic, Kraushappe, Letzel and Malmfors (1989), and Pidgeon, Hood, Jones, Turner, and Gibson (1992).

Therefore Hypothesis 3, that legislative demand is not affected by gender, has been constructed to test the findings of Barksdale and Darden (1972), Slovic, Kraushappe et al. (1989) and Pidgeon et al. (1992). H3:

There will be no significant difference in legislative demand by gender

Barksdale and Darden (1972) isolated age as a differentiating factor in legislative demand. However this is not reported in other surveys in the literature.

Hypothesis 4 has been

developed to test legislative demand by age. Due to the expected homogeneity of age in the sample, it is not expected that there will be significant variations of age to test Hypothesis 4. H4:

There will be significant difference in legislative demand by age

Summary

Examination of the risk perception, risk management, public policy and consumer behaviour field has identified three omissions in relation to legislative demand. First, the link between perceived risk and consumer demand for government interaction has been assumed,

but not tested ( Slovic et al., 1980; Kasperson, 1992; Patterson et al, 1992). Second, the literature has a series of untested assumptions concerning legislative demand for advertising controls on high risk products. Whilst several studies have uncovered support for higher levels of advertising control generally (Barksdale and Darden, 1972; Barksdale Perreault et al. 1982), few studies have specifically examined the relationship between perceived risk and legislative demand.

Finally, Barksdale and Darden's (1972) assertion that age is a

differentiating factor in legislative demand has not been supported by the literature reviewed. Consequently, the research will examine the relationship between perceived risk and legislative demand for controls over advertising of high risk products. This relationship will also be tested for differentiation between age and gender. The following chapter will set out the methodology to be used to test this relationship and its associated assumptions.

CHAPTER 3: METHODOLOGY

Chapter 3 outlines the methodological approach that this research will use to test the hypotheses derived from the literature. It will outline the suitability of a purpose built questionnaire to collect the data required for the study, and the size and nature of the sample. Finally the chapter examines the results of a pilot study conducted to test the draft questionnaire.

Quantitative v Qualitative Research Methods

There are two basic forms of data collection from which all techniques are derived. Quantitative methods assume that the data sought is objectively measurable, quantifiable and able to be translated into numbers in a meaningful fashion (Zikmund, 1988). Qualitative methods assume that data can not be described simply in terms of numbers and can only be understood in subjective measurements, by recording emotion, experience and other nonquantifiable data (Neuman, 1994)

Qualitative

Michal-Johnson (1993:175) describes qualitative data collection as 'an array of interpretive techniques which describe, decode, translate and come to terms with the meaning, not the frequency of an event'. Mason (1993) describes qualitative as being 'strong' for understanding the meanings people attach to their behaviour, discovering new things about phenomena and generating hypotheses. Arneson (1993) states that qualitative data collection offers the dual edged advantage of individualised outcomes, rather than the common outcomes provided by quantification.

It offers the advantages of providing in-depth, detailed and data rich

examination of events. However, its major disadvantage is its weakness in being unable to test hypotheses, and generate 'common outcomes' (Mason, 1993; Arneson, 1993). Qualitative

techniques include ethnographic studies, field research focus groups, indepth interviews, and historical comparative research (Neuman, 1994).

Quantitative data

Quantitative research methods collect data in the form of numbers (Neuman, 1994). The strength of this approach is its use of objective, scientific recording and testing procedures to test and generate hypotheses (Mason, 1993).

Further, quantitative techniques produce

standardised 'common' outcomes from the research, rather than individualised results, which enable the creation of generalised models (Arneson, 1993). The major disadvantage of the technique is the lack of depth of data generated from its techniques.

However, this

disadvantage is less applicable to research which seeks to study the frequency of an event rather than the reasons for that event occurring.

In selecting the appropriate methodological approach, Mason (1993:33) asks: which method(s) (a) increase validity, reliability and generalisability of data, [and] (b) maximise useful data.

This thesis sets out to test four hypotheses, in an objective, scientific and generalisable way. Consequently, quantitative data collection is better suited towards the thesis research objectives, insofar as it is better suited to hypothesis testing than qualitative research. Further, the research is concerned with measuring extent rather than the reasons for legislative demand.

Methodology The methodology used in this research is a purpose built questionnaire designed to adequately test the research propositions outlined in Chapter 2. The aim of the research is to examine a cross section of the population's risk perception and attitudes towards government intervention. The social survey technique is best suited to this research objective as it allows for the collection of a large amount of quantitative data for relatively low cost and involvement time (Neuman, 1994).

These considerations are important in light of the limited resources

available to the researcher. Further justification is provided from prior research which has used the questionnaire technique in investigating risk perceptions, consumer behaviour and consumer attitudes. In particular, hypotheses 3 and 4 are testing assertions by Barksdale and Darden (1972) that have been derived from findings conducted exclusively using the questionnaire technique. Consequently, in order to enhance the replication of that aspects of Barksdale and Darden's (1972) study in the current research, the research tools have been replicated. All attitude questions in the survey will be measured using Likert scales as they provide a fast, efficient prescaling method for data quantification which will facilitate the speed of data processing involved in the research (Neuman, 1994; Zikmund, 1988). As much of the prior research has used Likert scaling, the current research will attempt to follow prior techniques where appropriate for comparison purposes. Apart from the predominant use of the survey technique in the literature, the technical advantages of the questionnaire format are the secondary reasons for the adoption of this method. Such advantages include speed, ease of coding and quantifying data for analysis, and a high level of anonymity. The final advantage of the questionnaire is the reduction in interviewer bias than is generally present in both interview, or focus groups (Zikmund, 1988).

The disadvantages of using a survey are related to the lack of depth of questions that can be used combined with the limited number of questions that can practically be asked. Questionnaires allow for only a moderate number of questions, compared to indepth interviews or focus groups. Further, questionnaire based research is one-way communication and cannot explain misunderstandings or problems. For this research design, however, the survey is being used to gather a broad range of reactions to the issues of regulation demand, perceived risk and the consumer's belief in the levels of risk involved in the use of certain products. The research is not concerned with the underlying reasons for these events, which would require the greater depth of interview or focus group technique. As a result, the lack of depth is not relevant to the research which is demonstrating whether a relationship between perceived risk and legislative demand exists, and not the reasons for this relationship.

Sample

Sampling Technique

Sampling techniques are divided into two categories, probability and non probability (Neuman, 1994). Probability sampling is where an element, or groups of elements have an equal chance of being included in the sample (Adams & Schvaneveldt, 1991). This method is divided into random, systematic, stratified, and cluster sampling (Malhotra, 1993). Random sampling is where all of the elements have an equal chance of being included (Adams et al., 1991). Systematic sampling involves defining a population and selecting from the population at set intervals (deVaus, 1986). Stratification involves division of the population into 'strata' and random selection from the sub populations.

Finally, cluster sampling involves

multistaged random sampling from randomly assigned groups or 'clusters' of the population (Loether & McTavish, 1974).

Non probability samples are those techniques that do not determine the probability of an element being part of the sample (Adams & Schvaneveldt, 1991).

Commonly, these

techniques involve convenience, judgmental, quota and snowball sampling (Malhotra, 1993). Convenience samples are drawn from 'convenient elements' such as students, church groups or members of social organisations (Malhotra, 1993). Judgmental, or purposive, samples are those samples where the respondent is selected by the researcher (Neuman, 1994). Quota sampling uses similar techniques to judgmental sampling, although the number of people in each category is fixed, and the researcher continues sampling until the quotas are met (Malhotra, 1993; Neuman, 1994; deVaus, 1986). Finally snowball sampling uses initially identified subjects who can refer the researcher to additional respondents (Adams & Schvaneveldt, 1991). The major advantage of non probability sampling is the high speed and low cost associated with the technique (Malhotra, 1993).

However, the results derived

through these methods are not generalisable to wider populations (Neuman, 1994). Consequently, the use of non probability sampling involves a tradeoff of speed and cost against general applicability. A second advantage of non probability sampling is its ability to produce good estimates of population characteristics (Malhotra, 1993).

The research design involves the examination of a potential link between risk perception and legislative demand within a short time frame, and limited financial resources. Further, as an exploratory study, there is less requirement for the initial study to be able to be generalised to the wider community. This research is to establish whether a link exists between the two factors, not the strength of this proposed link. Further it is not intended to examine if the link is consistent between the sample and the general community. Considering these factors, non probability convenience sampling will be used because of its advantages of high speed and low cost combined with its suitability for exploratory research (Malhotra, 1993). Further, the primary disadvantage of the technique, its lack of generalisability, has already been accepted as a limitation of the research design.

Having established the sampling technique, the

following section will set out the determinants of the sample size.

Determining the sample size.

Scheaffer, Mendenhall and Ott (1990) define the formula for determining sample size, where N and B are known, σ2 is unknown, but can be derived from prior research, as:

Figure 3.1.

Formula for determining sample size 'n' n= Νσ2 (Ν−1)D + σ2

where

B2

D=

4 (Scheaffer, Mendenhall & Ott, 1990)

To obtain n, N(population) is defined as the number of university students at the Griffith university; B is defined as the bound on the error of estimation; and σ (variance) is derived from Calfee and Ringold's (1994) historical review of consumer attitudes towards regulation of advertising10. Consequently, N=16,436 (Ashenden & Milligan, 1994); B= 3.5% error of estimation; and σ = 19.65 (Calfee & Ringold, 1994). Figure 3.2.

Determining sample size 'n'

n=

16436x (19.65)2

(N=16,436)a σ = 19.65 B= 3.5 D=(3.5)2 4

(16,436−1)3.06 + (19.65)2

n = 125 (a: Ashenden & Milligan, 1994)

The questionnaire will use a sample of Griffith University students, from both postgraduate and undergraduate levels. Using the student population has several practical 10

See Appendix 3.1

advantages including convenience, speed and ease of accessibility. Further, Pidgeon et al. (1992) state that the participants in many risk perception studies 'are typically students'. Also Smith (1992), when conducting research on alcohol product warnings stated that students were a preferable study population because alcohol was found to be 'a highly important product' to the students. This is relevant to the current research as it examines risk perception of alcohol, amongst other products.

The main disadvantage to using the student population is the unrepresentative nature of the sample. Students are expected to be a more highly educated population than the general community, and will generally be from similar age groups.

Process

The questionnaire will be distributed at the beginning of a lecture, and then collected before the lecture commences. This will give the survey a higher response rate from having a 'captive' population and it will enable the conditions under which the questionnaire is completed to be more controllable. Further, conducting the questionnaire in a controlled environment should lead to a higher response rate than if the survey was conducted via a mail-out (Zikmund, 1988; Neuman, 1994).

Questionnaire Details

Product Selection

Nine products (vegetables, tobacco, alcohol, fried food, dairy products, red meats, olive oil, pork, and chicken) were selected for their varying degrees of risk involved in causing heart disease. The food products were selected from their positions within the National Heart Foundation's 'Food Pyramid' diagram (Appendix 3.2), with at least one product from each level of the pyramid. Fried food and red meats were selected from the 'eat least' category, with dairy products, olive oil, pork and chicken from the 'eat moderately' category and vegetables from the 'eat most' level. Consequently, the majority of the food products are from the 'eat moderately' category, indicating that they have a moderate level of risk associated with their consumption. 'Vegetables' have been selected as a 'control product', as they are expected to have no risk associated with their use, and no demand for government regulation of their marketing. An example of this section is illustrated below: Section 1: Product Use Questions The following is a list of products that people may consume. Considering your usual behaviour, please tell me how often, if at all, you use the following products. Please select only one response for each product: 3 times [product]

less than

never

daily

a week

weekly

weekly

monthly

yearly

used

1

2

3

4

5

6

0

Product use is examined since Slovic, Fischoff and Lichtenstein's (1980) assertion of the reliance on personal experience may be more prevalent in situations where the person may be familiar with the hazard, or believe that they have some degree of control over the risk. Further, people tend to believe that they are relatively immune to common hazards (Bettman et al., 1986). Consequently, the level of risk perception of a product with a high actual risk, that the consumer frequently uses, is expected to be lower than in a situation where a consumer has a low usage rate.

Payne's (1986) model indicates that consumers have

difficulty in making decisions between risk options where they could experience either a gain

or a loss (Bettman et al., 1986). Thus, the consumer is willing to use products with high risks if they believe certain factors, such as convenience or price, outweigh the risk. In every day living, the choice to use a motor vehicle is frequently cited as the 'classic' example of the consumer believing that the perceived risks of death, accident, or injury are outweighed by their control over those risks; the benefits of using the motor vehicle; and the personal experience of having controlled the risks in prior use of the motor vehicle (Bettman, 1986). Perceived risk of the use of the motor vehicle is not governed by actual risk. Rather it is affected by factors from both Payne (1985) and Slovic et al.'s (1980) models of risk perception. Risk perception of the nine products is examined in the following Section 2. Section 2: Product Risk Perception Risk Perception: General Risk of Product Use The following is a list of products that people may consume. To what extent, if at all, do you believe consumption of each product is harmful to a person's health. For each product, indicate whether you think that it is 1-very harmful, 2-somewhat harmful, 3-not too harmful, 4-not at all harmful or 0 - Don't know. Please circle only one response for each product

[product]

Very harmful 1

Somewhat harmful 2

Not Too harmful 3

Not at all harmful 4

Don't know 0

This question is a partial replication of Mazis, Morris and Swasy's (1991) telephone survey which evaluated consumer risk perception of the consumption of alcoholic beverages, tobacco products, dairy products and other consumer goods in their research.

This question is

included in the survey so that a comparison of average consumer risk perception of alcohol, tobacco, and dairy products between this sample, and Mazis et al.'s (1991) sample can be conducted. This section examined general risk perceptions, whereas the following section specifically examines the respondent's belief as to the product's risk of causing heart disease.

Specific Risk Perception: Perception of the Heart Disease Risk

The following is a list of products that people may consume. To what extent, if at all, do you believe that there is a risk of heart disease associated with the consumption of each product. For each product, indicate whether you think that it is 1- very high risk, 2-high risk, 3-moderate risk, 4-low risk, 5-very low risk, 6-no risk, or 0- don't know. Please circle only one response for each product.

[product]

very high risk 1

high risk 2

moderate risk 3

low risk 4

very low risk 5

no risk 6

Don't know 0

The second question specifically examines the consumer's perception of the level of risk involved in the consumption of the nine products with respect to heart disease. It will be used to examine the consumer's perceived risk of each product and in conjunction with the legislative demand table below to examine the relationship between level of government intervention sought and the risk perceived by the consumer. Section 3: Product Regulation - General The third section of the survey involves 18 questions which gauge the respondent's reactions to specific areas including the role of government, consumer sovereignty, the regulation of price, product, promotion, place and legislative demand. The questions for legislative demand are adapted from Barksdale and Darden's (1972) work for the purposes of replication in order to test Hypothesis 3. The questions are grouped by category in the discussion below. The following list of statements relate to product regulation. Please circle the number that you believe best suits your opinion:1-strongly agree, 2-agree, 3-unsure, 4-disagree, 5-strongly disagree and 0-no opinion. Please circle only one number for each question.

Role of government statement number 1 It is the government's responsibility to protect the consumer from products with a known risk of causing heart disease 7 The government should supply information to assist consumers to assess product risks 8 The role of the government is to protect the consumer 11 Government has a responsibility to ban products which are shown to cause heart disease 15 There should be a mandatory health warning label on any product with a risk of heart disease

The questions are designed to examine respondent's attitudes to the role of government in legislative control. The responses will be examined in relation to the specific legislative demand issues emerging from Section 4. Role of Government: Generic Legislative Demand

statement number 6 The individual has a right to choose to consume products with known health risks 8 The role of the government is to protect the consumer

Question 6 gauges the respondent's belief as to the role of government, compared to the rights of the consumer. This is the basic level of legislative demand, whether it is the government's responsibility to protect the consumer from risk or individuals responsibility to avoid that risk. Question 6 is the reverse of question 8, testing the respondent's belief of the extent to which the consumer has a right to consume products with health risks. Regulating Price, Product, Promotion and Place

statement number 3 The government should regulate advertising of products with known heart risks 4 The government should regulate the availability of products with known heart risks 16 Products with known risks of causing heart disease should be more heavily taxed than products without health risks 17 The government should set minimum health standards for all food products

Each aspect of legislative demand for control of the marketing mix is tested in generic form to determine the base level of approval for legislative intervention in the '4 P's' of marketing products with perceived risks. Negative Response statement number 5 There are too many controls on the marketing of products with known heart risks 9 The government should not use taxation to modify product demand 10 It is not the government's role to control advertising of products with health risks 18 Consumers should not be able to purchase products that may damage their health

The 'negative' questions are reverse questions from each of the major sections of legislative demand, role of government and consumer sovereignty. These questions have been included in the survey in an attempt to prevent an 'acquiescent response set' where the respondent

assumes the questions are arranged in the same manner, and gives the same response (deVaus, 1986). Further, the reverse questions provide the survey with a greater internal validity. Replication Questions Calfee and Ringold (1994): Consumer beliefs about advertising. statement number 2 The government should exercise more responsibility for regulating advertising, sales and marketing

Question 2 has been taken from Calfee and Ringold's (1994) analysis of six decades of advertising opinion research. It has been selected for a comparison of current demands for legislation, and historical levels of consumer demand. As a multi-faceted question it would be better asked in three separate questions, however, the original question is being followed to enable cross comparison. Barksdale and Darden (1972) Consumer attitudes towards marketing and consumerism.

statement number 12 The government should test competing brands of products and make the results of the tests available to the consumer 13 The government should set minimum standards of quality for all products sold to consumers 14 The government should exercise more responsibility for regulating the advertising, sales and marketing activities of manufacturers.

Hypothesis 3 (differences by gender) and Hypothesis 4 (differences by age) will be tested using a partial replication of Barksdale and Darden's (1972) research into consumer attitudes toward government regulation. In order to replicate the study as accurately as possible, the three questions are being used in the same format (Likert scale of strongly agree to strongly disagree) and in the same order as in the original study.

Section 4: Methods of Control: Generic

From the following list below please select the methods of government control you believe are appropriate for regulating products with known health risks. You may select more than one method. ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____

education programs, information campaigns product ban high taxes regulating product design, restricted availability restrictions on advertising controls over product promotion minimum age requirement government warning labels, no controls other: (please specify):

The ten options listed above have been drawn from Jeffrey's (1989) comparison of individual and population perspectives of risk behaviour. Each type of legislative demand raised by the literature has been represented in these questions. The 'No Controls' option has been included in the questionnaire based on the assumption that consumer sovereignty is a method of product risk control. The inclusion of an open ended question, the 'other: (please specify)' category, will allow respondents to recommend a method of product control that has not been either considered by the literature or appears to the respondent in a different form to those methods listed.

Section 5: Specific Controls over certain products In the following table there are several methods of government control over product marketing. Please circle the method, or methods of control you believe are most appropriate for each product (You may select more than one method for each product).

[product]

high tax 1

minimum advertising age prohibition 2 3

restricted availability 4

product ban 5

education campaigns 6

.... .... ...

no controls 0

Section 5 is a continuation of Section 4. All methods listed for the examination of attitudes towards generic risk will be included in the final survey, including an option for 'other controls' and 'no controls'. This has been constructed to examine the level of demand for specific controls for specific products. According to the literature, the level of legislative

demand for the generic product risk (Questionnaire Section 4), combined with perceived product risk (Questionnaire Section 2) should determine the specific product legislative demand (Questionnaire Section 5).

Section 6: Demographics Age: _______

Gender:

Female

Male

Age and gender characteristics will be collected to test Hypotheses 3 and 4 (legislative demand by gender and age).

Pilot Study

A pilot study was conducted using a sample of 20 people, aged between 19 and 56, with a mean age of 27 years. The pilot study questionnaire is set out in Appendix 3.3. The sample contained even distribution of male and female respondents. Feedback was received from several respondents (five male, three female), and a further two female respondents were given informal interviews to discuss any problems with the questionnaire. Initial analysis of the pilot study, using SPSS for Windows V.6.1., revealed difficulties in testing hypotheses 1, and 1A, the relationship between legislative demand and perceived risk. Further, several respondents reported difficulties with the product usage level in Section 1, in the post questionnaire interviews. Finally, Section 3 has been reordered and refocused as several negative response questions were placed too close to their respective 'positive' questions.

Hypothesis 1

Changes

In order to directly test Hypotheses 1 and 1A, Section 5, Levels of Government Control has been inserted. While this is an original question it is modelled on Mazis, Morris and Swasy's (1991) risk perception question (see Section 2, above).

Section 5: Levels of Government Control

The following is a list of products that people may consume. To what extent, if at all, do you believe that the level of government control over the advertising, sales and promotion of each product should change. Please indicate whether you think that it should 1- greatly increase, 2-partially increase, 3 - stay the same, 4 - partially decrease, 5 - greatly decrease or 0- have no controls

[product]

greatly increase 1

partially increase 2

stay the same 3

partially decrease 4

greatly decrease 5

no controls 0

The question has been included in the survey to facilitate the testing of Hypotheses 1 and 1A. Using the new questionnaire, correlations between Section 4 and Section 5 can be used to determine if a relationship exists between perceived risk and legislative demand. Further, the relationship between perceived risk and Section 5 can be tested in regards to both heart disease and general health risk (Section 2).

As a result of pretesting of Section 5, the list order of the nine products has been changed from 'vegetables, tobacco, alcohol, fried food, dairy products, red meats, olive oil, pork and chicken' to 'tobacco, red meats, chicken, fried food, dairy products, alcohol, olive oil, pork and vegetables'. This change was made after respondent complaints concerning the difficulty of recommending changes in regulation levels to a product with low risk, and low regulation levels. The major complaint was that the respondents found the question confusing as it was difficult to recommend changes to levels of government intervention where there were no existing advertising controls.

Product Usage Levels (Section 1)

Several respondents reported confusion as to which category of product usage they should select for products which they have used in the past, but no longer use. Further, the periods of use (daily, 3 times a week, weekly, less than weekly, monthly, yearly, never) where criticised by several respondents as being 'too vague', or 'difficult to relate to day-to-day life'. As a

result, the periods of use were changed to: daily; 3 times a week; weekly; between weekly and monthly; monthly; less than monthly; and don't know.

The changes in the depth of

examination of product usage levels, concentrating on the recency of use, should increase the relevance of Slovic et al.'s (1980) 'availability' heuristic. Section 3

Three questions have been removed from Section 3 following the analysis of the pilot study data. statement number 7

The government should supply information to assist consumers to assess product risks.

15 There should be a mandatory health warning label on any product with a risk of heart disease.

Questions 7 and 15 have been removed as they concern two specific types of regulatory control, which are examined in full in Section 6 (Generic Controls) and Section 7 (Specific Controls). Consequently the duplication has been removed. statement number 9

The government should not use taxation to modify product demand

Question 9 is redundant as a 'negative' question since Section 5 (Levels of Government Controls) Section 6 (Generic Controls) and Section 7 (Specific Controls) of the questionnaire also examine demand for regulatory control over price product promotion and place. The final questionnaire is given in Appendix 3.4.

Summary

The major limitations of an exploratory study is the strength of the research instrument. The research uses a purpose built questionnaire, constructed from replication studies and questions drawn from the literature. The survey was divided into seven sections: product use, heart risk perception, general opinions concerning product regulation, general risk perception,

methods of controls for products with known risks, and methods of controls for specific products. This survey design was tested in a pilot study conducted on 20 University of Queensland students. From that study, three main changes occurred: the addition of a specific legislative demand question, the alteration of the time frame for product usage, and the removal of three questions from general opinions concerning product regulation. Changes were also made to the ordering of the products, and questions in the general opinions of product regulation. The following chapter is a preliminary analysis of the data collected from the questionnaire.

CHAPTER 4: RESULTS

Chapter 4 reviews the statistical data collected from the main study outlined in Chapter 3. The chapter outlines the statistical frequencies for each section, including the preliminary analysis of the data collated for the four hypotheses. Finally, the results from the three replication questions are reviewed in comparison with the results of the original studies.

Demographics The survey was conducted by the distribution of 150 questionnaires. 132 surveys were returned (88% response), with 128 usable surveys (85% usable response rate). The university survey was conducted either at the immediate beginning of the class, or immediately following the completion. Gender distribution of the survey was 76 (57.6%) male and 56 (42.4%) female. The age of respondents ranged from 18 to 46, with a mean age of 25 (SD. 6.9). Respondent age was not evenly distributed with median age 23, and the majority 53.0% of the respondents between 18 and 23 years old. Product Usage Section 1 of the questionnaire established product usage rates for the nine products. Six alternative values were available: daily, 3 times a week, weekly, less than weekly monthly, yearly and never used. Dairy products (77.3%) and vegetables (70.3%) were the most common daily use products. Tobacco was the least used product, with 28.3% of respondents using the product. Table 4.1. summarises the product usage rates.

Product Usage

Table 4.1.

product

frequency

tobacco (never used) red meats 3 times a week chicken. weekly fried food weekly dairy products daily alcohol weekly olive oil 3 times a week pork weekly vegetables daily (Source: Current research)

% (71.7%) 43.0% 43.0% 35.9% 76.6% 26.6% 21.1% 33.6% 70.3%

Heart Risk Section 2 asked the respondents to give their belief as to the risk of heart disease that may arise from the ordinary use of the nine products. The response options ranged from very high risk to no risk, and included the option 'don't know'. Tobacco had the highest heart disease risk, (82% high or very high), followed by fried foods, (78% high or very high). Alcohol and red meats recorded moderate risk at 31.5% and 41.4% respectively. The remaining products recorded low to no risk. General Risk Section 4, 'General Risk', is a partial replication of Mazis, Morris and Swasy's (1991) telephone survey which was conducted to evaluate consumer risk perception of the consumption of alcoholic beverages, tobacco products, dairy products and other consumer goods. Respondents had the options of very harmful, somewhat harmful, not too harmful and don't know. In the current research, tobacco had the highest health risk with 99.2% of respondents rating it somewhat harmful to very harmful. Alcohol and fried foods both reported high health risks, with 78.6% and 83.2% selecting somewhat harmful or very harmful for the product risk. The remaining products were rated as not too harmful, or not at all harmful.

Legislative Demand

Levels of Government Control Section 5, 'Levels of Government Control', was constructed to establish levels of respondent demand for legislative controls over advertising, sales and promotion. Respondents had six options, greatly increase, partially increase, stay the same, partially decrease and no controls. Of the nine products, only two recorded significant demands for increased advertising control. Alcohol and tobacco recorded majority demand for increase, 53.4% and 58.8% respectively. Chicken, red meats, fried food, dairy products, olive oil, pork and vegetables all received majority support for 'stay the same'.

Government Controls: General

Section 6 examined consumer demand for a series of government controls. Education programs (89.1%), information campaigns (80.6%), government warning labels (82.9%) and minimum age requirements (53.5%) received majority support (see Table 4.2.). The total number of controls ranged from no controls to all types (10), with an average of four controls. Table 4.2.

Frequencies for Methods of Control Yes

education programs government warning labels information campaigns minimum age requirement restrictions on advertising controls over product promotion restricted availability high taxes regulating product design product ban no controls (Source: Current research)

89.1% 82.9% 80.6% 53.5% 48.1% 40.3% 34.1% 32.6% 24.0% 13.2% 0.8%

No 10.9% 17.1% 19.4% 46.5% 51.9% 59.7% 65.9% 67.4% 76.0% 86.8% 99.2%

General Opinions Concerning Product Regulation.

Section 3 has been divided into two subsections, opinions concerning regulations of products with heart risks, and opinions concerning the regulation of products with general health risks.

Heart Disease Risk

Examining the 'heart risk' questions first, only two of the six 'heart risk' questions received majority support, increases in advertising regulation (74%), and the responsibility of government to protect the consumer from heart disease risks (67.7%). Whilst advertising regulation was supported, respondents did not support increased taxation (35.2%), restriction of availability (42.5%), and product bans (20.5%). Opinions concerning whether there were too many controls on the marketing of products with heart risk were less clear cut, with 44.1% unsure, and 41.8% disagreeing with the statement. Only 8.7% agreed that the marketing of heart risk products was over controlled. These findings are summarised in Table 4.3. Table 4.3.

General opinions concerning regulation of products with Heart Disease Risk SA / Agree 67.7

Disagree/SD 26.0%

The government should regulate advertising of products with known heart risks.

74.0%

22.8%

4

The government should regulate the availability of products with known heart risks.

42.5%

44.9%

5

Products with known risks of causing heart disease should be more heavily taxed than products without health risks.

35.2%

47.2%

9

The government has a responsibility to ban products that are shown to cause heart disease.

20.5%

66.2%

8.7%

41.8%

1

It is the government's responsibility to protect the consumer from products with a known risk of causing heart disease.

3

14 There are too many controls on the marketing of products with known heart risks. (Source: Current research)

General Health Risks

There was a high level of support for the individual's right to choose products with health risks (90.6%), and over two thirds (71.7%) believing that the consumer had the right to purchase products that may damage their health. However there was some inconsistency between support for the rights of consumers, and role of government. Despite a high level of support for the consumer's right of choice, there was strong support (59.05%) for government to protect the consumer. Specifically, there was a high level of agreement for increased advertising controls (65.2%), government testing programs (75.6%), minimum health standards (83.5%) and minimum quality standards (89.0%). Increased government responsibility for regulating advertising, sales and marketing in general received 57.5% strongly agree/agree, whilst similar regulation of manufacturers received 50.4%. findings are summarised in Table 4.4. Table 4.4.

General Opinions concerning regulation of products with known health risks SA / Agree 57.5%

Disagree/SD 27.6%

The individual has a right to choose to use consumer products with known health risks.

90.6%

5.5%

7

The role of the government is to protect the consumer.

59.0%

22.8%

8

It is not the government's role to control advertising of products with health risks.

23.6%

65.2%

10 The government should test competing brands of products and make the results of the tests available to the consumer.

75.6%

12.6%

11 The government should set minimum standards of quality for all products sold to consumers.

89.0%

0.8%

12 The government should exercise more responsibility for regulating advertising, sales and marketing activities of manufacturers.

50.4%

27.5%

13 Consumers should not be able to purchase products that may damage their health.

21.2%

71.7%

15 The government should set minimum health standards for all food products. (Source: Current research)

83.5%

5.5%

2

The government should exercise more responsibility for regulating advertising, sales and marketing.

6

These

Hypothesis 1A H1A : The higher the perceived risk, the greater the legislative demand.

Hypothesis 1A examines the relationship between general risk perception variable (GR) and the legislative demand (LD). The relationship is tested by correlation of the two variables. The expected relationship between legislative demand and perceived risk relies on a high GR score (somewhat harmful, very harmful). Consequently, products with low perceived risk will not have a significant correlation. The four correlations that will be examined are tobacco (perceived on average in this study to be very harmful), alcohol, fried foods (somewhat harmful) and vegetables (not at all harmful). Significance for the correlation is at the .05 level.

To test H1A, the null hypothesis is set as "H0: That no significant relationship exists between perceived risk and legislative demand for each product where H0 : p >.05 and H1A: p<.05." Correlations for perceived risk and legislative demand had significance of .039 for tobacco, .003 for alcohol and .828 for fried food. Consequently, the null hypothesis for relationships between tobacco and alcohol can be rejected, demonstrating a relationship between perceived risk and legislative demand for alcohol and tobacco. In contrast, the correlations for fried food (.828) and vegetables (.604) indicate the null hypothesis should be accepted for these products. Hence, no statistically significant relationship exists between legislative demand and the perceived risk of fried foods and vegetables.

Hypothesis 2 H2: That there will be a demand for greater legislative control over the advertising, and other promotional activities of high perceived risk products.

Hypothesis 2 tests the relationship between perceived high risk products and the demand for controls of the advertising, and promotional activities. This will examined through a t-test between the general risk variable (GR), and the methods of government controls for specific product variables 'restrictions on advertising' (SCGC7), and 'control over product promotion'

(SCGC8). A requirement of the hypothesis is that the products tested be perceived as having a high risk, which reduces the field to tobacco, alcohol and fried food. Only tobacco, with a t-value of 2.37, had a significant relationship (.019) at the .05 level. Alcohol and fried food having no significant relationship between product risk and legislative demand for control over promotion, advertising and sales.

The same limitations apply to examining the relationship between perceived risk and demand for controls on product promotion as exist for controls on advertising. In contrast to advertising controls, tobacco had no significant relationship between risk and demand for controls on promotion. Only alcohol, with a t-value of 2.53, had a significant relationship (.014) at the .05 level.

Hypothesis 3 H3: There will be no significant difference in legislative demand by gender

Hypothesis 3 consists of two sections. The first part is an analysis of demand for levels of government control (Section 5) by gender. The second part is the analysis of responses to question 10, 11 and 12 from section 3. Both analyses will use the t-test of GENDER and the respective variables (LD, PR).

Eight of the nine products examined had no significant difference between male and female demand for government controls. Tobacco, with a t-value of 2.74, has significant difference (.007 at .01) in demand between males and females. Female respondents reported a higher level of demand (mean, 1.90) than males (2.42) for increased controls over tobacco. The three replication questions from from Barksdale and Darden (1972): the government should test competing brands and makes results of the testing available to the consumer; the government should set minimum quality standards for all products sold to consumers; the government should exercise more responsibility for regulating the advertising, sales and

marketing activities of manufacturers. All three questions had no significant differences in responses by gender with significance values between .532 and .810.

Hypothesis 4 H4: There will be significant difference in legislative demand by age

Hypothesis 4 is divided in the same manner as Hypothesis 3. However, Hypothesis 4 will be tested using correlations, not t-tests, because of the ordinal nature of the AGE data. No correlations occurred between age and legislative demand, with only tobacco (.066) approaching significance. Testing the second aspect of legislative demand resulted in no significant relationship between age and response to Questions 10 and 11. The relationship between Question 12, regulation of advertising, sales and marketing of manufacturers, and age was significant (.019 at the .05 significance).

Replication Studies

Mazis, Morris and Swasy's (1991): General Perceived Risk

Section 3 was a partial replication of Mazis, Morris and Swasy's (1991) telephone survey which evaluated consumer risk perception of the consumption of alcoholic beverages, tobacco products, dairy products and other consumer goods. Mazis et al. (1991) conducted two surveys between 1989 and 1991, where they collected respondents perceptions of risk associated with using alcohol, tobacco and dairy products, amongst other product. The research replicated the design of Mazis et al.'s (1992) survey for comparison of risk perceptions between the three studies. The first two columns of Table 4.5. summarise Mazis et al.'s 1989, and 1990 survey findings. The third column is the risk perception scores recorded from this research.

Table 4.5.

Percentage of responses rating the general risk of alcohol, tobacco and dairy products as 'very harmful' Mazis 1989

Product

Mazis 1990

Dann 1995

alcohol

49.8%

54.1%

26.7%

tobacco

81.6%

80.9%

79.4%

2.8%

2.1%

.8%

dairy products

(Source: Current research; Mazis et al, 1992)

Mazis et al.'s (1992) study also examined the combined percentages of the perception of tobacco and alcohol as 'very harmful' and 'somewhat harmful', which are summarised in Table 4.6. Table 4.6.

Product

Combined Percentage of responses rating the general risk of alcohol and tobacco as 'somewhat harmful' or 'very harmful' a Mazis 1989

Mazis 1990

Dann 1995

alcohol

88.7%

89.0%

78.6%

tobacco

94.7%

95.9%

99.2%

a: Combined percentages unavailable for dairy products (Source: Current research; Mazis et al., 1992)

Calfee and Ringold (1994): Consumer beliefs about advertising.

Question 2, Section 3 was taken from Calfee and Ringold's (1994) analysis of six decades of advertising opinion research for a comparison of current demands for legislation, and historical levels of consumer demand. Calfee and Ringold's (1994) historical review findings are summarised in Table 4.7.

Table 4.7.

Desire for greater government regulation of advertising sales and promotion

The government should exercise more responsibility for regulating advertising, sales and marketing

1971

1973

68

59

% responding 'agree' or 'strongly agree' 1975 1979 1982 1995 60

48

62

57.5

(Source: Current research; Calfee and Ringold, 1994: 235; Barksdale Perreault et al. 1982: 83)

Barksdale and Darden (1972) and Barksdale, Perreault et al. (1982) Consumer attitudes towards marketing and consumerism.

Questions 10, 11 and 12, Section 3, of the survey are replications of Barksdale and Darden's (1972), which itself was replication by Barksdale, Perreault et al. (1982). Barksdale et al.'s (1982) study of consumer attitudes was conducted in six different countries, Australia, Canada, England, Israel, Norway, and the United States to examine the international applicability of the original study. Differences and similarities between the responses from this research and the 1972, and 1982 studies are listed in Table 4.8. below. Table 4.8. Statement

Attitudes towards government regulation year of survey

strongly agree

agree

uncertain

disagree

strongly no disagree opinion

test competing brands

1972

28.2

45.5

8.5

14.1

3.7

0

of products and make

1982

28

40

8

20

4

0

the results of the tests

1995

24.4

53.5

9.4

10.2

2.4

2.4

The government should

available to the consumer

Attitudes towards government regulation (continued)

Table 4.8. Statement

year of survey

strongly agree

agree

uncertain

disagree

strongly no disagree opinion

The government should set

1972

25.1

48.6

9.9

13.6

2.8

0

minimum standards of

1982

35

49

4

10

2

0

quality for all products sold

1995

32.3

55.9

7.1

3.1

0.8

0.8

The government should

1972

25.1

43.2

12.4

16.7

2.5

0

exercise more responsibility

1982

17

45

15

20

3

0

for regulating the advertising,

1995

12.6

37.8

18.9

24.4

3.1

3.1

to consumers

sales and marketing activities of manufacturers. (Source: Current research; Barksdale and Darden, 1972:33; Barksdale Perreault et al. 1982: 83)

Summary

The chapter overviewed the statistical data concerned with each section of the survey, the relationships to be tested for the four hypotheses and the comparative frequencies for the replication studies. Chapter 5 will examine the implications and conclusions to be drawn from the results of this chapter.

CHAPTER 5: CONCLUSIONS AND IMPLICATIONS 'Ugh!' snarled the Wolf, as he limped through the brushwood with his tail between his legs, 'this is perfectly monstrous weather. Why doesn't the Government look to it?' - Oscar Wilde, The Star Child in Murray (1979).

This chapter analyses the results of the research in the context of risk perception, risk management, public policy and marketing theories. replication studies are summarised and discussed.

Each of the four hypotheses and Further, the implications from these

hypotheses are discussed for the fields listed above. Finally, the limitations and further research opportunities from the thesis are discussed.

Conclusions

Research Problem: Legislative Demand and Perceived Risk

Patterson et al. (1992) stated that when society identifies a potential risk, there are those people who seek government intervention to reduce that risk. This statement is at the core of the research problem, examining whether a relationship exists between risk perception and demand for government intervention in the marketing of consumer products with known or perceived health risks. The broad relationship between consumer perception and the role of government was defined as the 'legislative demand' concept. This concept incorporates the consumer's belief as to the role of government in any given situation, ranging from government established health standards, to the role of government in regulating the content of pay television, or access to pornographic material on the Internet. For the purposes of this research, the concept was restricted to the specific demand for controls on the promotion of products with perceived health risks.

To examine the relationship between perceived risk and legislative demand, four hypotheses were established testing for the existence of a relationship between perception of risk and demand for controls on advertising, sales and promotion of nine products (tobacco, red meats, chicken, fried food, dairy products, alcohol, olive oil, pork, and vegetables); the demand for controls on the advertising and promotion of perceived high risk products; and whether differences existed between legislative demand and the characteristics of age and gender. From these hypotheses, the findings of the research are summarised as follows:

• A relationship exists between products with high perceived risks, and the demand for increases in government controls of advertising, sales and promotion.

• No general relationship between risk and demand for advertising control and promotion control was found in the nine products tested. However, alcohol and tobacco had specific relationships between their perceived risk and demands for promotional and advertising controls.

• Legislative demand was independent of gender when examined in relation to 'generic' perceived risk. 'Generic' risk is the non specific form of risk stated as a 'known health risk'. In testing the nine specific products, tobacco recorded an unexpected relationship between gender and legislative demand. This was not consistent with either the literature, or the results of the eight other products.

• Measurement of the relationship between legislative demand and age was limited by the demographics of the sample, which was skewed towards the 18-23 age group. Despite this limitation, age was found have a significant relationship at the .05 level as a factor in determining levels of support for increased government responsibility for regulation of advertising sales and marketing activities of manufacturers.

In review, the majority of the thesis findings were consistent with the literature. Inconsistencies arose from the failure of the expected relationship in Hypothesis 2, and the unexpected findings of differential legislative demand for tobacco by gender.

Possible

reasons for these differences between the expectations of the literature and the results of the research are examined in the respective sections above.

H1:

Hypothesis 1

The consumer's perception of a product's risk will influence their desire for government intervention to reduce the level of risk associated with that product

H1A: The higher the perceived risk, the greater the legislative demand.

Hypothesis 1A was tested by correlation between the consumer's perception of the general risk of a product to their health (Section 4 of the survey) and their belief as to the levels of government controls over the advertising sales and promotion of that product (Section 5 of the survey). The hypothesis is dependent on the product having a high enough perceived risk to cause legislative demand to occur. Three products, tobacco (99.2%) fried foods (83.2%) and alcohol (78.6%), were rated between 'somewhat harmful' and 'very harmful'. As reported in Chapter 4, the relationships between perceived risk and legislative demand were only statistically significant for tobacco (.039 at .05) and alcohol (.003 at .01). This supports the existence of the relationship as stated in Hypothesis 1A for tobacco and alcohol. These findings are supported by the prior research of Slovic et al. (1980), which placed 'smoking' as being a 'dreaded' risk, which, as stated above, is the type of risk that people want the government to control.

The significance of this finding is that it demonstrates the existence of a link between perceived risk, and demand for increased government control over the advertising, sales and promotion of tobacco and alcohol. Slovic et al. (1980) and Kasperson (1992) speculated as to the existence of a relationship between risk and legislative demand. The results of Hypothesis 1 is to demonstrate the existence of such a link for these two products. From this, further

research can be conducted to test for the existence of the relationship between the perceived risks and desire for controls over other products, services and events.

H2:

Hypothesis 2

That there will be a demand for greater legislative control over the advertising, and other promotional activities of high perceived risk products.

Hypothesis 2 examines the specific relationship between the desire for government controls, and the types of controls that the consumer believes should be implemented. Jeffrey (1989) identified three levels of government intervention strategies: economic sanctions, passive protection from hazard, and control of promotion. Beales (1991) regards advertising control as the most common form of government intervention as it is the most visible aspect of protection. Finally, Calfee and Ringold (1994) identified a consistent level of demand for increased advertising controls. From this research Hypothesis 2 states that there will be a demand for greater legislative control of the advertising and other promotional activities of high perceived risk products. To test this relationship, t-tests were conducted between the product's perceived risks (Section 4 of the survey), and the specific demand for advertising and promotional controls (Section 7 of the survey). The results are discussed in Chapter 4. Only tobacco, (.019 at .05), had a significant relationship between perceived risk, and desire for advertising controls indicating that a relationship exists between tobacco and demand for advertising controls. Similarly, only alcohol, (.014 at .05), had a significant relationship between perceived risk and desire for controls on product promotion. As with tobacco, this indicates the existence of a relationship between alcohol and demand for promotion controls.

These results indicate the existence of the relationship between perception of risk, and the specific demands for each product. The absence of corresponding relationships between controls on advertising for tobacco and restrictions on promotion for alcohol reduces the general applicability of the hypothesis. Further, the absence of any relationship between fried food and advertising controls invalidates the hypothesis. Consequently, the hypothesis is only partially accepted.

This finding was not anticipated from the literature. However, upon examining Question 8, Section 3, 'it is not the government's role to control the advertising of products with health risk', the expectations of the literature were upheld.

Restating the question as 'the

government's role is to control the advertising of products with health risks', 65.2% of respondents supported the statement. The support for government control over advertising is consistent with Calfee and Ringold (1994). From the mixed results of Hypothesis 2, the research proposes that a relationship exists between perceived risk and demand for controls over promotion. However, further research beyond the scope of this research is required to test individual components of the promotional mix over a wider range of high risk products. Testing a wider range of products, and elements of the promotional mix may determine whether support exists for controlling promotion in general, or only specific components such as advertising. Further, Jeffrey (1989) and Beales (1991) state that advertising controls are a preferred method of government controls as they are a 'visible' method of control. It is possible, but uncertain from the research, that the preferred method of government controls may not be reflective of the consumer demand. This is proposed from the support given to government testing programs (75.6% strongly agree/agree), minimum health (83.5% strongly agree/ agree) and quality (89.0% strongly agree/agree) standards.

Consequently, further

research should examine consumer preferred methods of government controls, against the government's preferred methods of controls.

H3:

Hypothesis 3

There will be no significant difference in legislative demand by gender

Two significant findings have arisen from Hypothesis 3. First, for the majority of the products tested, Barksdale and Darden's (1972) assertion that there will be no significant difference in legislative demand has held true. Second, the difference in legislative demand for tobacco between male and female respondents is a new finding, and was not predicted by any of the research reviewed.

Hypothesis 3 was tested in two parts: a replication of Barksdale and Darden's (1972) research questions and an analysis of gender difference in legislative demand. There were no significant differences by gender on the three replication questions. In analysing legislative demand by gender, eight of the nine products had no significant difference. Tobacco recorded a significant difference in legislative demand by gender. The reasons for this are unknown, and not anticipated by the literature. On balance however, the hypothesis can be shown to have held for the three replication questions from Barksdale and Darden's (1972) research. The most significant aspect is the consistency of consumer attitudes across time and culture. Further, as the original study, and subsequent supporting work was conducted in America this finding indicates a possible consistency of consumer behaviour between Australian and American society, in terms of gender differences in legislative demand.

Whilst the research identified a difference in the legislative demand between gender, it does not indicate why this difference exists. Interpreting and investigating such a difference in legislative demand is beyond the scope of this research as this result had not been predicted by the literature. The resultant conclusion from this is that there is a need for further research into this apparent difference in legislative demand, and risk processing between males and females concerning tobacco products. This will be discussed in detail below.

H4:

Hypothesis 4

There will be significant difference in legislative demand by age

Hypothesis 4 was developed from Barksdale and Darden's (1972) study in which age was identified as being associated with consumer opinions. Whilst included in the hypotheses as an adjunct to Hypothesis 3, it was not expected to be validly tested. This is addressed in the research's limitations because of the expected homogeneity of respondent ages. Tonn et al. (1990) acknowledge that expectations of difference in risk perception and legislative demand by age are 'common sense', their research had similar limitations in variance of age.

Examining the demographics of the current research, the sample has an age range of between 18 and 46, with a mean age of 25 (SD 7 years). Further, the mode of the sample is 23, clumping the majority of the respondents in a five year bracket between 18 and 23. Consequently, as Tonn et al. (1990) stated, although the expectation of difference by age maybe 'common sense', it is unlikely to materialise, given the limited age range of the sample.

Having established these limitations, the absence of significant relationships between age and legislative demand was not unexpected. Only tobacco (.066) approached significance, with demand for the other products having no demonstrable relationship to age. There was, however, a relationship between age and responses to question 12, with a significance of .019, at .05 significance.

Question 12 examined whether the government should take more

responsibility for regulating the advertising, sales and marketing activities of manufacturers. Given the limitations imposed above, and the limited distribution of age, this finding does not actively prove the fourth hypothesis, but rather indicates that the relationship may exist. This indicates that there is a need for further testing of the hypothesis, using a sample with a larger, more evenly distributed spread of ages.

Replication Studies

The research conducted four replication studies in the questionnaire, from Mazis, Morris and Swasy (1992), Calfee and Ringold, (1994), Barksdale and Darden (1972) and Barksdale and Perreault et al. (1982). These replication studies were conducted to enable comparisons between the current study and the relevant historical data. Overall the results from the replication studies were mixed, with some areas supporting the original findings and other questions bearing no resemblance.

Mazis et al. (1992) : General perceived risk of a series of products.

The first replication study was a partial replication of Mazis et al. (1992) telephone survey which evaluated the perceived risk of a series of consumer products. This research used Mazis et al.'s (1992) question design to examine the perceived risks to the consumers health generally from the use of the nine products. This was only a partial replication, as only three of Mazis et al.'s (1992) products, alcohol, tobacco and dairy products, were reviewed in the current research. Table 4.5. Chapter 4. sets out the comparison of the responses for 'very harmful', and Table 4.6., Chapter 4 sets out the combined percentages rating alcohol and tobacco as 'somewhat harmful' or very harmful. In comparison, the sample had a higher overall perceived risk of the likelihood of harm from using tobacco, but a substantially lower perceived risk for alcohol. Mazis et al. (1992) survey was conducted to examine the effectiveness of recently implemented alcohol warning labels, and to determine whether these new labels had caused an increase in the perceived risk of the product. The recency of the warnings in Mazis et al.'s sample may be a contributing factor to the difference in level of perceived risk of alcohol. Perceptions of health risk from tobacco use were comparable.

Calfee and Ringold (1994): Advertising Sales and Controls

12 The government should exercise more responsibility for regulating advertising, sales and marketing

Calfee and Ringold's (1994) replication question in Section 3 examined whether the government should 'exercise more responsibility for regulating advertising, sales and marketing'. This question has also been replicated in a slightly altered form in the Barksdale and Darden (1972) replication study below.

In combining Calfee and Ringold's (1994)

overview of demand for government regulation from 1971 to 1979 with Barksdale Perreault et al. (1982) results, the average level of support (agree/strongly agree) was 59%. In the current

research, 58% of responses to question 2 were agree or strongly agree (Table 4.4). From this it can reasonably be concluded that the support for increased government responsibility for advertising sales and marketing is consistent with previous research. Further, this indicates that the demand for increased government controls is independent of cultural differences, having been constant over time and nationality.

Barksdale and Darden (1972) and Barksdale, Perreault et al. (1982). Consumer Attitudes to Government Regulations 10 The government should test competing brands of products and make the results of the tests available to the consumer 11 The government should set minimum standards of quality for all products sold to consumers 12 The government should exercise more responsibility for regulating advertising, sales and marketing.

Barksdale and Darden's (1972) study was replicated by Barksdale, Perreault et al. (1982) in an international replication study. For the purposes of this chapter only the findings from the Australian replication will be considered, along with the original study. Table 4.11, Chapter 4, sets out the comparative results from this research, and the two prior studies. The major findings from the replication studies has been the consistent increases in support for the government to set minimum quality standards (+15%).

In contrast support for

increased government controls over the advertising sales and promotion of consumer products has declined (-17.4%) over the same period. Support for product testing had a slight overall increase, despite declining in 1982 survey (+4.2% strongly agree/agree).

The decline in support for advertising controls, evidenced in both Hypothesis 2, and question 11 was not expected by the literature. Calfee and Ringold (1994) indicated that there had been a consistent level of demand for increased advertising regulation. However, this research, combined with the Barksdale studies (1972, 1982), indicates that the overall demand for controls on marketing is decreasing. At the same time, demand for alternative methods of controls is increasing, particularly the demand for the government to set minimum quality

standards for all products, and for government testing of competing brands. The rise in alternative approaches to government control may also be the cause of the decrease in demand for marketing regulation. The thesis proposes that whilst legislative demand for controls on advertising sales and marketing per se may be declining, that the overall level of legislative demand is constant. This is evidenced by increases in support for the alternative methods of testing and minimum standards. Further, this research indicates that support exists for advertising controls of products with health risks (Question 8, Table 4.4 Chapter 4). The role of government as the protector of the consumer was also supported (59% strongly agree/agree).

These findings demonstrate that legislative demand for advertising controls and government 'protection' of the consumer is still high. One possible explanation for the decline in support for the regulation of 'advertising sales and marketing activities' is the multifaceted nature of the question. As expressed in the methodology, one of the limitations of the question is that it would be better asked as three separate questions. Given the limitations of the research, legislative demand for advertising controls, the other sections of marketing and sales were not examined separately. Consequently, it is possible that decline in support for either form of regulation could have caused the decrease in the overall support. However this explanation will require further testing of levels of support for regulation of sales and marketing for verification. Such testing is not within the parameters of this research.

In summary, the research was an exploratory study to determine if a relationship existed between perceived risk and legislative demand. Whilst the study has several limitations, the proposed relationship between perceived risk and legislative demand does exist at least for tobacco and alcohol within the sample population. Further testing of the concept will determine if the relationship occurs with other products or activities. The research also validates Slovic's (1980) 'dread risk' statement that the greater the perceived risk, the greater the level of government intervention sought by the consumer.

Finally, Barksdale and

Darden's (1972) statement that legislative demand was independent of gender was generally

applicable to an Australian sample. Finally, the research revealed an unexpected relationship between gender and demand for increased controls over tobacco. The implications of these findings for public policy, risk perception, risk management and marketing are discussed below.

Implications for Theory

The research spans the fields of public policy, risk management, risk perception and marketing and as a result, this section will examine the implications for each of the fields separately.

Public Policy

The major implication for the public policy research field is the establishment of relationship between perceived risk and consumer demand for government intervention. This relationship had previously been implied by Kasperson Renn et al. (1988) and Kasperson (1992) in the Social Amplification of Risk model, which is detailed in Chapter 2. In essence, Kasperson et al. (1988) identify political pressures as part of the 'ripple effect', which leads to the possibility of regulatory constraints. Legislative demand, as evidenced by Hypotheses 1 and 1A, supports Kasperson et al.'s (1988) assertion of a link between perceived risk (interpretation and response) and possibility of regulatory constraints through political pressures (legislative demand).

Further, the existence of a relationship between perceived risk and demand for government intervention can be used by public sector policy analysts in regard to judging voter expectations. Should the relationship between risk and legislative demand remain constant, policy analysts can predict voter expectations of government intervention in scenarios where voters perceive the risk from an event or product to be unacceptable. This has potential to range from assessing voter expectations of government response to 'dreaded risks' such as

nuclear power to domestic issues of traffic safety. In essence, the practical application of legislative demand will be to predict the appropriate role of government from level of voter perceived risk. Application of the concept will also assist government to determine the expectations of voters in issues which are assessed by voters on instinct or 'gut reactions' rather than thought about rationally and calmly (Slovic, 1980).

Risk Management and Risk Perception

This dissertation ties together research from Barksdale and Darden (1972), Barksdale et al. (1982) Calfee and Ringold (1994), Kasperson et al. (1988), Kasperson (1992), Patterson et al. (1992), and Slovic et al. (1980), in establishing a link between risk perception and demand for government intervention. The research findings support assertions made by Slovic et al. (1980) as to the existence of a relationship between 'dread risk' and demand for government intervention. Also, the research reinforces Patterson et al.'s (1992) observation of the propensity of society to seek government intervention when it identifies potential risks.

Further, two models have been established from the literature reviewed, risk processing chain and the risk response model. As a result of the prior research of Slovic et al. (1980), Bettman et al. (1986), Tonn et al. (1990) and Jeffrey (1989), figure 2.1, Chapter 2, illustrates the simplified risk processing chain. An expanded version containing full heuristics and biases is contained in Appendix 2.1. Using the combined research, the model demonstrates the progression from the risk event through to legislative intervention, including the heuristics and biases that may impact upon cognitive processing of the risk information.

The second model arises as a subset of the Risk Processing Chain, the research established the risk response model (Chapter 2, Figure 2.2). The risk response model draws upon the risk perception research, and particularly Slovic's (1980) 'dread risk' concept in conjunction with the legislative demand concept.

The major feature of the model is to demonstrate the

progression from risk perception through to the risk response. Where the model differs from

prior research is in its distinction between potential risk responses from three levels of perceived risk. In Slovic et al.'s (1980) 'dread risk' model, the stated relationship is between high 'dread risk' and legislative demand. In the risk response model, legislative demand can arise either directly from high perceived risk or through the failure of personal risk management to reduce the risk to an acceptable level.

Marketing

The primary focus of the research was to examine the relationship between consumer perception of risk, and demand for controls over the marketing of perceived high risk products. The major outcome of the existence of this relationship is the need for marketing and marketers to control consumer perceptions of their products. One of the major implications of Kasperson's model (1992) is the 'rippling effect' which leads to legislative demand. Kasperson et al. (1988) outlined the effect of an incident involving the contamination of a batch of prestige label Swiss cheese where the 'rippling effect' extended from the specific brand of cheese to the entire industry. Further, a similar event occurred in Australia in 1995, with the Garabaldi Smallgoods food poisoning spreading from the individual company to the processed meats industry. The 'ripple effect' alone poses problems for marketing in terms of damage control and product perceptions. In combination with legislative demand, where demand for government regulation of the marketing of an entire industry may stem from the misfortune of a single company, the ripple effect poses a serious challenge for marketing.

The significance of the research is in identifying perceived risk as a component of legislative demand. This allows marketers to 'monitor' consumer perceptions of their product to be able to engage damage control campaigns prior to consumer risk perceptions reaching the level that instigates legislative demand. This is particularly valuable in industry which has perceived risk levels higher than the actual risk involved in the industry, such as nuclear power supplies. Further, legislative demand combined with Kasperson's (1992) 'rippling effect' recognises that general community perceptions of product risk must be examined,

rather than those in the producer's target markets. For example, legislative demand for controls over the sport of boxing stem from groups such as medical associations, rather than boxing's target market of sports fans. In this example, the target market will have a lower perceived risk of the sport, due to the heuristics of availability and overconfidence (see Appendix 2.1). In contrast, the groups outside of the target market, which will include those groups seeking government controls over the sport, will have a higher level of perceived risk. Consequently, marketing campaigns aimed to reduce the perceived risk would have to be focused on the groups with higher perceived risks of the product.

Limitations

There are two key limitations to the research; age distribution, and the sample. First, due to the expected homogeneity of ages in the sample, Hypothesis 4 was unable to be widely tested. Second, the members of the sample are not representative of the population in general. Although several of the replication studies had results comparable with research conducted on more representative samples, the general applicability of the research is limited. However, as this is exploratory research to determine if a relationship exists between legislative demand and risk perception, it was not intended to be generalised to society. Although the sample is nonrepresentative, it was sufficient for the purposes of this research to demonstrate the existence of the relationship. This relationship can be examined further using a more representative sample. Opportunities for further studies from this research are outlined below.

Further Research

The major area of further study is replication of this research using a larger range of products, activities and events with high perceived risks, with a variety of samples. Further, additional studies are needed to examine the apparent difference between male and female legislative demand for controls over tobacco products. First, these studies are required to test whether the difference is a factor of this sample, or if it is constant across a wider sample.

Second, should the relationship exist in a larger sample, further exploration of the reasons for the differentiation should be conducted. Finally, three areas of retesting have arisen from findings within the research: difference in legislative demand by age; the relationship between consumer preferred methods of control and government preferred methods; and finally a retest of the individual elements of sales marketing and promotional controls. Each of these areas returned results not predicted by the literature, and require further investigation which is beyond the scope of this research.

Appendix 2.1. Expanded risk processing chain

Appendix 3.1. Calfee and Ringold (1994) Frequencies

Calfee and Ringold (1994:234) reviewed consumer attitude surveys collected from 19711987, finding support for regulation ranging from 50 to 80% with an approximate mode of 70%. The table below lists the percentages for 'agree/strongly agree' for increased regulation over a period of 16 years. From these figures variance was calculated for use in the equation to determine sample size in Figure 3. 2. Calfee and Ringold (1994) Desire for More Government Regulation of Advertising, 1971-1987. 1974

79.00

1975

80.00

1976

80.00

1977

78.00

1978

71.00

1979

74.00

1980

71.00

1981

72.00

1982

67.00

1983

71.00

1984

72.00

1985

76.00

1986

67.00

1987

71.00

variance

19.65

mode

73.50

standard dev

4.43

(source: Calfee and Ringold, 1994)

Appendix 3.2. Heart Foundation Food Pyramid

Appendix 3.3. Pilot study questionnaire

Appendix 3.4. Questionnaire

Variable Labels

Appendix 4.1.

Section 1: Product Use Questions PU 1 PU 2 PU 3 PU 4 PU 5 PU 6 PU 7 PU 8 PU 9

tobacco red meats chicken. fried food dairy products alcohol olive oil pork vegetables

Value labels 1 daily 2 3 times a week, 3 weekly, 4 between weekly and monthly 5 monthly 6 less than monthly 7 don't use 99 missing value

Section 2: Heart disease risk of common products HR 1 HR 2 HR 3 HR 4 HR 5 HR 6 HR 7 HR 8 HR 9

tobacco red meats chicken. fried food dairy products alcohol olive oil pork vegetables

Value labels 1 very high risk, 2 high risk, 3 moderate risk, 4 low risk, 5 very low risk, 6 no risk, or 0 don't know 9 Missing value

Section 3: General opinions concerning product regulation PR1

It is the government's responsibility to protect the consumer from products with a known risk of causing heart disease.

PR2

The government should exercise more responsibility for regulating advertising, marketing.

PR3

The government should regulate advertising of products with known heart risks.

PR4

The government should regulate the availability of products with known heart risks.

PR5

Products with known risks of causing heart disease should be more heavily taxed than products without health risks.

PR6

The individual has a right to choose to use consumer products with known health risks.

sales and

PR7

The role of the government is to protect the consumer.

PR8

It is not the government's role to control advertising of products with health risks.

PR9

The government has a responsibility to ban products that are shown to cause heart disease.

PR10

The government should test competing brands of products and make the results of the tests available to the consumer.

PR11

The government should set minimum standards of quality for all products sold to consumers.

PR12

The government should exercise more responsibility for regulating the advertising, sales and marketing activities of manufacturers.

PR13

Consumers should not be able to purchase products that may damage their health.

PR14

There are too many controls on the marketing of products with known heart risks.

PR15

The government should set minimum health standards for all food products.

Value labels 1 strongly agree, 2 agree, 3 unsure, 4 disagree, 5 strongly disagree 0 no opinion. 9 Missing value

Section 4: General Risk of Product Use GR1 GR2 GR3 GR4 GR5 GR6 GR7 GR8 GR9

tobacco red meats chicken. fried food dairy products alcohol olive oil pork vegetables

Value labels 1 very harmful, 2 somewhat harmful, 3 not too harmful, 4 not at all harmful or 0 Don't know. 9 Missing value

Section 5: Levels of Government Control (Legislative Demand) LD1 LD2 LD3 LD4 LD5 LD6 LD7 LD8

tobacco red meats chicken. fried food dairy products alcohol olive oil pork

LD9

vegetables

Value labels 1 GREATLY INCREASE 2 PARTIALLY INCREASE 3 STAY THE SAME 4 PARTIALLY DECREASE 5 GREATLY DECREASE 0 NO CONTROLS 9 Missing value

Section 6: Methods of government control for products with known risks. GC1 GC2 GC3 GC4 GC5 GC6 GC7 GC8 GC9 GC10 GC11 GC12

education programs information campaigns product ban high taxes regulating product design restricted availability restrictions on advertising controls over product promotion minimum age requirement government warning labels no controls other: (please specify):

Section 7: Methods of control for specific products. SC1 SC2 SC3 SC4 SC5 SC6 SC7 SC8 SC9

tobacco red meats chicken. fried food dairy products alcohol olive oil pork vegetables

Value labels GC1 education programs GC2 information campaigns GC3 product ban GC4 high taxes GC5 regulating product design GC6 restricted availability GC7 restrictions on advertising GC8 controls over product promotion GC9 minimum age requirement GC10 government warning labels GC11 no controls GC12 Missing value

Section 8: Demographics AGE 99

Age Missing value

GENDER 1 Female

2 99

Male Missing value