2014

Teun Roks (360676) Jessica Pfeiffer (416452) Florine Van Den Biggelaar (333587) Babette Oostveen (333036)

[RESEARCH REPORT]
Can the decoy effect influence consumers to make healthier decisions?

Research Report 2014

Table of Contents Abstract ................................................................................................... 3 Introduction .............................................................................................. 4 Theoretical framework ................................................................................ 6 The decoy effect ..................................................................................... 6 Practical example .................................................................................... 7 The decoy effect in health decisions........................................................... 8 Methodology ............................................................................................10 Description and design experiment ...........................................................10 Procedure of the experiment and the survey ..............................................11 Results ....................................................................................................13 General information ...............................................................................13 Effect of manipulation .............................................................................13 Effect of gender .....................................................................................14 Discussion ...............................................................................................16 Implications .............................................................................................18 Limitations & Further research ....................................................................19 Limitations ............................................................................................19 Further research ....................................................................................19 Literature ................................................................................................21 Appendices ..............................................................................................24

2

Research Report 2014

Abstract The aim of this research is to investigate if the decoy effect can nudge people into making healthier decisions to combat obesity in today’s society. More specifically if this can influence consumers to choose a nonalcoholic beer instead of a regular alcoholic beer. We conducted an online experiment to test our hypothesis. The results reveal that the decoy effect on health decisions is ineffective for the population as a whole. However, our results show that a decoy has a significant effect on females, as they prefer the non-alcoholic target more in case a decoy is added to the choice set. For males this effect is not significant. The outcome of this research can be applied for establishments that want to focus more on the health of their customers.

3

Research Report 2014

Introduction
Obesity is a worldwide problem that affects over 1.4 billion adults (WHO, 2014). Despite the health risks that coincide with obesity, healthy lifestyles are not widely adopted yet. As a result, the rapid growing size of the problem causes obesity to be an epidemic. As obesity has a high impact on the rising healthcare spending (Thorpe et al., 2004), it is crucial that people are influenced to make healthier choices. The causes for being overweight or obese are abundant. Psychological factors such as anxiety or loneliness may affect food consumption, as well as the social and physical environment (Collins et al., 2009). For instance, Finkelstein et al. (2005) argued that the costs of not gaining weight are high in today’s society, due to the fact that information is not fully provided, which makes healthy choices difficult. Many people are not aware of the calories contained in alcoholic drinks and the fact that alcohol consumption can lead to an increase in food intake (NHO, 2012). A healthier alternative is the non-alcoholic beverage that contains less than half of the calories of alcoholic drinks (Voedingscentrum, 2014). Unfortunately, the consumption of non-alcoholic beverages is still far below the consumption of alcoholic drinks (Nationaal Bieronderzoek, 2014). Behavioural economics can be a specifically useful mean for looking into the problem, as it combines economic theories with a psychological perspective. With these insights, policymakers and managers can formulate problems and choice options in such a way that consumers are nudged into making a more desirable decision. Several of these behavioural insights could therefore also aid in solving the problem of obesity. In this research paper, we will focus on one theory: the decoy effect, which will be explained in the next chapter. We will investigate whether the decoy effect can nudge people into making a healthier decision, namely a non-alcoholic beer instead of a regular beer containing 5 percent alcohol. This gives rise to the following research question: Can the decoy effect influence consumers to make healthier decisions? The information that is derived from the outcome of the research question has both social and academic relevance. The decoy effect influencing healthy decision-making can be used by institutions such as the Ministry of Health when

4

Research Report 2014

constructing their policy in the battle against obesity. Commercial businesses can also play their part in this issue when constructing their product portfolio. Scientifically, the decoy effect has proven to be effective in various fields; not only within a range of tangible products, but also within elections and policies (Huber et al., 1982; Vedantam, 2007; Bateman et al., 2008). However, this effect has not yet been measured within the decision between an unhealthy product and a healthy alternative. This paper adds onto the theory about the decoy effect and shows that also choices for hedonic products such as alcoholic beer can be influenced by this phenomenon. In the next chapter, the decoy effect will be thoroughly explained. Secondly, the methodology of the experiment conducted to investigate this effect is described, followed by the results that were obtained. In the last section we will pay attention to the conclusions from these results and the implications of this information. The limitations of this study and suggestions for further research can also be found in this section.

5

Research Report 2014

Theoretical framework
The decoy effect
The proportionality model of Luce (1959) assumes that a new market offering will attract decision makers from their initial choices in proportion to the original shares of the existing items. This model suggests that any item introduced to the choice set will attract customers, as it expands the range of available choices. Another view on this phenomenon is stated by the similarity hypothesis (Tversky 1972). This hypothesis proposes that customers defect to items similar to existing offerings. According to this theory, a new item introduced within a choice set, will take disproportionately more share from those similar to it than those dissimilar. Both the proportionality model and similarity hypothesis have the assumption of regularity, which is also embodied in standard models of rational choice (Tversky and Simonson, 1993). The introduction of a new alternative cannot increase the probability of choosing a member of the original set. However, Huber et al. (1982) found that adding an asymmetrically dominated alternative, a decoy, can violate this regularity. The decoy effect explains the effects on choices between two options, the target and the competitor that results from the introduction of an additional option: the decoy. When the addition of such a decoy increases the choice share of the target, then an asymmetric dominance effect is said to have occurred (Huber et al., 1982). In asymmetrically dominated choices at least two options do not dominate each other and one (not both) of those options dominates a third option. In this explanation, we will take two dimensions into account, represented on the x-axis and the y-axis in figure 1 (figure copied from Huber et al., 1982). Examples of these dimensions are quality, price or taste. In the original choice set, only two items are present: the target and
Figure
1

the competitor. As can be seen, the competitor dominates the target on one dimension, but the target dominates the competitor on the other dimension. To

6

Research Report 2014

increase the choice share of the target, one can introduce a decoy into the original choice set that meets the requirements from the shaded area. The decoy must be inferior to the target on both dimensions, but has to be inferior to the competitor on only one dimension. Therefore, the decoy is also referred to as the asymmetrically dominated choice option. The decoy is added to the choice set to increase the choice share of the target, not to generate direct sales. Perceptual effects and decision-making processes are motivating factors for explaining the decoy effect. By adding a dominated third option, more emphasis is put on the unfavourable dimension of the target than on the favourable dimension. This results in a smaller deficit of the target in the unfavourable dimension. When more options fall in the dimension on which the target is superior, the perception on that dimension as most important may be enhanced (Huber et al., 1982). Consumers want to justify their choices when facing uncertainty in the decision-making process, in particular if they feel externally evaluated (Simonson, 1989). When facing uncertainty, the target may become more attractive through its superiority being unambiguous or as a compromise that combines the desirable attributes of the other choices (Bateman et al., 2008).

Practical example
Doyle et al. (1999) conducted a real world investigation regarding purchases of cans of baked beans in a supermarket. The researchers monitored weekly sales of two brands of beans, in cans of equivalent size. The first brand of beans, which we will call brand X, was cheaper than the leading brand but accounted for only 19 percent of the sales. The other brand of beans, brand Y, was market leader on the market for canned beans and more expensive than brand X, but accounted for the other 81 percent of the sales in the supermarket. Brand Y is in this case our competitor, while brand X is the target. After the initial monitoring, the researchers introduced a decoy item. The decoy was a smaller can of brand X, but sold at the same price as a large can of brand X. The week after, sales were monitored again. The results showed that no purchases of the decoy were made, but the sales of the target (the original can of beans of brand X) increased from 19 percent to 33 percent. This left leading brand Y with declining sales: from 81 percent without the decoy to 67 percent

7

Research Report 2014

after the decoy was introduced. The shift in purchased item from competitor to target by introducing a third item is the decoy effect at work.

The decoy effect in health decisions
The decoy effect as described above has been proven not only within a range of tangible product such as canned beans. Highhouse (1992) constructed an experiment within job candidates, where the addition of a decoy job applicant increased the preference for the targeted job candidate. Bateman et al. (2008) proved that the decoy effect can also be seen when a choice set involves policies. These experiments show significant proof for the effect of a decoy within a choice set. However, this effect is yet to be measured within healthy decisions. When applying this theory to health decisions, the target and the competing item have to be constructed carefully. Health policy tries to encourage consumers to make healthier decisions. The decoy effect is proven to shift preferences from competitor to the target, so in the case of health decisions the target has to be the healthier option. This experiment refers to non-alcoholic beer (containing 0.1 percent alcohol) as the healthier option, compared to alcoholic beer (containing 5 percent alcohol). Alcohol is the first dimension respondents can state their preference on. More than 80 percent of the Dutch population drinks a regular beer once a month whereas only 17 percent drinks non-alcoholic beer (Nationaal Bieronderzoek, 2014). The second dimension is price, as people prefer low prices to high prices. We will only take these two dimensions into account: volume, brand or quality of the beer will not be mentioned. In order to make the experiment work in line with the theory, the competing product has to dominate on one dimension, while the target product has to dominate on the other dimension. In the case of this experiment, the alcoholic beer will of course dominate on the alcohol preference dimension (y-axis). As a result, the non-alcoholic beer should dominate on the price dimension (x-axis), having a lower price than the alcoholic beer. This leaves a clear range for the decoy to be identified as such. The decoy has to dominate the competitor on one dimension, but cannot dominate the target. As a negative alcoholic percentage is not possible, the decoy therefore has to contain the same alcohol percentage as the target (0.1 percent). On the price dimension, the decoy has to dominate the competitor, thus has to be cheaper as the set price for the alcoholic beer. At the same time, the decoy has to be less attractive as the target

8

Research Report 2014

on this dimension, thus needs to be at a higher price than the non-alcoholic target beer. This is explained in figure 2.

Figure 2

With this information, we can construct the hypothesis that the experiment tests: H1: The addition of a non-alcoholic decoy to the choice set significantly increases the preference for the non-alcoholic target. Research shows that male beer drinkers more often than female beer drinkers, find non-alcoholic beer not tasty (Nationaal Bieronderzoek, 2014). Thompson and Thompson (1996) posited that low-alcohol beverages are more appealing to females than males. As gender experiences non-alcoholic beer differently we expect that the decoy effect in a non-alcoholic versus alcoholic setting is larger for females than males. Thus the following hypothesis is posited: H2: The addition of a non-alcoholic decoy influences the preferences of females more than it does for males.

9

Research Report 2014

Methodology

Description and design experiment

We choose for an artificial field experiment designed to test the influence of a decoy on the choice set of consumers. Due to lack of a suitable test environment - a restaurant or pub - we conduct an online survey. This has the benefit of obtaining a substantial number of respondents in a relatively short period of time. However, this makes the respondents aware that they are participating in an experiment. Furthermore, we chose for a between-subjects design of the survey, so we use one group for the choice set without the decoy and one group for the choice set with the decoy. This prevents any effects of priming that could bias the respondents. Also the decoy effect is often stronger in between-design experiment than in a within-design (Hayes et al., 2011). To make the experiment more realistic we shortly outline the environment in which the consumer is most likely to consume a beer. Moreover, we used images of beer to appeal to the imagination of the consumer. The more the situation is perceived to be realistic, the stronger the decoy effect is expected to be (Josiam and Hobson, 1995). The choice set does not include actual brand names as this can exert distinct influences on cognitive processes underlying consumer judgements and choices (Kim et al, 2006). Instead we chose to use the characters X, K, and C to represent the different brands. Huber et al. (1982), have formulated four strategies for the placement of a decoy within the choice set. In this experiment, the decoy can only deviate in the price dimension, because an alcohol percentage below 0 percent is not possible. The only strategy left to place the decoy is increasing the frequency of the dimension on which the target is superior. In the case of this experiment, the target is superior on the price dimension. An increase of the frequency in this dimension is said to result in an increase of the attention paid to this dimension (Huber et al., 1982). As shown in figure 2, the price of the decoy can be positioned within €2.5 (the price of the targeted non-alcoholic beer) and €5.- (the price of the competing alcoholic beer). According to Parducci (1974), adding alternatives within the range of others tends to spread out their distances on subjective category ratings. The price of the decoy will therefore be fixed at €4.5. Cutting the €2.5 price advantage between the target and the competitor in two parts

10

Research Report 2014

increases the subjective perception of this advantage, causing the €2.5 advantage to appear larger than when this would be presented as a whole. Furthermore, our design includes a question that could provide us explanations when any deviations might occur. We ask the respondents their preference for non-alcoholic beer in general, after they chose non-alcoholic beer in the choice set, in order to avoid a priming effect. As Iyengar and Kinder (1987) researched, the order questions appear in a survey can directly impact the responses that will be given.

Procedure of the experiment and the survey
We collect data through the online panel offered by the market research agency Metrixlab. Through this panel we survey 304 Dutch respondents of the age of 18 to 64. We target a diverse group, as we want to project preferences on alcohol and non-alcohol beer realistically. We explicitly do not target only students, as for them the use of alcoholic beverages is relatively high and thus the targeted nonalcoholic beer is not of any interest (Garretsen et al., 2008). The experiment consists of two different groups of equal size; each allocated to different questionnaires, so to say the treatment and control group. The control group receives the survey without the decoy option, to test the general preference between alcoholic and non-alcoholic beer. In the survey for the treatment group a third option is added to the choice set, the decoy, to see whether the preference for the targeted option can be increased. Both surveys start with a question whether the respondent drinks beer. For the respondents that answers negatively, the survey ends directly. When respondents answer positively, the next page asks the gender and age of the respondent (appendix A). For the control group the next page displays two options (random sequence): 1. 2. A non-alcoholic beer of brand K €2.50 A alcoholic beer of brand X of €5.-

On the second page of the survey we ask the respondents why they chose for the selected option. The treatment group has another choice set to choose from on the page after the gender and age question (appendix B). For the control group the second page displays three options (random sequence):

11

Research Report 2014

1. 2. 3.

A non-alcoholic beer of brand K of €2.50 A alcoholic beer of brand X of €5.A non-alcoholic beer of brand C of €4.50

On the third page of the survey we ask the respondents why they chose for the selected option. This option was especially useful after conducting a pilot study. Forty-seven percent of the respondents stated to dislike beer, which influenced their decision-making and increased their preference for non-alcoholic beer relatively. This information led to the addition of the first question to the survey, so respondents that do not drink beer would be excluded. In this way, data would not be distorted. The respondents that choose the decoy option will be excluded from the data.

12

Research Report 2014

Results
General information
In total 304 respondents participated in the survey. 152 participants filled in the treatment survey and another 152 respondents filled in the control survey. No participants were deleted from the final sample, as none of the respondents from the treatment survey chose the decoy. The age of the respondents ranges from 18 to 64 years with an average age of 43 years old (appendix C). Fifty-one percent of the 152 respondents in the control are male and forty-nine percent is female. In the treatment group this distribution is respectively fifty-three percent versus forty-seven percent (appendix D).

Distribution of gender between groups
51% 49% 53% 47% Male Female

Controlgroup

Treatmentgroup

Effect of manipulation
Research revealed that in the control group thirty-three percent of the respondents chose the targeted non-alcoholic beer (zero percent chose the decoy and sixty-seven percent chose the competition, appendix E). In the treatment group thirty-nine percent preferred the target (zero percent chose the decoy and sixty-one chose the competition). The difference between the mean of the treatment group and the control group is not significant (p=.283, appendix F). This means that for the total sample of respondents adding a decoy to the respondent’s choice set does not increase preference for the target; the nonalcoholic beer. Therefore the first hypothesis, ‘the addition of a non-alcoholic decoy to the choice set significantly increases the preference for the non-alcoholic target’, is rejected on a five percent significance level.

13

Research Report 2014

67%

Chosen option per group
61% Controlgroup (target vs competitor) 33% 39% Treatment (target vs competitor vs decoy)

Competitor - €5 Beer (5% alcohol)

Target - €2,50 Alcohol-free Beer

Effect of gender
The table below shows that females seem to prefer the non-alcohol beer more often than males. In total forty-three percent of the females chose the target option. Of the males only twenty-nine percent preferred the non-alcoholic beer (appendix G).

71%

Chosen option per gender (overall results)
57% 43% 29% Male Female

Competitor - €5 Beer (5% alcohol)

Target - €2,50 Alcohol-free Beer

Our research shows that for females there is a significant effect of adding a decoy to the choice-set of the respondents (p=.041, appendix H, appendix I). In the treatment group fifty-one percent of the females preferred the target whereas in the control group this was only thirty-five percent.

14

Research Report 2014

Chosen option per group (females only)
65% 49% 35% 51% Controlgroup Treatmentgroup

Competitor - €5 Beer (5% alcohol) Target - €2,50 Alcohol-free Beer

The male part of the respondents chose less for the targeted non-alcoholic beer when a decoy was added. In the control group thirty-one percent of the males preferred the target whereas in the treatment group this was only twenty-eight percent. However, this was not significant, implying that for males the decoy effect was not able to significantly influence preference (p=.617, appendix J, appendix K).

Chosen option per group (males only)
69% 73%

Controlgroup 31% 28% Treatmentgroup

Competitor - €5 Beer (5% alcohol) Target - €2,50 Alcohol-free Beer

With this information we can confirm the second hypothesis, ‘The addition of a non-alcoholic decoy influences the preferences of females more than it does for males’, on a five percent significance level. The answers given in the open questions do not provide any particular explanation for the chosen options. Furthermore, they do not deviate from the expected motivation as price, the amount of alcohol or the situation in which the beverage is consumed.

15

Research Report 2014

Discussion
The results of the survey lead to the discussion of several issues. Our research reveals that the decoy effect on health decisions is ineffective for the population as a whole. However, our results show that a decoy has a significant effect on females, as they prefer the non-alcoholic target more in case a decoy is added to the choice set. For males this effect is not significant. Current academic research provides no empirical evidence that females are more susceptible to the decoy effect in general. Although our data does not enable us to pinpoint the motivation behind this behaviour directly, we do argue that these results follow certain logic. There is evidence that argues that non-alcoholic beer is less likely to fall within the consideration set of males compared to females. Therefore females are more susceptible to the decoy effect in this experiment as females consider the decoy whereas males do not. Males drink higher quantity of beer than females and also drink beer more frequently (Nationaal bieronderzoek 2014). For most males beer is their number one choice beverage whereas females drink it on occasion. This implies that males have a higher preference for alcoholic beer and therefore consider nonalcoholic beer not as much as females do. Research shows that females prefer low-alcoholic beverages more than males do (Thompson and Thompson, 1996). The most important motives behind the consumption of low-alcoholic beverages are low calorie content and low alcohol content. The first motive helps to manage the consumer its weight and the latter prevents people from getting drunk. The most important driver behind the preference of low-alcohol beverages is low calorie content, because consumers associate it with being healthier (Chrysochou, 2013). Females have a stronger belief in the importance of healthy consumption than males. Also, they have a higher involvement in their weight control, and are more likely to avoid high-fat and high calorie products (Wardle et al., 2004). Chrysochou (2013) also concluded that low-alcohol beverages are considered more as a healthier alternative for alcoholic beverages and less as a substitute. Our data seems to confirm this as Warlde et al. (2004) show that females care more about weight management and therefore see non-alcoholic beer as a

16

Research Report 2014

substitute. Males on the other hand are not keen on weight management and do not consider non-alcoholic beer as a substitute. Females are more willing to give up certain aspect of the beverage for the right incentive. In this case females are more willing to give up alcohol for a lower price. Moreover, they are more susceptible to the decoy effect as alcoholcontaining beer is less valuable to them. Males on the other hand are not that easily manipulated by a non-alcoholic decoy as these beverages are not even in their consideration set. In advertisements regular beer is associated with masculinity (Strate, 1992). Thus, it is very likely that the social desirability bias is at play and has an influence on the choices of the male participants. Males tend to act according to social desirable behaviour and as a result only consider the alcoholic beer. In summary, we investigated if the decoy effect can nudge people into making healthier decisions. More specifically, whether this phenomenon can influence consumers to choose a non-alcoholic beer instead of a regular alcoholic beer. This study provides empirical evidence that females can be influenced through the decoy effect to make healthier decision, since they are more likely to consider non-alcoholic beer. This effect is not found for males.

17

Research Report 2014

Implications
As stated in the introduction of this paper, obesity is a worldwide epidemic, caused by various reasons. Because alcohol consumption is often associated with obesity, our results can help institutions such as the Ministry of Health construct policies. When designed properly, the decoy effect can influence health-based decisions in a positive way and can be implemented in the hospitality industry or supermarkets. As the results are only significant for females but not for males, the approach should be gender specific. To increase the health of females, one could want to stimulate drinking non-alcoholic beer instead of regular beer. This is interesting in the hospitality industry, especially for establishments that want to focus more on the health of their customers such as sports canteens and spa restaurants. Establishments would add a product to their offer that will not generate sales directly, but serve as a decoy. As the Ministry of Health has an educational role, they can provide businesses with information on how to add a decoy into the menu. Males will be more difficult to target, as there is no significant effect of adding a decoy in decisions on alcoholic versus non-alcoholic beer. Adding a decoy to menus of establishments with mostly male customers, such as sports bars, would be ineffective. However, to increase men’s health, it is still possible to use the decoy effect measured in this paper. As over 74% women do the grocery shopping for their households, this could be a way to influence the health of men (Te Pas, 2010). When a decoy non-alcoholic beer is added to the offerings of beer, our results show that women are more easily tempted to choose the nonalcoholic target instead of the alcoholic beer. This leads to a healthier consumption concerning beer for males. With these implementations, the decoy effect can eventually contribute in solving the problem of obesity.

18

Research Report 2014

Limitations & Further research
Limitations

Our research is limited as we were not able to conduct the experiment in a more real life setting i.e. with the help of a restaurant or pub where we could manipulate the menu. Conducting the experiment in a real life setting will increase the external validity of our research as it replicates more real life dimensions. Through this design the chosen option directly influences the consumers’ utility. Josiam and Hobson (1995) state in their research that the more the situation is perceived to be realistic, the stronger the decoy effect is expected to be. We therefore do not preclude that this unrealistic setting affects our results. Another limitation of this study is that the effect is only measured on two dimensions. We deliberately chose for two dimensions, to simplify the task for the respondents and to demonstrate the decoy effect in a clear way. However, this simplification reduces the applicability of the results in practice. In real life, preferences can come from various dimensions. This is not limited to price and alcohol percentage, but also quality, volume, taste and brand are possibilities. Finally our survey did not include a scale to measure preference in beer. By measuring preference we could have used these results to support the conclusions in our discussion, as this could be of influence on the final results.

Further research

With the discussion and limitations of this paper, some questions arise that are interesting for further research. To increase the applicability of decoy effects in healthy decisions, we propose to conduct this experiment with the inclusion of more product dimensions. Possible additional dimensions are taste, brand and volume that enhance the realistic setting of the experiment. Furthermore, it is of social and scientific use if the decoy effect is measured in other health based decisions. An example is to redo the experiment with the choice to consume an apple (healthy choice) or to consume a chocolate bar (unhealthy choice). The results of such an experiment will be especially helpful in constructing health policies by for instance the Ministry of Health.

19

Research Report 2014

As our results show a difference between males and females, the question arises whether there is a difference in the perception of a decoy based on gender. Various aspects of behavioural economic theory have confirmed dissimilarity between the perceptions of males and females. However, with respect to the decoy effect literature is still missing. Last, our research suggests that females consider non-alcoholic beer as a substitution for alcoholic beer whereas males do not. Further research should provide more empirical evidence for this finding and see whether this trade-off exists between other alcoholic and non-alcoholic beverages.

20

Research Report 2014

Literature
Bateman, I. J., Munro, A., & Poe, G. L. (2008). Decoy effects in choice experiments and contingent valuation: asymmetric dominance. Land Economics, 84(1), 115-127. Chrysochou, P. (2014). Drink to get drunk or stay healthy? Exploring consumers’ perceptions, motives and preferences for light beer. Food Quality and Preference, 31, 156-163. Collins J.C., J.E. Bentz (2009). Behavioural and psychological factors in obesity, Lancaster General Neuropsychology Specialists, vol. 4, no. 4, 124-127. Doyle, J, R,, D, J, O'Connor, G, M, Reynolds, and P, A, Bottomley, 1999, "The Robustness of the Asymmetrically Dominated Effect: Buying Frames, Phantom Alternatives, and In- Store Purchases," Psychology and Marketing 16 (3): 225-43, Finkelstein E.A., J.G. Trogdon, J.W. Cohen, W. Dietz (2009) Annual medical spending attributable to obesity: payer-and service-specific estimates, Health Affairs, 28, 822–831. Garretsen, H., Rodenburg, G., Van de Goor, L, Van den Eijnden, R., (2008). “Alcohol Consumption in The Netherlands in the Last Decade: Sharp Decreases in Binge Drinking, Especially among Youngsters.” Alcohol and Alcoholism, Vol. 43, No 4, pp. 477-480 Hayes, J., DePasquale, D., Moser, S., (2011). Asymmetric Dominance as a Potential Source of Bias in Hedonic Testing. Department of Food Science, College of Agricultural Sciences, The Pennsylvania State University, Food Quality and Preference 22 (2011) 559–566 Highhouse, S. (1996). Context-dependent selection: The effects of decoy and phantom job candidates. Organizational Behavior and Human Decision Processes, 65(1), 68-76. Huber, J., J. W. Payne, and C. Puto (1982). "Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis." Journal of Consumer Research, 9 (June): 90-98. Iyengar, S., Kinder D., (1987). News That Matters. Public Policy: The essential Readings, pp. 295-305.

21

Research Report 2014

Josiam, B. M., & Hobson, J. P. (1995). Consumer choice in context: the decoy effect in travel and tourism. Journal of Travel Research, 34(1), 45-50. Kim J., Park J., Ryu G. (2006). Decoy effects and Brands, Advances in Consumer Research, Vol 33 Luce, R. D. (1959). Individual Choice Behavior. New York: John Wiley and Sons. Nationaal Bieronderzoek (2014) Nationaal Bieronderzoek Nederland, een

kwantitatief onderzoek naar de consumptie en beleving van bier in 2014. http://www.nederlandsebrouwers.nl/sites/nederlandsebrouwers.nl/files/download s/nationaal_bieronderzoek_2014_-_website.pdf (viewed at 07/12/2014). NHO (2012) Obesity and alcohol: an overview,

http://www.noo.org.uk/uploads/doc/vid_14627_Obesity_and_alcohol.pdf (viewed at 07/12/2014) Parducci, A. (1974). Contextual effects: A range-frequency analysis. Handbook of perception, 2, 127-141. Simonson, I., (1989). “Choice Based on Reasons: The case of Attraction and Compromise Effects.” Journal of Consumer Research 16(2): 158-74 Simonson I., and A. Tversky (1992). "Choice in Context: Tradeoff Contrastand Extremeness Aversion." Journal of Marketing Research, 29 (August): 281-95. Strate, L., 1992. Beer Commercials: A Manual on Masculinity. In Steve Craig Men, Masculinity, and the Media. Sage Publications. Pp 78-92. Te Pas, H. (2010) ‘Steeds meer mannen in de supermarkt’,

http://www.distrifood.nl/Branche-Bedrijf/Algemeen/2010/11/Steeds-meermannen-in-supermarkt-DIS138142W/ (viewed at 13/12/2014). Thompson, N. J., & Thompson, K. E. (1996). Reasoned action theory: an application to alcohol-free beer. Journal of Marketing Practice: Applied Marketing Science, 2(2), 35-48. Thorpe, K. E., Florence, C. S., Howard, D. H., & Joski, P. (2004). The impact of obesity on rising medical spending. Health Affairs-Millwood va then Bethesda ma, 23, 280-283.

22

Research Report 2014

Tversky, A. (1972). “Elimination by Aspects: A Theory of Choice.” Psychological Review, 79 (July): 281-99. Tversky, A., and I. Simonson. 1993. "Context Dependent Preferences: The Relative Advantage Model." Management Science 39 (10): 1179-89. Vedantam, S. (2007) The Decoy Effect, or How to Win an Election. The Washington Post. Voedingscentrum (2014) Alcohol, http://www.voedingscentrum.nl/encyclopedie/alcohol.aspx (viewed at 07/12/2014). Wardle, J., Haase, A. M., Steptoe, A., Nillapun, M., Jonwutiwes, K., & Bellisie, F. (2004). Gender differences in food choice: the contribution of health beliefs and dieting. Annals of Behavioral Medicine, 27(2), 107-116. WHO (2014) Obesity and overweight, http://www.who.int/mediacentre/factsheets/fs311/en/ (viewed at 07/12/2014).

23

Research Report 2014

Appendices
Appendix A: Online survey page 1 (control and treatment group)

Appendix B: Online survey page 2 (treatment group)

24

Research Report 2014

Appendix C: Group vs Age

Age group treatcontrol Age 18-24 25-34 35-49 50-64 Total 21.1% 13.8% 21.1% 44.1% 100.0% ment 17.8% 13.8% 21.7% 46.7% 100.0% 19.4% 13.8% 21.4% 45.4% 100.0% Total

Appendix D: Group vs Gender

Gender

group treatcontrol What is your gender? Male Female 50.7% 49.3% 100.0% ment 52.6% 47.4% 100.0% Total 51.6% 48.4% 100.0%

25

Research Report 2014

Appendix E: Group vs Choice

Choice group control Choice €5 Bier €2.50 Alcoholvrij Bier Total 67.1% 32.9% 100.0% treat-ment 61.2% 38.8% 100.0% Total 64.1% 35.9% 100.0%

Appendix F: Independent Samples Test (Total) Independent Samples Test

Appendix G: Choice vs Gender

The table below shows that females seem to prefer the non-alcoholic beer more often than males. In total forty-three percent of the females chose the target option. Of the males only forty-three percent preferred the non-alcoholic beer. Choice * Gender What is your gender? Male Choice €5 Bier €2.50 Alcoholvrij Bier Total 71% 29% 100% Female 57% 43% 100% 64% 36% 100% Total

26

Research Report 2014

Appendix H: Choice vs Group (Females only)

Choice * Experiment (Females only) Experiment treatcontrol Choice €5 Bier €2.50 Alcoholvrij Bier Total ment Total

65.3% 48.6% 57.1% 34.7% 51.4% 42.9% 100.0% 100.0% 100.0%

Appendix I: Independent Samples Test (Females only)

Appendix J: Choice vs Group (Males only)

Choice * Experiment (Males only) Experiment treatcontrol Choice €5 Bier €2.50 Alcoholvrij Bier Total ment Total

68.8% 72.5% 70.7% 31.2% 27.5% 29.3% 100.0% 100.0% 100.0%

Appendix K: Independent Samples Test (Males only)

27