Exploring the relationship between ad appeal, purchase intention, and the willingness to pay for different types of organic products. Does a picture add more value than a thousand words?

(1)

Does a picture add more value than a thousand words?

Exploring the relationship between ad appeal, purchase

intention, and the willingness to pay for different types of

organic products.

By Frerick Hoeksma

Supervised by Prof. Dr. L.M. Sloot

2nd_{Supervisor: J.A. Koch}

Winschoterdiep 51A 9724 GJ, Groningen +316 40 14 74 83 f.hoeksma@student.rug.nl Student number: S3221709 University of Groningen

Faculty of Economics and Business MSc Marketing

(2)

2

Acknowledgements

Now that I have almost finalized my Master Thesis, a specific quote comes to my mind: “Good, better, best. Never let it rest. Until your good is better and your better is best.” As a motivated student it is a pleasure to be supervised by a professor who gives you an individual approach, which helps you to reach your full potential. For this reason, I would like to gratefully acknowledge Prof. Dr. Laurens Sloot for his advice during this Master Thesis. His expertise in the field of retail marketing & research is outstanding, which provided me the opportunity to ask for high qualitative feedback. Finally, I would to thank the experts who participated in my pretest, for helping me optimizing my advertisement appeals.

Frerick Hoeksma,

(3)

3

Management Summary

Despite growing consumer interest in organic products, retailers still discern a relatively low market share for organic food compared to conventional food. Prior research suggests that marketing campaigns should be initiated to make consumers more aware of the benefits of organic products, which ultimately may influence their purchase intention towards these products.

The aim of this study is to provide brand managers of organic food products an insight into which advertisement appeal is the most effective in influencing purchase intention and the willingness to pay (WTP) a premium price for organic food products. This study focuses on three types of advertisement appeals:

 Functional – factual arguments about the benefits of organic food products.

 Emotional – images that represent a pleasant scenery relating to the experiential value of the product.

 Combined – the simultaneous use of functional and emotional appeals in an advertisement.

Prior research suggests that the effectiveness of an advertisement appeal depends on

congruency with the advertised product. This would suggest that functional appeals are more effective for utilitarian products and emotional appeals are more effective for hedonic

products. However, the products in those studies were conventional products. This study answers the following research question:

Which type of ad appeal is the most effective in increasing a. purchase intention b. and WTP as this concerns c. the type of shopper and d. the type of product, and is this relationship positively confounded by the variables health consciousness and environmental concern among supermarket consumers?

To answer this question, a 4 (neutral, functional, emotional, and combined appeals) x 2

(utilitarian and hedonic products) between-subjects experiment was conducted by means of an online survey of 212 Dutch grocery shoppers in April 2018. The study demonstrates three theoretical contributions.

(4)

4

approximately 6.20% (€0.12). However, this result should be treated with caution as only 2.20% of the variance in WTP is explained by the functional advertisement (ŋ2_{= 0.022).}

Second, hedonic organic food products (e.g. chocolate) are perceived to be more utilitarian than their conventional alternatives. This may explain the third major finding of this study:

that the abovementioned theory of ad-product congruency does not apply to organic food products. For example, this study shows that WTP for a hedonic product (in this study, chocolate) increases by approximately 9.78% (€0.18) if it is advertised with functional product information. However, this result should be treated with caution as only 4.30% of the variance in WTP is explained by the functional advertisement (ŋ2 = 0.043). An explanation

may be found in earlier studies on motives for buying organic products. Prior research suggests that nutritional content is an important driver in the process of buying organic products. This may explain why a functional ad appeal, which contains arguments related to the organic benefits of an organic product, is more effective in influencing WTP than an emotional ad appeal, which does not contain such product-related arguments.

However, in the utilitarian condition, insignificant differences in the effectiveness of the different types of ad appeals were found.

In conclusion, the findings of this study provide a good basis for future research. A replication of this study is necessary in order to generalize its findings.

.

(5)

5

Does a picture add more value than a thousand words?

Exploring the relationship between ad appeal, purchase intention, and the willingness to pay for different types of organic products.

Abstract

Despite growing consumer interest in organic food, there is not much known about how these products can be advertised most effectively. This study examines which advertisement

appeal—neutral, functional, or emotional—is the most effective in influencing consumers’ buying behavior for two types of organic food products—utilitarian and hedonic. A 4 (neutral, functional, emotional, and combined appeals) x 2 (utilitarian and hedonic products) between-subjects experiment was conducted by means of an online survey of 212 Dutch grocery shoppers in April 2018. The results of this study show that advertisements that contain a functional appeal have a positive influence on the willingness to pay a premium price for organic food, especially for hedonic products. Furthermore, this study shows that hedonic organic food products (in this case, chocolate) are perceived to be more utilitarian than their conventional alternatives. This may explain the third major finding of this study: that the

theory of advertisement-product congruency—a theory that suggest that functional appeals are more effective for utilitarian products and emotional appeals are more effective for hedonic products—does not apply to organic food products. Brand managers of organic hedonic food products should advertise their products using functional appealing ads, as this increases consumers’ willingness to pay a premium price for such products. However, future research is necessary to explore the main drivers of the effectiveness of functional

advertisement appeals.

Keywords: advertisement appeals; organic food products; purchase intention; willingness to

pay premium; hedonic; utilitarian; grocery shoppers

1 INTRODUCTION

(8)

8

which is approximately 4% of the total food sales (Asif et al., 2018; Juhl, Fenger, & Thøgersen, 2017).

Compared to conventional food products, organic food products are cultivated with natural fertilizers, are less exposed to antibiotic-resistant bacteria, and have a slightly higher nutritional value, making them healthier than conventional foods (Popa et al., 2018).

Although organic food has both health and environmental benefits, products are primarily purchased for health rather than for environmental reasons (Asif et al., 2018; Lee, Chang, Cheng & Chen, 2018).

Growing consumer interest in healthy products and lifestyles has resulted in the availability of a wider selection of organic products on retail shelves. For example, in 2013, Aldi, a European supermarket chain, introduced the organic brand “SimplyNature” to their merchandise (Ellison, Duff, Wang, & White, 2016). Despite organic food having gained the interest of consumers, retailers still discern a relatively low market share for organic food compared to conventional food (Chekima, Igau, Wafa, & Chekima, 2017). Asif et al. (2018) suggest that marketing campaigns could be initiated to make consumers more aware of the benefits of organic products, which may influence the intention to purchase organic products. Advertisers can use different types of advertisement appeals (hereafter, ad appeal) to persuade consumers (Couwenberg, Boksem, Dietvorst, Worm, Verbeke, & Smidts, 2017).

However, there is a lack of research on how these appeals should be used in

advertisements for organic food products. Research by Ellison et al. (2016) suggests that food products that are labeled organic have different benefits depending on whether they are utilitarian or hedonic. This may suggest that a particular type of ad appeal has a different effect on the purchase intention for a utilitarian organic food product compared to a hedonic one. However, it is unknown which particular ad appeal is the most effective in influencing the purchase intention of grocery shoppers towards these two types of organic food.

Furthermore, there is no research that takes the shopping orientation of grocery shoppers (price-conscious shoppers vs. quality-/brand-conscious shoppers) into account as regards the advertising of organic food. This is interesting because the research of Ellison et al. (2016) suggests that differences in the image of retailers have an effect on the evaluation of organic food.

(9)

9 buying (utilitarian vs. hedonic).

The outcomes of this study provide supermarket chains such as Aldi (price-conscious target group) and Albert Heijn (quality-/brand-conscious target group) with tailor-made advice regarding which type of ad appeal is appropriate 1) for their supermarket format and 2) the type of product, and which is the most effective in influencing the purchase intention towards organic food products. The advice provided by this study will contribute to increasing sales of organic foods in supermarkets.

Therefore, the following research question is defined:

Which type of ad appeal is the most effective in increasing a. purchase intention b. and WTP as this concerns c. the type of shopper and d. the type of product, and is this relationship positively confounded by the variables health consciousness and environmental concern among supermarket consumers?

This study comprises six chapters. Following this introduction, a literature review discusses the variables relevant to this study. The relationship between these variables is presented in a conceptual model. In chapter 3, the research methodology is described, and thereafter the results are presented in chapter 4. In the final chapter, the conclusion contains a discussion of the most important findings, the scientific and managerial implications, and the limitations of this study, as well as suggestions for future research.

2 LITERATURE REVIEW

2.1 Conceptual Model

(10)

10

Figure 2.1

The moderation effect of shopping orientation and type of food on the relationship between ad appeal and purchase intention and the willingness to pay a premium price

2.1.1 Purchase intention

Consumers combine the information (e.g. provided by an advertisement) and their attitudes to an organic product in order to evaluate it and determine their purchase intention (Shafie & Rennie, 2012). Purchase intention can be defined as the likelihood that consumers will be willing to purchase a particular product in the future (Martins, Costa, Oliveira, Gonçalves, & Branco, 2018). Prior studies suggest that egoistic motives such as health consciousness are better predictors of the purchase intention towards organic foods than are altruistic motives such as environmental concerns (Asif et al., 2018; Zagata, 2012). As indicated above, Asif et al. (2018) suggest that marketing campaigns may be initiated to make consumers more aware of the benefits of organic products, which in turn influence the purchase intention towards organic products.

2.1.2 Willingness to pay a premium price

Organic food is sold at a premium price and such prices remain a barrier for consumers buying such products (Shafie & Rennie, 2012). Willingness to pay a premium price

(hereafter, WTP) is the additional price percentage consumers are willing to pay for organic food compared to the price of the conventional alternative (Shafie & Rennie, 2012). Prior research suggest that consumers are willing to pay a premium for organic foods over

(11)

11

lifestyles and attitudes do (Shafie & Rennie, 2012). Strengthening these positive attitudes towards organic products by, for example, advertising, may increase consumers’ willingness to pay a premium for them (Zagata, 2012).

2.1.3 Advertisement appeal

Retailers use ad appeals—the central idea of a message—to persuade consumers to buy organic products instead of often lower-priced conventional alternatives (Couwenberg et al., 2017). Prior research distinguishes three types of ad appeals: functional, emotional, and combined (Matthes, Wonneberger, & Schmuck, 2014).

An advertisement containing a functional appeal attempts to persuade consumers by using factual arguments about the benefits of the product attributes (Couwenberg et al., 2017; Matthes et al., 2014). Organic products are bought for their health-related benefits, so an advertisement containing a functional ad appeal should contain arguments related to these (Asif et al., 2018).

An advertisement containing an emotional appeal attempts to persuade consumers by using images that represent a pleasant scenery related to the experiential value of the product (Couwenberg et al., 2017; Matthes et al., 2014).

An advertisement can also contain a functional and an emotional appeal simultaneously; this is defined as a combined ad appeal (Matthes et al., 2014).

Research in another domain (non-food and non-organic) suggests that different kinds of ad appeals have different persuasive effects (Matthes et al., 2014). An advertisement that contains functional information may reduce uncertainty about the advertised brand or product, which may have a positive influence on purchase intention or WTP for the advertised organic product (Abernethy & Franke, 1996).

Prof. Dr. F. Stroebe, in his book, The Psychology of Advertising (2016), suggests that emotional ad appeals may evoke positive affect, which may be transferred to the advertised brand or product. Such as transfer may have a positive influence on purchase intention or WTP for the advertised organic product.

There are two contradictory perspectives regarding combined ad appeals. One suggests that combined ad appeals increase advertisement effectiveness (Pham, Geuens, & De

Pelsmacker, 2013); the other suggests that they decrease it (Small, Loewenstein, & Slovic, 2007). Despite these contradictory perspectives, I suggest that there exists a positive

(12)

12

depends on the type of organic product and the shopping orientation of the consumer, matters on which I elaborate in the subsequent sections of this chapter. This leads to the following hypotheses:

H1. Functional ad appeals have a significant positive effect on H1a. purchase intention and H1b.

the WTP premium for organic food products.

H2. Emotional ad appeals have a significant positive effect on H2a. purchase intention and H2b.

H3. Combined ad appeals have a significant positive effect on H3a. purchase intention and H3b.

2.1.4 Type of organic food

The organic food products sold by a retailer have either utilitarian or hedonic purposes (Kim, Lee & Choi, 2017). Purchases of utilitarian products such as margarine are more cognitively driven, instrumental, goal-oriented, and are primarily bought for their functional performance benefits, for example, providing energy (Antonides & Cramer, 2013; Loebnitz & Grunert, 2018). In contrast, hedonic products such as chocolate are often bought for the desire for pleasure (Antonides & Cramer, 2013; Loebnitz & Grunert, 2018). Hence, utilitarian and hedonic products evoke different kinds of goals.

(13)

13

combined ad appeals will be the most effective for hedonic organic food products. This lead to the following hypotheses:

H4. Functional ad appeals are more effective in influencing H4a. purchase intention and H4b.

the WTP premium for utilitarian organic food products.

H5. Combined ad appeals are more effective in influencing H5a. purchase intention and H5b. the

WTP premium for hedonic organic food products.

2.1.5 Shopping orientation: price consciousness vs. quality/brand consciousness

Different types of retailers (e.g. hard discounters vs. full-service retailers) carry organic food products in their merchandise. Prior research by Ellison et al. (2016) suggests that differences in the store image of the two types of retailers have an effect on the evaluation of organic products. Store image is the manner in which the store is conceptualized in the shopper’s mind in terms of functional and hedonic attributes (Ellison et al., 2016; Ngobo & Jean, 2012).

A hard discounter, such as Aldi, is characterized by having an every-day-low-pricing strategy, a narrow assortment, a lack of service, and an austere shopping environment (Ellison et al., 2016; Solgaard & Hansen, 2003). In contrast, a full-service retailer, such as Albert Heijn, can be characterized as having a high-low-pricing strategy, more hedonic attributes such as high quality and a high status orientation, and an exciting store atmosphere (Ellison et al., 2016; Solgaard & Hansen, 2003).

Hard-discount shoppers may therefore be characterized as more price-conscious shoppers. For price-conscious shoppers, if the price difference for the distinguishing features of a product is too large, they refrain from buying it (Campbell, DiPietro, & Remar, 2014; Lichtenstein, Bloch, & Black, 1988). This does not, however, mean that consumers pay the lowest price available; it means that they tend to pay a lower price when the cost of the distinguishing features of a more expensive alternative cannot be justified (Lichtenstein et al., 1988). Shoppers at a full-service retailer may be characterized as more

quality-/brand-conscious shoppers. This kind of shopper favors well-known brands and good-quality department stores (Bakewell & Mitchell, 2004).

Prior research suggests that the differences in store image can influence consumers’ willingness to pay a premium price for food (Ngobo & Jean, 2012). Furthermore, it is likely that differences between stores’ images have implications for the effectiveness of different types of ad appeals. I suggest that price- conscious shoppers are more influenced by

(14)

14

suggest that quality-/brand-conscious shoppers are more influenced by emotional ad appeals as these better match with the exciting shopping environment of quality/brand consciousness shoppers. This leads to the following hypotheses:

H6: For price-conscious shoppers, functional ad appeals are more effective in increasing H6a.

purchase intention and H6b. the WTP premium for organic food products.

H7: For quality-/brand-conscious shoppers, emotional ad appeals are more effective in

increasing H7a. purchase intention and H7b. the WTP premium for organic food products. 2.1.6 Health consciousness

The organic label on the packaging of a food product is used by consumers as a heuristic that indicates that the food is healthy (Lee et al., 2018). Consumers believe that food labeled organic is more nutritious and contains fewer calories than the conventional alternative (Asif et al., 2018; Ellison et al., 2016; Lee, et al., 2018). Due to the health associations of food labeled organic, it is likely that consumers with health-related beliefs are more influenced by advertisements for organic products than are consumers who do not have such beliefs. We refer to consumers with such beliefs as health-conscious consumers (Asif et al., 2018). Health-conscious consumers consciously weigh the costs and benefits of their health choices and can be defined by the degree to which their health beliefs are integrated into their daily life (Asif et al., 2018; Lee et al., 2018). This leads to the following hypothesis:

H8: Health consciousness among consumers positively confounds the relationship between ad

appeal and H8a. purchase intention and H8b. the WTP premium for food products labeled

organic.

2.1.7 Environmental concern

(15)

15

& Jones, 2012). Based on the prior research, I do not expect that environmental concern has a significant confounding relationship on ad appeal and the WTP premium for organic food products. However, I still test whether such a confounding relationship exists. This leads to the following hypothesis:

H9: Environmental concern among consumers positively confounds the relationship between

ad appeal and H9a. purchase intention and H9b. the WTP premium of food products labeled

organic.

3 METHODOLOGY

3.1. Data collection and sample characteristics

Data was collected by means of a web-based survey of 212 Dutch grocery shoppers who participated in a 4x2 between-subjects experiment in April 2018. The experiment was structured as ad appeal (functional, emotional, and combined) vs. types of organic food (utilitarian and hedonic). A between-subject experiment entails that the respondents are randomly assigned to each of the six manipulation conditions (Aronson, Wilson, & Brewer, 1998). In comparison to a within-participants design, random allocation has the advantage that causal relationships can be determined with a higher degree of certainty (Aronson et al., 1998). A previous study of ad appeals has shown that a web-based survey is a useful method for gathering this type of data (Matthes et al., 2014).

Convenience sampling is considered because of the minimal time and resources that were available (Malhotra, 2010). However, convenience sampling has certain limitations. Convenience samples may be biased because respondents may not fully represent the

population from which the sample has been drawn (Sousa, Zauszniewski, & Musil, 2004). For this reason, a single cross-sectional sample was drawn from grocery shoppers who were contacted at local supermarkets. Research participation was stimulated by giving away a €50 grocery shopping gift card to one randomly selected respondent from those who submitted a valid questionnaire. The survey included an “attention check” question that controls whether the respondents have read all the questions properly. Respondents who did not answer the attention check question properly were eliminated from the survey results. A more detailed explanation of the composition of the sample is discussed in sections 4.1 and 4.3. The survey was in the native language (Dutch) of the target group, which made it easier for the

(16)

16

3.2 Manipulation experiment

The online survey included a 4x2 between-subjects experiment in which the constructs “Type of ad appeal” and “Type of organic food” were manipulated. The underlying motive of the experiment is that I assume that an ad appeal (functional, emotional, or combined) is more effective when it matches the type of organic product (utilitarian or hedonic). The allocation of the eight conditions of the experiment is shown in Table 3.1.

The respondents to the survey were shown an advertisement based on one of the eight experimental conditions. To ensure the validity of the advertisements, the types of

advertisement appeal were classified in a pretest. An overview of the different advertisements used in the eight experimental conditions is shown in Appendix A.

TABLE 3.1

Conditions and manipulations NEUTRAL APPEAL FUNCTIONAL APPEAL EMOTIONAL APPEAL COMBINED APPEAL UTILITARIAN PRODUCT

Condition 1 Condition 3 Condition 5 Condition 7

HEDONIC PRODUCT

Condition 2 Condition 4 Condition 6 Condition 8

3.3 Measurement constructs

3.1.1 Procedure

The questionnaire consisted of three parts: a demographic profile (age, gender, and

occupation), an experiment, and the measurement of the constructs. To ensure the validity of the survey, scales and questions confirmed by other studies were adapted to measure the constructs. Furthermore, each respondent was compelled to respond to each statement, thereby contributing to the validity of the survey. A complete overview of the key constructs and their measurement is shown in Appendix B.

3.1.2 Measurement of WTP premium

(17)

17

direct questioning has several limitations. Firstly, the elicited WTP is often overestimated because of the lack of a reference point for making a realistic bid (Bett, Peters, Nwankwo, & Bokelmann, 2013). Furthermore, direct questioning requires a substantial cognitive task from respondents, which often results in a large number of non-responses (Klose, 1999).

One method of overcoming this problem is the use of payment cards. Payment cards overcome the limitations of direct questioning because they do not require respondents to make a bid without a reference point (Klose, 1999). Payment cards require respondents to indicate their maximum WTP from a range of given values by marking the amount closest to their WTP (Klose, 1999). However, the drawback of using payment cards is that they often cause a so-called range bias, meaning that the WTP may be influenced by the range of the amounts presented (Klose, 1999).

Another measurement technique that was considered is the bidding game, in which an auction system is simulated (Klose, 1999). The advantage of such a method is that it is closer to market situations (Klose, 1999). However the drawbacks of this method are the so-called starting point bias, which refers to the fact that the WTP may be influenced by the first bid of the auction process (Klose, 1999). Furthermore, the bidding-game procedure requires an interactive computer program, which was not available for this research.

Due to the limitations of direct questioning, it is decided to use the payment cards method to overcome these by providing respondents with a realistic reference point (Klose, 1999). The question was formulated as follows:

If an average 250g package of organic margarine costs €2.00 in an average supermarket, what do you think is a reasonable price for the package shown?

□ 50% cheaper □ 25% cheaper □ 10% cheaper □ same price □ 10% more expensive □ 25% more expensive □ 50% more expensive

3.3.3 Measurement of purchase intention

(18)

18

its close similarity with one of the other items (“I intend to buy this product”). Had the particular item not been deleted, the survey would have included three extra questions, which would have increased the length of the survey and, ultimately, decreased the quality of the responses (Galesic & Bosnjak, 2009). Furthermore, longer surveys lead to a higher percentage of incomplete responses (Marcus, Bosnjak, Lindner, Pilischenko, & Schütz, 2007). For these reasons, it was decided to delete one of the items of the original scale. The items used to measure the purchase intention construct are described in Appendix B.

3.3.4 Measurement of shopping orientation—Price consciousness

A scale confirmed by Campbell et al. (2014) and Lichtenstein et al. (1988) was adapted for measuring the price consciousness construct, which is one of the two moderating variables. The scale consisted of three items measured on a seven-point Likert scale with scale endpoints of 1 = totally disagree and 7 = totally agree. The Cronbach alpha (CA) score was calculated after the questionnaire had been conducted; no items could be deleted to increase the CA. This leaves a reasonable internal consistency, with a CA of 0.737. The items for measuring the level of price consciousness are described in Appendix B.

3.3.5 Measurement of shopping orientation—Quality/brand consciousness

A scale confirmed by Bakewell and Mitchell (2004) was adapted for measuring the quality/brand consciousness construct, which is the second moderating variable. The scale consisted of five items measured on a seven-point Likert scale with scale endpoints of 1 = totally disagree and 7 = totally agree. One item, a question on the respondent’s opinion on department and specialty stores, was deleted beforehand as it does not measure the extent of quality/brand consciousness. The remaining four items resulted in a reasonable internal consistency, with a CA of 0.715. The items for measuring the level of quality/brand consciousness are described in Appendix B.

3.3.6 Measurement of health consciousness

(19)

19

3.3.7 Measurement of environmental concern

A scale confirmed by Roberts and Bacon (1997) was adapted for measuring the

environmental concern construct, the second confounding variable. The items reflect how environmental concern enters the decision-making process. The original scale consisted of 12 items measured on a five-point Likert scale with scale endpoints of 1 = never true and 5 = always true. However, the scales used to measure the other constructs of this study consist of approximately three items. Using 12 items to measure the environmental concern construct would create an imbalance in the research. For this reason, the three items with the highest loading (all three items > .84) in the factor analysis in Robert and Bacon (1997) were used here to measure the environmental concern construct. The CA was calculated after the questionnaire was conducted. The CA could be increased by 0.026 if one item were deleted; however, it was decided to not delete this item due to the minimal increase in the CA. This leaves a high internal consistency, with a CA of 0.878. The three items for measuring the level of environmental concern are described in Appendix B.

3.3.8 Measurement manipulation check: hedonic level of the product

A scale confirmed by Voss, Spangenberg, and Grohmann (2003) was adapted for measuring the hedonic level of the products shown in the advertisements. The items reflect the extent to which consumers rated a product as being more or less hedonic. The original scale consisted of five items measured on a seven-point Likert scale with scale endpoints of 1 = totally disagree and 7 = totally agree. One item (“Thrilling”) was deleted beforehand because students who participated in the pretest indicated that it was highly similar to another item of this scale (“Excitement”) and its contribution would hence be minimal. After the

questionnaire was conducted, one item (“Excitement”) was deleted, which increased the CA from 0.787 to 0.801. The remaining three items resulted in a high internal consistency, with a CA of 0.801. The four items for measuring the hedonic level of the advertised products are described in Appendix B.

3.3.9 Measurement manipulation check: utilitarian level of the product

(20)

20

describe a food product. The remaining four items resulted in a high internal consistency, with a CA of 0.869. The four items for measuring the utilitarian level of the advertised products are described in Appendix B.

3.3.10 Measurement manipulation check: functional level of the ad

A scale confirmed by Jourdan (1999) was adapted for measuring the functional level of the advertisements used in the manipulation of the conditions. The items reflect the extent to which consumers rated an advertisement as being more or less functional. The original scale consisted of five items measured on a seven-point Likert scale with scale endpoints of 1 = totally disagree and 7 = totally agree. Two items were deleted beforehand because they were not suitable for measuring the respondents’ opinions on the appeal of the advertisement. One item was deleted because it was a question regarding competitors to the advertised brand. Another item was deleted because it was a question regarding respondents’ product

preferences. The remaining three items resulted in a high internal consistency, with a CA of 0.815. The three items for measuring the functional level of the advertisements are described in Appendix B.

3.3.11 Measurement manipulation check: emotional level of the ad

A scale confirmed by Jourdan (1999) was adapted for measuring the emotional level of the advertisements used in the experiment of this study. The items reflect the extent to which consumers rated an advertisement as being more or less emotional. The original scale consisted of four items measured on a seven-point Likert scale with scale endpoints of 1 = totally disagree and 7 = totally agree. One item, regarding the visual attractiveness of the ad, was deleted beforehand as students who participated the pretest indicated that the “visual attractiveness of the ad” could also be interpreted as a preference for colors, products, slogans, and the like. An incorrect interpretation of this question could have distorted the results of this scale. The three remaining items resulted in a high internal consistency, with a CA of 0.866. The three items for measuring the emotional level of the advertisements are described in Appendix B.

3.4 Pretest: classification of the ad appeals

(21)

21

and the combined ad appeal was a combination of the functional and emotional ad appeals. However, the experts indicated some points of improvements—these and their assessments are presented in Table 3.2.

TABLE 3.2

Pretest: classification of ad appeal REMARKS:

DR. M.C. LELIVELD

□ “Each advertisement appeal is clear, and the ad appeals are obviously different from each other.”

□ “Add an emotional quote/slogan to the emotional appeal condition.” ADJUSTMENT

MADE:

□ An emotional slogan was added to the emotional appeal condition.

DR. L. VOERMAN

□ “Make the letters in the functional advertisement bigger, so that they become more central to the ad.”

□ “The emotional factor is too conspicuous in the combined appeal condition.” ADJUSTMENTS

MADE:

□ The ratio of emotional/functional appeal in the combined was balanced by increasing the font size of the functional appeal, and by decreasing the size of the emotional picture used in the ad.

DR. M. MOEINI JAZANI

□ “Add a quote or a slogan to the emotional condition, as otherwise you have different effects because of the difference in levels of information (no slogan vs. slogan).”

□ “Substitute the current product (margarine) in the utilitarian condition with bread, because bread is a more utilitarian product than margarine.”

ADJUSTMENT MADE:

□ A slogan was added to the emotional condition.

□ The product used in the utilitarian condition (margarine) was not substituted with bread, because the advertised product in both the hedonic and utilitarian conditions was a packaged food product. Organic bread is sometimes not a packaged food product; this could result in variances in the DV through differences that are not controlled for in the experiment. For this reason, it is decided to decline this expert advice.

MSC JAN KOCH □ “Each advertisement appeal is clear, and the ad appeals are obviously different from each

other.” ADJUSTMENT

MADE:

□ None

3.5 Data analysis plan

(22)

22

Step 1: Cleaning the data

First, the raw data set was cleaned to remove inconsistencies (e.g. outliers) and missing values. Furthermore, responses of respondents who did not provide a valid answer to the attention check question were deleted from the survey data. Unnecessary variables (e.g. IP address information) that were automatically generated by Qualtrics were deleted from the dataset. Thereafter, the data was tested for normality to ascertain whether normal distributions could be assumed in the analysis.

Step 2: Testing the measurement models for validity and reliability

After the data had been cleaned of inconsistencies, a factor analysis (FA) was conducted to test the discriminant and convergent reliability of the particular constructs. A Varimax

rotation was used to minimize the number of variables that have high loadings on each factor. A Kaiser-Meyer-Olkin (KMO) and Barlett’s test was conducted to test whether the items correlate with one another. In addition to the FA, the internal consistency and reliability of the constructs was ascertained by means of the CA test.

Step 3: Testing the hypotheses

Following the checks for the validity and reliability of the constructs, the hypothesized model was tested. Thereafter, the hypothesized relationships, their significance levels, and statistical fit were analyzed.

4 RESULTS

In this chapter, the results of the steps discussed in chapter 3 are presented.

4.1. Cleaned data

(23)

23

For this reason it would not be appropriate to rely on children’s responses to measure the WTP for organic food products.

4.2. Checking for normality

A skewness and kurtosis analysis was applied to ascertain whether the DV scores (purchase intention and WTP) are normally distributed for functional, emotional, and combined ad appeals. Table 4.1 shows that all skewness and kurtosis residuals for the WTP variable were in the range -2/+2, which suggests that the data for WTP is normally distributed. Table 4.2 shows that all kurtosis values for purchase intention were in the range -2/+2. However, some skewness residuals for the purchase intention variable violate the assumption of normality. Table 4.2 shows that the data for the purchase intention variable is negatively skewed, which was expected. Asif et al. (2018) suggest that marketing campaigns have a positive influence on purchase intention towards organic products, which explains the high number of

observations indicating a positive purchase intention and the relatively few observations indicating a low purchase intention.

A statistical logarithmic transformation of the purchase intention variable was considered as the normality assumption is necessary to estimate confidence intervals and p-values in an unbiased manner (Schmidt & Finan, 2017). However, Schmidt and Finan (2017) suggest that violations of the normality assumption often do not noticeably impact on the results of parametrical tests. They furthermore suggest that a statistical logarithmic

transformation does not guarantee valid results and may lead to biased estimates. In addition, the few skewness residuals for purchase intention only slightly fail the normality assumption. For this reason, it was decided to not apply a statistical logarithmic transformation for the purchase intention variable. However, the results of the regression and MANOVA analyses should, as a result of the lack of normality, be interpreted with a certain degree of caution.

Table 4.1

Skewness and kurtosis (WTP)

CONSTRUCT SKEWNESS KURTOSIS

FUNCTIONAL –1.043 .540

EMOTIONAL –1.360 1.060

(24)

24

Table 4.2

Skewness and kurtosis (purchase intention)

CONSTRUCT SKEWNESS KURTOSIS

FUNCTIONAL –2.313 –.211

EMOTIONAL –1.603 –.988

COMBINED –2.697 –.186

4.3. Descriptive statistics

The sample consists of responses from 88 (41.5%) men and 124 women (58.5%) with an average age of 33.70 years (Mage = 33.70 SD = 14.94). The ratio of men to women in this

study is slightly different from a representative sample of Dutch grocery shoppers, which should comprise 30% men and 70% women (EFMI Business School, 2013). In addition, the average age of this sample is lower than the average age of the average Dutch grocery shopper (Mage = 46.9) (EFMI Business School, 2013). Furthermore, the majority of the

respondents (64.2%) identified themselves as primary shoppers. However, an optimal sample should comprise 100% primary shoppers (EFMI Business School, 2013). An explanation may be found in a prior study by Sarantopoulos, Theotokis, Pramatari & Doukidis (2016) that suggests that grocery shoppers enter supermarkets to fulfill diverse needs, for example buying a soft drink or a bag of potato chips for immediate consumption. I assume that these kind of shoppers indicate themselves not as primary shopper, because the main groceries are done by someone else in the household. The average respondent had completed a bachelor’s/HBO degree, and indicated that the total income of their household was average. The majority of the respondents could be categorized as non-users (44.3%) and light users (44.3%) of organic food products; only 11.3% of the respondents indicated that they were regular users of organic products.

Although responses were collected from grocery shoppers, as indicated above, the characteristics of the sample in this study are slightly different from a representative sample of Dutch grocery shoppers (Figure 4.3). For this reason, a replication of this study is

(25)

25

Table 4.3

Comparison of sample characteristics of this study with representative sample

CRITERIA REPRESENTATIVE SAMPLE OF

DUTCH GROCERY SHOPPERS

STUDY SAMPLE

AVERAGE AGE 46.9 33.7

GENDER 70.00% Female 58.5% Female

PRIMARY SHOPPER 100% 64.2%

4.4. Reliability and validity measurements

An FA was conducted to measure the reliability of the scales used in this research. Varimax rotation was applied to minimize the number of variables that have high loadings on each factor. An overview of the FA and CA scores are shown in Table 4.4. This table is described in more detail in the sections 4.3.1–4.3.5. The total output of the FA is shown in Appendix D.

Table 4.4 FA & CA measures

CONSTRUCT KMO BARLETTS’S

TEST NUMBER OF COMMUNALITIES BELOW .4 CUMLATIVE FACTOR LOADING EXCLUDED ITEMS CRONBACH ALPHA PRICE CONSCIOUSNESS .692 .000 0 61.102 0 .737 BRAND CONSCIOUSNESS .692 .000 0 61.102 0 .715 HEALTH CONSCIOUSNESS .752 .000 0 83.081 0 .898 ENVIRONMENTAL CONCERN .752 .000 0 80.957 0 .878 UTILITARIAN LEVEL _.754 _.000 ₀ _67.436 ₀ _.869 HEDONIC LEVEL _.754 _.000 ₀ _67.436 ₁ _.801 FUNCTIONAL APPEAL .745 .000 0 76.568 0 .815 EMOTIONAL APPEAL _.745 _.000 ₀ _76.568 ₀ _.866

4.4.1 Shopping orientation construct: price & quality/brand consciousness

Concerning the shopping orientation construct, the KMO and Bartlett’s test is significant (p <

0.05) with an KMO score of .692, which suggests that the use of an FA is allowed. The

(26)

26

Furthermore, the scree-plot shows two factors with an eigenvalue > 1, which confirms that the number of factors is 2.

The total cumulative variance of the two components is 61.10%, which is above the threshold of > 60.00%. In addition to the FA, the CA scores calculated for price

consciousness (α = .737) and quality/brand consciousness (α = .715) mean that the internal validity of the two factors is sufficient.

4.4.2 Health consciousness

Concerning the health consciousness construct, the KMO and Bartlett’s test is significant (p <

rotated extraction column, shown in Appendix D, indicates that all the items of the scale are one factor. Furthermore, the scree-plot shows one factor with an eigenvalue > 1, which confirms that the number of factors is 1.

The total cumulative variance of the component is 83.08%, which is above the threshold of > 60.00%. In addition to the FA, the CA score calculated for health consciousness (α = .898) means that the internal validity of the factor is high.

4.4.3 Environmental concern

Concerning the environmental concern construct, the KMO and Bartlett’s test is significant (p

< 0.05) with an KMO score of .712, which suggests that the use of an FA is allowed. The

rotated extraction column, shown in Appendix D, indicates that all the items of the scale are one factor. Furthermore, the scree-plot shows one factor with an eigenvalue > 1, which confirms that the number of factors is 1.

The total cumulative variance of the component is 80.96%, which is above the threshold of > 60.00%. In addition to the FA, the CA score calculated for environmental concern (α = .878) means that the internal validity of the factor is high.

4.4.4 Type of product: hedonic and utilitarian level

Concerning the product type construct, the KMO and Bartlett’s test is significant (p < 0.05) with an KMO score of .754, which suggests that the use of an FA is allowed. The rotated extraction column, shown in Appendix D, identifies two factors:

(27)

27

The total cumulative variance of the two components is 67.44%, which is above the threshold of > 60.00%. In addition to the FA, the CA scores calculated for the hedonic level

(α = .801) and the utilitarian level (α = .869) mean that the internal validity of the two factors

is high. As discussed in chapter 3, one item of in the hedonic-level scale was deleted to increase the CA score from .787 to .801. After deleting this item, the total cumulative variance increased from 67.44% to 72.72%.

4.4.5 Ad appeal: functional and emotional

Concerning the shopping orientation construct, the KMO and Bartlett’s test is significant (p <

rotated extraction column, shown in Appendix D, identifies two factors:  The three items measuring functional appeal are one factor; and  The three items measuring emotional appeal are one factor.

The total cumulative variance of the two components is 76.57%, which is above the threshold of > 60.00%. In addition to the FA, the CA scores calculated for functional appeal

(α = .815) and emotional appeal (α = .866) mean that the internal validity of the two factors is

high.

4.5. Manipulation check

4.5.1 Ad appeal: functional and emotional

(28)

28

Utilitarian condition (margarine):

- Advertisement 1 (functional ad appeal): A 2 (functional vs. emotional appeal) x 1 (utilitarian product)

ANOVA showed that respondents who viewed the functional ad indicated that it was significantly (p < 0.05) more functional than did respondents who viewed the emotional ad

(Mfunc_func = 4.28 vs. Mfunc_emo = 3.07). This means that the functional ad appeal in the

utilitarian condition was indeed perceived as being functional.

- Advertisement 2 (emotional ad appeal): A 2 (functional vs. emotional appeal) x 1 (utilitarian product)

ANOVA showed that respondents who viewed the emotional ad indicated that it was significantly (p < 0.05) more emotional than did respondents who viewed the functional ad

(Memo_emo = 3.84 vs. Memo_func = 4.80). This means that the emotional ad appeal in the

utilitarian condition was indeed perceived as being emotional.

- Advertisement 3 (combined ad appeal): Respondents who viewed the combined ad indicated

that it was average on both the functional and emotional scales (Mfun_comb = 4.25 vs. Memo_comb

= 4.09), which confirms that the combined ad appeal in the utilitarian condition was

perceived as being a combination of the functional and emotional ad appeals.

Hedonic condition (chocolate):

- Advertisement 1 (functional ad appeal): A 2 (functional vs. emotional appeal) x 1 (hedonic product)

ANOVA showed that respondents who viewed the functional ad indicated that it was significantly (p < 0.05) more functional than did respondents who viewed the emotional ad

(Mfunc_func = 4.691 vs. Mfunc_emo = 2.81). This means that the functional ad appeal in the

hedonic condition was indeed perceived as being functional.

- Advertisement 2 (emotional ad appeal): A 2 (functional vs. emotional appeal) x 1 (hedonic product)

ANOVA showed that respondents who viewed a functional ad and those who viewed an emotional ad indicated insignificant differences between the degree of emotional appeal of these types of advertisements (p = .417).

I assume that these insignificant differences may be explained by the emotional level of the advertised product (a chocolate bar). Parker, Parker, and Brothcie (2006) suggest that chocolate is a product that is used during emotional eating. The assumption is that this may have influenced the degree of emotional appeal in the functional advertisement.

- Advertisement 3 (combined ad appeal): Respondents who viewed the combined ad indicated

(29)

29

= 4.62), which confirms that the combined ad appeal in the hedonic condition is perceived as

being a combination of the functional and emotional ad appeals.

4.5.2 Product type: utilitarian and hedonic

All respondents randomly viewed a utilitarian (margarine) or a hedonic product (chocolate). Afterwards respondents were asked to indicate the degree to which they perceived the product as being more utilitarian or hedonic. High positive values for the items on the utilitarian scale would indicate that respondents viewed the product as more utilitarian. High positive values for the items on the hedonic scale would indicate that respondents viewed the product as more hedonic.

Respondents who viewed an ad with the utilitarian product (margarine) indicated the product was significantly more utilitarian than the respondents who viewed an ad with the hedonic product (chocolate) (Muti_uti = 4.31 vs. Muti_hed = 3.23).

However, respondents indicated insignificant differences (p = 0.103) between the hedonic levels of either product. Both products were perceived as being moderately low on the hedonic scale, which means that organic chocolate is perceived as being more utilitarian than conventional chocolate.

4.6. Advertisement recall

4.6.1 Utilitarian condition (margarine)

All respondents who viewed the advertisement for the tub of margarine (n = 111, 100%) remembered the product advertised. Furthermore, n = 96 (86.5%) respondents remembered the advertised brand (NEO BIO). This suggests that the brand and product recall for the advertisement in the utilitarian condition was high.

4.6.2 Hedonic condition (chocolate)

Almost all respondents who viewed the advertisement for the chocolate bar (n = 98, 97% of n

= 101) remembered the product advertised. Furthermore, (n = 86, 85.1%) respondents

remembered the advertised brand (NEO BIO). This suggests that the brand and product recall for the advertisement in the hedonic condition was high.

4.7. Control variables

An ANOVA was applied to ascertain whether the control variables have a significant

(30)

30

in the WTP for organic (p = .792), there are significant differences in their purchase intention towards organic food (Mnonusers = 4.45, Mlightusers = 5.03, Mregularusers = 5.35, SD = 1.43, p <

0.05). Hence, in the analysis that follows, the impact of user frequency on the relationship

with the several constructs is controlled for.

The following control variables did not have a significant influence on the dependent variables: gender, income, device on which the respondents completed the questionnaire, educational level, age, and household size. Hence, in the analysis that follows, the impact of these control variables on their relationship with the particular constructs is not controlled for.

4.8. Testing the hypotheses

In this section, the hypotheses discussed in chapter 3 are tested. First, the direct relationships between the IV and the DVs are tested (H1–H5). Thereafter, the moderating (H6–H7) and confounding (H8–H9) relationships are tested.

4.8.1 Advertisement appeal

First, the total scores of the purchase intention items were recalculated to generate an average score. Thereafter, dummy variables were created for the various conditions. A two-way MANOVA analysis was applied to check for significant differences in the means of the different types of ad appeals (functional, emotional, and combined). An ad appeal is effective in influencing the purchase intention towards an organic product when the mean of a

particular ad appeal is significantly higher than the neutral condition. The effects of the particular ad appeals are tested by the following hypotheses:

H1. Functional ad appeals have a significant positive effect on H1a. purchase intention and H1b.

H2. Emotional ad appeals have a significant positive effect on H2a. purchase intention and H2b.

H3. Combined ad appeals have a significant positive effect on H3a. purchase intention and H3b.

(31)

31

Table 4.5

Mean scores: purchase intention and WTP vs. advertisement appeal

Before the values of the two-way MANOVA can be interpreted, several assumptions need to be validated. First, the box’s test of the equality of covariance matrices needs to be

insignificant at a threshold of p > 0.001, which is the case (p=0.008). Hence, it can be assumed that the covariance matrices of the DVs are equal across groups. Furthermore, Levene’s test of the equality of error variances needs to be insignificant at a threshold of p >

0.05, which is the case (Purchase Intention p=.109; WTP p=.368). The values of the tests indicate

that the assumptions of the two-way MANOVA are not violated, which allows the interpretation of its output. First, the results for purchase intention are discussed and thereafter, the results for WTP.

Purchase intention: The results of the two-way MANOVA analysis show that compared with the neutral condition, the functional (p = .261), emotional (p = .078), and combined (p = .247) ad appeals result in an insignificant increase of purchase intention towards organic products.

WTP: The results of the two-way MANOVA analysis show a significant increase in the willingness of respondents to pay a premium price for organic food products once they had seen a functionally appealing ad (p < 0.05). The WTP of consumers in the functional condition was €0.12 (6.2%) higher than the consumers in the neutral condition (Mneutral =

€1.93, SD = .040; Mfunctional = €2.05, SD = .040).

However, the partial eta squared value (ŋ2 = 0.022) shows that only 2.2% of the variance in

WTP is explained by the functional ad appeal, which is relatively low.

(32)

32

point beyond which further advertising investment generates no extra revenue (Cao, Zhou, Xie, & Zhong, 2017). Future research should explore this threshold point for advertising organic food products.

The two-way MANOVA shows that compared with the neutral condition, the

emotional (p = .636) and combined (p = .934) ad appeals result in an insignificant increase of the WTP towards organic products.

Note that the analysis of hypotheses 1–3 is based on two advertised products—margarine and chocolate; the differences in the degree of hedonic and utilitarian level for these two products could result in other linear effects of the advertisement appeals on the DVs. These effects are tested in the next subsection.

4.8.2 Type of product

In this subsection, a two-way MANOVA analysis is applied to ascertain which advertisement appeal is most effective in influencing purchase intention and WTP for a utilitarian product (margarine) and a hedonic product (chocolate). An ad appeal is effective in influencing purchase intention towards an organic product when the mean of a particular ad appeal is significant higher than the neutral condition. First, the results for the utilitarian condition are discussed, and thereafter, those for the hedonic condition.

4.8.2.1 Utilitarian condition

First, a two-way MANOVA analysis is applied to check for significant differences in means between the constructs of the following hypotheses:

H4a. Functional ad appeals are more effective in influencing purchase intention towards

utilitarian organic food products.

H4b. Functional ad appeals are more effective in influencing WTP for utilitarian organic food

products.

(33)

33

Table 4.6

Mean scores: purchase intention and WTP (utilitarian condition)

insignificant at a threshold of (p > 0.001), which is the case (p = 0.169). Hence, it can be assumed that covariance matrices of the DVs are equal across groups. Furthermore, the Levene’s test of the equality of error variances needs to be insignificant at a threshold of p >

that the assumptions of the two-way MANOVA are not violated, which allows the interpretation of its output. First, the results for purchase intention are discussed, and thereafter, those for WTP.

Purchase Intention: The results of the two-way MANOVA analysis show that compared with the neutral condition, the functional (p = .188), emotional (p = .762), and combined (p = .758) ad appeals result in an insignificant increase in purchase intention towards utilitarian organic products (in this case, margarine).

WTP: The results of the two-way MANOVA analysis show that compared with the neutral condition, the functional (p = .240), emotional (p = .363), and combined (p = .194) ad appeals result in an insignificant increase in WTP for utilitarian organic products (margarine).

4.8.2.2 Hedonic condition

First, a two-way MANOVA analysis is applied to check for significant differences in means between the constructs of the following hypotheses:

H5a. Combined ad appeals are more effective in influencing purchase intention towards

(34)

34

H5b. Combined ad appeals are more effective in influencing WTP for hedonic organic food

products.

Prior to the detailed presentation of the hypothesized relationships, the means for each condition for purchase intention and WTP are shown in Table 4.7.

Table 4.7

Mean scores: purchase intention and WTP (hedonic condition)

insignificant at a threshold of p > 0.001, which is the case (p = 0.123). Hence, it can be assumed that covariance matrices of the DVs are equal across groups. Furthermore, the Levene’s test of the equality of error variances needs to be insignificant at a threshold of p >

that the assumptions of the two-way MANOVA are not violated, which allows the interpretation of its output. First, the results for purchase intention are discussed, and thereafter, those for WTP.

Purchase Intention: The results of the two-way MANOVA analysis show that compared with the neutral condition, the functional (p = .583) and combined (p = .090) ad appeals result in an insignificant increase in purchase intention towards hedonic organic products (in this case, chocolate). However, there is a significant negative relationship between emotional ad appeal and purchase intention towards hedonic organic food products

(p < 0.05). Remarkable here is that both the advertisements in which the product is central to

the advertisement indicate greater purchase intention (Mneutral = 4.97; Mfunctional = 4.74) than

those that contain an emotionally appealing picture (Memotional = 3.79; Mcombined = 4.23).

(35)

35

Prior research suggests that nutritional content is an important driver for buying organic products (Lee & Yun, 2015). This may explain why respondents in the neutral condition, in which the product is central to the ad, show high values for intention to purchase the

advertised product.

The neutral condition is used as a baseline in the analysis of the effectiveness of the three ad appeals (functional, emotional, and combined). High values in the neutral condition results in insignificant results for the three ad appeals for purchase intention towards organic food. However, the means of the two-way ANOVA test (Mfunctional = 4.74; Memotional = 3.79;

Mcombined = 4.23) suggest that a functional appeal is more effective in influencing purchase

intention towards hedonic organic food products. Future research could test this assumption by replicating this study.

WTP: The results of the two-way MANOVA analysis show that compared with the neutral condition, the emotional (p = 1.000) and combined (p = .232) ad appeals result in an insignificant increase of WTP for hedonic organic products (in this case, chocolate).

However, the results also show a significant increase of the willingness of respondents to pay a premium price for hedonic products (chocolate) once they had seen a functionally appealing ad (p < 0.05). The WTP of consumers in the functional condition was €0.18 (9.78%) higher than consumers in the neutral condition (Mneutral = €1.84, SD = .036; Mfunctional = €2.02, SD =

.059). However, the partial eta squared value (ŋ2_{= 0.043) shows that only 4.3% of the}

variance in WTP is explained by the functional ad appeal, which is relatively low. However, this does not automatically mean that the WTP increases by the same amount when a respondent sees an ad for the second time. As indicated above, Cao et al.’s (2017) threshold point beyond which advertising investment no longer generates extra revenue should be explored for the advertising of organic foods.

An explanation for the significant positive relationship of a functional ad appeal influencing WTP towards organic food products may be found in previous studies on motives for buying organic products. Prior research suggests that the nutritional content is an

(36)

36

4.8.3 Testing moderating confounding effects

In this subsection, the moderating effects proposed in hypothesis H6 are tested using a linear regression model. Furthermore, the confounding effects suggested in hypotheses H8 and H9 are tested by means of a Spearman correlation test (effect confounding variable on the IV) and a Pearson correlation test (effect confounding on the DVs).

H6: For price conscious shoppers, functional ad appeals are more effective in increasing WTP premium for organic food products.

H8: Health consciousness among consumers positively confounds the relationship between ad

appeal and WTP premium for food products labeled organic.

H9: Environmental concern among consumers positively confounds the relationship between

ad appeal and WTP premium for food products labeled organic.

Initially, the total scores of the moderating confounding variables are recalculated as an average score. Thereafter, the means of the IV, moderating variables, and confounding variables are subtracted in order to allow a better interpretation of the intercept and to

minimize the risk of multicollinearity. First, the direct effect of price consciousness and brand consciousness on the DV (WTP) was tested. No direct effects were found for either price consciousness (p = .107) and brand consciousness (p = .695). Thereafter, the moderating effects of price consciousness and brand consciousness on the relationship between the functional ad appeal and WTP were tested. Furthermore, the expected confounding variables (environmental concern and health consciousness) were tested for moderating effects on the relationship between functional ad appeal and WTP. No moderating effects were found for price consciousness (p = .762), brand consciousness (p = .540), environmental concern (p =

.719), or health consciousness (p = .396).

The confounding effects proposed in hypotheses H8 and H9 were tested by means of a Spearman correlation test (effect confounding variable on the IV) and a Pearson correlation test (effect confounding variable on the DVs). A variable is confounding when it has a direct relationship with both the IV and the DV. However, only insignificant relationships were found in this regard. A detailed description of the Spearman Correlation Test can be found in Table 4.6.

Note that H6a, H7a, H7b, H8a, and H9a could not be tested as no significant relationship was found between the IV and DV.

Exploring the relationship between ad appeal, purchase intention, and the willingness to pay for different types of organic products. Does a picture add more value than a thousand words?

Does a picture add more value than a thousand words?