MSc International Business and Management Faculty of Economics and Business
Master thesis
Scandals and their effect on purchase intentions: why
consumers stay with scandalous firms
S2765950 | Jesper Slijkhuis | j.slijkhuis@student.rug.nl June 2016
2 ABSTRACT
Based on the resource based view and transaction cost economics there are reasons why a firm engages in a scandal, while proven a scandal declines performance. Purchase intentions by consumers are expected decline due to a scandal. While expected to decline, there are consumers that keep buying from scandalous firms. This research tries to explain via the aspects: country of origin, trust, brand loyalty and the aftermath of a scandal why consumers still show purchase intentions towards a scandalous firm. Via a deductive approach with experimental vignette survey’s consumers will be presented with a scandal to see if their purchase intentions change. A T-test can see if purchase intentions change after a scandal and a regression analysis shows which aspects will influence purchase intentions. Results of this research indicate that purchase intentions are lower after a firm engages in a scandal. Brand loyalty and the aftermath of a scandal tend to positively influence purchase intentions after a scandal. The results of this research adds to the literature on scandals and consumer buying behaviour and creates the opportunity for firms to react and reduce the negative influence of a scandal.
3 ACKNOWLEDGEMENTS
This thesis marks the end of my masters program at the University of Groningen. The time spent at the University was challenging. I learned some valuable skills and insights which I hope to be using in my future endeavors.
I would like to thank a couple of people because without them I wouldn’t be able to finish this thesis. First I want to thank Dr. M. Astarlioglu for his guidance throughout this research which was a pleasant experience.
Secondly I am grateful towards my family and friends who have helped and supported me throughout my whole education. Without the freedom and support given by my family I wouldn’t have made to where I stand now.
4 INDEX
1. INTRODUCTION ... 6
2. THEORY ... 8
2.1 Resource Based View ... 8
2.2 Transaction Cost Economies ... 9
2.3 Purchase Intentions ... 11
2.4 Country level: Country of origin... 12
2.5 Firm level: Trust ... 14
2.6 Firm level: Brand loyalty ... 17
2.7 Firm level: Aftermath ... 19
3. METHODS... 21 3.1 Survey ... 21 3.2 Focus group ... 22 3.3 Dependent variable ... 23 3.4 Independent variables ... 23 3.5 Control variables ... 25 3.6 Hypothesis testing ... 25 4. RESULTS ... 26
4.1 Focus group findings ... 26
4.2 Descriptive statistics ... 27
4.3 Robustness checks ... 28
4.4 Correlations ... 30
5. DISCUSSION ... 35
5.1 Discussion: Purchase intentions ... 35
5.2 Discussion: Country of Origin ... 35
5.3 Discussion: Trust ... 36
5.4 Discussion: Brand loyalty ... 36
5.5 Discussion: Aftermath ... 37
6. THEORATICAL AND MANAGERIAL IMPLICATIONS ... 38
7. LIMITATIONS AND FURTHER RESEARCH ... 39
7.1 Limitations ... 39
7.2 Future research... 39
8. CONCLUSION ... 41
9. REFERENCES ... 42
5 LIST OF TABLES, APPENDICES AND ABREVIATIONS
Tables
Table 1: Trust dimensions ... 24
Table 2: Descriptive statistics: Survey part one ... 28
Table 3: Descriptive statistics: Survey part two ... 28
Table 4: Pearson Correlations regarding part one of the survey ... 31
Table 5: Pearson Correlations regarding part two of the survey ... 31
Table 6: Results of One sample T-Test ... 32
Table 7: Results of regression analysis on the variables of part one ... 33
Table 8: Results of regression analysis on the variables of part two ... 34
Appendices Appendix 1: Survey ... 48
Appendix 2: Focus group protocol ... 59
Appendix 3: Linearity ... 63
Appendix 4: Q-Q Plots ... 65
Appendix 5: Heteroscedasticity ... 67
Appendix 6: Breusch-Pagan test ... 68
Appendix 7: Normality ... 69
Appendix 8: Multicollinearity ... 70
Appendix 9: Reliability tests ... 71
Abbreviations
RBV - Resource Based View TCE - Transaction Cost Economics COO - Country of Origin
6 1. INTRODUCTION
Almost a year ago the world learned about the Volkswagen Dieselgate scandal. Cars from the Volkswagen AG group were more polluting in terms of CO2 emissions compared to the specifications. Several engine types that were built in the last 10 years contained software that could regulate nitrogen oxides emission. The software was able to reduce nitrogen oxides emissions when tested but on under normal use the engine would produce more CO2 than accepted. Researchers discovered that some cars emitted almost 40 times the normal and permitted levels of nitrogen oxides (Russell, Gates, Keller, Watkins, 2015).
A scandal can be seen as “an action or event regarded as morally or legally wrong and causing general public outrage” (Oxford Dictionaries, 2016). The Volkswagen case is a governance scandal in which management diluted and withheld information for its shareholders and stakeholders (Hotten, 2015). Media are always active during a scandal which are a huge influence on consumers (Ahluwalia, Burnkrant, Unnava, 2000). This negative publicity tend to have an effect on consumers, their behaviour and their attitude towards a scandalous firm, resulting in lower purchase intentions (Ahluwalia, et al. 2000).
When purchase intentions are lower, sales will be lower which in the end will firm performance, turnover and market share (Ahluwalia, et al. 2000). While these facts give a negative view on purchase intentions and firm performance after a scandal, there are consumers still willing to buy and use a product from a brand or firm engaged in a scandal. When looking at the Volkswagen case they sold 25.5% of all new cars sold in Europe in 2014 and 24.8% in 2015, which is the year in which the scandal happened (Lavell, 2016). While sales declined, these numbers indicate that there are aspects at play which make a consumer still willing to buy products from Volkswagen, even after a scandal. Other cases, like BP’s Deepwater Horizon scandal, show similar results as the Volkswagen case (BP, Annual Report, 2011). BP was encountered in an oil spill scandal in which it created the worst oil spill in American history (Rothman, 2015). In the BP case sales did not decline after the scandal, indicating that consumers were also still willing to buy products from BP (BP, Annual Report, 2011).
7 when talking about purchase intentions (Keller, 2013; Nagashima, 1970; Howard, Sheth, 1969). This research will fill the gap in the literature by offering a research on aspects which have an influence on purchase intentions after a scandal.
This research will discuss several aspects which are expected to influence purchase intentions after a firm is encountered in a scandal. The aspects of which this research is based on are the COO effect which is treated as a country level variable to capture country level aspects of purchase intentions and trust, brand loyalty as firm level aspects to capture purchase intentions that originate from the specifics of a certain firm. These aspects are chosen due to their elaborate use in the literature on purchase intentions to create a comparable situation with purchase intentions after a scandal. A less frequent aspect which is linked to purchase intentions in the literature is the aftermath of a scandal. During the aftermath of a scandal a lot of the negative attention occurs, how a firm handles the negative attention is among the more important factors influencing consumers buying behaviours (Ahluwalia, et al. 2000). Therefore it is expected that the aftermath, as a firm level indicator will play a role in the purchase intentions of consumers after a scandal. Therefore the following research question: Does a scandal negatively influence purchase intentions and how does the country of origin effect, trust, brand loyalty and the aftermath of a scandal influence purchase intentions after a scandal?
This research will offer implications for multinational enterprises (MNE) to counter the effects of a scandal and increase the performance of a firm after a scandal. The results and implications can MNE’s help to pro-actively create strategies how to deal with the effects of a scandal. This research uses a deductive approach with experimental vignette survey’s to see if consumers still have purchase intentions after a scandal and what aspects motivates them to keep having purchase intentions.
8 2. THEORY
Scandals tend to hurt firm performance due to their ability to create a lot of negative attention towards the firm, but firms are still being encountered in them (Ahluwalia, et al. 2000). To further understand why firms are encountered in scandals and to create a link between a scandal and purchase intentions the Resource Based View (RBV) will be used as well as the theory of Transaction-cost economics (TCE). According to both these theories a sustainable competitive advantage is a way for a firm to create higher performance (Kuhn, Ashcraft, 2003). Both theories try to create a competitive advantage from a different perspective, the RBV from a perspective of resources within a firm and TCE from a perspective of transactions in which a firm or consumer engages in.
2.1 Resource Based View
According to the RBV, firms exist because they can provide economies of scale and scope which originates from the resources the firm has (Kuhn, Ashcraft, 2003). According to this theory all firms are seen as a unique bundle of tangible and intangible resources and capabilities, which are acquired and developed over time (Esteve-Pérez, Mañez-Castillejo, 2008). These resources are results from strategic choices and commitments and determine the firms performance (Esteve-Pérez, Mañez-Castillejo, 2008). When the resources of the firm are valuable and a few or no other firm has them (rare), and if other firms find it too costly or difficult to imitate these resources, then the firm controlling these resources have a source of a sustainable competitive advantage. (Kozlenkova, Samaha, Palmatier, 2013).
In the light of scandals, the RBV provides an insight in scandals via organizational cultures that foster and enable deceptive practices (Kuhn, Ashcraft, 2003). Kuhn and Ashcraft (2003) discuss that a possible outcome of aligning actors and structures to create a competitive advantage can create a high market success as well as do damage to the firm. it is possible that company cultures, routines, and identities that produce valuable core capabilities run counter to those that produce ethically responsible workplaces (Jackall, 1988).
9 valuable, rare and inimitable resources can come from a morally or illegal source (and therefore have the ability to create a scandal) as long as they create a competitive advantage. 2.2 Transaction Cost Economies
The economic view of TCE starts with the question why firms exist compared to simply have markets in which goods and services are traded (Kuhn, Ashcraft, 2003). TCE tries to explain why a certain transaction favors a certain particular form of transaction instead of others. The basic principle of TCE explains that individuals or firms like to conduct transactions in the most economical way possible (Thompson, Yuanyou, 2005). In a market which is perfect the coordination of transactions go without problems and are cost-free, due to the fully competitive nature in prices and salaries (Williamson, 1971). Actual markets are not perfect because all the transaction have costs. Due to this imperfection the market will possess different situation and circumstances in which transactions take place. Under some circumstances, transaction cost are lower if a transaction takes place in a market (Thompson, Yuanyou, 2005). In other situations, transaction cost will be lower if managers from a firm coordinate the transaction (Thompson, Yuanyou, 2005). Through a TCE lens, a firm can be seen as “a nexus for contracting relationships characterized by divisible residual claims on assets and cash flows (i.e., ownership), which actors can generally sell without the permission of other contractors” (Jensen, 2000).
10 In the light of a scandal, according to the TCE a competitive advantage can be created by choosing options which have the lowest transaction costs (Kuhn, Ashcraft, 2003). Choosing strategies or creating relationships that have the lowest transaction, while legal or illegal, can therefore be seen as a viable strategy to apply in creating a competitive advantage (Kuhn, Ashcraft, 2003). Strategies that are not in line with law and regulations creates the possibility to trigger a scandal. In other words, the costs to engage in a scandal can be lower than to do transactions in a legal sense which doesn’t have the ability to trigger a scandal. Bounded rationality and opportunism can also play a role due to the fact that managers can make decisions while not able to see the consequences or make choices based on self-interest, therefore making a decision that can harm the firm (Simon, 1997).
According to TCE, The attributes of an transaction are the frequency, uncertainty and asset specificity (Williamson, 1981). Transactions can be rare or frequent, have low or high uncertainty, or involve specific or non-specific assets (Williamson, 1981). Uncertainty of a transaction arises from the difficulty to predict the other party’s actions in the transaction due to opportunism and bounded rationality, the higher the uncertainty the higher the transaction costs. (Thompson, Yuanyou, 2005). The frequency of a transaction indicates how rare or frequent the transaction takes place. (Thompson, Yuanyou, 2005). The more frequent the transaction takes place the less transaction costs will arise due to the fact less costs are to be made (for example search costs) to find different offerings (Dahlman, 1979). The asset specificity refers to ability of the assets to be used by alternative users or put to alternative use. The higher the asset specificity the harder it is to put it to alternative use and therefore result in higher transaction costs (Thompson, Yuanyou, 2005).
11 Both RBV and TCE show motivations why a firm can be encountered in a scandal. TCE also has the possibility to expand to consumers and explain why consumers stay or leave certain firms for their purchases before or after a scandal. To fill the gap in the research, this research will expand on why consumers still have purchase intentions towards a firm after that firm is encountered in a scandal. The following subjects are seen as important determinants to fill this gap.
2.3 Purchase Intentions
Purchase intentions are an part of consumer behaviour towards a brand or a firm. (Howard, Sheth 1969). Purchase intention can be defined as the probability that customers in a certain purchasing situation choose a certain brand over another brand active in a certain product category (Crosno, Freling, Skinner, 2009). Howard and Sheth created one of the earlier theories of purchase intentions (Laroche, Chankon, Lianxi , 1996). The tried to explain why people chose a certain brand, because they assume that this choice is not random but systematic. March and Simon (1958) also claimed that buying behaviour is rational and that it is contained within a person’s bounded rationality.
The decision to buy a product from a certain brand is formed by a set of motives, several alternative course of actions and decision facilitators (Howard, Sheth, 1969). Motives are focussed on the product class of a product and reflect the underlying needs of the buyer (Howard, Sheth, 1969). The alternatives are the various other brands that are a possibility to satisfy the needs of a consumer (Howard, Sheth, 1969). Decision facilitators are a set of rules a consumer uses to match his motives and his means of satisfying his needs (Howard, Sheth, 1969). These facilitators are influenced by the information from the buyer’s environment and from experience buying and using a brand. (Howard, Sheth, 1969).
12 of a firm being in a scandal will influence the decisions in the buying behavior and therefore purchase intentions of a consumer.
When purchase intentions are lower, sales figure will be lower which in the end hurts firm performance. As mentioned in the introduction, year-on-year sales of Volkswagen AG group declined from 25.5% to 24.8% due to the scandal they were encountered in. Another but different scandal is the scandal of Tiger Woods, which was engaged in personal scandal where he was accused of being a serial cheater in his love life (Rohof, 2015). The performance of this athlete and the effects of the scandal resulted in lower purchase intentions with its main sponsor Nike (Farrell, Karels, Monfort, McClatchey, 2000). Another study by Knittel and Stango (2010) found confirming results with the ones of Farrell et al. (2000). These results proved that after the scandal of Tiger Woods was made public, Nike’s market value fell significantly (Rohof, 2015). Due to these facts it is expected that purchase intentions will decline when a scandal have taken place therefore the first hypothesis will be: Hypothesis 1: A scandal will result in lower purchasing intentions towards a firm in which the scandal took place.
Purchase intentions reflect the buying behavior of consumer. Compared to sales numbers, purchase intentions is an imperfect measure to predict actual sales, and therefore performance of a firm (Morwitz, Steckel, Gupta, 2007). Purchase intentions originate from consumers that are choosing or may choose in the future to buy or not to buy (Howard, Sheth 1969). Therefore intentions still have the ability to be influenced by a scandal, media or marketing or other variables. This makes purchase intentions more suited in this research than sales numbers to compare the purchasing behavior of consumers before and after a scandal.
2.4 Country level: Country of origin
13 Romeo, 1992). COO is often described as ‘Made In’ or ‘Manufactured in’ influencing the image of a country and therefore influencing a consumers evaluation of a product or brand that is made in a certain country (Chapa et al., 2006).
The stereotyped images about countries are held by consumers as information sources to identify and judge products from a certain country (Suh. Smith. 2008). The evaluation, the feelings and attitudes of consumers towards products and brands mostly based on opinions from their own and from the public (Aaker, 1991). These feelings and the attitudes consumers have towards a product or brand are made via brand associations (Aaker, 1991). Brand associations are the informational nodes linked to the brand node in memory that contains the meaning of the brand for consumers, which includes the perception of brand quality and attitudes towards the brand (Keller, 2013). The evaluation of the stereotype that consumers attach to products of a specific country is therefore a result of brand associations. The COO effect provides an attribute summary of a product or brand within a decision making process (Suh. Smith, 2008). The COO effect is therefore helpful in product evaluating purposes, especially when no other information is known about a product category. The stereotyped images can come from different backgrounds; economic, history, political, everything that makes up the characteristics of a certain country (Suh. Smith, 2008).
Different studies claim that the COO effect can influence the evaluation of a product and brand to the extent it can change the purchase intentions of a consumer (Crosno et al., 2009). Crosno et al. (2009) state that the COO effect can be used as a cognitive shortcut when other information of a product is scares. Therefore the COO effect is of importance when choosing a brand or product, certainly when there is low product knowledge (Crosno et al., 2009). Solomon (2004) states that COO has an influence on the decision making process as well, as in the assessment of products. Keller (2013) states that a consumer can have a positive association with a brand or a negative association with a brand. COO can be seen as a part of such an association and therefore able to influence behaviour towards a brand or product. The definition of Samiee (1987) adds to the this claim and states that COO is: "any influence, positive or negative, that the country of manufacture might have on the consumer's choice processes or subsequent behavior".
14 1982), stating that a products “home country” is the country in which it is manufactured. In this research the former counts, the geographical location of the headquarters indicate the origin of a firms products.
While studies confirm that COO does exist in purchasing behaviour, there are studies that argue COO doesn’t exist (Gregory, Ross, 1994). If COO would exist, these studies claim that it’s only has a minor significance on purchase decisions (Gregory, Ross, 1994). While some of these studies deny the existence of COO, the majority of the studies confirm of COO existence (Gregory, Ross, 1994). According to the statements of Gregory and Ross (1994) the magnitude and mechanism of COO still seems to be unresolved.
Linking COO effect to a scandal it is not expected that the COO effect will suffer due to a scandal. The COO effect is an image and stereotyping of a complete country’s strength and weaknesses in production and marketing, not an image specifically of the firm in which the scandal occurred (Roth, Romeo, 1992). As Suh and Smith (2008) mentioned, the stereotyped images can come from different backgrounds; economic, history or political. According to this theory, it is not expected that a favorable COO effect will suffer due to a scandal and therefore it is expected that the favourable COO will have an effect on purchase intentions after a scandal. Therefore the hypothesis:
Hypothesis 2: A favourable Country of Origin effect will have a positive effect on purchase intentions after a scandal.
Related to purchase intentions, COO effect is able to capture a product or brand evaluation, based on the overall perception and stereotypes of a country which influence decision making processes related to purchase intentions. (Roth, Romeo, 1992; Solomon, 2004). Due to the fact COO effect is not expected to be harmed by a scandal and the fact that the COO effect can connect on an individual level with consumers, it is chosen as a variable in determining purchase intentions after a scandal.
2.5 Firm level: Trust
15 Trust is a large concept that has been researched in different studies and contexts. Trust is therefore measured and defined differently in different contexts of studies. Rousseau, Sitkin, Bur and Camerer (1998) define trust as the psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behavior of another. (Rousseau et al., 1998). According to Delahaye Pain (2003) trust can be seen as a relationship between individuals. From a more elaborative point of view Romano (2003) discusses four characteristics which are at the core of trust.
The first characteristic Romano discusses is that trust needs to be seen as an attitude (Jones, G. R. George, J.M., 1998). Trust can be seen as an attitude because it can be found in the thoughts, feelings and the intentions in behavior of people. (Clark, Payne, 1997). The second characteristic of trust is the social background of trust. Trust is embedded in a social context because it shows someone’s attitude towards another party in a given situation. (Romano, 2003). The social context can be explained by a trustor that is directed toward a recipient, a target, within a setting, the context. Trust is an evaluation of the perceived influence of the trustor, target and context on the outcome of a certain situation created between them. (Romano, 2003). Simply said, trust is concerned with how the interests of a person are influenced when a trustor, target and context have an interacting effect. (Bhattacharya, Devinney, Pillutla, 1998). The influence on the interests is the outcome of the social context and the communicational behavior within that context (Delahaye Paine, 2003).
16 and interact with, another party (Romano, 2003). While trust can be functional it doesn’t mean that trust has to be productive and the lack of trust counterproductive. (Jones, George, 1998). Trust is a subjective experience and is intended to work for the trustor’s own best interests. (Romano, 2003) While it serves the owners best interests it can make choices that won’t be beneficial to a goal or target within a certain environment.
The characteristics of Romano (2003) explain and dissect trust to give a better understanding on trust. This research is primarily interested in what trust does with purchase intentions; therefore there need to be a link of trust and consumer behavior. Howard and Sheth (1969) were one of the first to add the construct of confidence to purchase intentions (Laroche, Sadokierski, 1994). Confidence is seen as “the buyer’s degree of certainty that his or her evaluative judgement of the brand is correct” (Howard, 1989). Bennet and Harrell (1975) interpreted this definition in two ways: it may refer to the buyer’s confidence in the brand or it may refer to the buyer’s confidence in his/her ability to judge or evaluate attributes of the brands (Bennett, Harrell, 1975). The latter meaning of confidence can be seen as the perception of risks underlying decision making (Laroche, Chankon, Lianxi, 1996). This is also in line with a statement of Howard and Sheth (1969) saying that the confidence is the buyer’s belief that she/he can estimate the payoff of purchasing a particular brand. Urbany, Dickson and Wilkie (1989) follows up on the literature of consumer confidence with the conceptualization of two types of confidence, knowledge confidence and choice confidence (Laroche, Chankon, Lianxi , 1996). Knowledge confidence reflects as consumer’s certainty regarding what is known about the brands under consideration, such as available attributes, importance of these attributes and the performance of the brands (Urbany, Dickson, Wilkie 1989). Choice confidence reflects a consumer’s certainty regarding which brand to choose (Urbany, Dickson, Wilkie 1989). Results of the study of Laroche, Chankon, Lianxi, (1996) indicate that there is a relationship between the confidence in brands and purchase intentions. Therefore trust can be seen as a determinant of purchase intentions.
17 2013). This may sound contradicting to the claim that trust isn’t necessarily influenced by a scandal but a scandal doesn’t necessarily influence decision facilitators for which a consumer chose the firm or the brand in the first place (Howard, Sheth, 1969). In other words, The set of rules consumers use to match their motives and their means of satisfying their needs doesn’t have to be influenced by a scandal (Howard, Sheth, 1969).
From the TCE perspective, high trust means low transaction costs because of low uncertainty (Thompson, Yuanyou, 2005). When decision facilitators of a consumer about a firm or product are not influenced by a scandal, doing a transaction somewhere else can mean higher transaction cost due to increased uncertainty and risk when engaging with another firm (Thompson, Yuanyou, 2005). Due to the theory of having higher transaction costs when buying from another firm it is expected that consumers stay with the current firm, even though they are encountered in a scandal. Therefore trust that has been accumulated over the time by a firm can be helpful to keep consumers having purchase intentions (Keller, 2013). Therefore the hypothesis:
Hypothesis 3: Favourable Trust will have a positive effect on purchase intentions after a scandal.
2.6 Firm level: Brand loyalty
Brand loyalty can be defined as a deeply held commitment to rebuy or patronize a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brand-set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior (Oliver 1999). The definition consists of two parts, one is behavioral and the other is attitudinal. The behavior that originates from brand loyalty is repurchasing products or services from a similar brand, also called the purchase loyalty (Chaudhuri, Morris, 2001). Attitudinal brand loyalty includes a degree of commitment to a brand that is due because of the unique value associated with a brand (Chaudhuri, Morris, 2001). For a firm high brand loyalty among its costumers leads to higher performance due to the price premiums firms can ask for the products and due to having a higher market share (Oliver 1999).
18 recognize a brand they are able to consider a brand when making a purchase intention, when it is not recognized they won’t (Baker W, J, & Nedungadi, 1986). Brand awareness creates an association with which a consumer can recognize a certain brand. (Muhammad, Muhammad, Hafiz, Usman, Noor, Muhammad, Saleh, 2013). When consumers repeatedly buy a product from a similar brand, brand loyalty is created (Chaudhuri, Morris, 2001).
While replicated purchases can indicate brand loyalty it is not the same because brand loyalty does not only facilitate replicated purchases but also the frequency of purchases (Muhammad et al. 2013). Consumers who are loyal to a brand will make purchases in a confident and trustful way without considering. The trust in a brand is the willingness of the average consumer to rely on the ability of the brand (and its products) to perform its stated function (Chaudhuri, Morris, 2001). This type of trust is seen as a calculative process based on the ability of a brand to meet its obligations and on an estimation of the costs versus rewards it is interesting staying in the relationship (Doney, Cannon, 1997).
Muhammad et al. (2013) conducted a study that shows that brand loyalty contributes in the purchase intentions of consumers The results show that brand loyalty has a significantly positive association on the purchase intentions of consumers (Muhammad et al. 2013). They conclude that purchase intentions can be enhanced by creating brand loyalty. This indicates that brand loyalty is indeed influencing purchase intentions without the presence of a scandal. TCE theory can explain why consumers may have purchase intentions after a scandal. Brand loyalty is chosen as a variable because it can be substituted by frequency of a transaction mentioned in the TCE theory (Thompson, Yuanyou, 2005). The more frequent the transaction takes place due to favorable brand loyalty, the less transaction costs will arise due the trust to make purchases without considering (Keller, 2013). According to TCE switching brands or firms results in higher transaction costs due to search costs (Dahlman, 1979). When switching firms or brands, the trust to pick the brand without considering that is built via brand loyalty needs to be rebuilt, which can be seen as search costs.
19 expected that brand loyalty will have a positive effect on purchase intentions after a scandal, therefore the hypothesis:
Hypothesis 4: Favorable Brand loyalty will have a positive effect on purchasing intentions after a scandal.
2.7 Firm level: Aftermath
The aftermath of a scandal is the period of time that occurs after a scandal in which the signs and results of the scandal become apparent. The Oxford Dictionaries (2016) define aftermath as “the consequences or after-effects of a significant unpleasant event”. The aftermath of a scandal often leads to a situation where often a lot of change occurs for a firm (Argawal, Cooper, 2016). An example for such change is a large stock decline after a scandal is announced to the public (Palmrose, Richardson, Scholz 2004). When taking the example of accounting scandals, the research of Palmrose et al. (2004) indicates that there are different responses on a scandal and different ways to handle a scandal. In the case of accounting scandals this means different ways of restating the financial sheets, different ways in explaining why they need to be restated and for how much they need to be restated (Palmrose et al, 2004).
Previous governance changes due to scandals happened internally and externally of the firm. The Sarbanes-Oxley act is a governance reform act that had a large external impact on firms, putting laws and guidelines into action to change governance in firms (Agrawal, Cooper, 2016). Internally, scandals can have serious consequences for the top management (Agrawal, Cooper, 2016). These changes in management can lead to costs or benefits for the firm (Palmrose, et al. 2004).
20 of Palmrose, Richardson and Scholz (2004) indicating that a large stock price drop occurs when a firm is encountered in a scandal. Fourth, firms that experience a distress in finance tend to change their management faster (Agrawal et al. 1999).
While there are reasons for management replacement, changes in management will only be expected to happen when the benefits will exceed the cost (Agrawal et al. 1999). Scandals do not always create managerial change within a firm (Agrawal et al. 1999). Therefore consequences for management are not always similar. Results of the research by Agrawal and Cooper (2016) state that when a restatement on the balance sheet takes place to resolve financial scandals it doesn’t have to mean that there is an increase in managerial turnover due to general bad governance within the company.
As mentioned in the introduction, a scandal creates a lot of negative attention, how a firm handles the negative attention is among the more important factors influencing consumers buying behaviours (Ahluwalia, et al. 2000). Therefore it can hypothesized that there is a link between management replacement in the aftermath of a scandal and purchase intentions after a scandal. When keeping to the theory of Argawal et al. (1999) reputational gains can be a reason for management replacement. Reputational gains will create a better image for a firm or brand (Keller, 2013). This will in its turn create more confidence among consumers in a brand or firm (Keller, 2013). More confidence will lead to more purchase intentions as mentioned by the research of Howard and Sheth (1969). Keeping to this theory it is expected that consumers respond to governance changes in a firm, especially management replacement after a scandal took place. Therefore:
21 3. METHODS
To put the hypotheses to the test, a deductive approach is used. This approach is used because of the amount of theory concerning the variables and to avoid any risk of making wrong assumptions and therefore wrong hypotheses. To answer the hypotheses a quantitative approach is being used via surveys. Surveys will gather the data that fits this research and its research question instead of using second-hand data. The results of these surveys will be put to use in a T-test and regression analysis to test the hypotheses.
3.1 Survey
The surveys used for this research will be designed similar to a vignette study. Vignettes are short stories about individuals or situation from different sources and can be used to test people’s perceptions, beliefs and attitudes (Hughes, 1998). Participants are often asked what they would do in a certain situation, based on the story told by the vignette (Hughes, 1998). Due to the use of a vignette this research can be classified as an experimental study (Nosanchuk, 1972). Second-hand data is available but not in conjunction with each other as necessary for this research, therefore surveys are used to create data that fit’s this research. An experimental vignette creates the opportunity to measure purchase intentions before and after a scandal in a similar place and time. This isn’t possible without taking this experimental route because other methods need to be measured in two places in time to be able to compare purchase intentions and would include a real scandal. Due to not having the ability to time a scandal, a vignette survey makes this an appropriate experimental method for this research. The survey will consist of two parts with questions and a Likert answering scale of seven possibilities from completely disagree to completely agreeing. The survey can be found in Appendix 1: Survey. In part one of the survey people are being asked questions about a Scandinavian MNE. Questions will be about the COO, the trust in the MNE, their brand loyalty and their purchase intentions. After this, a vignette will discuss a scandal in which the Scandinavian firm supposed to be engaged in. Part two of the survey will start with the same questions as used in part one of the survey. To test how the aftermath of the scandal effects purchase intention’s, questions regarding the aftermath will only be asked in part two of the survey. With part one and part having similar questions, answers can be compared to see if purchase intentions decrease after a scandal and which variables affect purchase intentions before and after a scandal (Hughes, 1998).
22 The Scandinavian firm has been chosen for the vignette because of having a history without a business scandal that attracted wide attention. Therefore respondents have a clean perception on the firm. The scandal that is mentioned in the vignette is being made up to create a situation where respondents will hear about the scandal for the first time. Creating such a situation will enable a real and honest reaction on the scandal because for all respondents it will be the first time they will be confronted with the scandal.
This survey was targeted on consumers. The sample group that has been surveyed consisted of 130 respondents. In total, 61 respondents finished the survey, which makes the sample size for this research 61. This indicates that there was a dropout rate of 53,1%. This may be due to the longevity of the survey or an translation issue. The survey was in English, distributed among mostly Dutch respondents. The survey wasn’t translated to Dutch due to the fear of translation issues. The survey was digital and therefore easy to distribute via social media. To gain access to social media respondents had to have access to a computer or smartphone and an internet connection. To buy or rent these appliances a respondent had to show purchasing behaviour to gain access to them. Therefore it can be expected that all respondents can be seen as consumers with purchase behaviours. Social media also creates the opportunity for participants to share the survey in their social environment. Sharing the survey is possible due to the fact that all people using social media can be seen as consumers and are therefore within the target group of the survey.
3.2 Focus group
To further explore the themes of scandals and purchase intentions a focus group is organised. The focus group is designed in a similar sense as discussed in the book of Morgan (2013). For more practical details about the design and setting of the focus group the paper by Krueger (2002) is used.
23 regarding the aftermath are used in conjunction with theories discussed in the theory to create survey questions regarding the subject.
To start off, the group received a hand out with the protocol of the focus group (Appendix 2: Focus group protocol). While not in line with the literature of Morgan (2013), the focus group consisted of people with different educational backgrounds to promote diversity within the discussion. All people are consumers with purchase intentions, a diversified group would therefore lead to different answers and a broad perspective on the questions asked. In the results part of this research the findings of the focus group will be presented.
3.3 Dependent variable
The dependent variable for this research is purchase intentions. As mentioned in the theory, purchase intentions can be defined as the probability that customers in a certain purchasing situation choose a certain brand of a product category (Crosno et al., 2009). A goal of this research is to show that there are less purchase intentions after a scandal than before a scandal. To test if this relationship exists questions related to purchase intentions will be asked in part one and part two of the survey. Questions used in the survey are based on questions used in earlier research on purchase intentions (Spears, Singh, 2004). The dependent variable will be used similar in both parts of the survey.
3.4 Independent variables
The first independent variable is COO. Questions for this variable, in both part one and two of the survey, represents the respondents overall perception or stereotype that they attach to products of a specific country, based on their views of the country’s strengths and weaknesses in production and marketing (Roth and Romeo, 1992). The questions related to COO will discuss the image on the people and the products of the country (Parameswaran, Pisharodi, 1994). These questions together will form the variable COO and will represent the respondents overall perception.
24 Competence Belief that an entity has the ability to do what it says it will do
Integrity Belief that an entity is fair and just
Dependability/Reliability Belief that an entity will do as it says what it has promised
Openness and Honesty Belief that the information shared is enough accurate in a sincere and communicated appropriately
Identification Belief that there are common goals, norm, values and beliefs with the entity
Commitment Belief that the both entities believe the relationship is worth spending energy on to maintain
Table 1: Trust dimensions (Delahaye Paine, 2003).
In this research these specific dimension are proposed to explain the trust dimensions between a consumer and a firm and the products/services the firm selling. These dimensions are chosen because they create the possibility to align and classify questions with the respective dimension.
25 Brand loyalty is the third independent variable that will be measured in both part one and two of the survey. Brand loyalty can be explained as a deeply held commitment to rebuy or patronize a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brand-set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior (Oliver 1999). The questions related to the variable of brand loyalty are extracted from a research by Bobâlcă, Gătej and Ciobanu (2012) in which a brand loyalty measure is developed.
The fourth independent variable is the aftermath of a scandal. This variable is only used in part two of the survey to see how respondents react on governance changes mentioned in the vignette as a result of a scandal. The questions related to this topic are extracted from the focus group discussing the topic of scandals and their effects on purchase intentions. The questions that are asked in the survey are based on the answers given in the focus group regarding management replacement. The focus group concluded that replacing management, while maybe ceremonial or not, will have an influence on the purchase intentions of consumers. It indicates that the firm has seen the problems and tries to own up to it. Therefore the questions asked in the survey will focus on the replacement or non-replacement of the CEO and board members.
3.5 Control variables
To control for certain aspects the control variables of age, gender, work experience and highest obtained degree will be inserted in the regression. These controls are used because of their ability to see if people from different places in a society react different to the questions asked. (Frölich, M. 2008). In this case gender is used as a dummy variable where male is 1 and female is 2.
3.6 Hypothesis testing
26 4. RESULTS
The following chapter discusses the results and will test the hypotheses. This part will start off with the findings from the focus group, the descriptive statistics followed by robustness checks, a correlation table and the regression analysis itself.
4.1 Focus group findings
Following are findings from the focus group regarding the research subject. To start of participants were asked what they expected to be the effect of a scandal on with purchase intentions, respondents all agreed that a scandal would create a loss in brand image, it would create negative perceptions of consumers about the brand and quality and a decline in purchase intentions. These statements are in line with Hypothesis 1 which states that: A scandal will result in lower purchasing intentions towards a firm in which the scandal took place. Although confirmative, participants added that the effect of a change in their purchase intentions would depend on the size and type of the scandal.
When asked about the COO effect participants answered that it would influence purchase intentions but a scandal wouldn’t necessarily change its influence. A quote: “German cars are known for their quality…the Volkswagen scandal doesn’t mean that Opel does a worse job regarding quality”. Followed by “I don’t think: oh, because Volkswagen has a scandal I won’t buy a Opel anymore, I need to switch to a Japanese car” This is in line with the main reasoning for hypothesis 2, saying that a favorable COO won’t be affected by a scandal and therefore still has the positive effect on buying behavior.
Regarding trust after a scandal, participants agreed with each other trust would be lower towards a brand or firm. One participant added: “It makes me feel as if I am not treated as an equal human being”. Participants concluded that a decrease in trust would depend how bad the reasons why they trusted a firm or brand in the first place, would be damaged. If damaged, they expected that consumers wouldn’t trust a firm after a scandal. When not damaged, there wouldn’t be a reason not to trust the firm. This is in line with the decision facilitators theory of Howard and Sheth (1969).
27 related to giving the company a second change. In this context, brand loyalty would lead to that second change and therefore increase purchase intentions after a scandal. A quote by a participant adding to this statement: “I always bought Adidas predator football shoes, after a bad line of shoes with poor quality is still purchase them because I always liked the comfort before they had a bad line of shoes”.
While the former questions were of confirmatory nature to see if the reasoning of consumers fit the theories, the questions related the aftermath are used as input for the survey questions regarding the subject. Participants answered they expected that management replacement would have an effect on purchase intentions. They mentioned that, while ceremonial, it feels righteous that someone gets the blame and therefore help the scandal to be resolved. The participants mentioned that a management replacement wouldn’t necessarily solve the scandal but would still looks good towards consumers, as if the scandal is being resolved. This is in line with literature by Agrawal et al. (1999) which state that management replacement takes place to regain reputational capital. When no management replacement would take place participants felt like the firm wouldn’t own up to its wrong doings. Participants mentioned that a quick reaction would be the best because it would decrease the amount of negative attention by the media. The section of the independent variables will discuss which aspects regarding the aftermath are used as the independent variable’s survey questions.
4.2 Descriptive statistics
In Table 2 and Table 3 the descriptive statistics of both part one of the survey as part two of the survey are presented. The tables represent the number of respondents, minimum, maximum, mean and standard deviation of each survey part respectively. Compared to part one of the survey, part two’s dependent variable, purchase intentions, has a lower mean, which means that respondents have lower purchase intentions after they read of the alleged scandal then they did before the scandal. To confirm its significance, a T-test will be used when testing the hypothesis. The standard deviation in part two is also smaller than in part one. This means that respondents of part two are more agreeing with each other regarding purchase intentions because the data points lay closer to the mean than in part one of the survey.
28 N Minimum Maximum Mean Std. Deviation
Dependent Variable Purchase intentions 61 4,00 7,00 5,9057 ,87341 Independent Variables Country of Origin Trust 61 61 3,00 2,57 7,00 7,00 5,2404 5,3466 ,73639 ,73746 Brand Loyalty 61 1,50 7,00 4,9549 ,94704 Control Variables Age 61 18 71 29,08 11,671
Work experience (in years) 61 0 34 8,89 8,337
Gender 61 1 2 1,62 ,489
Highest obtained degree 61 1 8 5,90 1,338
Valid N 61
Table 2: Descriptive statistics: Survey part one (Source: Author)
N Minimum Maximum Mean Std. Deviation Dependent Variable Purchase intentions 61 4,00 7,00 5,5492 ,85491 Independent Variables Country of Origin 61 2,33 6,44 5,0164 ,78798 Trust 61 1,86 6,71 4,6557 ,97009 Brand Loyalty 61 1,25 7,00 4,7910 ,97807 Aftermath 61 2,71 6,86 4,6956 ,93622 Control Variables Age 61 18 71 29,08 11,671
Work experience (in years) 61 0 34 8,89 8,337
Gender 61 1 2 1,62 ,489
Highest obtained degree 61 1 8 5,90 1,338
Valid N 61
Table 3: Descriptive statistics: Survey part two (Source: Author)
4.3 Robustness checks
29 makes is that the data needs to be linear. To see if the data of the dependent and independent variable is linear, a scatterplot (Appendix 3: Linearity) of every independent variable linked to their respective dependent variable is made. While the data is dispersed in the scatterplots, eyeballing these graphs doesn’t indicate other patterns that would interfere with linearity. Due to this fact the data can be seen as linear. To see if there are significant outliers in the dataset Q-Q Plots are made (Appendix 4: Q-Q Plots). Eyeballing these graphs doesn’t indicate any significant outliers. There is one observation outside of the linear line drawn in the plots but the observation are still in line with the rest of the observations instead of laying in the far corners of the graphs, therefore this assumption can be accepted.
Berry (1993) indicates that the data inserted in a regression should not show heteroscedasticity, but homoscedasticity. To test if the data shows any heteroscedasticity scatterplots are made for each of the dependent variables containing the standardized residuals and standardized predicted values which can be seen in Appendix 5: Heteroscedasticity. Eyeballing these scatterplots, no patterns can be found which would indicate heteroscedasticity. To further examine if heteroscedasticity is present, a Breusch-Pagan test has been applied on the data. The Breusch-Pagan test, tests if the estimated variance of the residuals are depend on the independent variables. When this is the case this would indicate heteroscedasticity (Breusch, Pagan, 1979). As can be seen in Appendix 6: Breusch-Pagan test, the dependent variable of both the parts of the survey have a non-significant F-value. In this case the null hypothesis can be accepted and it is therefore accepted that the data shows homoscedasticity instead of heteroscedasticity.
30 multicollinearity are presented. As for the variables of part one and two there are no indications of multicollinearity due to the fact the VIF’s of these variables are all lower than 5. The highest VIF is 3,043 which indicates there are no problems with multicollinearity. The last robustness check is done to test if all the survey items and comprised sum variables are reliable. Cronbach’s Alpha is a measure which is used to measure internal consistency of the survey items. Simply said, it checks how closely related the items measuring a certain variable are as a group. The Cronbach Alpha tests indicated that all questions where reliable and higher than the threshold of α=0.7. One sum-variable (PurchaseIntentions2) scored a α=0.651 on the test. A survey item could be deleted to lift the α=0.651 over the threshold of α=0.7 to make the sum-variable reliable. In this case the item wasn’t deleted due to the fact that the exact same sum-variable with the same survey items is used in part one of the survey. In part that sum-variable scored above the α=0.7 threshold which indicates that it is reliable. Deleting an item from the variable in part two would mean that the item would be removed from the same sum-variable in part one, in order to keep part one and part two comparable. In order not to delete a sufficient reliable item in part one the choice is made to keep the survey item in part two, accepting the α=0.651 reliability In part two. In Appendix 9: Reliability tests, the results of the tests can be found.
4.4 Correlations
31
Table 4: Pearson Correlations regarding part one of the survey (Source: Author)
Table 5: Pearson Correlations regarding part two of the survey (Source: Author)
When looking at the correlations table of part two of the survey (table 5) it is noticeable that Trust 2 and Brand Loyalty 2 are significantly correlated (at p < 0.01) with Purchase Intentions 2. The variables Country of Origin 2 and Aftermath are also significantly correlated (at p < 0.05) with Purchase Intentions 2. Thus far these results are in line with several hypotheses.
4.5 Hypothesis testing
32 compare the means of both the dependent variable of part one (PurchaseIntentions1) and the dependent variable of part two (PurchseIntentions2) to see if there is a significant change between them. When a significant difference is detected it will indicate that respondents have different purchase intentions after a scandal compared to before a scandal.
Table 6: Results of One sample T-Test, testing PurchaseIntenions2 on the mean value of PurchaseIntentions1 (Source: Author)
Table 6 show the results of the T-test. The t-value resulting from the analysis is -3,257 which is significant at .002 (p < 0.05) with a 95% confidence level. Therefore the null hypothesis can be accepted which states that the group of PurchasIntentions1 is not the same as the group of PurchaseIntentions2. The mean difference is -,35652 which indicates that the mean variable purchase intentions decreased by -,35652 after respondents read about the scandal in the survey. Combining these results indicates that hypothesis 1 is supported.
33 Hypothesis 2 is formulated as: A favourable Country of Origin effect will have a positive effect on purchase intentions after a scandal. As can be seen in the model 2 of Table 8 there is no significant relationship between purchase intentions and the COO effect. Therefore hypothesis 2 is not supported. Also notable in Table 8 is that the coefficient that expresses the relationship of trust and purchase intentions isn’t significant. Therefore, hypothesis 3 formulated as: Favourable trust will have a positive effect on purchase intentions after a scandal is not supported.
Hypothesis 4 is formulated as: Favorable brand loyalty will have a positive effect on purchasing intentions after a scandal, is represented by the variable brand loyalty in both Table 7 and in Table 8 it is positively significant with purchase intentions. This means that when brand loyalty increases it will positively increase purchase intentions with 0.388 with a significance of (p < 0.001). In both model 2 of Table 7 and 9 this relationship is traceable. Table 8 shows that when brand loyalty increases it will positively increase purchase intentions with 0.321 with a significance of (p < 0.01). The results of show support for hypothesis 4.
34
Table 8: Results of regression analysis on the variables of part two. Dependent variable: PurchaseIntentions2
Hypothesis 5, which is formulated as: A favorable Aftermath of a scandal will have a positive effect on purchase intentions after a scandal, can also be supported with the results presented in table 8. The effects of the aftermath as discussed in the theory, will positively increase purchase intentions with 0,260 with a significance of (p < 0,01). This result shows support for hypothesis 5.
35 5. DISCUSSION
The discussion will elaborate on the findings that are presented in the results. The findings will be interpreted to be able to further elaborate on the results.
5.1 Discussion: Purchase intentions
The goals of this research was to investigate the relationship between a scandal and purchase intentions resulting from the scandal. This research confirms that purchase intentions decline after a scandal due to the support for hypothesis 1 which states that: A scandal will result in lower purchasing intentions towards a firm in which the scandal took place. The results supporting hypothesis 1 are in line with literature which expects that a scandal will influence purchase intentions negatively just as the negative attentions of a scandal leads to lower purchase intentions (Ahluwalia et al. 2000; Howard, Sheth, 1969). The focus group conducted for this research also shares this opinion. Also looking at the motivation for this research, the Volkswagen AG emission scandal and the results regarding sales performance of Volkswagen mentioned in the introduction, it was expected that hypothesis 1 is supported.
5.2 Discussion: Country of Origin
Hypothesis 2 discusses the effect of the COO effect on purchase intentions after a scandal. The results from part one and part two of the survey both don’t show any significant relationship in the regression. The results contradict literature which states that COO has an effect on purchase intentions (Solomon, 2004). The results of the survey show that respondents have a positive perception on the COO effect of the Scandinavian country from where the products originate from. (means of the COO variable, 5,2404 for part one and 5,0164 for part two, respectively). The positive COO effect doesn’t seem to influence purchase intentions before or after a scandal, which is not in line with expectations. Gregory and Ross (1994) don’t deny the existence of the COO and its effect on purchase intentions but state that it isn’t significantly influencing purchase intentions. They state that COO has a major difficulty to be isolated from all other information cues impinging consumer decision making processes (Gregory, Ross, 1994). These results can mean that the statements of Gregory and Ross (1994) have more significance to them than expected, compared to the majority of the literature stating that COO has an effect on purchasing behaviour (Crosno et al., 2009).
36 wasn’t completely covered and therefore harder to identify as a part of a consumers decision making process (Gregory, Ross, 1994). To address this, a follow up study could use a more elaborate approach that captures COO to better understand where its stands in the consumer decision making process.
5.3 Discussion: Trust
The third hypothesis is based on trust and their effect on purchase intentions. Both part one and part two of the survey didn’t show any relationship with purchase intentions. This result is also not expected and is not in line with the reasoning for hypothesis 3, the TCE theory and the main reasoning of the focus group. The dimensions of Delahaye Paine (2003) that are represented trough the questions aren’t seen as significant in relationship with purchase intentions. According to the theory of Howard and Sheth (1969) and answers from the focus group trust doesn’t necessarily have to be harmed when a scandal doesn’t affect the reasons (decision facilitators) why a consumer trusted the firm or brand in the first place. What is apparent in the data is that the mean for the trust variable drops significantly (from 5,3466 in part one to 4,6557 in part two). These results can be interpreted as a favorable trust towards the Scandinavian firm before the scandal and damaged trust after reading about the scandal. Keeping to the theory, these results can be interpreted as an indication that decision facilitators for the respondents to trust the firm are harmed among respondents which accordingly to Howard and Sheth (1969) leads to lower trust. In the end, lower trust in a firm would lead to higher transaction costs according to the TCE theory (Thompson, Yuanyou, 2005). The higher transaction costs and the damaged decision facilitators can therefore be interpreted as the reason for not having significant purchase intentions after a scandal.
5.4 Discussion: Brand loyalty
37 create low transaction costs, which can be interpreted as the consumer knows what he is getting therefore doesn’t have to worry about additional things, which can lead to additional transaction costs.
From a broader theoretic perspective the results can indicate that a scandal can create brand loyalty in itself. Brand loyalty originates from brand awareness; high brand awareness means that consumers recognize a brand under different circumstances (Keller, 2013). When expanding on the results it can be assumed that when a firm gets a lot of attention due to a scandal it can create brand awareness. Brand awareness is the start towards gaining brand loyalty. It can be assumed, keeping to the theory of Keller (2013), that a scandal can therefore be a first step to create brand loyalty among costumers. Due to the scope of this research, this question cannot be answered and needs to be addressed in further research topics.
5.5 Discussion: Aftermath
Hypothesis 5 is only tested in part two of the survey due to the fact it discusses the aftermath of a scandal. The questions were related to management replacement and originated from the focus group that was organised for this research. When interpreting the results it can be said that after a scandal, purchase intentions increase due to a favourable aftermath. The regressions analysis shows that there is a significant positive relationship between the replacement of management and purchase intentions. This means that replacing the management will have a positive effect on purchase intentions.
38 6. THEORATICAL AND MANAGERIAL IMPLICATIONS
This research offers opportunities for both managerial and theoretical implications. The first implication for managers will be to approach loyal consumers after a scandal. According to the results of this research, consumers that are loyal to a brand will have purchase intentions after a scandal. Therefore it is important to create a loyal consumer base before a scandal occurs. Keller (2013) states several suggestions to create brand loyalty. First of all, the experience with the product/service on offer should be able to exceed consumer’s expectations. Numerous studies have shown that high quality brands tend to perform better than lesser quality brands (Keller, 2013). After that Keller discusses that to create “brand resonance”, the ultimate level of relationship a brand can have with its consumer. Brand resonance is characterized in terms of intensity (the depth of the psychological bond that consumers feel, as if they are “in sync” with the brand) and the level of activity (the actual loyalty and repurchases) (Keller, 2013). the characterizations can be broken down into four categories: Behavioral loyalty (How often and how much a consumer rebuys the brand), attitudinal attachment (the personal attachment a consumer has with a brand, consumers should see the brand as something special), sense of community (a consumer can identify itself with the community of a brand) and active engagement (the fact that consumers are willing to invest time, money and other resources beyond the purchase). To reach brand resonance firms should use the brand resonance model as described above, as a road map and guidance to create a brand with high brand loyalty (Keller, 2013).
39 7. LIMITATIONS AND FURTHER RESEARCH
As with every research, this research also has limitations that should be noted. Next to limitations this research offer subjects for further research.
7.1 Limitations
Using the experimental vignette study also means accepting its limitations. The respondents all read a similar vignette and is therefore similar to constant-variable-valuable-vignette (CVVV) (Cavanagh, Fritzsche 1985). Uncontrolled respondent projections is the first limitation that can be found in this research. Respondents need to have facts before they can make a judgement. Due to the fact the vignettes are short; respondents can invent facts to come to a good judgement on the questions that follow (Cavanagh, Fritzsche 1985). The facts that a responded invents can be seen as a variable that is outside the control of the researcher. Choices made on the questions that follow from the vignette can be from different sources than what is tested and described by the vignette (Skipper, Hyman, 1993).
A second limitation which is applicable on this particular vignette study is the fact that the evaluation method is not measured. This means that while respondent’s answers are recorded, the way they concluded the answered is not being measured (Hyman, Steiner, 1996) to get the data on how respondents come to a conclusion on the answers, another study needs to be done among the same respondents to show why respondents chose for a certain answer.
As a third critique on using the method of vignette studies is the creation of a social desirability bias (Fernandes and Randall 1992). A social desirability issue occurs when respondents will answer questions in a way that others will look favourable upon them. A scandal is mostly seen as a negative and bad thing to occur, it is often about lying and deceiving (Ahluwalia et al. 2000). Therefore people may answer the question in a way that is favored by others, in a socially accepted way. Punishing or denying a scandalous firm is therefore a more favorable reaction than a positive reaction. In the case of this research it can lead to denying a firm purchase intentions due to such favorable reactions, while respondents individually may not care.
7.2 Future research
40 is a first step towards brand loyalty (Keller, 2013). Future research can expand on the results of this research, stating that brand loyalty will create purchase intentions among consumers after a scandal, by researching the claims by Keller (2013) and testing the proposed theory. Results of that research will indicate if a scandal is able to create brand loyalty.
Another topic lays in line with the management replacement theories after a scandal. The results of this research discusses management reforms as a reason to increase purchase intentions after a scandal. For future research it will be interesting to see why consumers keep having purchase intentions after a scandal due to management reform. Further research can focus on this why question based on several reasons indicated by the theory of Agrawal et al. (1999) (as discussed in the literature review) for management replacement and discover on which aspects of management replacement respondents react on.
In table 10, the aftermath questions are separated, separating the questions about the board and the CEO. Aftermath is the sum variable of all the questions related to the aftermath, as used in the regression. Aftermath CEO is a sum variable of all the questions related with the change in CEO after a scandal and Aftermath Board is a sum variable containing all the questions related to a change the change of a board after a scandal.
Table 10: Descriptive statistics of the sum variable Aftermath.
