Honesty : too much of a good thing? : what happens when a company discloses negative information about its products?

(1)

Honesty: Too Much of a Good Thing?

What happens when a company discloses negative information about its

products?

MSc Thesis Business Administration

Track Marketing

Date August 14th, 2015

Author Nadine de Ruiter

Student Number 6040748 First Supervisor J. Demmers

(2)

Statement of Originality

This document is written by Student Nadine de Ruiter who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no source other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

It seems counterintuitive for marketers to add negative information to their advertisements. However, several researchers have suggested that product-related messages containing both negative and positive information, so-called two-sided messages, can be even more effective compared to messages containing solely positive information. Including negative information in advertisements leads consumers to attribute this behavior to unselfish intentions of the advertiser. The current study contributes to existing literature on transparency by investigating the relation between the valence of disclosed information and willingness to pay. Furthermore, the influences of trust and disclosure source on this relation are researched. Analyses of data of 250 respondents have confirmed a positive relation between valence of disclosed product information and WTP. When the information is disclosed by a brand the relationship took the form of an inverted U-curve, and when it is disclosed by an online review it showed a linear form. Although consumers have increased levels of trust for companies who disclose extremely negative product information, WTP will decrease at some point due to negative effects on quality perception. This adds to current literature by suggesting there exists a certain tipping point in the positive effects of brands disclosing negative information in advertisements: something as ‘too much honesty’ exists.

(5)

1. Introduction

The rise of new communication technologies, most eminent the Internet, has made it almost impossible for organisations to hide information from consumers (Rogers & Peppers, 2012). Media and business analysts constantly scrutinize businesses, while legal constraints force companies to share information concerning their current and future projects (van Riel, 2000). Normal consumers now have access to extensive amounts of information on which they can base more informed purchasing decisions (Labrecque, 2013). This has lead consumers to hold more power over companies than ever before. As Rogers and Peppers (2012) put it: “From WikiLeaks and the Arab Spring to a cable TV repair-man asleep on your couch or an airline’s luggage handler mistreating bags, people will find things out.” And as it has become increasingly easy to share information with others all across the globe, negative brand or product information will travel fast and far. The odds of consumers finding out negative brand or product information are not in favour of the companies anymore (Fournier & Avery, 2011). Companies often react to these changing relationships by consciously crafting their outward messages (Noble, Noble & Adjei, 2012). Transparency has become an essential dimension of organizations (Christensen, 2002).

Transparency is a relevant field to study as it has the potential to influence perceptions of consumers concerning marketers’ attempts at persuasion. It is an opportunity for companies to change the perception that consumers hold of them, from being seen as ‘opponents’, to being perceived as more cooperative and friendly. Full transparency implies sharing not only positive brand of product information, but also uncovering negative information in their advertising, so called two-sided messages. Two-sided messages is of theoretical interest within the transparency literature, suggesting that advertising containing both positive and negative product information can enhance source credibility and result in higher ratings of brand attitude, than in one-sided advertising containing solely positive product information (Golden & Alpert, 1987; Belch, 1981; Kamins & Marks, 1987). Knowledge of the persuasive works of two-sided communication can increase our expertise

(6)

and grasp on basic persuasive processes between marketing and consumers (Crowley & Hoyer, 1994).

Although the field of two-sided messages has its roots in social psychology (e.g., Lumsdaine & Janis, 1953; Hovland, Lumsdaine & Sheffield, 1949), more recently this area has also been studied broadly in consumer research. However, a view on the literature shows few attempts to providing structural guidelines to the use of two-sided messages in advertisement. Empirical evidence shows that two-sided messages can exert a positive influence on source credibility (Kamins & Marks, 1987; Eisend, 2006), but also that negative information can have negative effects on consumers’ expected product benefits and their intentions to buy the product (Settle & Golden, 1974; Stayman, Hoyer & Leon, 1987). To increase profound knowledge of the use of two-sided messages as an advertising tool, it is important to investigate what the ‘boundaries’ of these communications are; does a trade-off occur between the positive effects of two-sided messages on source credibility and the negative effects on attitude? This leads to the main question of this paper; where do the boundaries of disclosing negative information in two-sided messages lie? Is there a tipping point where the positive effects of the valence of disclosed information included in two-sided messages on willingness to pay diminish?

The current study adds to existing literature by providing specific guidelines to the use two-sided messages in advertising. By assessing the area in which the disclosure of negative product information leads to the abovementioned positive effects of two-sided messages, this paper offers a practical framework for marketers. Applying the concept of two-sided messages to advertising is a big opportunity for marketers/managers, as the current study shows it leads to a higher willingness to pay. Careful attention must be paid however to the amount and type of negative information they include in the messages.

(7)

2. Literature Review

Essentially, when a firm is transparent, it means that it is perceived to be open and honest with its stakeholders (Dapko, 2012). By being more transparent, companies attempt to reduce customers’ scepticism and avoid asking consumers to trust a ‘black box’ (Kirby, 2012; Rogers & Peppers, 2012). Rewards for transparent companies come in the form of increased favourable attitudes, and higher levels of trust and purchase intentions (Dapko, 2012). Ways for managers to increase perceptions of transparency are: offering stakeholders easy access to company and product information, providing opportunities for mutual conversations, and by sharing negative information about itself, or its products (Dapko, 2012; Christensen, 2002).

Traditional marketing communications present brands and/or products in a favourable way by presenting positive features or associating the brand with positive symbols (Crowley and Hoyer, 1994; Eisend, 2006). In doing so, marketers attempt to influence consumers’ brand attitudes and preferences (Crowley & Hoyer, 1994). On the surface, it seems counterintuitive for marketers to add negative information to their advertisements, as this would lead to a negative representation of the advertised product (Eisend, 2006). Information integration theory (Anderson, 1971) confirms this view, suggesting that positive information increases valuation of products, while negative information decreases consumers’ valuation. However, several researchers have suggested that product-related messages containing both negative and positive information, so-called two-sided messages, can be even more effective compared to one-sided messages containing solely positive information (Golden & Alpert, 1987; Kamins & Mark, 1987; Kamins & Assael, 1987; Pechmann, 1992). The first hypotheses of the current study will help determine the effect of the valence of disclosed product information, ranging from very negative to very positive, on the price consumers are willing to pay for the product:

H1 There is a positive relation between the valence of disclosed information and willingness to pay (WTP).

(8)

Several researchers have suggested that the effect of disclosing negative information is also influenced by who the ‘messager’ is, in other words: the disclosure source (Demmers, 2015; Smiths, De Houwer & Nosek, 2013). Although an information source generally becomes more credible when it discloses negative or unique information (Eisend, 2006), this effect might be even stronger when an independent source such as a customer, discloses the product information, rather than the company itself. The finding that source credibility is lower when positive information is enclosed by a brand itself rather than an independent source confirms this view (Smiths, De Houwer & Nosek, 2012). These findings suggest that different levels of trust explain the relationship between transparent information and the disclosure source.

Customer reviews tend to be perceived as highly credible and trustworthy because this third-party information is seen as completely separate and disregarding of the involved company’s intentions (Chen & Xie, 2008). Source credibility is high, irrespective of the information valence: whether a customer review discloses extremely negative or extremely positive product information, consumers who read the review are expected to trust the review, seeing as the review is completely independent of the brands’ intentions (Chen & Xie, 2008). Therefore, as more positive product information is disclosed in a customer review is, consumers’ WTP is expected to increase (see Figure 1). This leads to the following hypothesis:

H2 The relation between the valence of disclosed information and willingness to pay is moderated by disclosure source, such that:

a. When the information is disclosed by another source, the relation is positive and linear, as shown in Figure 1.

(9)

Several researchers have suggested that, when a brand discloses negative product information, the increases in source credibility may go at the expense of other key variables, such as attitude (Eisend, 2006; Crowley & Hoyer, 1994). Seeing that negative information can have negative effects on attitudes and/or consumers’ expectations, too much disclosure of negative information by a brand in two-sided messages could have detrimental effects on the persuasiveness of the advertisement (Anderson, 1971). In other words, a trade-off might occur between the positive effects of two-sided messages on source credibility, and the negative effects on attitude (Eisend, 2006; Crowley & Hoyer, 1994).

However, when a brand discloses solely very positive information in their advertisements, the brand is expected to lose credibility. Consumers might perceive extremely positive advertisements as ‘too good to be true’, and experience a decreased trust in the brand (Smith et al., 2013). This suggested concept of an optimal range of negative information disclosure in two-sided messages has not been studied before, albeit prior research provides considerable confirmation. For example, disclosing very negative information concerning an important attribute results in credibility gains but has a disadvantageous effect on subjects’ willingness to pay for the product (Stayman, Hoyer & Leon, 1987). In a study by Settle and Golden (1974), increases in source credibility (as the result of disclosing negative information in two-sided messages) were counterbalanced by losses in expected product benefits.

To recapitulate: when disclosed information is too negative, the negative effect of the unfavourable information on attitude towards the product is detrimental to the positive effects of brand transparency. Consumers’ WTP is then expected to be low. The disclosure of mildly negative information however, increases a brands’ credibility and is expected to increase consumers’ WTP. When an advertisement contains too positive information, the ad loses credibility and consumers lose trust in the brand (Smith et al., 2013). The persuasion power of the ad is then low, and WTP decreases. The abovementioned studies suggest that when product information is disclosed by a brand, the relationship between information valence and WTP takes the form of an inverted U-shape (see Figure 2).

(10)

b. When the information is disclosed by the brand, the relation takes an inverted U-shape, as shown in Figure 2.

Figure 2. A curvilinear relationship between two variables

The found positive effects of two-sided messages are most consistent on source credibility (e.g. Golden & Alpert 1987; Kamins & Marks 1987; Smith & Hunt 1978; Settle & Golden, 1974; Stayman, Hoywer & Leon, 1987). Attribution Theory has directed most of the studies on two-sided advertising, and contributes to the clarification of the effect of two-sided messages on source credibility (Kelley & Michela, 1980; Weiner, 1985). Attribution is the process by which people explain others’ behavior by assigning causes to it (Friestad & Wright, 1994). Applied to advertising, attribution theory assumes that consumers will attribute product claims either to the advertiser’s eagerness to sell products, or to real product features communicated by a truthful advertiser. Generally however, consumers tend to distrust brands (Friestad & Wright, 1994). Brand trust is of great importance to brands, as it is a major predictor of consumer behavior, such as customer loyalty and willingness to pay. (Chaudhuri & Holbrook, 2001). In prior research, trust is often described as ‘the willingness to be vulnerable to the actions of the other party’ (Mayer, Davis & Schoorman, 1995; Barney & Hansen, 1994). Applied to marketing, this infers that including negative information in advertisements leads consumers to attribute this behavior to helpful, unselfish intentions of the advertiser, concluding that he is ‘telling the truth’ (Crowley & Hoyer, 1994). This leads to the final hypothesis of the current study:

H3 The relation between the valence of disclosed information and willingness to pay, moderated by disclosure source, is mediated by trust.

(11)

3. Method & Data

3.1 The Sample

The sample (N = 275) consisted of 136 women (49.5%) and 139 men (50.5%), with an average age of 28 years old, varying from 18 to 65 (SD = 9.77). Most of the respondents had received education on university level (38.5%) and HBO level (28.4%). Participation in the survey was deliberate and completely anonymous. Respondents were approached through online personal social networks and in public libraries.

3.2 Procedure

An online pre-test was conducted to assess the effect of the experimental manipulation in the different treatment groups, vouching for high internal validity in the main study. The respondents (N = 35) consisted of 20 women (57.1%) and 15 men (43.9%) with an average age of 26.3 (SD = 6.7), varying from 20 to 54 years of age. Most of the respondents had an education level of WO (40%) and HBO (37.1%). Respondents were shown advertisements of ‘HEMA bycicle lamps’, containing the average burning time of such lamps (70 hours). They were asked to indicate burning times they perceived as ‘very negative’ of ‘very positive’. The manipulation of information valence in the main study is done accordingly to the results of the pre-test (see Table 1).

Table 1. Results of the pre-test

Information Valence Product Information Very

negative Mildly negative Neutral Mildly positive Very positive Control HEMA Bicycle lamps Burning time (Hours) 3 35 70 105 138 -

The main study was conducted using an online questionnaire. The data is hereby directly linked from the questionnaire to the database, implying no need to manually enter

(12)

data. This saves time and significantly reduces the possibility of errors in the database (Smith, 1997).

The experimental design is a 5 x 2 + 1 factorial between-subjects design with five levels of negative information manipulation and two levels of disclosure source manipulation. In the control group with baseline treatment, subjects were not exposed to any manipulation at all. Respondents were shown an advertisement of ‘HEMA bicycle lamps’ and asked how much they were willing to pay for them. Respondents were randomly exposed to one of the 11 treatments containing different advertisement texts referring to the burning time of the lamps (typically 70 hours), varying from very negative (3 hours) to very positive product information (138 hours). Depending on the treatment group, this information was disclosed by either the HEMA brand itself, or by a hypothetical online reviewer. Afterwards they answered four multiple-choice questions regarding their trust the brand.

3.3 Measures

Willingness to pay, the dependent variable, was measured by directly asking respondents to fill in the maximum price in Euros they were willing to pay. This measured the respondents’ hypothetical willingness to pay, as no real money was involved.

Message sidedness was manipulated by altering the amount of negative/positive product information included in the advertisements, in terms of burning time of the bicycle lamps. The typical burning time of 70 hours is displayed in every treatment group and according to the treatment group a review of the product shows a burning time varying from 3 hours (very negative) to 138 hours (very positive). These burning times are based on the results of the pre-test (see Table 1). Figure 3 shows examples of a treatment group in which the message sidedness is mildly negative, displaying a burning time of 35 hours.

Disclosure source was manipulated by varying the product information source in the advertisements. In treatment 1 to 5, additional product information was disclosed by the HEMA brand itself, whereas in the treatment groups 6 to 10 this information was disclosed by a hypothetical online reviewer named Saskia from the website www.kieskeurig.nl. In

(13)

treatment group 11, the control group, no additional product information was shown. See Figure 3 for examples of the different treatment groups.

Figure 3. Manipulation of disclosure source; product information disclosed by HEMA (left) and by online reviewer (right)

Brand trust was measured using a four-item index based on a seven-point Likert scale varying from 1 (very strongly disagree) to 7 (very strongly agree). An example item is: ‘I trust this brand’. The four items are selected from an existing trust-scale designed by Chaudhuri and Holbrook (2001). A test on the reliability of the scale showed a Crohnbach’s Alpha of .97, meaning the scale is highly reliable.

4. Results

Errors in the data were adjusted by deleting respondents with missing values. Also, certain answers were adjusted for analysis purposes, for example: the word ‘nothing’ was changed to a number 0.

The data on willingness to pay was well modelled by a normal distribution; the histogram of standardised residuals showed normally distributed errors, as did the P-P plot of standardized residuals, indicating dots that were on average close to the line. Table 2 provides an overview of the correlations, means and standard deviations of the main variables.

(14)

Table 2. Means, Standard deviations and Pearson’s Correlations among the main variables of the study (N=251) M (SD) ₁ ₂ ₃ ₄ 1. Willingness to pay 2.63 (1.65) - 2. Valence of information 2.99 (1.41) .292** - 3. Disclosure source .64 (.64) .142* .009 - 4. Trust 5.48 (1.42) .255** .057 -.026 -

* Correlation is significant at p < .05 (two-tailed) ** Correlation is significant at p < .01 (two-tailed)

H1 There is a positive relation between the valence of disclosed information and willingness to pay (WTP).

Based on the results of a correlational analysis, the valence of disclosed information is related to willingness to pay (r = .29, p = .000). The outcome of a simple linear regression confirms this relation by showing that a significant proportion of the variation in WTP is predicted by the valence of disclosed information. In other words, the valence of the disclosed information in advertising messages can be perceived as a predictor of customers’ WTP (F (1, 250) = 23.25, p = .000). WTP is higher when disclosed product information is positive rather than negative, hereby confirming the first hypothesis (see Table 3).

Table 3. Summary of linear regression analysis for willingness to pay (N=251)

B SE (B) ß t Sig. (p)

Valence of Information .349 .072 .292 4.822 .000

A planned contrasts test was performed to investigate this relationship further. In order to do so a one-way ANOVA test was applied to reconfirm the significant relationship

(15)

between information valence and WTP (F (5, 274) = 5.34, p = .000). The control treatment (where no product information was shown whatsoever) was then compared to the other 5 levels of information valence, ranging from highly negative to highly positive through a test of planned contrasts. Although the hypothesis is confirmed, no significant contrast between the control group and all other treatment groups together was found (t = .54, p = .591).

However, significant contrasts were found between the treatment groups where the disclosed information was highly negative versus highly positive, with a WTP mean difference of 1.58 Euros (p = .000). Also, a significant contrast was found between the highly negative versus the mildly positive information groups, with a WTP mean difference of 1.09 Euros (p < .05). See Figure 4 below for a visual plot of H1.

Figure 4. Mean of WTP (in Euros) at different levels of information valence

H2 The relation between the valence of disclosed information and willingness to pay is moderated by disclosure source.

After centering the variables disclosure source and information valence and computing a source-by-valence interaction term (Aiken, West & Reno, 1991), the two predictors and the interaction were tested in a simultaneous regression model. The results of a one-way ANOVA suggest a significant interaction effect of information valence and source,

(16)

on the dependent variable WTP (F (2, 250) = 14.71, p = .000). This confirms the expected effect that the relation between information valence and WTP is different for different types of disclosure source (ß = 0.16, p < .05).

Table 4. Summary of multiple regression analysis for willingness to pay (N=251)

Note. R 2 = 0.11

a. When the information is disclosed by another source, the relation is positive and linear.

A regression including a curvilinear relation is conducted to assess whether a linear model, rather than a curvilinear model, better explains the relation between information valence and WTP, when the information is disclosed by another source.

Indeed, a significant linear relation between the valence of disclosed information and WTP, when the disclosure source is another source than the brand, is confirmed by a linear regression (F (1, 125) = 13.34, p = .000). Also, the linear model proved to explain the relation between information valence and WTP, better than the curvilinear relation (F (2, 124) = 7.01, p < .01).

B SE (B) t (Sig.) p

Constant 1.583 .237 6.691 .000

Disclosure Source (centered) .316 .156 2.029 .043

Valence of Information (centered) .273 .078 3.484 .001

(17)

In other words, when another source (rather than the brand) discloses negative or positive product information, this leads to respectively low and high WTP. The hypothesis is hereby confirmed, (ß = .31, p = .000).

Table 5. Summary of linear and curvilinear regression for WTP, when information is disclosed by another source than the brand

(N = 127)

b. When the information is disclosed by the brand, the relation takes an inverted U-shape.

Table 6. Summary of linear and curvilinear regression for WTP, when information is disclosed by the brand (N = 124) Model B SE (B) ß t (Sig.) p Linear .307 .097 .277 3.183 .002 Valence of Information _Quadratic _.839 _.492 _.758 _1.704 _.091

A curve-estimation suggests a linear relationship between WTP and disclosure source

(F (1,123) = 1.123, p = .002). A non-linear regression shows that the quadratic model is also applicable to the relation, but slightly less significant (F (2,123) = 5.684, p = .004). In other

Model B SE (B) ß t (Sig.) p

Linear .349 .108 .311 3.653 .000

Valence of Information

(18)

words, when the brand discloses very negative information, WTP is low, but when the brand discloses very positive information, WTP is also low. The hypothesis is partly confirmed.

H3 The relation between the valence of disclosed information and willingness to pay, moderated by disclosure source, is mediated by trust.

A regression of valence, disclosure source and an interaction variable (source*valence) on the mediator variable: trust, shows that only the interaction variable has significant effect on trust (F (1, 250) = 16.53, p = .000). Therefore, no mediation effect of trust exists on the relation between valence and source on WTP, because neither valence (F (4, 274) = 1.27, p = .282) nor source (F (1, 274) = .002, p = .966) explains the variance in trust sufficiently.

However, a one-way ANOVA shows a significant interaction effect of information valence and disclosure source on trust (as the dependent variable) has been found (F (4, 250) = 17.08, p = .000). So, although trust does not mediate the relationship between information valence, disclosure source and WTP, when used as a dependent variable, trust certainly is influenced by information valence and the disclosure source. In other words, depending on the information source (HEMA or online customer review), the effect of information valence on trust differs. This result will be further elaborated on in the discussion section. See Figure 5 for a visual plot of these findings.

(19)

Figure 5. Mean of trust at different levels of information valence

5. Conclusion & Discussion

Transparency has become an essential dimension of organizations (Christensen, 2002). Transparent companies are rewarded in the form of favourable attitudes, higher levels of trust and higher purchase intentions (Dapko, 2012). Sharing negative information about a companies’ own products is one way of increasing perceptions of transparency (Dapko, 2012; Christensen, 2002). Product-related messages containing both negative and positive information have proved even more effective compared to messages containing solely positive information (Golden & Alpert, 1987; Kamins & Mark, 1987; Kamins & Assael, 1987; Pechmann, 1992). These types of messages are called two-sided messages (Eisend, 2006). The underlying mechanism of the described gains of two-sided messages lies in the perceptions of a brands’ honesty: in the literature also called source credibility (e.g. Golden & Alpert 1987; Kamins & Marks 1987; Smith & Hunt 1978; Settle & Golden, 1974; Stayman, Hoywer & Leon, 1987). Attribution Theory clarifies this more clearly; attribution is the process by which individuals explain the causes of behaviors or events around them (Fiske & Taylor, 1991). Applied to advertising, the theory assumes that consumers will try to explain a company’s behavior by attributing its product claims either to an eagerness to sell products, or to actual product quality communicated by a trustworthy advertiser. When an

(20)

advertisement shows not only positive product claims but also a certain amount of negative information, this might lead consumers to attribute this information to unselfish and helpful intentions of the company. The conclusion that the company is ‘telling the truth’ has positive consequences for consumers’ attitudes and their willingness to pay (Crowley & Hoyer, 1994).

The aim of the current study was to provide practical guidelines to marketers who aim to offer two-sided advertisements in an attempt to increase their transparency. As one can imagine, when a company discloses too much negative product information in two-sided messages, consumers might still perceive that company as being trustworthy and unselfish, but will not be very eager to buy the product, due to low perceptions of product quality. Common sense thinking suggests there must be a certain tipping point in the relationship between information valence and WTP where the positive effects of two-sided messages diminish due to perceptions of low quality/negative attitudes. The following questions were set up in order to provide guidelines to transparent companies.

The main question was whether the valence of transparent product-information has an effect on customers’ willingness to pay (WTP), and if so; in what form. As expected, information valence had a significant effect on WTP. Surprisingly however, no significant contrast was found between the control group and the other treatment groups. A possible explanation is the small range of measured WTP. By using such a relatively cheap product (bicycle lamps) the different amounts that were reported on WTP lay extremely close to one another, and in total ranged from around zero to 3.5 Euros. If a more expensive product were used, differences in WTP would be clearer, and potentially even significant.

Furthermore, disclosure source was investigated as a potential moderator in this relationship. Expected was that when a brand, in this case HEMA, acts as disclosure source of transparent information, something as ‘too much honesty’ exists. In other words, when the brand discloses too much negative product-information (in an attempt to gain the benefits associated with acting transparent), the WTP will decrease rather than increase, due to negative effects on attitude. Also, when the brand discloses too much positive information, WTP will also decrease rather than increase, due to negative effects on source credibility. The

(21)

relationship between the valence of information and WTP, when the information is disclosed by a brand, will then fit the form of an inverted U-shape (see Figure 1). This was indeed confirmed by a non-linear regression.

When the disclosure source of transparent information is another source, in this case an independent online customer review, there exists no such thing as ‘too much honesty’. Customer reviews are commonly perceived as highly credible, therefore it was expected that when the review discloses negative and positive product-information, WTP was respectively low and high. The relationship between the valence of information and WTP, when the information is disclosed by another source (customer review), will then fit the form of a positive linear model (see Figure 2). A linear regression was supportive of this hypothesis.

An explanation for the change in relationship between information valence and WTP for different disclosure sources lies in perceived trustworthiness of the ‘messager’. Customer reviews are perceived as highly credible because they are perceived as completely separate from the concerned company (Chen & Xie, 2008). According to Attribution Theory consumers will then attribute the information in the online review to unselfish and helpful behavior of the reviewer, also when the review is highly positive (Fiske & Taylor, 1991). However, when a company itself discloses only highly positive product-information in advertisements, consumers might attribute this to an urge to sell rather than trusting the information.

The underlying mechanism in the relationship between information valence, disclosure source and WTP, was expected to be different levels of trust. Although a clear mediation effect of trust was not supported by the analysis, an interesting discovery was made in the process of analysing the data. Depending on the disclosure source (brand of online customer review) the relationship between information valence and trust changes.

Unfortunately these trust levels do not correspond to different levels of reported WTP. This lacking effect can be explained by the measure of WTP in the current study (explained further in the limitations section).

(22)

Although results so far show that applying them can have huge positive impact on consumers’ attributes and purchase intentions, literature on two-sided messages is scarce. Further studies might increase the generalization of results and refine the phenomenon of two-sided messages in the transparency and marketing theory.

Marketeers, for now remember: honesty is the best policy, but beware: there are boundaries.

5.1 Limitations & Directions for Future Research

Several limitations should be noted. First, the measurement of willingness to pay might have been too unspecific. Respondents were simply asked how many Euros they were willing to pay for the bicycle lamps in an open-ended answer box, without notifying them of the actual market price of the product. Therefore it remains unclear what participants individually use as ‘anchor’ price: an acceptable price with which they compare their WTP in their minds. Disclosing an actual market price of the product might have framed the WTP a bit, but would have made it very clear if participants paid more or less than average. The way in which WTP was measured has negative consequences for the external validity of the study; seeing as only ‘hypothetical’ WTP was measured, involving no actual money or payments, respondents might have answered differently then they would have had in a real setting.

Also, the way in which information valence is operationalized through altering levels of burning hours is very specific. It is not clear if the results of the current study would be similar to studies in which they ‘apply’ different types of transparency, such as transparency concerning the allocation of resources to farmers in a coffee brewing company, or the labour conditions in the fabrics of a retail company.

As mentioned earlier in the discussion, a limitation of this study is the use of a relatively cheap product (bicycle lamps). The reported WTP ranged from zero to 3.5 Euros, which is an extremely small range. This probably explains the lack of several significant contrasts between the treatment groups on WTP. A more expensive product would have yielded for a wider range on WTP and more specific and clearer contrasts between the

(23)

treatment groups. Also, more expensive products often require more purchase involvement and yield higher perceptions of risk. Future studies could compare different experiments where products differ as well in price, but also on product type.

And last, the set up of the current study ignored potential effects of personality variables on WTP. Former disappointing attempts at predicting consumer behavior through the use of personality traits/characteristics prevented me from adding such research to the current study (Wells & Beard, 1973, Engel, Blackwell, & Miniard, 1986; Kakkar & Lutz, 1981).

Future research should focus more on the influence of personality variables on the relationship between information valence and WTP. In particular, self-monitoring (Snyder, 1974) and need for cognition (Haugtvedt, Petty & Cacioppo, 1982) seem to have potential in the attempt to explain consumer behavior. Individuals who score high on need for cognition enjoy cognitive work and elaborate more on information they are shown (Haugtvedt, Petty & Cacioppo, 1982). Highly negative of positive information contained in two-sided messages might therefore have a different effect on people who have a high need for cognition compared to people with low need for cognition, who tend to avoid effortful thinking and might be persuaded easier. This idea is derived in the context of the Elaboration Likelihood Model of persuasion (Haugtvedt, Petty & Cacioppo, 1981): a general model for understanding the effects of persuasive communications.

(24)

References

Aiken, L. S., West, S. G., & Reno, R. R. (1991). Multiple regression: Testing and interpreting interactions. Sage.

Anderson, N. H. (1971). Integration theory and attitude change. Psychological Review, 78(3), 171.

Barney, J. B., & Hansen, M. H. (1994). Trustworthiness as a source of competitive advantage. Strategic management journal, 15(S1), 175-190.

Belch, G. E. (1981). An examination of comparative and noncomparative television

commercials: The effects of claim variation and repetition on cognitive response and message acceptance. Journal of Marketing Research, 333-349.

Chaudhuri, A., & Holbrook, M. B. (2001). The chain of effects from brand trust and brand affect to brand performance: the role of brand loyalty. Journal of marketing, 65(2), 81-93.

Chen, Y., & Xie, J. (2008). Online consumer review: Word-of-mouth as a new element of marketing communication mix. Management Science, 54(3), 477-491.

Christensen, T. L. (2002). Corporate communication: The challenge of

transparency. Corporate Communications: An International Journal, 7(3), 162-168. Crowley, A. E., & Hoyer, W. D. (1994). An integrative framework for understanding

two-sided persuasion. Journal of Consumer Research, 561-574.

Eisend, M. (2006). Two-sided advertising: A meta-analysis. International Journal of

Research in Marketing, 23(2), 187-198.

Engel, J. F., Blackwell, R. D., & Miniard, P. D. (1986). Consumer Behavior. New York: Dryden Press

Fiske, S. T., & Taylor, S. E. (1991). Social cognition, 2nd. NY: McGraw-Hill, 16-15.

Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with persuasion attempts. Journal of consumer research, 1-31.

(25)

Golden, L. L., & Alpert, M. I. (1987). Comparative analysis of the relative effectiveness of one-and two-sided communication for contrasting products. Journal of

Advertising, 16(1), 18-68.

Haugtvedt, C. P., Petty, R. E., & Cacioppo, J. T. (1992). Need for cognition and advertising: Understanding the role of personality variables in consumer behavior. Journal of

Consumer Psychology, 1(3), 239-260.

Hovland, C. I., Lumsdaine, C. I., & Sheffield, F. D. (1949). Experiments on mass communication. Princeton: Princeton University Press.

Kakkar, P., & Lutz, R. J. (1981). Situational infuluences on consumer behavior: A review. In H. H. Kassarjian & T. S. Robertson (Eds.), Perspectives in Consumer Behavior, 204-214. Glenview, IL: Scott, Foresman.

Kamins, M. A., & Marks, L. J. (1987). Advertising puffery: the impact of using two-sided claims on product attitude and purchase intention. Journal of Advertising, 16(4), 6-15.

Kamins, M. A., & Assael, H. (1987). Two-sided versus one-sided appeals: A cognitive perspective on argumentation, source derogation, and the effect of disconfirming trial on belief change. Journal of Marketing Research, 29-39.

Kelley, H. H., & Michela, J. L. (1980). Attribution theory and research. Annual review of

psychology, 31(1), 457-501.

Labrecque, L. I., Mathwick, C., Novak, T. P., & Hofacker, C. F. (2013). Consumer power: Evolution in the digital age. Journal of Interactive Marketing, 27(4), 257-269.

Lumsdaine, A. A., & Janis, I. L. (1953). Resistance to “counterpropaganda” produced by

one- sided and two-sided “propaganda” presentations. Public Opinion

Quarterly, 17(3), 311-318.

Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of management review, 20(3), 709-734.

(26)

483.

Pechmann, C. (1992). Predicting when two-sided ads will be more effective than one-sided ads: The role of correlational and correspondent inferences. Journal of Marketing

Research, 441-453.

Peppers, D., & Rogers, M. (2012). Extreme Trust: Honesty as a Competitive Advantage. Penguin.

Settle, R. B., & Golden, L. L. (1974). Attribution theory and advertiser credibility. Journal of

Marketing Research, 181-185.

Smith, R. E., & Hunt, S. D. (1978). Attributional processes and effects in promotional situations. Journal of Consumer Research, 149-158.

Smith, C. B. (1997). Casting the net: Surveying an Internet population. Journal of

Computer-Mediated Communication, 3(1), 0-0.

Smith, C. T., De Houwer, J., & Nosek, B. A. (2013). Consider the Source Persuasion of Implicit Evaluations Is Moderated by Source Credibility.Personality and Social

Psychology Bulletin, 39(2), 193-205.

Stayman, D., Hoyer, W. D., & Leon, R. (1987). Attribute importance in discounting product features in advertising. In American Marketing Association, Summer Educators’

Conference, Toronto.

Van Riel, C. B. (2000). Corporate communication orchestrated by a sustainable corporate story. The expressive organization, 157-81.

Wells, W., & Beard, A. (1973). Personality and consumer behavior. In S. Ward & T. S. Robertson (Eds.), Consumer Behavior: Theoretical Sources, 141-199. Eaglewood Cliffs, NJ: Prentice-Hall.

Weiner, B. (1985). An attributional theory of achievement motivation and emotion.

(27)

Appendix

Table . An overview of the different treatment groups

The valence of disclosed information

Disclosure source N

Treatment 1 Highly negative Brand 25

Treatment 2 Negative Brand 25

Treatment 3 Neutral Brand 26

Treatment 4 Positive Brand 24

Treatment 5 Highly positive Brand 24

Treatment 6 Highly negative Other Source 25

Treatment 7 Negative Other Source 26

Treatment 8 Neutral Other Source 26

Treatment 9 Positive Other Source 24

Treatment 10 Highly positive Other Source 26

Honesty : too much of a good thing? : what happens when a company discloses negative information about its products?