Trust a brand or trust a friend?
The effect of brand familiarity, risk, attitude and word-of-mouth on brand extension choice.
Céline Meurs
Student number: 11111151 January 26th 2017
Thesis final version
MSc. Business Administration – Marketing Track University of Amsterdam
Statement of originality
This document is written by Student Céline Meurs, 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 sources 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.
Table of content
ABSTRACT ... 3
1. INTRODUCTION ... 4
1.1.BACKGROUND... 4
2. LITERATURE REVIEW ... 7
2.1.CONSUMER EVALUATION AND CHOICE ... 7
2.2.BRAND EXTENSIONS ... 8
2.3.PERCEIVED RISK... 9
2.4.WORD-OF-MOUTH ... 10
2.5.THEORY OF REASONED ACTION ... 11
2.6.CONCEPTUAL MODEL ... 12
3. DATA AND METHOD ... 16
3.1.RESEARCH DESIGN ... 16
3.2.SAMPLE ... 17
4. TELEVISION PRODUCT CATEGORY ... 17
4.1.MEASURES ... 17 4.2.RESULTS ... 19 4.2.1. Data ... 19 4.2.2. Statistical procedure ... 19 4.2.3. Hypotheses testing ... 24 4.2.4. Discussion ... 25
5. WASHING MACHINE PRODUCT CATEGORY ... 25
5.1.MEASURES ... 25 5.2.RESULTS ... 27 5.2.1. Data ... 27 5.2.2. Statistical procedure ... 27 5.2.3. Hypotheses testing ... 31 5.2.4. Discussion ... 32
6. GENERAL DISCUSSION AND CONCLUSIONS ... 32
Limitations ... 33
Theoretical and practical implications ... 33
7. REFERENCES ... 35
Abstract
This empirical study focused on predictors of brand extension choice for durable products. Using data from two experiments (N=209 and N=208) with consumers in the Netherlands, this study examined the effects of relative brand familiarity in product group, word-of-mouth, risk and attitude on brand extension choice. The results indicate that both relative familiarity and word-of-mouth are important predictors of brand extension choice and that these effects are partially mediated by attitude. The results suggest that managers of less familiar brands can use word-of-mouth as a tool to increase sales when extending to a new product category, however the use of this tool is limited as the impact of word-of-mouth became weaker as a product became more expensive.
1. Introduction
1.1. Background
In today’s marketplace, consumers are regularly exposed to familiar brands that operate in different product categories (Milberg, Goodstein, Sinn, Cuneo, Epstein, 2013). Brands often extend into new product categories, as this is an important strategic tool for growth, where companies leverage their current customer base and their brand image (Carter & Curry, 2011). By leveraging parent brand knowledge, a new product’s identity can quickly be established when a new product is linked to an already known product (Aaker & Keller, 1990). Brand extensions can lead to positive spillover effects on the parent brand and a favorable attitude towards the extension product, which have great strategic value for a firm’s growth and revenue (Chun, Park, Eisingerich, MacInnis, 2015). Through the favorable evaluations of brand extensions, firms expand their current market base and revenue while they expand their future markets and revenue by the added associations to the parent brand and consumers’ growing acceptance of future brand extensions (Chun et al., 2015). As these outcomes are very important, it is critical to understand the factors that induce them so firms can optimize these when extending into a new product category.
While there are many forms of brand extensions, they often classified as being either horizontal or vertical. A product extension is classified as vertical when the extension is introduced under an existing brand in the same product category yet differentiating in price and quality (Kim, Lavack, Smith, 2001). On the other hand, product extensions are classified as horizontal when the extension is introduced in a similar product class or an entirely different product class than the core brand category (Kim et al., 2001). Horizontal extensions can either be a line extension, where the current brand name is used to enter a new market segment in the same product category (Pitta & Katsanis, 1995), or a franchise extension,
where a new product category is entered using the current brand name (Pitta & Katsanis, 1995). This research focuses on franchise extensions, as durable products are most often introduced as that type of extension.
The extant research on brand extension focused on the predictors of brand extension success and this interest is rising (e.g. Milberg et al., 2013; Volckner & Sattler, 2006; Aaker & Keller, 1990). Volckner & Sattler (2006) found that fit between the parent brand and the extension category is the most important of the known predictors of brand extension success, while Aaker & Keller (1990) found the quality of the parent brand is also a leading driver of brand extension evaluation success.
However, it is important to note the context of these researches, as most of these effects were found when consumers evaluate extensions without other brand choices being present (Milberg et al., 2013), whilst consumers actually consider the extensions in the presence of many other brand options in a retail environment (Milberg, Sinn & Goodstein, 2010). In absence of these alternatives, consumers tend to be positively biased and judge brands more positively than when compared to a situation with different brand options (Kardes, Sanbonmatsu, Cronley, & Houghton, 2002). For this reason, providing different brand options is not just another variable, it is an essential element that must be considered for research that will advance theory and inform practitioners in a way that mimics the real world situation (Ailawadi & Keller, 2004)
Milberg et al. (2013) found that the relative familiarity of a brand compared to competitor brands within a product category was the only significant predictor when it came to brand extension choice. This is because although consumers had a positive attitude towards relatively less familiar brands, they also perceived them to be more risky. When asked to choose one of the three brand offerings, consumers were very likely to choose the more familiar brand option, because relatively more familiar brands were perceived to be less risky.
This preference for more familiar brands when risk was high indicates that familiar brands serve as risk alleviators. Campbell & Goodstein (2001) concluded that risk is an important construct affecting consumers’ preferences, choices and behaviors directly, and that consumers try to reduce purchase risk through different strategies when buying new products.
Although brand extensions were often shown to reduce the purchase risk associated with buying a new product (Delvecchio & Smith, 2005), Milberg et al. (2013) found that brand extensions were perceived as more risky when competitors were relatively well known within a product category. It must be noted that Milberg et al. (2013) used durables as product category, which are typically associated with high risk.
While a lot of past literature focused on brand extension evaluations in terms of attitudes (e.g. Carter & Curry, 2011), Garbarino and Edell (1997) found that choice is a better predictor of consumers’ behavior than the evaluation of attitudes. Asking consumers to provide evaluations of products lead to very different outcomes than when asking for choice preferences (Campbell & Goodstein, 2001). Asking consumers to make a choice between brand alternatives was found to activate a sense of purchase risk associated with the product, which better mimics real-world scenarios, while no such effect was found with evaluations only (Dowling & Staelin, 1994). Choice was found to be a more precise measure than attitudes, because small changes between brands can lead to a different being chosen, while the attitudes towards the brands only marginally differ from each other (Campbell & Goodstein, 2001). For that reason, this research included choice as a dependent variable together with attitude.
One important aspect Milberg et al. (2013) overlooked is that consumers are often confronted with opinions of other consumers before making a purchase decision, which is referred to as word-of-mouth (WOM). Academics extensively researched WOM and it was found to be more persuasive than advertising or even the consumers’ own attitude (East,
Hammond, Lomax, Robinson, 2005). As WOM was found to be a major determinant of consumer behavior, it is important to take this into account. This research assesses whether the effect of relative familiarity on choice is still significant when considering WOM. The main question of this research is:
How do relative familiarity, risk, attitude and WOM affect brand extension choice? This research adds to existing literature because it reassesses the effect of relative familiarity on choice by also considering WOM, as this directly affects consumers’ decisions. This study extends the brand extension literature by examining another factor that influences the success of brand extensions that has previously been overlooked, which provides new insights that marketing managers can apply when considering a brand extension. The results better reflect the real-world retail environment as two key factors, different brand options and choice, were taken into account that have previously not received adequate attention in the literature. This research adds practical value to marketing literature because it provides insights into under which circumstances WOM can be used as a tool to increase sales. Also, as only a limited number of studies focused on the receiver of WOM (Sweeney, Soutar, Mazzarol, 2014) this research provides extra insights on what happens after WOM is received.
2. Literature review
2.1. Consumer evaluation and choice
Consumers judge the quality of products based on a systematic process of acquisition, evaluation and integration of product information or cues available to them (Ahmed, Johnson, Yang, Kheng Fatt, Sack Teng, Chee Boon, 2004). A cue can be extrinsic or intrinsic and is defined as all informational stimuli available to the consumer before consumption (Ahmed et.
al., 2004). Examples of extrinsic are brand and price, while intrinsic cues can be taste and design (Rao & Monroe, 1989). When consumers lack information about the product, they often rely on the brand name to infer its quality (Szybillo & Jacoby, 1974).
2.2. Brand extensions
In the case of brand extensions, consumers rely on a variety of extrinsic cues to assess brand extension performance (Milberg et al., 2010; Park, Milberg, Lawson, 1991). One of these cues is perceived fit between a parent brand and an extension product category. It is suggested that when perceived fit decreases, the perceived purchase risk increases, which leads to more negative extension evaluations (Milberg et al., 2010). Consumers are likely to infer that the positive associations related to the parent brand are also present in the extension when perceived fit is high, leading to more positive extension attitudes (Aaker & Keller, 1990). Secondly, past research indicated that consumers high-quality brands with lower purchase risk, therefore evaluating extensions coming from high-quality parent brands more favorably (Erdem & Swait, 2004). Finally, another important cue found to reduce purchase risk and influence consumer preferences was the familiarity of a brand within a product category (Heilman, Bowman & Wright, 2000; Milberg et al., 2010). Aaker & Keller (1990) found that parent brand familiarity reduced risk for an extension and enhanced extension value.
These previous findings suggest that extension fit, parent brand quality and an extension’s relative brand familiarity influence extension attitude and choice partly because they reduce risk perceptions. Although all three predictors were found to have a significant effect on attitude and risk, relative familiarity was the only significant predictor1 when came to choice between brand options, partly because more familiar brands reduced risk
1
perceptions (Milberg et al., 2013). Figure 1 shows the effects of relative familiarity on evaluations and choice found in Milberg et al. (2013).
Figure 1.
2.3. Perceived risk
Bauer (1960) viewed risk perception as arising from unanticipated and uncertain consequences of an unpleasant nature resulting from the product purchase (Lim, 2003). The idea of whether objective risk exists depends on a researcher’s philosophical perspective, yet either way an objective measure of risk is difficult to obtain in contrast to subjective or perceived risk, which can easily be measured (Mitchell, 1999).
Consumer research identified and measured different sources of perceived risk; such as financial, performance, social, psychological and physical (Campbell & Goodstein, 2001). Perceptions of risk can be derived from one or from a combination of these sources. Stone and Gronhaug (1993) found that of the various dimensions of perceived risk, the financial and psychosocial dimensions of risks captured the majority of the overall risk perceptions. Financial risk concerns how greatly the product choice will impact the individual shopper’s finances and ability to make other purchases, while psychosocial risk refers to how the
purchase decision affects opinions of other people about the shopper (Stone & Gronhaug, 1993). Perceived risk often leads to a variety of risk handling activities (Dowling & Staelin, 1994) and affects consumer behavior in a variety of ways, such as the use of WOM information, new product adoption, brand loyalty and reliance on well-known brands (Erdem, 1998).
2.4. Word-of-mouth
As WOM is informal advice that is passed from one consumer to the other, and is usually interactive, swift and lacks commercial bias (East, Hammond, Lomax, 2008). Consumers often share their purchase experiences with other people after their consumption of a specific product and/or service (Cheng, Lam, Hsu, 2006). Andersen (1998) found that WOM is given more often at high levels of satisfaction and dissatisfaction and WOM can be divided in positive (PWOM), in which brand choice is encouraged, or negative (NWOM), in which brand choice is discouraged (East et al., 2008).
Research about the impact of positive and negative information yielded mixed findings (East, Uncles, Romaniuk, Lomax, 2016). The negativity bias shows that negative information often has a stronger impact on attitude and cognition than positive information (Baumeister, Bratslavsky, Finkenauer, Vols, 2001; Rozin & Royzman, 2001). Adding to this, negative information was shown to have a strong impact on judgement and decision (East et al., 2016). The greater effect of NWOM relates to the gap between a receiver’s expectations and the incoming information, as negative information is more dissident of existing expectations, which are most often positive (Skowronski & Carlston, 1989). In contrast to these results, previous research (East et al. 2008; Sweeney et. al., 2014) found that purchase intention is more often changed by PWOM than by NWOM. However, these results must be reassessed as the research of East et al. (2008) used absolute changes rather than proportional,
and because behavior is either yes or no, only changes moving intentions towards either end of the behavioral intention scale (0-1) are actually meaningful (East et. al., 2016).
2.5. Theory of Reasoned Action
Ajzen & Fishbein formulated the theory of reasoned action (TRA) in 1980 (figure 1), and it is theory of human behavior describing the influences on an individual’s behavioral decision (Cheng et al., 2006). The TRA has both a predictive and an explanatory power, which provides insights to better understand behavior and choices (Petrovici, Ritson, Ness, (2004), and meta-analysis has shown that it predicts behavior extremely well (Sheppard, Hartwick, Warshaw, 1988). The TRA concerns volitional behavior that is directly influenced by a person’s intention to perform a behavior, which is a latent variable depending on the attitude towards that behavior and subjective norms such as social pressures (Petrovici et al., 2004).
In the research of Milberg et al. (2013), attitude and behavior were measured. Milberg et al. (2013) did not measure behavioral intentions; instead they measured behavior in the form of brand choice. They did not isolate subjective norms, referring to a person’s feeling of social pressure to perform or not perform the intended behavior (Ajzen & Fishbein, 1980, p.6.). This concerns the perceptions of how family and friends think about the outcome of the behavior called the normative belief, and to what extent individuals adapt their behavior to social reference groups, which is weighted by the motivation to comply (Ajzen & Fishbein, 1980). Subjective norms together with attitude determine behavioral intention, which in turn determines actual behavior (see figure 2).
Figure 2.
2.6. Conceptual model
In this research, the framework of the TRA is used to predict the factors that influence the behavior of brand extension choice. According to the TRA, behavior is the outcome of behavior intention, which is underpinned by attitude and subjective norms. These subjective norms are presumed have an influence on a buying decision, yet it is often difficult to document its effects due to the fact that norms are rarely explicitly communicated (Martensen & Mouritsen, 2014). This lead previous research to criticize the subjective norms construct for being insufficient (Armitage & Conner, 2001) and consequently, meta-analytic studies found subjective norms to be a weak predictor of consumer behavior (Armitage & Conner, 2001).
On the other hand, several studies found that WOM has a greater impact on consumer behavior than subjective norms. This is because WOM reflects the sharing of knowledge and experiences, where informative evaluations of brands are considered trustworthy as other consumers lack commercial motives (Martensen & Mouritsen, 2014). For this reason, this research replaced the weak subjective norms predictor with the more powerful WOM-predictor. It must be noted that the research of Milberg et al. (2013) was performed in isolation of the opinions of other consumers, and this current research tested if their findings still uphold when considering the impact of WOM on choice.
In the TRA, attitude is determined by behavioral beliefs and evaluations of outcomes (Ajzen & Fishbein, l980). First of all, when forming an attitude about brand name products, consumers take into account extrinsic cues available to them. In the case of purchase behavior, consumers often base their perception of the quality of products from familiar brands on the belief that they have high profits because they have many satisfied customers, and can therefore spend more money on advertising (Macdonald & Sharp, 2003). Milberg et al. (2013) found that the positive attitude towards familiar brands is relative, where respondents had better attitude towards relatively more familiar brands than relatively less familiar competitor brands in a product category. In line with the TRA, respondents chose the brand that they had the most positive attitude towards. This leads to the following hypothesis:
H1: Relative familiarity positively affects (a) attitude and (b) choice; so that more familiar brands are evaluated higher and chosen more often than less familiar brands.
Second of all, in addition to these behavioral beliefs mentioned above, consumers evaluate the outcomes of behavior when forming an attitude about behavior such as product purchase. Bauer (1960) viewed risk perception as arising from unanticipated and uncertain consequences of an unpleasant nature resulting from the product purchase (Lim, 2003). For durable products like consumer electronics, appliances and automobiles, a wrong purchase decision leads to high sunk costs (Laurent & Kapferer, 1985). These potential high sunk costs increase the associated negative consequences, which increases perceived risk associated with durables (Oglethorpe & Monroe, 1987). Brand familiarity is important because consumers may need the reassurance of an established brand name to reduce risk (Aaker & Keller, 1990). In the case of brand extensions, Milberg et al. (2013) found that relative familiarity is the most important predictor of attitude, partially because of the risk involved. Relatively more familiar brands are perceived to be less risky than the competitor brands and for that reason
they are evaluated higher and more likely to be chosen. This leads to the following hypotheses:
H2: Relative familiarity negatively affects perceived risk; so that more familiar brands are perceived to be less risky than less familiar brands.
H3: Perceived risk negatively affects brand extension (a) attitude and (b) choice.
The research of Milberg et al. (2013) did not measure intentions; instead they measured actual behavior in the form of choice. The brand respondents chose was the brand respondents had the strongest intention to buy, which was the brand respondents had the most positive attitude towards. Behavioral intention is therefore conceptualized as choice in this research. This is in line with the TRA, as attitude towards brands are positively associated with behavioral intention, which in turn is positively associated with choice (Sheppard et al. 1988) and this leads to the following hypothesis:
H4: Brand extension attitude positively affects brand extension choice.
In this research, WOM replaced subjective norms in the TRA, subsequently WOM is expected to have an influence on the consumer’s brand attitude and consumer’s brand choice. While PWOM is known to have a positive effect, NWOM has a negative effect. As this research focused on brand extensions, this leads to the following hypotheses:
H5: PWOM (NWOM) positively (negatively) affects brand extension (a) attitude and (b) choice.
With regard to brand familiarity, it is expected that unfamiliar brands benefit more from PWOM than familiar brands. This is because PWOM may reduce the perceived risk associated with unfamiliar brands (Sundaram & Webster, 1999). On the other hand, NWOM may negatively impact unfamiliar brands more than familiar brands, because the perceived
risk involved with an unfamiliar brand is further magnified by NWOM. This leads to the following hypotheses:
H6: PWOM (NWOM) negatively (positively) affects perceived risk. Figure 3 shows the hypotheses in the conceptual model.
3. Data and method
3.1. Research design
This research was of explanatory nature and to answer the research question, a quantitative study was conducted as an experimental vignette study through a survey. The survey had a 2 (brand extension is more familiar than competitor brands / brand extension is less familiar than competitor brands) x 2 (brand extension will receive PWOM / brand extension will receive NWOM) between-subjects factorial design. Two durable product categories were tested, the first was televisions for which the three brand options were priced around €250, and the second was washing machines for which the three brand options were priced around €350.
Each survey started with an expression of gratitude for the respondents’ participation and ensured their privacy. Respondents were first asked to give their gender, their age and educational background to check if the sample was representative of the population. Respondents were then randomly assigned to one of the four groups by Qualtrics. In terms of pricing, the extension brand was always positioned in between the alternative brands, as done in the research of Milberg et al. (2013).
The survey started out with a photo of a brand-less television / brand-less washing machine and asked the respondents to image that they want to buy a 32-inch television / standard washing machine. In the “brand extension is more familiar than alternatives” group, respondents saw an image of an advertisement of the brand extension, to make sure that they are more familiar with the brand extension brand than the other two brand options. In the “brand extension is less familiar than alternatives” group, respondents first saw an image of two advertisements of the other two brand options, to make sure that they were more familiar with the alternative brands than the extension brand. Respondents then evaluated their attitude
and the perceived risk for each of the alternatives. Respondents then chose one of the three brand options. Respondents then either received PWOM or NWOM from a friend about the extension brand, and after this they were asked to evaluate their attitude and perceived risk again for all three brands again and to choose one of the three brand options again. All respondents were asked about their motivation to comply.
3.2. Sample
The population of interest in this research is Dutch consumers. The population is very large and the sampling frame is unknown, this research therefore conducted non-probability convenience sample. Respondents were sought through email and Facebook. Also, demographic information was gathered in the survey such as age, gender and educational background. The survey was spread until at least 200 respondents had completed the survey for each of the two product categories tested.
4. Television product category
4.1. MeasuresThe survey asked respondents about their gender (nominal variable), age (ratio variable) and their educational background (ordinal variable). The independent variable relative familiarity was nominal (more / less). Six electronics brands were selected and pre-tested to ensure that they were indeed more familiar or less familiar. The brands Motorola and Archos do not produce televisions and were therefore be used as the brands for the television brand extension. In the pre-test (N=22), brand familiarity was measured on a seven-point scale with the adjectives “extremely unfamiliar-extremely familiar”. The more familiar brand extension was Motorola (M=6.36, SD=.58), compared to the other brand options Lenco (M=2.18, SD=.66) and Finlux (M=2.23, SD=.61). The less familiar brand extension was
Archos (M=2.27, SD=.63), compared to the other brand options Sony (M=6.59, SD=.50 and Philips (M=6.64, SD=.49). The brand extension was priced at €250 while one other brand option was priced higher at €270 and the other brand option lower at €230, see figure 4. Figure 4.
The other independent variable WOM was either PWOM:
“A friend heard that you were planning on buying a television and tells you: “The X television is the best television I've ever owned. It's really easy to use, and I haven't had a single problem with it. I would strongly recommend you to buy it.” adjusted from Herr, Kardes & Kim (1991).
Or NWOM:
“A friend heard that you were planning on buying a television and tells you: “The X television is the worst television I've ever owned. It's really hard to use, and I've had nothing but problems with it. I would strongly recommend you not to buy it." adjusted from Herr, Kardes & Kim (1991). In the brand extension is more familiar condition, Motorola replaces the X, while in the brand extension is less familiar condition Archos replaces the X.
To ensure that the WOM was indeed perceived to be positive or negative, the PWOM and NWOM messages were pre-tested (N=22) on a seven-point scale with the adjectives “extremely negative-extremely positive”, where PWOM had an average of 6.23 (SD=.61) and NWOM had an average of 1.77 (SD=.61). The pre-test results are found in appendix A. Television More familiar Less familiar
1. Other brand option Lenco €270 Philips €270 2. Brand extension Motorola €250 Archos €250 3. Other brand option Finlux €230 Sony €230
The dependent variable attitude toward the brand was measured directly on a seven-point scale by the adjectives “dislike very much – like very much”. Risk was measured directly as in Milberg et al. (2013): “I perceive Brand X to be” 1 = ‘not at all risky’, 7 = ‘extremely risky’. Choice was discreetly measured, where respondents chose one of the three brand options. The control variable is motivation to comply and was also directly measured as in Ajzen (2006) (“When it comes to which brand I should choose, I do what my friends think I should do” 1 = ‘not at all’, 7 - ‘very much’).
4.2. Results
4.2.1. Data
After the data was gathered through a Qualtrics survey, it was prepared for analysis in SPSS. A total of 209 surveys were fully completed and as all answers were forced in the survey, there was no missing data. The average age of the respondents was 23 years (SD = 2.23) of which 60% were female and 40% male and 78% were university students.
4.2.2. Statistical procedure
First, the direct attitude measure was recoded from 1 to 7 into -3 to 3 and gender was recoded into males (0) and females (1). Then, dummy variables were created for relative familiarity, where 1 represented a more familiar brand extension than the alternatives and 0 represented a less familiar brand extension that the alternatives. After this, a dummy variable was created for WOM, where 1 represented PWOM and 0 represented NWOM. Also, dummy variables were created for choice before WOM and choice after WOM, where 1 represented the brand extension and 0 represented the other brand options. The mean, standard deviation and correlations were calculated in SPSS and are found in Table 1.
In the condition before receiving WOM, all correlations were in line with the expectations of this research. Relative familiarity had a positive correlation with attitude,
r(207) = .33, p<.001 and a positive correlation with choice r(207) = .77, p<.001. As expected, relative familiarity had a negative correlation with risk, r(207) = -.53, p<.001 and risk had a negative correlation with attitude, r(207) = -.34, p<.001 and a negative correlation with choice r(207) = -.41, p<.001. And finally, attitude had a positive correlation with choice, r(207) = .42, p<.001. In the condition after receiving WOM, most correlations were still significant yet a bit weaker. Relative familiarity had a positive correlation with attitude, r(207) = .19, p<.001 and with choice r(207) = .40, p<.001. However, the correlation of relative familiarity and risk, r(207) = .19, p>.23, became insignificant in this condition. Risk still had a negative correlation with attitude, r(207) = -.45, p<.001 and with choice r(207) = -.32, p<.001, while attitude still had a positive correlation with choice, r(207) = .59, p<.001. Finally, PWOM had a positive correlation with attitude, r(207) = .69, p<.001 and positive correlation with choice r(207) = .55, p<.001, while PWOM had a negative association with risk, r(207) = -.51, p<.001. The whole model is tested using AMOS and the results are seen in table 2 and figure 5. As the Chi-square was insignificant χ 2(2, N = 209) = 1.66, p=.44 and the RMSEA = 0.000 and the CFI = 1.000, the model had a good fit (Hu & Bentler, 1999).
Television group: N=209. For Gender (GEN), 0 = male, 1 = female. EDU = Education. MTC = Motivation to comply. RFD= Relative familiarity dummy.
WOMD= Word-of-mouth dummy. AttB= Attitude before WOM. RiskB= Risk before WOM. ChoiceB= Choice before WOM dummy. AttA: Attitude after WOM. RiskA= Risk after WOM. ChoiceA= Choice after WOM dummy. * p<.05. (2-tailed) ** p<.01 (2-tailed)
Table 1. Variables
Correlations
M (SD) Age Gen Edu MTC. RFD WOMD AttB RiskB ChoiceB AttA RiskA ChoiceA
Age 23.40 (2.23) - Gen .60 (0.49) -.17* - Edu 1.33 (0.72) -.01 .01 - MTC 4.96 (1.33) .10 .04 -.01 - RFD .48 (0.50) .01 .06 .05 -.03 - WOMD .51 (0.50) -.05 .05 .17* .01 .00 - AttB .29 (1.09) -.11 .10 .06 .02 .33** .04 - RiskB 4.15 (1.29) .09 -.06 -.04 .04 -.53** -.81 -.34** - ChoiceB .39 (0.49) .02 .07 .09 -.18 .77** .02 .42** -.41** - AttA 4.11 (1.58) -.08 .11 .15* .04 .19** .69** .40** -.29** .18** - RiskA 4.20 (1.83) .04 .13 -.18** -.00 -.13 -.51** -.41 .36** -.09 -.45** - ChoiceA .32 (0.47) -.09 .11 .14* -.02 .40** .55** .28** -.25** .44** .59** -.32** -
Table 2: Path analysis television group Structural equation model
Variable B Std. Error β C.R. p-value Result
RF Risk -.51 .21 -.14* -2.38 .02 H2 accepted
Gender Risk .64 .22 .17** 2.92 .00
Age Risk .04 .05 .05 .83 .40
WOM Risk -1.86 .21 -.51*** -8.77 .00 H6 accepted
RF Attitude .51 .15 .16*** 3.37 .00 H1a accepted
Gender Attitude .25 .16 .08 1.56 .12
Age Attitude -.02 .03 -.03 -.59 .55
WOM Attitude 1.98 .18 .63*** 11.25 .00 H5a accepted
Risk Attitude -.10 .05 -.12* -2.09 .04 H3a accepted
Attitude Choice .09 .05 .29*** 4.15 .00 H4 accepted
Risk Choice .01 .02 .05 .78 .43 H3b rejected
RF Choice .33 .05 .35*** 6.90 .00 H1a accepted
WOM Choice .34 .07 .37*** 5.01 .00 H5b accepted
*** Significant at the p<.001 level (2-tailed) **Significant at the p<.01 level (2-tailed) * Significant at the p<.05 level (2-tailed)
Figure 5.
B [S.E.] * p<.05, ** p<.01, ***p<.001
4.2.3. Hypotheses testing
The results from AMOS showed that relative familiarity significantly affected attitude (β=.16, p <.001) and choice (β = .35, p<.001), therefore both H1a and H1b are accepted. The results from AMOS showed that relative familiarity significantly affected risk (β= -.14, p=.017). According to these results, H2 is accepted. The results showed that perceived risk significantly affected attitude (β= -.12, p=.037), however, risk did not significantly affect choice (β=.05, p=.432). For this reason, H3a is accepted but H3b is rejected. The results showed that attitude positively affected choice (β=.29, p<.001) and H4 is therefore accepted. The results showed that PWOM positively affected attitude (β=.63, p<.001) and choice (β =
.37, p<.001), which means H5a and H5b are accepted. Finally, the results showed PWOM negatively affected risk perceptions (β = -.51, p<.001), which means H6 is accepted.
4.2.4. Discussion
Overall the model was good in predicting brand extension choice as all but one hypothesis are accepted. The results indicate that WOM is equally as powerful as relative familiarity in predicting brand extension choice and is therefore an important aspect to consider. WOM and relative familiarity are both found to have strong direct effects on choice and indirect effects that are partially mediated through attitude. Contrary to the expectation, risk did not have a significant effect on choice when considering WOM. Although relative familiarity and WOM did affect risk, the results indicate that the effect of perceived risk on choice is fully mediated through attitude.
According to the literature, the potential high sunk costs increase perceived risk associated with durables (Oglethorpe & Monroe, 1987), however the finding of this research implies that the potential sunk costs associated with buying the television are not high enough to cause people to choose based on risk alone. It is interesting to further research this by testing a second product group to assess if a product with a higher price indeed causes risk to be an important predictor of choice.
5. Washing machine product category
5.1. MeasuresAs in the other survey, respondents were asked about their gender (nominal variable), age (ratio variable) and their educational background (ordinal variable). The independent variable relative familiarity was nominal (more / less). For the washing machine product category, six electronics brands were selected and pre-tested to ensure that they were indeed
more familiar or less familiar. The brands Sony and Lenco do not produce washing machines and were therefore used as the brands for the television brand extension. In the pre-test (N=21), brand familiarity was measured on a seven-point scale with the adjectives “extremely unfamiliar-extremely familiar”. The more familiar brand extension was Sony (M=6.43, SD=.51), compared to the other brand options Indesit (M=2.14, SD=.66) and Beko (M=1.95, SD=.39). The less familiar brand extension was Lenco (M=2.05, SD=.50), compared to the other brand options Samsung (M=6.57, SD=.51) and Bosch (M=6.38, SD=.59). The pre-test results are found in Appendix A. In this case, the brand extension and the two other brand options were priced €100 higher than in the previous experiment, see figure 6.
Figure 6.
Washing machine More familiar Less familiar
1. Other brand option Indesit €370 Samsung €370 2. Brand extension Sony €350 Lenco €350 3. Other brand option Beko €330 Bosch €330
The other independent variable WOM was either PWOM:
“A friend heard that you were planning on buying a washing machine and tells you: “The X washing machine is the best washing machine I've ever owned. It's really easy to use, and I haven't had a single problem with it. I would strongly recommend you to buy it.” adjusted from Herr, Kardes & Kim (1991).
Or NWOM:
“A friend heard that you were planning on buying a washing machine and tells you: “The X washing machine is the worst washing machine I've ever owned. It's really hard to use, and I've had nothing but problems with it. I would strongly recommend you not to buy it." adjusted from Herr, Kardes & Kim (1991). In the brand extension is more familiar
condition, Sony replaced the X, while in the brand extension is less familiar condition Lenco replaced the X.
The dependent variable attitude toward the brand was measured by the mean of three seven-point scales anchored by the adjectives "good-bad," "dislike very much-like very much," and "pleasant-unpleasant." with a Cronbach’s alpha of 0.92 (Mitchell, 1986). Risk was measured directly as in Milberg et al. (2013): “I perceive Brand X to be” 1 = ‘not at all risky’, 7 = ‘extremely risky’. Choice was discreetly measured, where respondents will choose one of the three brand options. The control variable was motivation to comply and was directly measured as in Ajzen (2006) (“When it comes to which brand I should choose, I do what my friends think I should do” 1 = ‘not at all’, 7 - ‘very much’).
5.2. Results
5.2.1. Data
After the data was gathered through a Qualtrics survey, it was prepared for analysis in SPSS. A total of 208 surveys were fully completed and as all answers were forced in the survey, there was no missing data. The average age of the respondents was 23 years (SD = 2.13) of which 53% were female and 47% male and 76% were university students.
5.2.2. Statistical procedure
First, the counter-indicative item of the attitude scale was recoded from 7 to 1 into 1 to 7. The attitude scale had a Cronbach’s alpha of 0.91, which means it was reliable. Then, the values are recoded from 1 to 7 into -3 to 3 and the mean of the three items of the attitude scale was calculated. Gender was recoded into males (0) and females (1). Then, dummy variables were created for relative familiarity, where 1 represented a more familiar brand extension than the other two brand options and 0 represented a less familiar brand extension that the other two brand options. After this, a dummy variable was created for WOM, where 1 represented
PWOM and 0 represented NWOM. Also, dummy variables were created for choice before WOM and choice after WOM, where 1 represented the brand extension and 0 represented the other two brand options. The mean, standard deviation and correlations were calculated in SPSS and are found in Table 3.
In the condition before receiving WOM, all correlations were in line with the expectations of this research. Relative familiarity had a positive correlation with attitude, r(206) = .33, p<.001 and a positive correlation with choice r(206) = .77, p<.001. As expected, relative familiarity had a negative correlation with risk, r(206) = -.53, p<.001 and risk had a negative correlation with attitude, r(206) = -.34, p<.001 and a negative correlation with choice r(206) = -.41, p<.001. And finally, attitude had a positive correlation with choice, r(206) = .42, p<.001. In the condition after receiving WOM, the correlations were still significant yet a bit weaker. Relative familiarity had a positive correlation with attitude, r(206) = .19, p<.001 and with choice r(206) = .40, p<.001. Relative familiarity and risk have a negative correlation, r(206) = .19, p>.23. Risk still had a negative correlation with attitude, r(206) = -.45, p<.001 and with choice r(206) = -.32, p<.001, while attitude still had a positive correlation with choice, r(206) = .59, p<.001. Finally, PWOM had a positive correlation with attitude, r(206) = .69, p<.001 and positive correlation with choice r(206) = .55, p<.001, while PWOM had a negative correlation with risk, r(206) = -.51, p<.001.
The whole model was tested using AMOS and the results are seen in table 4 and figure 7. As the Chi-square was insignificant χ 2(2, N = 208) = 3.17, p = .20 and the RMSEA = 0.000 and the CFI = 1.000, the model had a good fit (Hu & Bentler, 1999).
Washing machine group: N=208. For Gender (GEN), 0 = male, 1 = female. EDU = Education. MTC = Motivation to comply. RFD= Relative familiarity
dummy. WOMD= Word-of-mouth dummy. AttB= Attitude before WOM. RiskB= Risk before WOM. ChoiceB= Choice before WOM dummy. AttA: Attitude after WOM. RiskA= Risk after WOM. ChoiceA= Choice after WOM dummy. * p<.05. (2-tailed) ** p<.01 (2-tailed)
Table 3. Variables
Correlations
M (SD) Age Gen Edu MTC. RFD WOMD AttB RiskB ChoiceB AttA RiskA ChoiceA
Age 23.36 (2.13) - Gen .53 (0.50) -.16* - Edu 1.36 (0.71) .04 -.08 - MTC 5.07 (3.03) -.10 .09 -.01 - RFD .51 (0.50) .02 .15* -.01 .05 - WOMD .50 (0.50) -.06 .03 .02 -.02 -.01 - AttB 4.51 (1.07) .04 .05 .02 -.03 .61** .06 - RiskB 3.80 (1.52) -.05 -.11 -.01 -.03 -.64** -.03 -.65** - ChoiceB .42 (0.50) -.01 .15* -.06 .05 .79** -.06 .63** -.68** - AttA 4.12 (1.57) -.00 .03 .04 -.09 .30** .68** .44** -.34** .30** - RiskA 4.07 (1.92) .03 -.09 -.06 .04 -.25** -.69** -.28** .30** -.21** -.82** - ChoiceA .34 (0.47) -.03 .00 -.01 -.05 .43** .50** .34** -.35** .41** .66** -.62** -
Table 4: Path analysis washing machine group Structural equation model
Variable B Std. Error β C.R. p-value Result
RF Risk -.97 .18 -.25*** -5.30 .00 H2 accepted
Gender Risk -.14 .19 -.04 -.75 .78
Age Risk -.01 .04 -.01 -.28 .45
WOM Risk -2.66 .18 -.69*** -14.73 .00 H6 accepted
RF Attitude .50 .12 .16*** 4.04 .00 H1a accepted
Gender Attitude -.16 .18 -.05 -1.37 .17
Age Attitude .01 .03 .02 0.45 .65
WOM Attitude .85 .17 .27*** 5.14 .00 H5a accepted
Risk Attitude -.49 .04 -.60*** -11.08 .00 H3a accepted
Attitude Choice .09 .03 .29*** 3.28 .00 H4 accepted
Risk Choice -.04 .02 -.18* -2.03 .04 H3b accepted
RF Choice .29 .05 .30*** 5.70 .00 H1a accepted
WOM Choice .17 .07 .18** 2.46 .00 H5b accepted
*** Significant at the p<.001 level (2-tailed) **Significant at the p<.01 level (2-tailed) * Significant at the p<.05 level (2-tailed)
Figure 7.
B [S.E.] * p<.05, ** p<.01, ***p<.001
5.2.3. Hypotheses testing
The results from AMOS showed that relative familiarity significantly affected attitude (β=.16, p <.001) and choice (β=.30, p<.001), therefore both H1a and H1b are accepted. The results from AMOS showed that relative familiarity significantly affected risk (β= -.25, p<.001). According to these results, H2 is accepted. The results showed that perceived risk significantly affected attitude (β = .60, p<.001) and risk significantly affected choice (β= -.18, p<.05). For this reason, H3a and H3b are accepted. The results show that attitude positively affected choice (β=.30, p=.001) and H4 is therefore accepted. The results showed that PWOM positively affected attitude (β=.27, p<.001) and choice (β=.18, p=.01), which
means H5a and H5b are accepted. Finally, the results show PWOM negatively affected risk perceptions (β= -.69, p<.001), which means H6 is accepted.
5.2.4. Discussion
As all hypotheses are accepted, the results from this product group are in line with the expectations from the literature. The main difference between the two product categories was the price, where the washing machine brands were priced €100 higher than the television brands. Contrary to the prior results from the television category, risk did significantly affect choice in the washing machine category. This indicates that the high price of the washing machine leads to higher potential sunk costs, which in turn causes people to be more risk averse and choose a brand with lower risk. Another difference compared to the television product category was that relative familiarity had a stronger effect on choice than WOM in the case of washing machines. This can again be explained by the higher price, causing people to rely more on a brand name than on WOM. However, although relative familiarity had a stronger effect on choice, these results showed that WOM is still an important aspect to consider. Overall the model was good in predicting brand extension choice.
6. General discussion and conclusions
After the results from both product categories the research question can be answered: How do relative familiarity, risk, attitude and WOM affect brand extension choice? First of all, WOM and relative familiarity both have strong direct effects on choice and strong indirect effects that run through attitude. It is important to note that the effect of WOM became weaker compared to relative familiarity when the price of the product became higher, as in the washing machine product category. In both cases, attitude had a strong direct effect on choice. Relative familiarity and WOM were found to affect perceived risk, however, the results
indicate that the effect of perceived risk on choice is fully mediated through attitude in the case of the less expensive televisions, while the effect is partially mediated through attitude in the case of the more expensive washing machines. Risk was only found to have a significant effect on choice when the price of a product was relatively high.
Limitations
This research had a few limitations. First of all, this research was conducted in an experimental setting as opposed a real-life situation. Also, most respondents were university students with an average age of 23 years old and they were therefore not a proper representation of the average Dutch consumer. Adding to this is the limited sample sizes N=209 and N=208, making the results less externally valid. Another limitation is that the respondents were presented with three brand options while in real-life situations a consumer can be confronted with many more brand options. In addition, consumers often base their decision on product features, but in this research they were only presented with a price and a brand name. Another limitation is that this research only measures hypothetical choice, not real-world choice, therefore making the results less externally valid.
Theoretical and practical implications
This research adds important insights to the brand extension literature. The findings of this research support the research of Milberg et al. (2013) that relative familiarity is an important predictor of brand extension choice and also further extends their findings as WOM is found to be an important predictor of brand extension choice. This is not only an interesting theoretical implication, but also a practical one because it implies that although less familiar brands are at a relative disadvantage compared to familiar brands, they can use WOM to obtain an advantage. As brand familiarity is hard to gain and can take a long time to establish, this means that less familiar brands can resort to using WOM as tool to increase brand
extension sales. Brand managers can actively promote PWOM through setting up referral systems that reward people giving PWOM for their products with discounts and credit. Implementing a WOM rewards system is relatively easy and it gives less familiar brands a good chance to compete with the more familiar brands. It must be noted that the effectiveness of WOM is dependent on the price of a product and becomes less useful as the price of a product increases. This research implies that people are willing to trust on the opinion of friends, however, this willingness is not limitless. As the price of a product increased, respondents relied more on the brand name than on WOM.
This research influences future discussions in the literature in two ways. First of all, it backs the notion of Milberg et al. (2013) that experimental settings need to include more ‘real-world’ factors such as different brand options, brand choice and WOM. Without these aspects, the applicability of experimental findings to reality is limited. Secondly, it invites future research to look for other factors that bring experiments closer to reality, thus yielding better insights in how to manage the success of a brand extension.
It would be interesting to further research if WOM has a stronger effect on choice for durable product categories with high social risk such as watches or sunglasses. Future research could assess is these findings uphold when the respondent has the choice between more than three brand options. Also, future research could explore what other factors contribute to the success of a brand extension that have previously been overlooked.
7. References
Aaker, D.A., & Keller, K.L. (1990). Consumer evaluations of brand extensions. Journal of Marketing, 54, 27–41.
Ailawadi, K.L., & Keller, K.L. (2004). Understanding retail branding: Conceptual insights and research priorities. Journal of Retailing, 80, 331–342.
Ahmed, Z.U., Johnson, J.P., Yang, X., Kheng Fatt, C., Sack Teng, H., & Chee Boon, L. (2004) Does country of origin matter for low-involvement products? International Marketing Review, 21(1), 102-120.
Ajzen, I. (2006). Constructing a TpB Questionnaire: Conceptual and Methodological Considerations September, 2002 (Revised January, 2006)
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.
Armitage C.J., & Conner, M. (2001). Social Cognitive Determinants of Blood Donation, Journal of Applied Social Psychology, 31(7), 1431-1457.
Baumeister, R.F., Bratslavsky, E., Finkenauer, K., & Vols, K.D. (2001). Bad is stronger than good. Review of General Psychology, 5(4), 323–370.
Campbell, M.C., & Goodstein, R.C. (2001), The Moderating Effect of Perceived Risk on Consumers’ Evaluations of Product Incongruity: Preference for the Norm. Journal of Consumer Research, 28, 439-449.
Carter, R.E., & Curry, D.J. (2011). Perceptions versus performance when managing extensions: New evidence about the role of fit between a parent brand and an extension. Journal of the Academy of Marketing Science, 34, 1–17.
Cheng, S., Lam, T., & Hsu, C.H.C. (2006). Negative Word-of-Mouth Communication Intention: An Application of the Theory of Planned Behavior” Journal of Hospitality and Tourism Research 30(1), 95-116
Chun, H.H., Park, C.W., Eisingerich, A.B., & MacInnis, D.J. (2015). Strategic benefits of low fit brand extensions: When and why?. Journal of Consumer Psychology, 25(4), 577-595. DelVecchio, D., & Smith, D.C. (2005). Brand-extension price premiums: The effects of
perceived fit and extension product category risk. Journal of the Academy of Marketing Science, 33(2), 184–191.
Dowling, G.R., & Staelin, R. (1994). A model of perceived risk and intended risk-handling activity. Journal of Consumer Research, 21, 119–134.
East, R., Hammond, K., Lomax, W., & Robinson, H. (2005). What is the Effect of a Recommendation?. The Marketing Review, 5, 145-157.
East, R., Hammond, K., & Lomax, W. (2008). Measuring the impact of positive and negative word of mouth on brand purchase probability. International Journal of Research in Marketing. 25(3), 215–224.
East, R., Uncles, M.D., Romaniuk, J., & Lomax, W. (2016). Measuring the impact of positive and negative word of mouth: a reappraisal. Australasian Marketing Journal, 24, 54-58. Erdem, T. (1998). An Empirical Analysis of Umbrella Branding. Journal of Marketing
Research, 35 (8), 339–351.
Erdem, T., & Swait, J. (2004). Brand credibility, brand consideration, and choice. Journal of Consumer Research, 31, 191–198.
Garbarino, E.C., & Edell, J.A. (1997). Cognitive effort, affect, and choice. Journal of Consumer Research, 24, 147–158.
Heilman, C. M., Bowman, D., & Wright, G. P. (2000). The evolution of brand preferences and choice behaviors of consumers new to a market. Journal of Marketing Research, 37, 139– 155.
Herr, P.M., Kardes, F.R., & Kim, J. (1991). Effects of Word-of-Mouth and Product Attribute Information on Persuasion: An Accessibility-Diagnosticity Perspective. Journal of Consumer Research, 17, 454-462.
Hu, L.T., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1-55.
Kardes, F.R., Sanbonmatsu, D.M., Cronley, M.L., & Houghton, D.C. (2002). Consideration set overvaluation: When impossibly favorable ratings of a set of brands are observed. Journal of Consumer Psychology, 12(4), 353–361.
Kim, C.K., Lavack, A.M., & Smith, M. (2001). Consumer evaluations of vertical brand extensions and core brands. Journal of Business Research, 52, 211-222.
Lim, N. (2003). Consumers’ perceived risk: sources versus consequences. Electronic Commerce Research and Applications, 2(3), 216-228.
Martensen, A., & Mouritsen, J. (2014). How ad-talk and word-of-mouth influence consumers’ responses. Innovative Marketing, 10(2) 58-69.
Macdonald, E., & Sharp, B. (2003). Management perceptions of the importance of brand awareness as an indication of advertising effectiveness. Marketing Bulletin, 14(1), 1-11. Milberg, S.J., Goodstein, R.C., Sinn, F., Cuneo, A., & Epstein, L.D. (2013). Call back the
jury: Reinvestigating the effects of fit and parent brand quality in determining brand extension success. Journal of Marketing Management, 29(3), 374-390.
Milberg, S.J., Sinn, F., & Goodstein, R.C. (2010). Consumer reactions to brand extensions in a competitive context: Does fit still matter? Journal of Consumer Research, 37(3), 543– 553.
Mitchell, A.A. (1986) The Effect of Verbal and Visual Components of Advertisements on Brand Attitudes and Attitude Toward the Advertisement. Journal of Consumer Research, 13, 12-24.
Mitchell, V.W. (1999). Consumer perceived risk: conceptualisations and models. European Journal of Marketing, 33(1), 163 - 195.
Oglethorpe, J.E., & Monroe, K.B. (1987). Risk Perception and Risk Acceptability in Consumer Behavior: Conceptual Issues and an Agenda for Future Research, AMA Winter Marketing Educators' Conference, ed. Russell W. Belk et al., Chicago, IL: American Marketing Association, 255-260.
Park, C.W., Milberg, S.J., & Lawson, R. (1991). Evaluation of brand extensions: The role of feature similarity and brand concept consistency. Journal of Consumer Research, 18, 185–193.
Petrovici, D.A., Ritson, C., & Ness, M. (2004). The Theory of Reasoned Action and Food Choice. Journal of International Food & Agribusiness Marketing, 16(1), 59-87.
Pitta, D.A., & Katsanis, L.P. (1995). Understanding brand equity for successful brand extension. Journal of Consumer Marketing, 12(4), 51 - 64.
Rao, A., & Monroe, K. (1989). The effect of price, brand, name and store name on buyers’ perceptions of product quality: an integrative review. Journal of Marketing Research, 26(3), 351-368.
Rozin, P., & Royzman, E.B. (2001). Negativity Bias, Negativity Dominance, and Contagion. Personality and Social Psychology Review, 5(4), 296-320.
Sheppard, B., Hartwick, J., & Warshaw, P. (1988). The Theory of Reasoned Action: A Meta-Analysis of Past Research with Recommendations for Modifications and Future Research. Journal of Consumer Research, 15(3), 325-343.
Skowronski, J.J., & Carlston, D.E. (1989). Negativity and extremity biases in impression formation: a review of explanations. Psychological Bulletin, 105(1), 131–142.
Stone, R.N., & Gronhaug, K. (1993). Perceived risk: Further considerations for the marketing discipline. European Journal of Marketing, 27(3), 372–394.
Sundaram, D.S., & Webster, C. (1999). The Role of Brand Familiarity on the Impact of Word-Of-Mouth Communication on Brand Evaluations. Advances in Consumer Research (26), 664-670.
Sweeney, J., Soutar, G., & Mazzarol, T. (2014). Factors enhancing word-of-mouth influence: positive and negative service-related messages. European Journal of Marketing, 48(1/2), 336–359.
Szybillo, G.J., & Jacoby, J. (1974). Intrinsic versus extrinsic cues as determinant of perceived product quality. Journal of Applied Psychology, 59(1), 74-88.
Volckner, F., & Sattler, H. (2006). Drivers of brand extension success. Journal of Marketing, 70, 18–34.
Webb, T.L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychological Bulletin. 132(2), 249–268.
8. Appendix A
Pretest results television
N Min Max Mean S.D.
Lenco 22 1 3 2.18 .66 Motorola 22 5 7 6.36 .58 Finlux 22 1 3 2.23 .61 Sony 22 6 7 6.59 .50 Archos 22 1 3 2.27 .63 Philips 22 6 7 6.64 .49 PWOM 22 6 7 6.23 .61 NWOM 22 1 3 1.77 .61
Pretest results washing machine
N Min Max Mean S.D.
Indesit 21 1 3 2.14 .66 Sony 21 6 7 6.43 .51 Beko 21 1 3 1.95 .39 Samsung 21 6 7 6.57 .51 Lenco 21 1 3 2.05 .50 Bosch 21 5 7 6.38 .59
