The relationship between multiple
irrelevant attributes and product
evaluation: The moderating effect of
product category and involvement
Name: Nalini Bhagwandin Student number: 5819296
Study/Track: Business Administration, Marketing Thesis supervisor: Dr. K. Venetis
Date & Place: 5 January 2015, Amsterdam Thesis: Master thesis
2 Abstract This research investigates the effect of irrelevant attributes on products evaluation. We further propose a moderating effect of product category (utilitarian/hedonic) and
involvement (low/high) on the relationship between irrelevant attributes and product evaluation. Results from a 2x2x2 design experiment showed that even though consumers’ attitude and their intention to purchase the product decrease with the addition of irrelevant attributes to the product, they are still willing to pay significantly more for products including irrelevant attributes. Furthermore, analysis showed that product category moderated the relationship between irrelevant attributes and product evaluation on the dimensions purchase intentions and willingness to pay. That is, consumers are willing to pay relatively more for hedonic products than utilitarian products but their intention to purchase the product decreases (as opposed to utilitarian products) with the addition of irrelevant attributes. No moderating effect of involvement was found. Theoretical and practical implications of the findings are discussed. Also, limitations of the study and directions for future research are provided.
Keywords: irrelevant attributes, product differentiation, marketing strategy, involvement, hedonic products, utilitarian products.
1. Introduction
Product differentiation is a classic marketing strategy which occurs when consumers are offered a product that is relevant, unique and of perceived value to them compared to similar products from rivals (Porter, 1980). Traditionally, successful differentiation can only occur if one offers an objectthat is perceived as relevant and of value (Cahill, 1996 and Porter, 1985). This traditional marketing strategy can play a role in achieving a competitive advantage. That is, the competitive advantage will occur if the differential aspects of the
3 product are perceived to have a superior benefit over other competing products (Dumovic & Knowles, 2008 and Gebauer et al., 2011). However, it is often a time consuming, complex and expensive process for a company to create and add attributes to a product that can lead to product differentiation (Brown & Carpenter, 2000). But, what if you could differentiate your product without investing redundant time and effort into discovering novel and salient attributes? What if there exists a so called short cut to product differentiation?
This is where the concept meaningless differentiation falls into place. A company can create irrelevant attributes in an attempt to differentiate the offering by including a fairly meaningless element or feature to a product (Brown and Carpenter, 2000). Broniarczyk and Gershoff (2003) have defined an irrelevant attribute as a feature added to a product, that is distinct from similar products’ attributes, but which offer no significant performance benefit. However, according to these authors, it is not a requisite that the irrelevant attribute should have no possible relationship to objective quality. Broniarczyk and Gershoff (2003) suggest that in the context of irrelevant attributes, the relationship should be relatively weak between the irrelevant attribute and product performance compared to the relationship between obviously important attributes and product performance.According to Brown and Carpenter (2000), a trivial or irrelevant attribute concerns either an objectively irrelevant aspect that does not provide any physical, chemical or technical benefit to the product or an attribute whose contribution to the product is so little that it must be considered as irrelevant. Brown and Carpenter (2000) discuss several strategies to offer an irrelevant attribute such as: introducing an unrelated premium (e.g. the offering of a tube of toothpaste with the purchase of a cd), incorporating a rare element to a product (e.g. adding ‘flaked coffee crystals’ to coffee which have no effect on the taste), or creating a novel association with the brand (e.g., Volkswagen have ‘Fahrvergnuegen’). In this study, we follow the definition of Broniarczyk and Gershoff (2003) and apply irrelevant attributes that have a relatively weak relationship
4 with product performance. Furthermore, we add irrelevant attributes to products that provide no meaningful performance benefit.
From a consumer’s perspective, the only way to discover the irrelevance of a certain attribute is to consult objective sources such as consumer protection organizations (e.g. Radar) or experts (Albrecht et al., 2011). As regards the legal framework, companies are permitted to use irrelevant attributes in their brand communication as long as they truly incorporate the advertised substances into the products. It is not important how consumers describe the performance or quality of a particular attribute as long as the product lives up to consumers’ expectations concerning the products quality and performance (Albrecht et al., 2011). Therefore, it is not surprising that many companies have put effort into developing and designing new products with irrelevant attributes to differentiate the product. For example, La Prairie’s Cellular Radiance Concentrate Pure Gold differentiates itself from other nourishing serums by the addition of 24-karat gold extracts in one of their products. Although the product contains gold, the inclusion of gold is not likely to have a positive effect on the skin since the concentration of the particles is quite small. The golden particles merely reflect the light which is rather a temporary cosmetic effect. It is not likely that the product stimulates the maintenance of collagen and elasticity of the skin which would result in a better looking skin which the product claims to achieve with, among others, the addition of the golden particles (futurederm.com; Öko-2005). In other words, the addition of gold has no medically proven effects on product performance (i.e. help prevent the breakdown of collagen and improve elasticity and firmness of the skin). Yet, the product is successful and people are willing to spend €540 for a 30ml bottle of crème. What triggers us during the decision making process to purchase such products?
Several studies have indicated that consumer decisions can be positively influenced by attributes that have no direct relationship with enhanced product quality (e.g. Carpenter et
5 al., 1994; Broniarczyk & Gershoff, 1997; Brown & Carpenter 2000; Sun, 2010 and Albrecht et al., 2011). In fact, previous research has demonstrated that such irrelevant components can even be of influence when consumers are aware of the fact that these attributes contribute nothing to the actual product (Brown & Carpenter, 2000; Broniarczyk & Gershoff, 1997). Several explanations have been provided in the marketing literature for the effects of irrelevant attributes on product evaluation.
First of all, irrelevant attributes influence the choice of consumers through their
uniqueness in the choice set (Carpenter et al. 1994 and Albrecht et al., 2011). That is,
consumers perceive the product including the irrelevant attribute as superior because other products do not possess this particular feature. Another explanation might be the salience
effect which indicates that the irrelevant attribute draws the attention away from other
important attributes (Hutchinson & Alba, 1991). The salience effect occurs because consumers rarely endeavour to test hypotheses about many of the products they use and therefore do not make an objective and deliberate choice when comparing products (Hoch & Deighton, 1989). For example, consumers are often distracted by advertising and packaging (Hutchinson & Alba, 1991; Van Osselaer, et al., 2004). In the context of the salience effect, the irrelevant attribute could be for example emphasized in the advertisement or on the package. Therefore, the existence of an irrelevant attribute might guide the attention away from the actual differentiating features (Albrecht et al., 2011).
However, another line of research suggests that adding irrelevant attributes to your product is not always a guarantee for successful product differentiation (e.g. Simonson et al.,1993 and Meyvis & Janiszewski, 2002). According to Simonson, Nowlis & Simonson (1993), consumers might interpret irrelevant attributes as an indicator for poor quality and relate this to products that have to compensate other areas in which they lack something. Another theory that explains the poor effect of the addition of irrelevant attributes is the
6
dilution effect. The dilution effect has been a popular discussed subject in marketing literature
(Nisbet et. Al., 1989; Gurhan-Canli et al., 1998; Keller, 2003 and Gierl & Heuttl, 2012). The dilution effect occurs when people are presented to diagnostic information, knowledge that is relevant in making a certain judgment or decision, and non diagnostic information,
knowledge that is useless in making a certain judgment or decision (here: irrelevant
attributes). The dilution effect is founded around the idea that when people are presented to both types of information, non diagnostic information may lead to less extreme judgments (e.g., De Dreu et al., 1995; Fein & Hilton, 1992). The dilution effect indicates that irrelevant product information may weaken consumers’ belief in a product’s ability to deliver a desired benefit and therefore may be valued less than products excluding irrelevant attributes
(Meyvis & Janiszewski , 2002 and Gierl & Heuttl, 2012). A third explanation for the negative effect of irrelevant attributes on product evaluation might be the perseverance effect which indicates that consumers stick to their opinion and beliefs. That is, consumers will continue to value a product which they are familiar with over a new product including irrelevant
attributes regardless the attractiveness of the new product (Carpenter et al. 1994 and Albrecht et al., 2011).
Brown & Carpenter (2000) replicated their previous study (1994) and found that the direction (positive or negative) of the influence of the irrelevant information depends on the size of the choice set which was also supported by Broniarczyk and Gershoff (2003). Thus, the alternating results of previous research do not contradict each other, but rather disclose the different instrumental approaches used in the research design. We will replenish the work of Brown & Carpenter (2000) by investigating the antinomy created by both lines of research from previous literature in a more profound way by investigating under which conditions consumers make certain choices by putting our primary focus on consumer involvement, multiple irrelevant attributes and different product categories .
7 Involvement is an elaborately discussed and researched topic in marketing- and consumer behaviour literature. However, there is to the best of our knowledge, only one study that sheds light on the role of involvement in the context of meaningless differentiation (Sun, 2010). To fill this gap in the literature, this study intends to explore to what extent consumer involvement influences the decision making process with regards to products including irrelevant attributes. We expect that the product evaluation of low and high involved consumers to be different since they process information and thus irrelevant attributes in a distinct way.
.Furthermore, this study provides a better understanding for managers by including two other factors that have, to the best of our knowledge, not been examined before in this context, namely the inclusion of multiple irrelevant attributes and product categories (utilitarian or hedonic). That is, we will investigate if the positive effect of irrelevant attributes will remain with the addition of multiple irrelevant attributes or if it will decrease because of the lack of credibility amongst consumers as a result of the obviousness of the irrelevant attributes. Furthermore, we will research if the nature of the product category (utilitarian or hedonic) has any influence on the product evaluation of products including irrelevant attributes. That is, irrelevant attributes might not be a valued addition to all products. We expect a difference in the product evaluation of hedonic and utilitarian products since these products differ in their nature and irrelevant attributes usually have a stronger association with the nature of hedonic products.
For marketing researchers, it is necessary to explore and clarify the appropriate product strategy for introducing an irrelevant attribute. However, almost all of the above mentioned literature concerns the evaluation of a product in a consumer’s choice set. In this research, consumers will not be exposed to multiple similar products to draw a conclusion and therefore will not be influenced by other factors besides the actual product with the
8 irrelevant attributes.
Therefore, we contribute to the marketing literature by examining the moderating role of involvement and product category on the relationship between (multiple) irrelevant attributes and product evaluation by confronting respondents with one product at the time.
2. Conceptual model and hypothesis development
The following figure represents the framework of this study. This model proposes that there is a relationship between the amount of irrelevant attributes added to a product and product evaluation. Furthermore, this model proposes an interaction effect of involvement and product category on this relationship. In the next subsections we discuss the nature of these relationships and the proposed hypotheses.
Involvement: - high - low Product evaluation: - Attitude - Purchase intentions - Willingness to pay Irrelevant attributes: - 0 Irrelevant attributes - 1 irrelevant attributes - 2 irrelevant attributes Product category: - Hedonic - Utilitarian Product evaluation: - Attitude - Purchase intentions - Willingness to pay
9
2.1 Irrelevant attributes
From a rational perspective, it would be logical to ignore irrelevant information in advertisements since they do not carry any relevant information and therefore cannot be evaluated in a rational way. However, previous research has shown that irrelevant attributes will not be ignored (Brown & Carpenter, 1994; Sun 2010 and Broniarczyk et al., 1997). Next to the arguments that were presented in the introduction of this paper, there are several other theories that suggest that products including irrelevant attributes will be valued despite their irrelevance.
One of the reasons why we not ignore irrelevant information could be explained by the spill-over effect. The spill-over effect takes place when certain attributes rub off on the product. That is, the attributes transport information to the product by evoking certain associations. For example, in this research diamond flakes & mother of pearl extracts are added to the product shampoo and natural bees wax & vitamin b12 to the product mascara. Even though these attributes have no effect on the product performance, one might still associate natural bees wax with natural, thick and full lashes and vitamin b12 with hair growth and therefore longer lashes. The attributes diamond flakes and mother of pearl extracts can be associated with a (natural) glow or shine of the hair. Therefore, one might infer that these products are of better quality than products lacking these attributes because the irrelevant attributes elicit associations that are linked to positive product performance. Another explanation for processing irrelevant attributes could be the informativeness principle of communication theory. According to the informativeness principle of
communications theory, the goal of communicating is to inform or communicate something that is unknown (Clark, 1985; Brown & Carpenter, 1994). According to this theory, a message can contain two elements, i.e. a semantic part (the message's literal meaning) and a pragmatic part (the reason for communicating). If the semantic part is uninformative, one will
10 focus on the reason for communication: the pragmatic part (Gruen- feld & Wyer, 1992 and Brown & Carpenter, 1994). Since irrelevant attributes do not communicate actual valuable information, the consumer would focus on the pragmatic part and might wonder why the communicator is spending money, effort or time into communicating this message in the first place. Therefore, the receiver will imply that the message and thus the irrelevant attributes must be informative and true. Thus, according to this theory, an irrelevant attribute can become relevant to consumers because it conveys pragmatic information.
Previous authors have stated that scarcity and limited editions are related to
perceived expensiveness (Lynn, 1989; Balachander & Stock, 2009). That is, products that are perceived as exclusive will be likely to be more expensive and valuable. Since the irrelevant attributes are unique and rare, we predict that people will be willing to pay more for products including irrelevant attributes because of the uniqueness of the product.
Previous researchers have also found that the number of arguments can affect the way we make decisions (Harkens & Petty, 1981; Petty & Cacioppo, 1983; Chebat et al., 1987; Pentony, 1990 and Green et al., 1990). These authors have stated that, when consumers are presented to multiple arguments, the likeliness of them being persuaded by these arguments increases which can result in a positive change in their behaviour. Furthermore, the previous mentioned arguments that support the outcome of positive product evaluation as a
consequence of the addition of irrelevant attributes do not exclude the effect of multiple irrelevant attributes. Therefore, we propose that a product will be evaluated more positively with the addition of more irrelevant attributes because these products will become more unique, more positive associations are rubbed on the product and the communicator sends out a message that contains more pragmatic information. Since our research is based on
unobvious irrelevant attributes that are linked to possible enhanced product performance, we expect a small increase of product evaluation for products including one irrelevant attribute
11 and a significant increase of product evaluation for products including two irrelevant
attributes since the effect will be stronger.
Previous marketing researchers have investigated consumer’s product evaluation by examining attitude towards the product (Engel et al., 1990; Berens et al., 2005, Albrecht et al., 2010; Sun, 2010), purchase intentions (Dodds et al., 1991; Berens et al., 2005, Albrecht et al., 2010; Sun, 2010) and wiliness to pay (Brown and Carpenter, 2000). Therefore, based on previous research which suggests that the existence of irrelevant attributes increase product evaluation, we propose that:
H1: There will be a positive relationship between the number of irrelevant attributes added to a product and product evaluation:
H1a: Adding one irrelevant attribute to a product will bring about a more positive attitude, higher purchase intentions and willingness to pay among consumers toward this product, i.e. products including one irrelevant attribute will be valued more than products excluding irrelevant
H1b: Adding two irrelevant attributes to a product will bring about a significantly more positive attitude, purchase intentions and willingness to pay among consumers toward this product, i.e. products including two irrelevant attributes will be valued significantly more than products excluding irrelevant attributes and products including one irrelevant attribute.
2.2 The effect of product category (utilitarian vs. hedonic) on product evaluation
Product characteristics have been frequently discussed as an important factor in the marketing literature (e.g. Chen et al., 2009). Products can be distinguished along various dimensions. Multiple researches in the marketing literature have classified products on the basis of utilitarian- and hedonic characteristics (e.g. Geuens et al., 2006; Drolet et al., 2007; Okada, 2005 and Bottemly & Doyle, 2006). In the marketing literature, there are many terms
12 that are used interchangeablyfor these two concepts such as the terms practical and useful that refers to utilitarian products and luxury-, sensory-social- and pleasure based products that refer to hedonic products. In the present research, we limit ourselves to the terminology of hedonic and utilitarian products and adopt the conceptualizations of Park (1986). In our context, utilitarian products can be viewed as products that fulfil a need for problem solving or problem prevention (Park et al., 1986) and therefore are of a practical or useful nature. Examples of this category are paper tissues, laundry detergent, shampoo and aspirins. In this research we consider hedonic products as products that fulfil a need for personal expression, convey status and are related to sensory pleasure (look, taste or smell nice). Therefore, these products are more pleasure and experiential based. Examples of hedonic products are
perfume, chocolate and a majority of cosmetic products.
Previous literature suggests that people value utilitarian and hedonic products in a different way (Park et al., 1986; Palazon et al., 2013 and O’Curry, 2001). Thus, the nature of the product category (utilitarian or hedonic) can have different connotations for the decision-making process in a differentiation context. According to Levitt (1980), a product consists of four tiers which are the generic product, the expected product, the augmented product and the potential product. The generic product is described as the base of the product that is necessary and fundamental to make the product successful. The expected product tier consists of the minimum (anticipated) demands of the customer. In the augmented phase there is room to excel and surpass customers’ expectations. In our case, one could propose that there is room in the augmented phase for trivial differentiation by adding irrelevant attributes to the product. Levitt (1980) points out that not all customers will be attracted to all products that are
amplified by the addition of items in their augmented tier. Because hedonic products are more pleasure and experiential based, we suggest that there is more room (in the augmented tier) for irrelevant attributes as opposed to utilitarian products that are focused on problem
13 solving and practical aspects of the product.
Previous research has also suggested that consumers who are motivated by hedonic needs or are searching for hedonic products are more willing to process information since they are looking for emotional worth and entertainment and therefore spend more time and effort in to picking out a product. On the other hand, utilitarian shoppers are more
concentrated on completing the task in an efficient way (i.e. get the right product for the right price) and therefore are less inclined to process more information (here: irrelevant attributes) since their rationality plays an important role in the decision making process (Teller et al., 2008; Barbin et al., 1994 and Jones et al., 2006). Therefore, it is highly possible that
consumers shopping for hedonic products give greater weight to products including irrelevant attributes, since it is more directly linked to the sensory experience, whereas consumers shopping for utilitarian products attach less value to products including redundant
information (here: irrelevant attributes), since this hinders them to complete shopping tasks efficiently.
According to Gierl and Huettl (2012), hedonic products are more likely to carry hedonic features such as art because of the spill-over theory that was described earlier. In our context, hedonic products would be more likely to have a transfer of information (here: irrelevant attributes) because both stimuli, the hedonic and irrelevant attributes, can be classified on the same dimension (Hirschman & Holbrook, 1982). On the other hand, this transfer of information will probably not occur for utilitarian products since consumers are less likely to cognitively control the transfer of luxury conations. That is, in the case of utilitarian products, the stimuli, hedonic (here: mother of pear extracts and diamond flakes) and irrelevant attributes, are not classified on the same dimension since the attributes are more luxurious than the nature of the product, which is practical.
14 valued significantly more positive than utilitarian products including more irrelevant
attributes. More particularly, we expect that the product evaluation of hedonic products including irrelevant attributes will increase and that of utilitarian products will decrease because the nature of hedonic products is associated with irrelevant attributes and because consumers of utilitarian products will only be ‘hindered’ by irrelevant attributes. We expect this effect to be stronger with the addition of two irrelevant attributes. Therefore, we propose that product category will moderate the relationship between irrelevant attributes and product evaluation:
H2: The effect of irrelevant attributes on product evaluation is contingent on product category, with the effect of irrelevant attributes on product evaluation (attitude, purchase intentions and willingness to pay) being significantly higher in the context of hedonic products and significantly lower in the context of utilitarian products, i.e. hedonic products including one irrelevant attributes will be valued significantly more than hedonic products excluding irrelevant attributes and hedonic products including two irrelevant attributes will be valued significantly more than hedonic products excluding or including one irrelevant attribute. Utilitarian products including one irrelevant attributes will be valued significantly lower than products excluding irrelevant attributes and utilitarian products including two irrelevant attributes will be valued significantly lower than utilitarian products excluding or including one irrelevant attribute.
2.3 The effect of involvement on product evaluation
In the past, various authors have showed that significant differences can exist in product evaluation as a result of different levels of involvement (e.g., Petty et al., 1983; Petty & Cacioppo,1984; Chaudhuri, 2000; Moaz & Tybout, 2002). Although the term involvement
15 has been an elaborately discussed topic in the marketing literature, in the field of consumer behaviour, product involvement broadly has been described as a person's perceived relevance of the focal object based on three major properties: consumer's inherent needs, interests, and values (Bloch & Richins, 1983; Andrews et al., 1990; Zaichkowsky, 1994 and Sun, 2010). According to Traylor (1981), product involvement reflects recognition that a certain product category may be more or less central to people's lives, their sense of identity, and their relationship with the rest of the world. Furthermore, Celsi & Olson (1988) described involvement as a concept that reflects an individual’s goals and basic values. Similarly, Greenwald & Leavitt (1984) have stated that there is a consensus that high involvement is related to personal relevance or importance which is also supported by Petty & Cacioppo (1984). In summary, product involvement is subject to one’s personal experiences and life. In this study, we focus on high product involvement, in which the product is central to the consumer’s life and low product involvement, in which the product doesn’t fulfil an important role in the consumer’s life.
Research in persuasion has consistently found that consumers in high product involvement situations process information in a different way than those in low product involvement situations (e.g., Chaudhuri, 2000; Moaz & Tybout, 2002). The Elaboration Likelihood Model (ELM) developed by Petty & Cacioppo posits that consumers pursue either a central route or a peripheral route when processing information (Petty et al., 1983; Petty & Cacioppo, 1984). Consumers who are high involved with a product are likely to follow the central route to persuasion, in which consumers process the necessary cognitive effort to evaluate arguments. In these situations, consumers will be more inclined to focus on highly diagnostic cues such as information about the product and relevant features in order to make an evaluation of the product. The central route followers tend to ignore simple product cues such as price and brand name. On the other hand, consumers who are low involved with a
16 product are more likely to follow the peripheral route to persuasion in which consumers evaluate products based on a careless analysis of simple and salient cues such as price or celebrity endorsement. Therefore, in purchase decisions, consumers in high product involvement situations tend to search for more product information (e.g., detailed attribute information) and make more product comparisons to insure product quality and value. In contrast, consumers in low product involvement situations may rely more on salient and readily available cues to make product inferences (Nijssen et al., 1995).
Previous literature have suggested that number of arguments in a message affects the way people think and to what extent they agree with a certain matter because they have more to think about (Petty & Cacioppo, 1984; Chaiken, 1980). Petty & Cacioppo (1984) have found that the number of arguments can serve as a peripheral cue. That is, consumers who follow the peripheral route and are low involved may be persuaded more easily if multiple arguments (here: attributes) are presented to them. Since they are low involved and possess over less knowledge about the product (than people who follow the central route) more information will be considered as the truth and products will become more valuable with the addition of more peripheral cues. Therefore, peripheral route followers would be more inclined to accept a message in a communication context (here: product persuasion in an advertisement). Since the brand name and packaging of the presented products are fictitious in our study, these aspects should not function as a salient cue since these are nonexistent and couldn’t be compared to a similar products in the experiment. Therefore, consumers who follow the peripheral route would focus on new information: irrelevant attributes.
Petty & Cacioppo (1984) have also found that high involved consumers who follow the central route are less susceptible to weak arguments since they can distinguish them from legitimate arguments or information. According to this theory, high involved consumers who are involved with the product and posses over prior knowledge about the product will
17 evaluate the products with irrelevant attributes less positive since they are more likely to distinguish the irrelevant from relevant information. In contrast, consumers in low involvement situations who follow the peripheral route are more likely to evaluate the product including irrelevant attributes more positively since they will assume that the more arguments and thus information a product entails the more credible and better the product is. Petty & Cacioppo (1984) suggest that this effect would be intensified with the addition of more arguments. We therefore expect a more positive product evaluation amongst high involved consumers for products containing more irrelevant attributes than low involved consumers.
Since we apply unobvious irrelevant attributes it might be difficult to detect the irrelevant nature of one irrelevant attribute. Therefore, we propose that there will not be a significant moderating effect of product involvement with the addition of one irrelevant attribute. That is, we expect that both high and low involved consumers will positively respond to the addition of 1 irrelevant attribute since the relevance is not obvious. Low involved consumers will receive the irrelevant attribute as a peripheral cue and high involved consumers will receive the irrelevant attribute as desired extra information. However, when consumers are exposed to two irrelevant attributes we expect that these peripheral cues (i.e., irrelevant attributes) will result in a more positive product evaluation for low involved consumers but in a less positive product evaluation for high involved consumers since they will be more likely to detect the irrelevance of the attributes.
Because the level of involvement is associated with consumer behaviour (Xuemei and Luiz, 2011; Petty & Cacioppo, 1984) and consumers in high product involvement situations tend to search for more valid product information and strong arguments (irrelevant attributes are not valid and strong arguments) to guarantee product value, and consumers in low involvement situations tend to focus on salient cues (here: irrelevant attributes), this study therefore
18 assumes that the level of product involvement will moderate the relationship between
products with two irrelevant attributes and product evaluation. We propose that:
H3a: The addition of one irrelevant attribute will bring about a more positive product evaluation for both low and high involved consumers, i.e. both low and high involved consumers will have a significantly more positive attitude, higher purchase intentions and willingness to pay towards products including one irrelevant attribute than products excluding irrelevant attributes.
H3b: The effect of two irrelevant attributes on product evaluation (attitude, purchase intentions, willingness to pay) is contingent on product involvement, with the effect of two irrelevant attributes on product evaluation being more significant in the context of low involved consumers than in the context of high involved consumers, i.e. low involved
consumers will have a significantly higher product evaluation towards products including two irrelevant attribute than products excluding or including one irrelevant attribute.
High involved consumers will have a significantly lower product evaluation towards products including two irrelevant attributes than products excluding or including one irrelevant
attribute.
Research design and procedure
With this study we research the effect of irrelevant attributes on product evaluation.
Also, we investigate if there is a moderating effect of involvement and a moderating effect of product category on the relationship between two irrelevant attributes and product evaluation. In order to investigate these effects, which are researched in an experimental design, two preliminary tests were conducted prior to the experiment.
The purpose of the first pre-test was to determine hedonic and utilitarian products. Participants considered purchasing a variety of 6 products (deodorant, shampoo, laundry
19 detergent, massage oil, perfume and mascara) and stated whether their decisions would be motivated by the fulfilment of functional or hedonic needs. The same participants were also asked how involved they were with all six products. The products that were rated the most hedonic and utilitarian and that showed the highest dispersion in the level of involvement were selected and used in the experiment.
A second pre-test was administered to a group with the purpose to identify irrelevant
attributes. In this test, the respondents were presented to several irrelevant attributes related to the selected hedonic and utilitarian products and were asked about the relevance of these attributes. The attributes that scored the lowest on relevance were selected as the irrelevant attributes for each product in the experiment.
The information that was gathered from the preliminary steps was applied in the experimental online study which was a 3 (0 irrelevant attributes/ 1 irrelevant attribute/ 2 irrelevant
attributes) x 2 (utilitarian/hedonic) x 2 (low involvement/high involvement) design. The purpose of this experiment was to investigate the effect of irrelevant attributes on product evaluation. Product evaluation was measured by three separate dimensions; attitude, purchase intentions and willingness to pay. Furthermore, we wanted to investigate the effect of the number of irrelevant attributes on product evaluation. Therefore, three separated groups were used in this study that were exposed to either none, one or two irrelevant attributes.
Subsequently, we investigated the moderating effect of involvement and the moderating effect of product category (hedonic/utilitarian) on the relationship between irrelevant attributes and product evaluation. To test the moderating effect of involvement, each group was asked about their involvement with the selected products and all participants were exposed to both a hedonic and a utilitarian product to examine the moderating effect of product category.
20 PRELIMINARY TESTS
Sample
For the first pre-test, 38 participants (18 men and 20 women ranging in age from 23
to 58 and with different educational levels) were asked to identify hedonic and utilitarian products. For the second pre-test, 43 respondents (20 men and 23 women, ranging in age from 23 to 54 and with different educational levels) were asked to rate the importance of each attribute.
Measures
Hedonic and utilitarian products. Participants stated whether their decisions would be motivated by the fulfilment of functional or hedonic needs using a scale with endpoints ‘completely functional’ (1) and ‘completely hedonic’ (7). This method was adapted from Leclerc, Schmitt & Dubé (1994) and Bottomley and Doyle (2006). We deliberately did not change the word ‘functional’ to utilitarian in the survey due to the fact that these terms are used interchangeably in the literature, are consistent in meaning and the word ‘functional’ is more likely to occur in the vocabulary of the participants. We provided the respondents with the definitions of functional- and hedonic products to diminish perception errors. Utilitarian products were stated as, ‘products that fulfil a need for problem solving or problem
prevention’. Hedonic products were defined as those ‘that fulfil a need for personal expression, convey status, attain social approval or sensory pleasure (look, taste or smell nice), variety or stimulation’. The two definitions were selected from Park et al. (1986).
Level of involvement. Participants were asked how involved they were with the products using a reduced version of the involvement scale of Zaikowsky’s PII (1984) that was constructed and tested by Mittal (1994). The involvement scale consisted of 5 items that were all measured on a 7-point Likert scale. Respondents were asked to indicate
21 how important each product was. The items ranged from ‘important - unimportant’, ‘of no concern to me– of concern to me’1, ‘means a lot to me – means nothing to me’, ‘significant - insignificant’ and ‘matters to me – does not matter to me’. The reduced involvement scale of five items was found to be highly reliable in the pre-test (laundry detergent: α = .916,
shampoo: α = .929, perfume: α = .867, mascara: α=.968).
Irrelevant Attributes. Respondents were asked to rate the importance of each attribute, using a 5-point Likert scale ranging from “not relevant at all” to “very relevant”. We included several relevant attributes for each product to diminish suspicion among respondents about the purpose of the survey. The questioned attributes were made up and have, to the best of our knowledge, no influence on product performance and therefore are irrelevant.
Results
First pre-test: Identifying hedonic and utilitarian products. The two most
utilitarian and two most hedonic products were selected for the subsequent involvement test. The selected utilitarian products were laundry detergent (M= 2.33) & shampoo (M= 2.76) and the selected hedonic products were perfume (M= 5.55) and mascara (M= 5.17) (see table 1 for all results). A repeated measures ANOVA was used to demonstrate that the four products differed significantly. However, Mauchly's Test of Sphericity indicated that the assumption of Sphericity had been violated, χ2 (14) = 37.752, p = .001, and therefore, a Greenhouse-Geisser correction was applied. The Greenhouse-Greenhouse-Geisser test determined that the four products differed significantly from each other, F(3.608, 133.499) = 28.945, p = .001. Post hoc tests using the Bonferroni correction revealed that the utilitarian products were
statistically significantly different from the hedonic products (p < .001). In other words, the
1 Item was reversed coded
22 selected functional products indeed belonged to a different category than the selected hedonic products.
With regards to involvement, we selected the products which showed the highest variance in the level of involvement because we will research involvement as a moderating variable in the experimental design and are therefore interested in products with a high diversity in involvement. Therefore, even though laundry detergent (M= 2.33) and perfume (M= 5.55) were rated more utilitarian and hedonic than shampoo (M= 2.76) and mascara (M= 5.17), we selected the functional product that had the highest dispersion in involvement which was shampoo (SD= 1.86) and the hedonic product that had the highest dispersion in involvement which was mascara (SD = 2.33) (see table 2 for all results).
Second pre-test: identifying trivial attributes for the selected products. The attributes that scored the lowest on relevance were selected as the irrelevant attributes for each product. The attributes mother of pearl extracts and diamond flakes were selected for the utilitarian product shampoo and the attributes vitamin b-12 and natural beeswax were
selected for the hedonic product mascara (see table 3 and 4 for all results).
MAIN ANALYSIS
Sample. The sample size was investigated across the treatment groups (see table 5). The complete sample of the experimental study consisted of 283 respondents that were mainly students. 69 respondents were dropped because of incomplete answers resulting in a final sample size of 214 respondents (group exposed to 0 irrelevant attributes: n=84, group exposed to 1 irrelevant attribute: n=66 and group exposed to 2 irrelevant attributes: n=64). It seems that the high bounce rate was a consequence of a technical error since 85 percent (n=59) of the incomplete surveys were quitted at the same question. Of the 214 respondents who completed the survey, 25% were male and 75% female. The respondents’ age ranged from 18 to 65 while 79 % (n=166) of them were 18 – 27 years of age. Subjects were
23 randomly assigned to a group. A comparison of the three groups shows that these do differ to some extent from each other with respect to the characteristics gender, age, and level of education. Especially group 1 has more male respondents and relatively younger respondents than the other two groups.
Table 5: demographics sample group
Sample size
Gender % Gender # Highest completed education
Age (18 -27) Group 0 I.A. n= 84 Male:29%
Female: 71% Male: n=25 Female: n=59 High School: 34% Bachelor: 46% Master: 20% 71.5%
Group 1 I.A. n=64 Male: 18% Female: 82% Male: n=12 Female: n=52 High School: 29% Bachelor: 50% Master: 21% 86.5%
Group 2 I.A.’s n=66 Male: 28% Female: 72% Male: n=18 Female: n=48 High School: 28% Bachelor: 47% Master: 25% 78.8%
I.A. = irrelevant attribute
Measures
Hedonic and utilitarian products. Participants stated whether the products were considered functional or hedonic using a scale with endpoints ‘completely functional’ (1) and ‘completely hedonic’ (7). This method was adapted fromLeclerc, Schmitt & Dubé (1994) and Bottomley and Doyle (2006) and was used the same way as in the pre-test.
Level of involvement. Participants were asked how involved they were with the products using a reduced version of the involvement scale of Zaikowsky’s PII (1984) that was constructed and tested by Mittal (1994) as in the pre-test. The reduced involvement scale of five items was found to be reliable. Cronbach's alphas for the involvement items of
mascara and shampoo were .69 and .88, respectively.
Attitude. The scale for brand attitude consisted of one item assessing overall brand evaluation on a 10-point scale ranging from ‘I like the brand very much’ to ‘I dislike the
24 brand very much’ as applied by Hartman and Apaolaza- Ibáñez (2012).
Purchase Intentions. Participants rated their likelihood on purchasing the products on a traditional 5-point Likert scale ranging from ‘definitely will not buy, probably will not buy, might/might not buy, probably will buy’ to ‘definitely will buy’ (Kalwani & Silk, 1982; Jamieson & Bass, 1989 and Hartman & Apaolaza-Ibáñez, 2012).
Willingness to pay. Respondents’ willingness to pay was measured on a continuous scale. Participants were presented the average price of the selected products (shampoo: €2.50 and mascara €10.00) and were asked to enter the price they were willing to pay for the products in euro’s.
Advertising stimuli.
The online experiment simulated the consumer’s confrontation with the selected products. The products were presented with a branded packaging and description of the product. To avoid systematic biases that could arise from potential influences of existing brands, fictitious brands were used across all advertisements and products packages were created by the author. The subjects in the control group were presented to an advertisement of a utilitarian and a hedonic product excluding irrelevant attributes (figure 1), other subjects were presented to a utilitarian and a hedonic product including either one (figure 2) or two irrelevant attributes (figure 3). To enhance the nature of the hedonic and utilitarian products, the hedonic product was presented in a colour that is associated with luxury (i.e. gold) while the utilitarian product was presented in a functional colour (blue) (Melillo, 2006; Bottomley, 2006). With the exception of the integration of the irrelevant attribute, the visual and textual designs were identical across all products to diminish influence of other elements such as colour or design. Besides the visual advertisement, the respondents were also exposed to the description box of the fictive brands. Similar to the visuals, the description boxes were
25 identical for all products except for the information regarding the irrelevant attributes.
Manipulation Check
Product category. We asked all respondents in the experimental design to give their opinion about the nature of the products (utilitarian/hedonic). We provided them with the same definitions that were used in the pre-test and asked them on a scale of 7 how functional or hedonic (1 = functional, 7= hedonic) they found the products. In line with the results of the pre-test, there was a distinction between the utilitarian and hedonic product (t (426) = -19.26,
p < .001). That is, the respondents considered shampoo as a utilitarian product (M = 2.31) and
mascara as a hedonic product (M = 5.30).
Involvement. Involvement was measured again in the experimental design to measure the interaction effect of involvement on the relationship between products and product evaluation. In contrast to our pre-test (see table 2, shampoo SD=1.86, mascara
SD=2.33), the level of involvement was not as dispersed and differed across the group. That
is, from the entire sample, 79% were low involved with shampoo and 57 % of the sample was low involved with mascara.
Table 6: Mean and standard deviation for product involvement (on a scale of 1 – 7) 0I.A. 1I.A. 2I.A. Involvement M SD M SD M SD
Shampoo 3.13 1.36 2.81 1.14 2.71 1.44
26 Graph 1: Product involvement (on a scale of 1 – 7)
Table 7: Number of participants with low and high involvement across research groups
Number of participants 0I.A. 1I.A. 2I.A.
Low shampoo 60 53 57
High shampoo 24 11 9
Low mascara 46 44 33
High mascara 38 22 33
low = 1 – 3.9 (on a scale of 1 – 7) high= 4 – 7 (on a scale of 1 -7)
Results Main Analysis
Descriptive date. Table 8 displays the means and standard deviations for all the measures categorized by treatment group (group exposed to 0, 1 or 2 irrelevant attributes). As expected, differences exist between the groups across the outcome variables. Also, as
mentioned before, differences were found between the groups on involvement for both products. Since the respondents were mainly students, it is not surprising that the average age is similar across treatment groups.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 In volvemen t
Number of Irrelevant Attributes
Shampoo Mascara
27
Table 8 Means and standard deviations for the study measures
0 I.A. 1 I.A. 2 I.A.’s
Variable M SD M SD M SD
Attitude shampoo 5.19 1.81 5.14 2.23 4.98 1.76
Attitude mascara 5.25 2.15 4.73 1.79 4.83 2.28
Purchase intentions shampoo 3.49 1.15 3.53 1.21 3.61 1.02 Purchase intentions mascara 4.19 1.77 2.89 0.99 2.79 1.09 Willingness to pay shampoo 2.73 1.29 2.72 1.58 3.35 1.70 Willingness to pay mascara 7.77 4.39 8.78 4.40 10.28 5.03 Involvement shampoo 3.13 1.36 2.81 1.14 2.71 1.44 Involvement mascara 3.97 1.58 3.71 1.39 4.01 4.50
Age 27.02 6.37 25.62 6.41 26.44 7.04
Correlation analysis. Pearson correlations were calculated between all outcome measures (see Table 9). A significant weak negative correlation was found between overall attitude and overall purchase intentions (r = -0.27, p<0.001) and between overall attitude and willingness to pay (r = -0.11, p<0.005). Furthermore, a significant weak positive correlation was found between overall purchase intentions and overall willingness to pay (r = 0.22,
p<0.001). Since we are also interested in the separate product evaluations of product
category, we also calculated the Pearson correlations between all outcome measures per product category (see table 10 and 11). As regards to the utilitarian product shampoo, a significant strong negative correlation was found between attitude and purchase intentions (r = -0.63, p<0.001) and between attitude and willingness to pay (r = -0.35, p<0.001).
Furthermore, a significant weak positive correlation was found between purchase intentions and willingness to pay (r = 0.27, p<0.001). With regards to the hedonic product shampoo, a significant strong negative correlation was found between attitude and purchase intentions (r = -0.4, p<0.001) and between attitude and willingness to pay (r = -0.55, p<0.001). In line
28 with previous results, a significant weak positive correlation was found between purchase intentions and willingness to pay (r = 0.36, p<0.001). Based on these results, the outcome variables are significantly correlated with the each other but the direction differs. That is, there is a negative correlation between attitude and purchase intentions and willingness to pay. On the other hand, there is a positive correlation between purchase intentions and wiliness to pay.
Table 9 Pearson correlations between overall attitude, overall purchase intentions and overall willingness to pay (independent variables).
Variable 1. 2. 3.
1. Attitude 1
2. Purchase intentions -.27* 1
3. Willingness to pay -.11** .22* 1
*p<.05, **p< .01
Table 10 Pearson correlations between attitude, purchase intentions and willingness to pay for the utilitarian product shampoo
Variable 1. 2. 3.
1. Attitude 1
2. Purchase intentions -.63* 1
3. Willingness to pay -.35* .27* 1
29 Table 11 Pearson correlations between attitude, purchase intentions and willingness to pay for the hedonic product mascara
Variable 1. 2. 3.
1. Attitude 1
2. Purchase intentions -.46* 1
3. Willingness to pay -.55* .36* 1
*p< .01
Effect of irrelevant attributes on product evaluation. We first investigated the main effect between products including and excluding irrelevant attributes and product evaluation. Since respondents were asked about their involvement in shampoo and mascara separately in the surveys, we combined the separate outcome variables into three main outcome variables, i.e. overall attitude, overall purchase intentions and overall willingness to pay.
Levene’s test verified the equality variances in the samples (homogeneity of variances) (p > .05). A significant main effect (see table 12) was found between irrelevant attributes and purchase intentions (F(2,215) = 7.17, p = .001) and willingness to pay (F(2,215) = 5.86, p = .002) but no significant difference was found between irrelevant attributes and the dimension attitude (F(2,215) = .467, p = .628).
Table 12: Main effect of irrelevant attributes (I.A.) on product evaluation
0I.A 1I.A. 2I.A.
Dimensions product evaluation M SD M SD M SD
Attitude 5.22 1.94 4.93 2.02 4.91 2.05
Purchase Intentions 3.83 1.51 3.21 1.14 3.19 1.11
Willingness to pay 5.30 4.12 5.75 4.48 6.77 5.09
30 of irrelevant attributes (M = 5.22, SD = 1.94; M = 4.93, SD = 2.02; M = 4.91, SD = 2.05) and were therefore not supported on this dimension (attitude).
With regards to purchase intentions, planned contrasts and a Tukey post-hoc test revealed that the group that was not exposed to irrelevant attributes significantly differed (M = 3.83, SD = 1.51) from the group that was exposed to one irrelevant attribute (M = 3.21, SD = 1.14) and the group that was exposed to two irrelevant attributes s (M= 3.19, SD = 1.11) on this dimension of product evaluation (p = .005), but there was no significant difference between the group that was exposed to one irrelevant attribute and two irrelevant attributes (p = 1,00). As opposed to hypothesis H1, people were more inclined to purchase products excluding irrelevant attributes than products including irrelevant attributes. In fact, purchase intention decreased with the addition of irrelevant attributes. Therefore, these results are not supporting for hypotheses H1a and H1b for the dimension purchase intentions.
As regard to willingness to pay, there was no significant difference between the group that was not exposed to irrelevant attributes and the group that was exposed to one irrelevant attribute ( p = 1,00) but there was a significant difference between the group that was not exposed to irrelevant attributes (M= 5.30, SD =4.12) and the group that was exposed to two irrelevant attributes (M= 6.77, SD= 5.09) (p=.004). Also, there was a significant difference between the group that was exposed to one irrelevant attribute (M= 5.75, SD= 4.48) and the group that was exposed to two irrelevant attributes(M= 6.77, SD= 5.09.70) (p= .047) on this dimension.Therefore, products including two irrelevant attributes were valued significantly more than products excluding irrelevant attributes and products including one irrelevant attribute on this dimension. These results are in line with hypothesis H1a and H1b and therefore support the first hypotheses regarding the dimension willingness to pay.
31 The moderating effect of product category.
To investigate the interaction effect of product category on the relationship between irrelevant attributes and product evaluation, data were analyzed using the macro Process (A. Hayes) in SPSS hence the variables were centred and multicollinearity was avoided. In order to interpret the interaction effect and plot the interaction graphs, repeated measures were used since the macro Process does not provide the significance region for categorical moderators (here: product category).
A significant interaction effect of product category (see table 13 and 14) was found between irrelevant attributes and purchase intentions F(2, 215) = 21.95, p < .001 and willingness to pay F(2,215) = 3.78, p = .024. No significant interaction of product category was found between irrelevant attributes and attitude F(2,215) = 1.05, p = .352. If we take a closer look at these results on the dimension attitude (see graph 2), we see that attitude towards the
utilitarian product shampoo decreases with the addition of irrelevant attributes which is in line with the direction of hypothesis 2. As regards the hedonic product shampoo, people s’ attitude decreased as well with the addition of one irrelevant attribute which is not in line with our proposition. However, attitude has a slight increase for hedonic product including two irrelevant attributes (as opposed to products including one irrelevant attribute) but these results do not exceed the product evaluation of product excluding irrelevant attributes. Since there was no moderating effect of product category on this dimension and not all results
32 Table 13: Moderating effect of product category
0I.A. 1I.A. 2I.A.
Product evaluation M SD M SD M SD
Attitude Shampoo 5.19 1.81 5.14 2.23 4.98 1.76
Attitude Mascara 5.25 2.15 4.73 1.79 4.83 2.28
Purchase Intentions Shampoo 3.49 1.15 3.53 1.21 3.61 1.02 Purchase Intentions Mascara 4.19 1.77 2.89 0.99 2.79 1.09 Willingness to pay Shampoo 2.73 1.29 2.72 1.58 3.35 1.70 Willingness to pay Mascara 7.77 4.39 8.78 4.40 10.28 5.03
Graph 2: moderating effect of product category - attitude
With regards to purchase intentions (see graph 3) , planned contrasts and a Tukey post-hoc test revealed that the group that was not exposed to irrelevant attributes (Shampoo: M= 3.49,
SD= 1.15; Mascara: M= 4.19, SD= 1.77) significantly differed from the group exposed to one
irrelevant attribute (Shampoo: M= 3.53, SD= 1.21; Mascara: M= 2.89, SD= 0.99) and the group that was exposed to two irrelevant attributes (Shampoo: M= 3.61, SD= 1.02; Mascara:
M= 2.79, SD= 1.09) on purchase intentions (p = .003). There was no significant difference
between products containing one irrelevant attribute and two irrelevant attributes on this
0 1 2 3 4 5 6 0 1 2 Attitud e
Number of Irrelevant Attributes
Shampoo Mascara
33 dimension (p = 1.00). Despite the significant differences between groups, the moderating effect of product category on the relationship between irrelevant attributes and purchase intentions was in the opposite proposed direction. That is, people were less inclined to buy the hedonic product mascara including irrelevant attributes but more inclined to purchase the utilitarian product shampoo including irrelevant attributes. Since there was no significant interaction effect of product category and the results are not in the proposed direction, hypothesis H2 was not supported on this dimension (purchase intentions).
Graph 3: Moderating effect of product category - purchase intentions
As regards to willingness to pay (see graph 4), the group that was not exposed to irrelevant attributes (Shampoo: M=, 2.73, SD= 1.29; Mascara: M= 7.77, SD= 4.39) only differed from the group that was exposed to two irrelevant attributes (Shampoo: M= 3.35, SD= 1.70; Mascara: M= 10.28, SD= 5.03) (p = .002) and not from the group that was exposed to one irrelevant attribute (p = .713). The group that was exposed to one irrelevant attribute did significantly differ from the group exposed to two irrelevant attributes (p = .031). The results
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 Purch ase In tentions
Number of Irrelevant Attributes
Shampoo Mascara
34 disclose that people were willing to pay more for hedonic products including one irrelevant attribute and significantly more for hedonic products including two irrelevant attributes, which is in line with hypothesis H2. However, the same holds for utilitarian products, which is in conflict with hypothesis H2. Since there was no significant interaction effect of product category and the results are not all in the proposed direction, hypothesis H2 was not
supported on this dimension (willingness to pay).
Graph 4: Moderating effect of product category - willingness to pay
The moderating effect of involvement
In the second analysis involvement was examined as a moderator of the relation between irrelevant attributes and product evaluation. As in the previous moderation analysis, the macro Process (A. Hayes) was used so that the variables were centred and
multicollinearity was avoided. Analysis revealed that there was no significant interaction effect of the level of involvement on the relationship between irrelevant attributes on any of the three indicators for product evaluation: attitude (b = .052, 95% CI [-.106, .210], t = .649,
p = .517), purchase intentions (b = -.008, 95% CI [-.168, .147], t = -.121, p = .903) and 0 2 4 6 8 10 12 0 1 2 Will iingnes s To Pa y
Number of Irrelevant Attributes
Shampoo Mascara
35 willingness to pay (b = -.049, 95% CI [-.204, .106], t = -.624, p = .534).
Since the manipulation check revealed that the level of involvement was not consistent with the results of the pre-test, we investigated the dispersion of involvement amongst the different groups (see table 6 & 7 and graph 1). As the distribution of involvement is quiet evenly distributed for the hedonic product mascara as opposed to the distribution of involvement regarding the utilitarian product shampoo , we investigated the moderating effect of
involvement for each product category separately (see table 15) with the focus on the hedonic product mascara. Analysis revealed that there was no moderating effect of involvement on the relationship between irrelevant attributes included to utilitarian products and any of the outcome variables of product evaluation (attitude, p=.819, purchase intentions, p=.490 and willingness to pay, p=.342). However, simple slopes and graphs reveal the direction of the results (see in annex graph 5, 6 and 7 & simple slopes 1, 17 and 18 for all results for the utilitarian product shampoo).
As regards to the hedonic product mascara, no moderating effect was found on the relationship between irrelevant attributes and the outcome variables attitude (p= .076), purchase intentions (p= .832) and willingness to pay (p=.345) either. Even though the moderating effects were also not significant, simple slopes (table 19) and interaction graph 8 still reveal for which level of involvement results are significant and the direction of the results. The results of the simple slopes disclose that involvement will only moderate the relationship between irrelevant attributes included to hedonic products and attitude when it is low (b= -0.560, 95% CI [-1.080, -0.039], t = -2.120, p = .035). Graph 8 reveals the direction of the interaction effect: consumers who are high involved will have a slightly increased positive attitude towards products including more irrelevant attributes as opposed to low involved consumers who will have a significant decreased positive attitude towards products including more irrelevant attributes. Since there was no significant interaction effect of
36 involvement and the results were not in line with the proposed direction, hypothesis H3a and H3b were not supported on this dimension (attitude – mascara).
Graph 8: Moderating effect of involvement - attitude (mascara)
As regards to purchase intention, the results of the simple slopes (see table 20) disclose that involvement will only moderate the relationship between irrelevant attributes included to hedonic products and purchase intention when it is low (b= 0.703, 0, 95% CI [1.054, -0.351], t = -3.942, p <.001) and medium (b= -0.703, 0, 95% CI [-0.961, -0.501], t = -235,
p<.001). That is, low involved consumers are overall more inclined to purchase the product
than medium and high involved consumers which is in line with the proposed direction. However, all consumers’ purchase intentions decrease with the addition of more irrelevant attributes. Since there was no significant interaction effect of involvement and the results were not in line with the proposed direction, hypothesis H3a and H3b were not supported on this dimension (purchase intentions – mascara).
0 1 2 3 4 5 6 7 0 1 2 Attitud e Mascar a
Number of Irrelevant Attributes
Low Medium High
37 Graph 9: Moderating effect of involvement on purchase intentions (mascara)
With regards to willingness to pay, the results of the simple slopes (see table 20) and graph 10 disclose that involvement will only moderate the relationship between irrelevant attributes included to hedonic products and willingness to pay when it is low (b= 1.571, 95% CI [0.516, 2.623], t = 2.931, p = .003) and medium (b=1.182, 95% CI [0.451, 1.923], t = 3.182, p = .001. That is, in line with our proposition, low involved consumers are overall willing to pay more for the hedonic product mascara including more irrelevant attributes. However, this also holds for high involved consumers which is not in line with our proposition. Since there was no significant interaction effect of involvement and the results were not all in line with the proposed direction, hypothesis H3a and H3b were not supported on this dimension
(willingness to pay – mascara).
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 1 2 Purch ase In tentions Mascar a
Number of Irrelevant Attributes
Low Medium High
38 Graph 10: Moderating effect of involvement on willingness to pay (mascara)
Since mascara is a product used by females, the moderating role of involvement was also investigated with only the female respondents as sample group (n=160) since this would be a more realistic representation of the population. However, no moderating effect of
involvement was found on either of the dimensions of product evaluation: attitude (b= 0.097, 95% CI [-0.243, 0.437], t = 0.563, p = .574), purchase intentions (b= -0.003, 95% CI [-0.203, 0.196], t = 0.034, p = .972) and willingness to pay (b= 0.284, 95% CI [0.880, 0.311], t = -0.943, p = .346) and simple slope results revealed that the results were in the same direction as for the complete sample size (including males).
Discussion and implications
Past research on meaningless differentiation suggests that adding an irrelevant
attribute to a product could either lead to positive or negative product evaluation. The answer to whether irrelevant attributes are of added value is contingent on different factors such as the definition of irrelevant attributes and the experimental design as seen in previous research. Therefore, this paper endeavoured to filter situations in which adding irrelevant attributes would have a significant positive effect on product evaluation.
0 2 4 6 8 10 12 14 0 1 2 Will ingnes s To Pa y Mascar a
Number of Irrelevant Attributes
Low Medium High
39 Main analysis
The main analysis showed that even though consumers’ attitude and purchase
intentions decreased by the addition of irrelevant attributes, these people were still willing to pay more for these products. The difference between the products including one and two irrelevant attributes was rather small which could be explained by the fact that we applied irrelevant attributes that were rated ‘least irrelevant’ in our pre-test. The majority of the irrelevant attributes used in previous research were ambiguous. These ambiguous attributes could lead to a more extreme product evaluation than the present research since the
differences between the products including and excluding irrelevant attributes are greater as a consequence of stronger arguments (here: ambiguous attributes) (Petty & Cacioppo, 1984). Consistent with previous research (Carpenter, 1994), consumers are willing to pay more for products including irrelevant attributes. An explanation for the decrease in attitude and purchase intentions and the increase of willingness to pay could be the exclusiveness of the products (Lynn, 1989; Balachander & Stock, 2009). That is, the products including the irrelevant attributes might be perceived as rare since they do not exist in the actual market. These products might be interesting to a niche group but not to the mass. Therefore, even though people’s (the mass) attitude and purchase intentions decrease, they still acknowledge the increased monetary value of (niche) products including irrelevant attributes.
Another explanation for the decrease in attitude and purchase intention might be explained by the dilution effect that was introduced earlier. Gierl & Heuttl (2012) have stated that in the case of the presence of neutral non-diagnostic information, positive evaluations deteriorate and negative evaluations improve. Since the irrelevant attributes in this study are neutral non-diagnostic information and combined with positive non-diagnostic information, attitude and purchase intentions might be lower as a consequence of this combination of information. Lastly, attitude is a behavioural dimension that is close to the self and linked to the long term.
