Effects of Anthropomorphism, Other-oriented perfectionism and Empathy
on the Perception of Flawed products.
Master Thesis, MSc Marketing Management
Table of Contents
Introduction ... 3
Theoretical Background and Hypotheses Development ... 6
Anthropomorphism ... 6 Other-oriented Perfectionism ... 8 Empathy ... 10 Methodology ... 13 Results ... 15 Discussion ... 20
Limitations and Future Directions ... 23
Conclusion ... 23
References ... 24
Abstract
Through quantitative research and the collection of data through an online panel, the effect of anthropomorphism on flawed products was investigated. Furthermore, this study addressed the concepts of other-oriented perfectionism and empathy as possible influences on this effect. This research was conducted in the context of the rising waste of only superficially flawed products in the supply chain. Here, anthropomorphism offers a possible solution that might change the consumers’ perceptions of these products. However, this study could not establish more than a marginal effect of the concept. Therefore, other concepts, which might interfere with
anthropomorphism are mentioned in the discussion and present opportunities for further research.
Introduction
In the context of the growing demand for sustainable products and the also growing
discussion about environmental issues, the topic of wasting resources in the supply chain has drawn increasing attention over the past years, both in research and the media (Aschemann-Witzel, de Hooge, Amani, Bech-Larsen & Oostindjer, 2015; Chan & Wong, 2012; Lundblad & Davies, 2016; Nixon, 2015; White, Lin, Dahl & Ritchie, 2016; Young, Jirousek & Ashdown, 2004). The problem hereby is that resources are not used to their full potential because goods never reach the end of the supply chain or are then discarded by the end consumer. This issue is especially wasteful
considering that most products which get discarded are fully functioning but simply do not fulfil the consumers’ expectations of aesthetic perfection. Looking at the food industry as an example, a third of the world’s produces of fruits and vegetables never get consumed, either because they are lost during production and processing or because retailers and consumers discard them as waste (Royte, 2016). One big factor here is that actors in the supply chain are unwilling to sell or purchase
imperfect foods (Aschemann-Witzel et al., 2015; Buzby & Hyman, 2012). The percentage of food waste is highest for fruits and vegetables and most of it happens for aesthetic reasons only, meaning that the produce is not spoiled but simply scratched or discolored (Royte, 2016). The same problem goes for superficial packaging damages (e.g., a dented can) which do not even concern the product itself. In another study, conducted by MeadWestvaco Corporation, results showed that 75% of customers would not buy frozen food if the package was damaged. Here, even if the product itself is still fully usable and edible, people see the exterior damage as a potential source of contamination and consequently forego the purchase (White et al., 2016).
The problem of minor appearance flaws is also present in the clothing industry as another example. Here, millions of tons of fabric are discarded each year, simply because they are dyed the wrong color (Murray, 2016; Young, 2016). The issue of wasting perfectly usable products is especially important, keeping in mind that this industry is one of the biggest polluters in general. For both the food and the clothing industry, discarded goods end up as landfill, creating
environmental and economic problems (Buzby & Hyman, 2012). Hence, to improve the current situation, it is necessary to tackle this problem of waste from all different angles.
The following research will do this, starting from the end of the supply chain, by investigating the end consumers’ perceptions of flawed products. If these perceptions can be
change the perceptions of consumers by humanizing products and brands through giving them human features. For this topic, anthropomorphism could be an appropriate strategy because it raises the care and concern for a product (Ahn, Kim, Aggarwal, 2014). If this is also the case for flawed products, anthropomorphism could support the acceptance of these flaws and retailers and
producers would not be forced to discard them anymore.
In the event of anthropomorphism, products resemble humans and are perceived by the observer as having a presence of mind which makes it more likely for them to treat the product or campaign differently (Waytz, Cacioppo & Epley, 2010). According to Chandler and Schwarz, humanizing a product leads consumers to focus less on quality and functionality and more on information that is valued in the interpersonal domain which then leads to a positive affect when interacting with the product (2010). However, this study was conducted on products that the respondents already owned, therefore, it is necessary to see whether this positive effect could also hold for purchasing products, especially when flawed. Positive effects of anthropomorphism on product perceptions have been confirmed by research, however, studies on flawed products do not exist yet. Hence, this research will firstly investigate if anthropomorphism can also improve the purchase intentions for flawed products.
Additionally, certain conditions, which might possibly moderate or mediate the
anthropomorphism effect, were looked at to explain how anthropomorphism changes the perception of flawed products. Once a product is anthropomorphized and therefore seen more as a human, consumers are likely to apply social interaction norms and to form a relationship with it (Chandler & Schwarz, 2010). Hence, two concepts will be looked at which have been widely researched as influencers of real social interaction, perfectionism and empathy. Both concepts are established in the domain of social psychology and are determinants of how people interact with others in a social environment. Therefore, it will be investigated whether these concepts also hold for humanized products and what impact they have on the flaw perception. In particular, this means that this study will also look at the possible moderating effect of other-oriented perfectionism on product liking and purchase intentions and whether the product evoked empathy in the respondent.
Theoretical Background and Hypotheses Development
Anthropomorphism
Anthropomorphism, as derived from the Greek words anthrōpos for human and morphē for shape or form, is a concept by which individuals ascribe human characteristics to nonhuman entities. The notion to anthropomorphize already extends back centuries and is still largely present today. Many examples of the phenomenon can be found in arts and storytelling across cultures and centuries. One early example is the Sphinx of Giza, famously showing a lion with a human head, which is believed to have been created around 2540 BC. Also in folklore, it was common to
attribute animals with human characteristics to tell stories like in the Grimm’s fairytales or Aesop’s fables. Nowadays, we encounter anthropomorphism in Disney movies, products and brands or by having our new computer applications introduced to us by a virtual office paperclip.
The phenomenon of anthropomorphism has been documented as simply being part of the human mind which entails that observers transfer their knowledge about themselves and others onto nonhuman entities which would, for example, also explain the similarities between gods and
believers (Lesher, 1992). The use and attempted understanding of anthropomorphism is present in many disciplines. From theology to psychology and the arts, many scholars have stated that people are inclined to see nonhumans as humanlike (Epley et al., 2007). Despite this everyday presence and the possible importance for marketing strategy, the use of this concept as an actual research dynamic has only taken off recently, changing from a simple description of the accuracy of
anthropomorphism to an account of its process (Epley et al., 2007; MacInnis & Folkes, 2016; Wan & Aggarwal, 2015).
The concept of anthropomorphism, seeing a nonhuman entity as humanlike, allows research to look at the interaction between an observer and a nonhuman agent as a human to human kind of interaction (Wan & Aggarwal, 2015). This is because people who anthropomorphize use the same social norms, mental processes and neural systems which are also involved when interacting with another human being (Epley et al. 2007; Wan & Aggarwal, 2015). For some individuals, this even goes as far as to build relationships with brands and products as a way of substituting for real social connections (Epley, Waytz, Akalis & Cacioppo, 2008).
if it is perceived as having emotions and thoughts of its own which are then felt or understood by the observers themselves. Essentially, anthropomorphizing a product is simple as there are numerous ways; for example, MacInnis & Folkes (2016) define three by which people
anthropomorphize forms: by having human-like features (e.g., a car grill that seems to resemble a face), a human-like mind (e.g., computers which are perceived as having their own intentions) and/or a human-like personality (e.g., a brand that seems friendly). This means countless possibilities to anthropomorphize flawed products if it does turn out to be a positive concept in regard to product liking and purchase intentions.
While anthropomorphism is already pervasive across time and cultures and generally comes rather natural to people (Guthrie, 1993), the SEEK model of Epley et al. describes three main concepts which make it more likely for people to anthropomorphize. Meaning, “Sociality,
Effectance, and Elicited Agent Knowledge”, the model states that the general tendency to humanize is facilitated by an individual's knowledge of people as a cognitive element and two motivational factors. The cognitive response to transfer human knowledge onto a nonhuman agent can be very easily triggered. For example, simply the design of a car grill can lead to an automatic response to access the available human schema, making it seem like the car has a face (Wan & Aggarwal, 2015). Then, depending on the design, the observers have different reactions to the product. In the consumer context, anthropomorphized products were generally evaluated as seeming more
intelligent, competent and trustworthy, however, liking only increased if the nonhuman agent was perceived as having no negative traits (Delbaere, McQuarrie, and Phillips, 2011; Gong, 2008). For this study, it means that the product should be designed as neutral as possible, to have no
interference between positive character traits of the product and general anthropomorphism on liking. In general, we can expect that anthropomorphism will enhance target evaluation and
therefore also purchase intensions (Hur et al., 2015). For the main question of this study, we simply want to see whether this positive effect of anthropomorphizing products also holds for flawed ones, hence, the first of three hypotheses will be the following:
Conceptual Model H1:
Keeping in mind that the likability of anthropomorphism is dependent on certain
characteristics, the following two paragraphs will mention two concepts from social psychology which might also influence the perception of anthropomorphism. As this research is mostly
concerned with the issue of how anthropomorphism can improve the perception of flawed products, the two concepts are related to people’s lay beliefs regarding the acceptance of others. Firstly, given that anthropomorphism creates a human to human like interaction, it is appropriate to look at these aspects of social psychology because this domain of psychology is depended on the presence of others. This brings novelty to the research of anthropomorphized products, as in past research, these two concepts were mostly applied to actual human to human interactions. One exception here is the research on empathy as a human trait which has been effectively used to design service robots, which evoke and show empathy to increase the observer’s approval of them (Złotowski, Proudfoot, Yogeeswaran & Bartneck, 2015). If this effect can also be found for regular products, it could greatly improve the acceptance of flawed, anthropomorphized products. Contrarily, high perfectionism will most likely decrease the acceptance of flaws.
Other-oriented Perfectionism
Firstly, a concept which might explain how accepting people are of mistakes is perfectionism. Given that anthropomorphism creates a human to human like interaction, it is appropriate to look at this social interaction concept in understanding how people perceive flaws in other people and therefore, anthropomorphized products.
perfectionism as a multidimensional concept and is very extensive in measuring the underlying factors as positive and negative, active and passive or healthy and unhealthy parts of the concept (Lundh, 2004; Lynd-Stevenson & Hearne, 1999; Slade & Owens, 1998; Stumpf & Parker, 2000).
Firstly, addressed by Horney (1950) and Adler (1956) in the 1950s, perfectionism was initially seen as a clinical construct which could explain individuals’ tendencies to suffer from mental health issues. Later, Hewitt and Flett (1991) and Frost, Marten, Lahart and Rosenblate (1990), developed measurements of perfectionism which were also applied to student bodies and found to be significant. Therefore, perfectionism is not just a concept that applies to clinical populations (Cox, Enns & Clara, 2002; Hewitt, Flett, Turnbull-Donovan & Mikail, 1991). As this particular study concerns the consumer as a target observer, it is important to note that
perfectionism exists on a spectrum in normal (non-clinical) populations. From a more general perspective, the concept simply tells us what individuals expect from themselves and others around them (Hill, Zrull & Turlington, 1997). So, for the normal population, perfectionism is a personality trait, characterized by striving for flawlessness and setting high standards for oneself and/or for others which leads to an overly critical judgement of performance and behavior (Frost et al. 1990; Hewitt & Flett 1991; Stoeber, 2015).
Furthermore, as perfectionism has been defined as having many dimensions, the right aspect needs to be chosen for this study. Following Hewitt and Flett (1991), perfectionism was divided into self-oriented perfectionism, other-oriented perfectionism and socially prescribed perfectionism. Keeping in mind that we consider an interaction between a human observer and an
factor. As has been mentioned, flawed products are generally less liked by consumers, however, if perfectionism as part of social psychology is related to products, a person high in other-oriented perfectionism would expect humanized products to be even more perfect than the non-humanized one. Consequently, they would like flawed products even less if they are anthropomorphized. Hence, other-oriented perfectionism would negatively moderate product liking and purchase intentions for anthropomorphized products, forming the next hypothesis:
H2: The level of other-oriented perfectionism will negatively moderate purchase intentions and product liking towards a flawed product, especially for anthropomorphized products.
Conceptual Model H2:
Empathy
This means that empathy can be shown by actually feeling the other person’s emotional state or by imitating the other person’s behavior and expressions in the affective approach (Stotland, 1969, Leite et al.,2013). Stephan and Finley (1999) define the two types of empathy as either “taking the perspective of another person” or as “emotional responses to another person that either are similar to those the other person is experiencing or are a reaction to the emotional experiences” (pp. 730). For this study, the first type of showing empathy is more likely as an emotional response to a product is unlikely. However, the consumer might be able to put oneself in a flawed,
anthropomorphized product’s shoes.
Moreover, for this particular research, the ambiguity of defining empathy is not a real issue, rather, the effects which empathy can have on the purchase intentions of flawed products have to be looked at. According to Batson, Batson, Todd & Brummett (1995), empathy refers to other-oriented feelings, which are “empathic feelings[,] includ[ing] sympathy, compassion, tenderness, and the like” (pp. 621). Consequently, through empathy, people are less likely to reject others and empathic concern causes an increase in helping (Batson, 1991; Stephan & Finlay, 1999). This can be because of altruistic reasons or guilt, for example (Ahn et al., 2014; Batson, 2011, Batson & Shaw, 1991; Batson et al., 1995). In the research of Ahn et al. (2014), anticipatory guilt was found to be a significant mediator on compliance with campaigns if they were anthropomorphized. According to Hoffman (2000), guilt is also a part of how people learn to empathize as children because they internalize moral obligations, obliging them to consider others’ feelings (Eisenberg & Morris, 2001). For anthropomorphized products, this could mean that consumers might feel guilty about rejecting a flawed product, as they would also feel guilty when rejecting a “flawed” human. For the goal of raising purchase intentions of flawed products without further devaluing the product with price reductions, evoking empathy in the observer might trigger feelings like sympathy and care towards the product, leaving the consumer with a guilty feeling if they reject it for that minor flaw. However, there are also other parts to empathy, besides guilt. Consumers could also work from an altruistic point of view, if they feel sadness or pity towards the flawed product, it will also make rejection less likely. Whether the reason is guilt, pity or altruism, the flaw itself is still seen as a negative thing, not something positive and unique. For this research, it will therefore only be investigated if empathizing with the product, in whatever way, makes it indeed less likely that consumers reject it. This effect should only hold for the anthropomorphized condition, as otherwise, the observer will not apply social interaction mechanisms to the product. On that account, the application of anthropomorphism makes it possible for the observers to feel empathic towards the product because it is perceived as having own emotions and thoughts.
et al., 2014; Chandler & Schwartz, 2010). Furthermore, a positive effect of evoking empathy was found for increasing the approval of robots and could further explain how anthropomorphism works. However, this research was not concerned with flawed products. Finally, for this research, it is important that we measure empathy in the given situation and not as a general trait. Hence, situational (or state) empathy will be measured that simply describes how the consumer felt towards the product at the moment of evaluating it. This can be different from the respondents’ level of feeling empathic on a general scale, as situational empathy is also related to the emotional arousal that the respondent feels at that time (Eisenberg, Fabes, Murphy, Karbon, Maszk, Smith, O’Boyle & Suh, 1994).
Therefore, this study will see whether the positive effect of anthropomorphizing products works for flawed ones as well and if this effect works through empathy. Given the elements and effects of empathy, it can be predicted that product liking and purchase intentions might be
improved for a flawed, anthropomorphized product, simply because people are less willing to reject them, leaving us with the final hypothesis:
H3: When anthropomorphism is present (vs. not present), consumers who show more situational empathy towards the flawed product will have higher purchase intentions and product liking.
Conceptual Model H3:
Methodology
The predictions were tested in one study. Through Amazon Mechanical Turk, an online panel, anonymous data was collected from a random sample of 267 (mostly American) respondents (47.6% females, Mage = 33). Each respondent received an online questionnaire which had to be fully completed. The study followed a 2 (anthropomorphized vs. non-anthropomorphized) x 2 (flaw vs. no flaw) between-subject design. All respondents were randomly assigned to one of the four conditions and had to evaluate a product, in this case a black outdoor backpack. Each participant was shown a picture of the backpack and a short description of its functions as can be seen in Figure1:
In the anthropomorphism condition, the description was changed from third-person to first-person to humanize the product. Here, the backpack is introduced with: “Hi, my name is BeBe! I am a backpack. I have enough space to carry all of your belongings! Moreover, I have a built-in power bank, so you can charge your devices on the go!”. Using human features, like in this case, personal speech, is a common strategy to anthropomorphize objects and to evoke a feeling of human-to-human interaction (Aggarwal & McGill, 2007; Chandler & Schwartz, 2010; Epley et al., 2008). The flaw condition was achieved by simply adding a differently colored zipper to the backpack’s right side (pictures of this and the other two conditions can be found in appendix A). This counts as a minor design flaw which will not affect the functionality of the backpack. In the flaw conditions only, participants were also asked to indicate how noticeable they think the variation is and how much they are bothered by it.
After being showed the picture of the backpack, participants then had to indicate their purchase intentions (“How likely are you to purchase this backpack if you saw it in a store?”) and liking for the product. The latter was measured by several questions, regarding the general liking for
the backpack, the expected quality and the expected enjoyment of wearing it. All of these questions were measured with a 7-point bipolar scale, ranging from “1 = Not at all” to “7 = Very much” and can also be found in the full study design in the appendix.
The same measurement style was used for the five empathy related questions which were shown subsequently. As it was important to measure situational empathy and not a general
tendency,the questions were specific to the backpack (e.g., “Did you imagine this product to have feelings upon first reading the description?“ and “Do you feel empathic towards this product?”). These questions were adapted from other studies which used situational empathy and will simply show whether the respondent felt empathic towards the backpack (Carré, Stefaniak, D’Ambrosio, Bensalah & Besche-Richard, 2013; Escalas & Stern, 2003; Haegerich & Bottoms, 2000).
Furthermore, all respondents were asked how regularly they use backpacks. This measurement can later be used as a control variable. Additionally, in the flaw conditions only, participants were also asked to indicate how noticeable they think the variation was and how much it bothered them.
Contrarily to empathy, other-oriented perfectionism had to be measured as a general personality trait. Here, the other-oriented perfectionism sub scale of the Multidimensional
Perfectionism Scale by Hewitt & Flett (1991) was used. According to Stoeber, the sub scale of the MPS also performed better than comparable scales, measuring the same concept (2016). This scale has been shown to be significant for both clinical and non-clinical studies on almost the same level [⍺s = .80 (clinical), .74 (non-clinical)] and shows high reliability and validity (Cox, Enns & Clara, 2002; Hewitt, Flett, Turnbull-Donovan & Mikail, 1991). The sub scale for other-oriented
perfectionism uses a 7-point Likert scale ranging from 1 (disagree) to 7 (agree) and includes 15 items, which include for example: “Everything that others do must be of top-notch quality” and “If I ask someone to do something, I expect it to be done flawlessly”.
Finally, the study included two control scales. Firstly, a scale which was adapted from Chin, Sims, Clark & Lopez (2004), which assessed how likely the respondents are to anthropomorphize products in general (9 items, e.g., “I would talk to my car”; “I would not talk to my computer”; “People who talk to their car are normal”). Secondly, respondents were asked how accepting they are of imperfections (6 items: e.g., “It is normal for humans to be imperfect.”; “I accept my friends’ imperfections”). The items of the latter scale were separated in a variable that measured the
Results
To check if the hypotheses can be confirmed, firstly, a single dependent variable is created. For that aim, a correlation matrix is conducted, including the four variables of liking and purchase intentions. This matrix, which can be found in appendix C, shows that all measurements for liking and purchase intention are highly, positively correlated. Additionally, a reliability analysis of the four items resulted in a Cronbach’s alpha of .876. Hence, purchase intention, liking, expected quality and expected enjoyment of the product are computed as one final variable, creating a sound indication which is, from now on, referred to as Liking (dependent variable).
Next, it is important to check for data anomalies. While there are no missing values, the creation of boxplots with the dependent variable and the flaw and anthropomorphism conditions indicates five outliers. However, a look at the respondents’ answers also shows that they were rather extreme but not implausible, according to the literature. Hence, no data was emitted and the final sample included 267 respondents.
Furthermore, controls have to be set up to make the formation of a cause and effect more credible. Firstly, the variable of backpack usage was used as a control variable for it to not influence the study’s results as we want to know how this particular (flawed) backpack is evaluated and not how someone generally evaluates backpacks. As someone who never wears backpacks might evaluate them less favorably in general, backpack usage will be used as the first covariate.
Secondly, likelihood to anthropomorphize was used as a second covariate. As was established in the model by Epley et al., the likelihood to anthropomorphize depends on several factors and these personal factors should not influence this research (2007). However, the variable firstly had to be constructed by adding the results of the corresponding scale and creating a mean value. This resulted in a highly reliable measurement (reliability analysis: 15 items; ⍺=.849). Additionally, a second correlation matrix showed high correlations between the covariates and the dependent variable (BackpackUsage: r = .381, p < .000; AnthroLikelihood: r = .244, p < .000).
flaw and anthropomorphism conditions. Firstly, a univariate analysis of variance was conducted to show the different estimated means for the dependent variable. The results can be seen in Figure 2:
Figure 2)
The sequentially conducted ANCOVA showed that both covariates were significant (BpUsage: F = 37.862, p < .000; TotalAnthroLikelihood: F = 12.495, p < .000). Looking at the figure above and the results (appendix D), two things can be concluded from this: firstly, anthropomorphism does appear to increase the level of Liking, Anthro (M = 5.13, SD = 1.19), NoAnthro (M = 4.88, SD = 1.19).However, the first hypothesis cannot be confirmed as neither the anthropomorphism variable, F(1, 267)=3.004, < p=.084, nor the interaction term of
anthropomorphism and flaw, F(1, 267=0.957, < p=.329, are significant on Liking. Nonetheless, it can be argued that anthropomorphism has a marginally significant main effect with a positive direction. Secondly, it is quite surprising that Liking for flawed products (M = 5.04, SD = 1.20) is higher than Liking for unflawed products (M = 4.97, SD = 1.19), regardless of whether
anthropomorphism is present.
Thus, acceptance of machine-made and human mistakes were also taken in as covariates. Here, we can assume that someone who is high in acceptance of machine-made mistakes is probably less bothered by the product flaw and someone who is high in the acceptance of human mistakes might be more acceptant of the flaw in the anthropomorphism condition. However, results of the ANCOVA show no significance of these covariates (humans: F(1, 267)=0.193, < p=.661 and
5.2008 4.7388 4.968 5.0597 5.0149 5.0373 5.1297 4.8769 4.5 4.6 4.7 4.8 4.9 5 5.1 5.2 5.3
Total Liking and Purchase Intention Means
No flaw Flaw Anthro No Anthro
machines: F(1, 267)=0.640, < p=.424) and no overall change in the model (Appendix E). While there is a marginal effect of anthropomorphism on product liking and purchase intention, there is not enough significance to reject the null hypothesis. Hence, it can be rejected that when a flawed product is anthropomorphized, consumers have higher purchase intentions and product liking than when a flawed product is not anthropomorphized.
Hence, another ANCOVA was conducted, including the level of other-oriented perfectionism as third factor. This might give us an indication of why the flawed product was perceived that well and it will also lead us to the second hypothesis.To conduct this ANCOVA, the responses for other-oriented perfectionism first had to be added into an average value for each respondent and a median split was performed to create a categorical variable since the original continuous variable cannot be used in an ANCOVA. The new dummy variable was labeled as 0 = low in other-oriented perfectionism and 1 = high in other-oriented perfectionism. Again, the covariates were backpack use and likelihood to anthropomorphize. Except for the covariates
(<p=.001) and the marginal effect of Anthro (<p=.092), no effects were significant (results table can be found in appendix F). The interaction term between Anthro and OOP was not significant (F(1, 267)=0.042, < p=.851), as well as the interaction between Flaw and OOP (F(1, 267)=1.392, < p=.278). The estimated means of Liking can be seen in Figure 3:
Figure 3) 4.5 4.6 4.7 4.8 4.9 5 5.1 5.2 5.3
Anthro No Anthro Flaw No Flaw
Here, it is apparent that people low and high in OOP have very different views on the conditions, with people low in OOP clearly preferring the flawed product (no flaw: M = 4.75, SD = 1.20 and flaw: M = 5.11, SD = 1.19. However, there is almost no difference regarding the anthropomorphism condition. Here, we can already say that the assumption that respondents high in OOP will respond even more negatively to the anthropomorphism condition is not confirmed as there is no apparent difference between the conditions. However, they do show a slightly lower mean for Liking in the flawed condition (no flaw: M = 5.18, SD = 1.16 and flaw: M = 4.94, SD = 1.20).
This ANCOVA gave a good overview of the means of the different conditions, however, due to the median split, some data also might have been lost as a value close to the median is considered the same as one at the far end of the spectrum. Hence, to see whether OOP is a valid moderator on Liking and if it interacts with both flaw and anthropomorphism (as was proposed in hypothesis two) a regression analysis was conducted, which allows for the use of OOP as a continuous variable. Firstly, a reliability analysis for the 15-item OOP scale was conducted which resulted in a Cronbach’s Alpha of 0.344. Considering that this is an established scale, that is a very low value. Looking at the generated inter-item correlations, items 9 (“If I ask someone to do something, I expect it to be done flawlessly”), 11 (“The people who matter to me should never let me down”), 12 (“I respect people who are average”) and 13 (“It doesn’t matter when someone close to me does not do their absolute best”) showed high negative correlations and might be deleted. The revised scale then had a Cronbach’s Alpha of 0.816 which is above 0.7 and therefore a sufficient value. However, the deleted questions do not show a common theme as some are directed to close friends and others are more general.
Therefore, two regressions will be conducted, one that includes the revised scale and then a second which includes factors. Firstly, for the regression including the revised scale, the
independent values and covariates (backpack usage and likeliness to anthropomorphize) were mean-centered to facilitate interpretation of values and to forego problems of multicollinearity. The results indicate a significant model but with a low R-value (R2 =.203, F(9,266)=7.284, p<.000; Appendix G). However, apart from the marginal effect of Anthro (β = .350, p<.067) and the covariates, no variables or interactions were significant. Neither for OOP (β = -.018, p<.908), the interaction of OOP and Flaw (β = -.008, p<.974), nor the interaction of OOP and Anthro (β = .149, p<.455). Secondly, a factor analysis was also conducted to see if the scale had any underlying themes which could be used as separate factors since it could be that people have different
corresponds to the type of relationship. For example, there is no difference between friends, acquaintances or others as all factors include questions concerning each. Still, factor one could be summarized as criticizing others, factor two as expectations about mistakes and factor three as expectations about others’ improvement. Factor four only included the second question and was therefore left out. Including these factors in a multiple regression resulted in a few significant effects (R2 =.250, F(17,266)=4.870, p<.000: Appendix I). Firstly, the covariates were, again, significant (BPusage: β = .220, p<.000; TotalAnthro: β = .188, p<.001). From the factors, the first showed a marginal main effect (β = -.454, p<.064), a significant interaction with anthropomorphism (β = .814, p<.023), and a marginal effect for the flaw interaction (β = .612, p<.072) and the
interaction with both (β = -.919, p<.065). Furthermore, there were significant effects of the third factor and the interaction with flaw (β = -.479, p<.047) and both anthropomorphism and flaw (β = .741, p<.030). Regarding the first factor, it can be said that people who have high expectations in others and want them to strive to better themselves and do their best show a negative main effect on Liking, especially in the condition where both flaw and anthropomorphism are present. If it is just the one or the other, then the effect is positive. Hence, anthropomorphism and flaw are seen as positive but when it occurs at the same time, the product is evaluated less positive. This confirms the hypothesized effect that a flawed product is especially negatively evaluated when it is
anthropomorphized, in other words, when the social concept of perfectionism is applied. However, this only holds for the first factor. Regarding the third, people who tend to criticize others and their friends show a negative effect on Liking when a flaw is present but a positive effect when both anthropomorphism and flaw exist. For the second factor, no effects are significant. Looking at the first factor only, the marginal and significant interaction effects mean that Factor1 is a moderator but that a marginally significant, negative main effect on Liking also exists. Factor3, on the other hand, shows complete moderation for the interaction of both anthropomorphism and flaw and on the flaw effect.
This leaves the concept of empathy and the third hypothesis, stating that when
anthropomorphism is present, consumers who show more situational empathy towards the flawed product will have higher purchase intentions and product liking. Before checking this assumption, another correlation matrix was conducted to see how the different empathy measures correlated, which showed highly positive and significant values for all (Appendix J). Additionally, a reliability analysis of the five items resulted in a Cronbach’s Alpha of 0.910. Therefore, whether the
anthropomorphism leads to higher purchase intentions and liking, a regression analysis, following model 8 of Andrew F. Hayes’ process macro, was conducted (the detailed output can be found in appendix K). The results indicated that neither the indirect nor the direct effect were significant. The effects (a-paths) of Anthro (β = .3135, SE = .2491, p<.2094), Flaw (β = .0028, SE = .2463, p<.9909) and the interaction between both (β = .2835, SE = .3531, p<.4228) were not significant, as well as the effects of all three on Liking (Anthro: β = .2967, SE = .1831, p<.1063; Flaw: β = .2594, SE = .1805, p<.1518; AnthroxFlaw: β = -.3202, SE = .2591, p<.2176). The only significant effect was the b-path effect of Empathy on Liking (β = .2042, SE = .0454, p<.000). Hence, there is no apparent mediation of empathy. Thus, hypothesis 3 has to be rejected. Nevertheless, the effect of Empathy on Liking is positive, meaning that respondents who felt more empathic towards the product, also indicated more liking and purchase intention. The results and possible limitations are being discussed in the following section.
Discussion
This research had one main goal, to see whether anthropomorphizing flawed products can increase people’s liking towards them. For this research, it is difficult to draw conclusions in this regard as most effects were either non-significant or only marginally significant. What can be said is that both the use of backpacks and the likelihood to anthropomorphize were significant on Liking in all analyses. Additionally, anthropomorphism itself had a positive marginally effect in almost all of them. However, none of the hypotheses could be confirmed at a significant level.
Another problem with the results of this study is that the anthropomorphism effect could only be established at a marginal level, which cannot be considered as truly significant. This could be due to other concepts which might interfere with the effect of anthropomorphism. For one, a possible interference might be the difference between hedonic and utilitarian products. The former is mostly bought and consumed for enjoyment and pleasure purposes and can be exciting or even thrilling, while the latter are acquired to satisfy basic needs and practical uses and are more of a necessity, however, products can also show considerations of both (Dhar & Wertenbroch, 2000; Voss, Spangenberg & Grohmann, 2003). As elements of consumption, they are part of how customers form attitudes about products (Voss et al., 2003). For example, it is easier to charge a price premium for hedonic products, owners keep them for longer and are less likely to sell them, suggesting that they are generally more valued by consumers (Dhar & Wertenbroch, 2000; Park & Mowen, 2007). In this study, the backpack could be characterized as either. On one hand, it
resembles a utilitarian product because it fulfils rather functional tasks (“storage capacity and a built-in power bank for charging various devices”). This might have had an impact on the
respondents’ answers as attitudes towards utilitarian products apparently differ from those towards hedonic products, meaning that results might have been different for a more hedonic product (e.g., a perfume). On the other hand, this cannot be clearly said at this point because the backpack also includes hedonic attributes, like the design, and in a different study, a backpack was said to incorporate both attitudes respectively, however, that backpack clearly went further in design than the backpack in this study (Lin, 2007). If we were to control for this factor, it is to be expected that a utilitarian product with a flaw will be evaluated less favourably as it is expected to fulfil a
practical, necessary task flawlessly. If the consumer perceives an eventual product flaw as an actual flaw, he or she will doubt if that task can be fulfilled. On the other hand, as hedonic products are mostly bought for pleasure, a flaw might not bother the consumer as much. For the hedonic, flawed product, we can also expect a positive influence of anthropomorphism. For the utilitarian product, it might be dependent on whether the way it is anthropomorphized convinces the consumer that the respective task can be accomplished. To conclude this point, the product attributes might have interfered with the study results.
promotion (approach) or prevention (avoidance) focus (Lin, 2007). For this research, it means that respondents with a prevention focus might be faster to reject flawed products as they show more anticipated regret, in other words, they want to avoid a negative outcome, like choosing a (truly) flawed product and regretting the purchase (Schwartz, 2004). Furthermore, the study by Lin (2007) states that people with a promotion focus respond more to an advertisement with a gain frame and consumers with a prevention focus more to a loss frame. In this study, the backpack description is worded following a gain frame method since positive outcomes are offered to the consumer, instead of, for example, saying that one will miss out by not buying this backpack. This gain frame appeals to people with a promotion focus (Lin, 2007). At this point, we can assume that the respondents in this study, as a collective, show more of a promotion focus as this focus is associated with the North American culture and almost all respondents were American (Hamamura, Meijer, Heine, Kamaya & Hori, 2009). Eventually, this fit between promotion focus and gain frame could have interfered with the study and explain why respondents were less bothered by the flaw than expected.
Especially, and hereby we return to the previous point, if the backpack was perceived as a more hedonic product because consumers with an approach motivation perceive the value of products higher if it is hedonic and presented in a gain frame (Lin, 2007).
Another factor, which could be the reason for the non-significant results, is related to both the product and the consumer, as it describes the brand or product relationship. In some instances, even when products are not anthropomorphized, some people form relationships with them or the brand, the same way as they would with human agents (Aggarwal, 2004; Fournier, 1998). Here, the perceived competence of the product comes into play, which is, in this research, most likely
Limitations and Future Directions
Naturally, this research was also subjected to some limitations. Firstly, due to the fact that the data was collected through an online panel, we have less control over the data collection than in a
controlled laboratory study. This means a disadvantage as we cannot control extraneous factors like, for example, how long someone took to fill out the survey and how much attention they actually paid to it. Given that the results of our data were inconclusive, it could be worth it to repeat the study in an actual lab setting where these variables can be controlled for.
Additionally, to see if respondents would actually prefer the flawed version over the unflawed one, future research should follow a study design with a conjoint analysis which actually shows the flawed and unflawed conditions at the same time, making it possible for the consumer to compare them. This future research could establish the impact of flaw and anthropomorphism and also the interaction between both. Moreover, if this main effect can be established, future research might also confirm empathy as a mediator tor evaluating anthropomorphized products. Concerning empathy, it might also be necessary to have two further conditions, including one where empathy is manipulated to be more present.
Finally, the unreliable results of the other-oriented perfectionism scale were quite
unexpected. This might be solved in the future by only using positively worded statements or other perfectionism scales which tackle the underlying factors of the concept (Frost et al., 1990, Hewitt et al., 1991, Lundh, 2004, Stoeber, 2016).
Further studies, concerning the points mentioned in the discussion, will also be necessary. Here, it needs to be clearly differentiated between hedonic and utilitarian products, a loss and prevention focus, if the product is presented in a gain or loss frame and/or more as a partner or servant. Following these recommendations, further research might actually result in a sound cause and effect relationship.
Conclusion
References
Adler, A. (1956). The neurotic disposition. In H. L. Ansbacher & R. R. Ansbacher (Eds.), The individual psychology of Alfred Adler (pp. 239-262). New York: Harper.
Aggarwal, P. (2004). The Effects of Brand Relationship Norms on Consumer Attitudes and Behavior. Journal of Consumer Research, 31(1), 87–101.
Aggarwal, P. & McGill, A. L. (2007). Is That Car Smiling at Me? Schema Congruity As a Basis for Evaluating Anthropomorphized Products. Journal of Consumer Research, 34(4), 468–479.
Aggarwal, P. & McGill, A. L. (2012). When Brands Seem Human, Do Humans Act Like Brands? Automatic Behavioral Priming Effects of Brand Anthropomorphism. Journal of Consumer
Research, 39(2), 307–23.
Ahn, H. K., Kim, H. J. & Aggarwal, P. (2014). Helping Fellow Beings: Anthropomorphized Social Causes and the Role of Anticipatory Guilt. Psychological Science, 25(1), 224–9.
Allen, R., & May, S. (2015). Encountering Anthropomorphism. Performance Research, 20(2), 1-3.
Batson, C. D. (2011). Altruism in humans. Oxford: Oxford University Press.
Batson, C. D., Batson, J. G., Todd, R. M., & Brummett, B. H. (1995). Empathy and the collective good: Caring for one of the others in a social dilemma. Journal Of Personality And Social
Psychology, 68(4), 619-631.
Batson, C. D., & Shaw, L. L. (1991). Evidence for Altruism: Toward a Pluralism of Prosocial Motives. Psychological Inquiry, 2(2), 107-122.
Buzby, J. C., & Hyman, J. (2012). Total and per capita value of food loss in the United States. Food Policy, 37(5), 561-570.
Carré, A., Stefaniak, N., D'Ambrosio, F., Bensalah, L., & Besche-Richard, C. (2013). The Basic Empathy Scale in Adults (Bes-A): Factor Structure of a Revised Form. Psychological Assessment, 25(3), 679–91.
Chandler, J. and Schwarz, N. (2010). Use Does Not Wear Ragged the Fabric of Friendship: Thinking of Objects As Alive Makes People Less Willing to Replace Them. Journal of Consumer Psychology, 20(2), 138–145.
Chin, M. G., Sims, V. K., Clark, B., & Lopez, G. R. (2004). Measuring Individual Differences in Anthropomorphism towards Machines and Animals. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 48(11), 1252 – 1255.
Delbaere, M., Phillips, B. J., & McQuarrie, E. F. (2011). Personification in advertising: Using a visual metaphor to trigger anthropomorphism. Journal Of Advertising, 40(1), 121-130.
Dhar, R., & Wertenbroch, K. (2000). Consumer choice between hedonic and utilitarian goods. Journal of Marketing Research, 37, 60–71.
Dong, P. & Aggarwal, P. (2016). Partner or Servant? When Relationship Type Affects Trait Expectations and Evaluations of the Brand. NA - Advances in Consumer Research, 44, eds. Page Moreau and Stefano Puntoni, Duluth, MN: Association for Consumer Research, 118-122.
Eisenberg, N., Fabes, R. A., Murphy, B., Karbon, M., Maszk, P., Smith, M., O'Boyle, C., & Suh, K. (1994). The Relations of Emotionality and Regulation to Dispositional and Situational Empathy-Related Responding. Journal of Personality and Social Psychology, 66(4), 776–97.
Eisenberg, N., & Morris, A. S. (2001). The Origins and Social Significance of Empathy-Related Responding. A Review of Empathy and Moral Development: Implications for Caring and Justice by M. L. Hoffman. Social Justice Research, 14(1), 95-120.
Epley, N., Waytz, A. & Cacioppo, J. (2007). On seeing human: a three-factor theory of anthropomorphism. Psychological Review, 114(4), 864-886.
Epley, N., A. Waytz, S. Akalis, and J. Cacioppo (2008). When we need a human: Motivational determinants of anthropomorphism. Social Cognition, 26(2), 143-155.
Escalas, J., & Stern, B. (2003). Sympathy and Empathy: Emotional Responses to Advertising Dramas. Journal of Consumer Research, 29(4), 566-578.
Fournier, S. (1998). Consumers and Their Brands: Developing Relationship Theory in Consumer Research. Journal of Consumer Research, 24(4), 343–53.t
Frost, R. O., Marten, P., Lahart, C., & Rosenblate, R. (1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449–468.
Gong, L. (2008). How social is social responses to computers? The function of the degree of
anthropomorphism in computer representations. Computers In Human Behavior, 24(4), 1494-1509.
Guthrie, S. (1993). Faces in the clouds. 1st ed. New York: Oxford University Press.
Haegerich, T., & Bottoms, B. (2000). Empathy and Jurors' Decisions in Patricide Trials Involving Child Sexual Assault Allegations. Law and Human Behavior, 24(4), 421-448.
Hamamura, T., Meijer, Z., Heine, S., Kamaya, K. & Hori, I. (2009) Approach--Avoidance Motivation and Information Processing: A Cross-Cultural Analysis. Personality and Social Psychology Bulletin, 35(4), pp. 454–462.
Hewitt, P. L., & Flett, G. L. (1991). Perfectionism in the self and social contexts: conceptualization, assessment, and association with psychopathology. Journal Of Personality And Social Psychology, 60(3), 456-70.
Hewitt, P. L., Flett, G. L., Turnbull-Donovan, W., & Mikail, S. F. (1991). The Multidimensional Perfectionism Scale: Reliability, validity, and psychometric properties in psychiatric samples. Psychological Assessment, 3, 464–468.
Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52, 1280–1300.
Higgins, E. T., Friedman, R. S., Harlow, R. E., Idson, L. C., Ayduck, O. N., & Taylor, A. (2001). Achievement orientation from subjective histories of success: Promotion pride versus 83 prevention pride. European Journal of Social Psychology, 31, 3–23.
Hill, A. P., Stoeber, J., Brown, A., & Appleton, P. R. (2014). Team perfectionism and team performance: a prospective study. Journal of Sport & Exercise Psychology, 36, 303–315.
Hodges, R. J., Buzby, J. C., & Bennett, B. (2011). Postharvest losses and waste in developed and less developed countries: opportunities to improve resource use. The Journal Of Agricultural Science, 149(1), 37-45.
Hoffman, M. L. (2000). Empathy and Moral Development: Implications for Caring and Justice. Cambridge, U.K: Cambridge University Press.
Homey, K. (1950). Neurosis and human growth. New York: Norton
Hur, J. D., Koo, M., & Hofmann, W. (2015). When Temptations Come Alive: How
Anthropomorphism Undermines Self-Control. Journal Of Consumer Research, 42(2), 340-358.
Leite, I., Pereira, A., Mascarenhas, S., Martinho, C., Prada, R., & Paiva, A. (2013). The influence of empathy in human-robot relations. International Journal Of Human - Computer Studies, 71(3), 250-260.
Lesher, J. H. (1992). Xenophanes of Colophon: Fragments. Toronto: University of Toronto Press.
Lin, H. F. (2007). Framing hedonic and utilitarian product attributes in advertisements: The impact of regulatory fit on persuasion. The Pennsylvania State University, Master Thesis. Available at: https://etda.libraries.psu.edu/files/final_submissions/2267
Lundblad, L., & Davies, I. A. (2016). The values and motivations behind sustainable fashion consumption. Journal Of Consumer Behaviour, 15(2), 149-162.
Lundh, L. G. (2004). Perfectionism and Acceptance. Journal Of Rational-Emotive And Cognitive-Behavior Therapy, 22(4), 251-265.
Lynd-Stevenson, R. M., & M. Hearne, C. (1999). Perfectionism and depressive affect: the pros and cons of being a perfectionist. Personality And Individual Differences, 26(3), 549-562.
MacInnis, D. J., & Folkes, V. (2016). Humanizing Brands: When Brands Seem to Be Like Me, Part of Me, and in a Relationship with Me. Marshall School of Business, Working Paper No. 17-8. Available at SSRN: https://ssrn.com/abstract=2890236
Nixon, R. (2015). Food Waste Is Becoming Serious Economic and Environmental Issue, Report Says, The New York Times, (February 2015), (accessed March 30th, 2016), [available at https://www.nytimes.com/2015/02/26/us/food-waste-is-becoming-serious-economic-and-environmental-issue-report-says.html]
Park, S., & Mowen, J. C. (2007). Replacement Purchase Decisions: On the Effects of Trade-Ins, Hedonic Versus Utilitarian Usage Goal, and Tightwadism. Journal of Consumer Behaviour, 6(2-3), 123–131.
Puzakova, M., Kwak, H. & Rocereto, J. (2009). Pushing the envelope of brand and personality: Antecedents and moderators of anthropomorphized brands. Advances in Consumer Research, 36, 413-420.
Royte, E. (2016). How ‘Ugly’ Fruits and Vegetables Can Help Solve World Hunger, National Geographic, (March 2016), (accessed March 30th, 2016), [available at
http://www.nationalgeographic.com/magazine/2016/03/global-food-waste-statistics/].
Schwartz, B. (2004). The Paradox of Choice: Why More Is Less. New York, N.Y.: Ecco/HarperCollins.
Shoss, M. K., Callison, K., & Witt, L. A. (2015). The Effects of Other-Oriented Perfectionism and Conscientiousness on Helping at Work. Applied Psychology, 64(1), 233-251.
Slade P. D., & Owens R. G. (1998). A dual process model of perfectionism based on reinforcement theory. Behavior Modification, 22(3), 372-90.
Stephan, W. G., & Finlay, K. (1999). The Role of Empathy in Improving Intergroup Relations. Journal Of Social Issues, 55(4), 729-743.
Stoeber, J. (2014). Multidimensional perfectionism and the DSM-5 personality traits. Personality And Individual Differences, 64, 115-120.
Stoeber, J. (2015). How Other-Oriented Perfectionism Differs from Self-Oriented and Socially Prescribed Perfectionism: Further Findings. Journal of Psychopathology and Behavioral Assessment, 37(4), 611-623.
Stoeber, J., Mutinelli, S., & Corr, P. J. (2016). Perfectionism in students and positive career planning attitudes. Personality And Individual Differences, 97, 256-259.
Stotland, E. (1969). Exploratory investigations of empathy. In: L. Berkowitz (Eds.), Advances in Experimental Social Psychology, vol. 4: 271-314. New York: Academic Press.
Stumpf, H., & Parker, W. D. (2000). A hierarchical structural analysis of perfectionism and its relation to other personality characteristics. Personality And Individual Differences, 28(5), 837-852.
Voss, K. E., Spangenberg, E. R., & Grohmann, B. (2003). Measuring the hedonic and utilitarian dimensions of consumer attitude. Journal of Marketing Research, 40, 310–320.
Wan, J., & Aggarwal, P. (2015). Befriending Mr. Clean: The Role of Anthropomorphism in Consumer-Brand Relationships. In: S. Fournier, M. Breazeale, & J. Avery (Eds.), Strong Brands, Strong Relationships. Abingdon: Routledge.
Wan, E. W., Chen, R. P., & Jin, L. (2017). Judging a Book by Its Cover? The Effect of Anthropomorphism on Product Attribute Processing and Consumer Preference. Journal of Consumer Research, 43(6), 1008-1030.
White, K., Lin, L., Dahl, D. W., & Ritchie, R. J. B. (2016). When Do Consumers Avoid Imperfections? Superficial Packaging Damage as a Contamination Cue. Journal Of Marketing Research, 53(1), 110-123.
Young, C. (2016). Clothing Industry Remains One Of The Largest Environmental Polluters In The World, Twisted News, (September 2016), (accessed March 30th, 2016), [available at
http://www.twisted.news/2016-09-26-clothing-industry-remains-one-of-the-largest-environmental-polluters-in-the-world.html].
Young, C., Jirousek, C., & Ashdown, S. (2004). Undesigned: A Study in Sustainable Design of Apparel Using Post-Consumer Recycled Clothing. Clothing And Textiles Research Journal, 22(1-2), 61-68.
Appendix A) Conditions
Condition 2) “This is a BeBe backpack. It has a substantial storage capacity and a built-in power bank for charging various devices. Please note, on the pocket on the right side of the backpack, the color of the zipper is slightly different from others: It is gray instead of black (see rightmost photo).”
Condition 4) “Hi, my name is BeBe! I am a backpack. I have enough space to carry all of your
belongings! Moreover, I have a built-in power bank, so you can charge your devices on the go! Please note, on my pocket on the right side, the color of my zipper is slightly different from the others: It is gray instead of black (see rightmost photo)”
B) Study Design
(Liking) How much do you like this backpack?
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Not at all:Very much (1)
m m m m m m m
(Purchase intention) How likely are you to purchase this backpack if you saw it in a store?
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Not likely:Very
likely (1)
m m m m m m m
(Liking-enjoyment) How much would you enjoy wearing this backpack?
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Not at all:Very much (1)
m m m m m m m
(Liking-quality) To what extent do you think this backpack is of high quality?
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Not at all:Very much (1)
(Empathy-feelings) Did you imagine this product to have feelings upon reading the description? 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7) Not at all:Very much (1) m m m m m m m
(Empathy-perspective) Did you imagine yourself taking the backpack's perspective upon first reading the description? 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7) Not at all:Very much (1) m m m m m m m
(Freewill)Do you think this product has free will/agency?
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Not at all:Very much (1)
m m m m m m m
(Empathy) Do you feel empathic towards this product?
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Not at all:Very much (1)
m m m m m m m
(Similarity) In your opinion, do you have anything in common with this backpack?
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Nothing at all:Very much (1)
m m m m m m m
(Flaw-Prominence) How noticeable in your opinion is the variation in zipper colour? – ONLY FLAW CONDITION 1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7) Not at all noticeable:Very noticeable (1) m m m m m m m
(Flaw-bother) Does the variation in zipper colour bother you? – ONLY FLAW CONDITION
1 (1) 2 (2) 3 (3) 4 (4) 5 (5) 6 (6) 7 (7)
Not at all:Very much (1)
OOP Scale:
I am not likely to criticize someone for giving up too easily. (1) It is not important that the people I am close to are successful. (2) I seldom criticize my friends for accepting second best. (3) Everything that others do must be of top-notch quality. (4)
I have high expectations for the people who are important to me. (5) I do not have very high expectations for those around me. (6)
I can’t be bothered with people who won’t strive to better themselves. (7) I do not expect a lot from my friends. (8)
If I ask someone to do something, I expect it to be done flawlessly. (9) I cannot stand to see people close to me make mistakes. (10)
The people who matter to me should never let me down. (11) I respect people who are average. (12)
It doesn’t matter when someone close to me does not do their absolute best (13) It does not matter to me when a close friend does not try their hardest. (14) I seldom expect others to excel at whatever they do. (15)
Imperfection acceptance scale: I would talk to my computer (1) I would not talk to my car (2)
I would praise my computer when it performed properly (3) I would name my computer (4)
I would name my car (5)
People who talk to their computer are lonely (6) People who talk to their car are normal (7) People who talk to their computer are bored (8)
People who talk to their car are psychologically disturbed (9)
Anthropomorphism likeliness scale:
It is normal for humans to be imperfect. (1)
Humans are naturally different from one another. (2) Machine-made objects should be uniform/consistent. (3)
It is normal for machine-made objects to be different from one another. (4) I accept my friends’ imperfections. (5)
C) Correlation Matrix Liking Correlations
TotalPurchase TotalLiking TotalLikingenjoy TotalLikingquality TotalPurchase Pearson Correlation 1 .704** .761** .529** Sig. (2-tailed) .000 .000 .000 N 267 267 267 267 TotalLiking Pearson Correlation .704** 1 .714** .631** Sig. (2-tailed) .000 .000 .000 N 267 267 267 267 TotalLikingenjoy Pearson Correlation .761** .714** 1 .558** Sig. (2-tailed) .000 .000 .000 N 267 267 267 267 TotalLikingquality Pearson Correlation .529** .631** .558** 1 Sig. (2-tailed) .000 .000 .000 N 267 267 267 267 **. Correlation is significant at the 0.01 level (2-tailed).
D) ANCOVA Flaw - Anthro
E) ANCOVA human – machine covariates Source Type III Sum of Squares df Mean Square F Sig. Corrected Model 76.394a 7 10.913 9.285 .000 Intercept 30.120 1 30.120 25.626 .000 Bpusage 43.312 1 43.312 36.850 .000 TotalAnthro 12.759 1 12.759 10.855 .001 Imp_humans .227 1 .227 .193 .661 Imp_machines .753 1 .753 .640 .424 Flaw 1.079 1 1.079 .918 .339 Anthro 3.444 1 3.444 2.930 .088 Flaw * Anthro 1.133 1 1.133 .964 .327 Error 304.417 259 1.175 Total 7063.313 267 Corrected Total 380.810 266 a. R Squared = .201 (Adjusted R Squared = .179)
F) ANCOVA Flaw – Anthro – AverageOOP
G) Regression Coefficients – Clean OOP scale
Model
Unstandardized Coefficients Standardized Coefficients
B Std. Error Beta t Sig.
1 (Constant) 4.758 .133 35.816 .000 Flaw .257 .189 .107 1.357 .176 Anthro .350 .190 .146 1.837 .067 AnthroxFlawInteraction -.242 .271 -.088 -.895 .372 CleanOOP_c -.018 .154 -.014 -.116 .908 COOPFlaw -.008 .254 -.004 -.033 .974 COOPAnthro .149 .199 .090 .748 .455 COOPAnthroFlaw -.221 .312 -.094 -.710 .478 BPU_c .209 .034 .352 6.064 .000 TotalAnthro_c .182 .055 .190 3.298 .001 a. Dependent Variable: Total Liking and Purchase
H) Rotated Factor Matrix
Factor 1: 5, 7, 12, 13, 14 and 15 Factor 2: 4, 6, 8, 9, 10
Factor 3: 1, 3, 11
F1: People who have high expectations in others and want them to strive to better themselves and do their best F3: people who tend to criticize others and friends
I) Regression coefficients – Factors ANOVAa Model Sum of Squares df Mean Square F Sig. 1 Regression 95.028 17 5.590 4.870 .000b Residual 285.782 249 1.148 Total 380.810 266 Model Unstandardized Coefficients Standardized Coefficients B Std. Error Beta t Sig.
J) Correlation Matrix Empathy
Totalfeelings Totalperspective Totalempathy Totalfreewill Totalsimilarity Totalfeelings Pearson Correlation 1 .817** .632** .615** .552** Sig. (2-tailed) .000 .000 .000 .000 N 267 267 267 267 267 Totalperspective Pearson Correlation .817** 1 .698** .710** .601** Sig. (2-tailed) .000 .000 .000 .000 N 267 267 267 267 267 Totalempathy Pearson Correlation .632** .698** 1 .740** .720** Sig. (2-tailed) .000 .000 .000 .000 N 267 267 267 267 267 Totalfreewill Pearson Correlation .615** .710** .740** 1 .701** Sig. (2-tailed) .000 .000 .000 .000 N 267 267 267 267 267 Totalsimilarity Pearson Correlation .552** .601** .720** .701** 1 Sig. (2-tailed) .000 .000 .000 .000 N 267 267 267 267 267 K) Mediation Output ************************************************************************** Model = 8 Y = DV_Likin X = Anthro M = AllEmpat W = Flaw Statistical Controls: CONTROL= TotalAnt BPU_c Sample size 267 ************************************************************************** Outcome: AllEmpat Model Summary R R-sq MSE F df1 df2 p ,3108 ,0966 2,0292 5,5828 5,0000 261,0000 ,0001 Model
Anthro ,3135 ,2491 1,2584 ,2094 -,1770 ,8040 –a1 Flaw ,0028 ,2463 ,0114 ,9909 -,4822 ,4878 –a2 int_1 ,2835 ,3531 ,8029 ,4228 -,4118 ,9789 –a3
TotalAnt ,2059 ,0709 2,9061 ,0040 ,0664 ,3454 BPU_c ,1287 ,0444 2,8970 ,0041 ,0412 ,2161 Product terms key:
int_1 Anthro X Flaw
************************************************************************** Outcome: DV_Likin Model Summary R R-sq MSE F df1 df2 p ,5062 ,2562 1,0894 14,9288 6,0000 260,0000 ,0000 Model
coeff se t p LLCI ULCI constant 4,8044 ,1280 37,5272 ,0000 4,5523 5,0565 AllEmpat ,2042 ,0454 4,5027 ,0000 ,1149 ,2935 – b Anthro ,2967 ,1831 1,6205 ,1063 -,0638 ,6572 – c1 Flaw ,2594 ,1805 1,4374 ,1518 -,0959 ,6147 – c2 int_2 -,3202 ,2591 -1,2361 ,2176 -,8303 ,1899 – c3 TotalAnt ,1481 ,0527 2,8080 ,0054 ,0443 ,2520 BPU_c ,1812 ,0331 5,4815 ,0000 ,1161 ,2464 Product terms key:
int_2 Anthro X Flaw
******************** DIRECT AND INDIRECT EFFECTS ************************* Conditional direct effect(s) of X on Y at values of the moderator(s):
Flaw Effect SE t p LLCI ULCI ,0000 ,2967 ,1831 1,6205 ,1063 -,0638 ,6572 1,0000 -,0235 ,1832 -,1285 ,8979 -,3842 ,3372 Conditional indirect effect(s) of X on Y at values of the moderator(s):
Mediator
Flaw Effect Boot SE BootLLCI BootULCI AllEmpat ,0000 ,0640 ,0568 -,0362 ,1852 AllEmpat 1,0000 ,1219 ,0565 ,0248 ,2533 ---
Indirect effect of highest order product: Mediator
Expectation H1
0 0.5 1 1.5 2 2.5 3No Flaw Flaw Anthro/No Flaw Anthro/Flaw
Study Design
•2 (anthropomorphism vs. no anthropomorphism) x 2 (flaw vs. no flaw) between-subject design
•Online survey on Amazon mechanical Turk, designed with Qualitrics
•Subjects randomly assigned to account for individual differences à internal validity
Conditions
No Flaw/No Anthro condition: “This is a BeBe backpack. It has a substantial storage capacity and a built-in power bank for charging various devices.” Flaw/Anthro condition: “This is a BeBe backpack. It has a substantial storage capacity and a built-in power bank for charging various devices. Please note, on the pocket on the right side of the backpack, the color of the zipper is slightly different from others: It is gray instead of black (see rightmost photo).”Conditions II
No Anthro/No Flaw condition: “This is a BeBe backpack. It has a substantial storage capacity and a built-in power bank for charging various devices.” Anthro/Flaw condition:Measurements
Anthropomorphism: Independent dummy variable (present or not present) Flaw: Independent dummy variable (present or not present) Empathy: 4 questions (empathy, feelings, perspective, freewill) à 1 IV Dependent variables: Liking (3 items: enjoyment, general liking, quality) and purchase intention (asked simultaneously with showing picture) à 1 DVOOP: 15-item scale – some reverse coded à IV (asked at the end)
Methods of Analysis
H1: ANCOVA - the Flaw, Anthropomorphism and interaction (categories) effects on the dependent variable (Liking - ordinal) H2: Multiple regression H3: Mediation Macro by Hayes Model 8Results
ANCOVA H1: 4.5 4.6 4.7 4.8 4.9 5 5.1 5.2 5.3Anthro No Anthro Total Total
Total Liking and Purchase Intention Means
