2021
The Influence of Artistic
Symbolism on Product
Evaluations
A DEEPER LOOK AT THE ART INFUSION EFFECT: HOW
CONGRUENCY BETWEEN ARTWORKS AND PRODUCTS CAN
IMPACT CUSTOMER PERCEPTIONS
ANNET REIJ
The Influence of Artistic Symbolism on Product Evaluations
A deeper look at the art infusion effect: How congruency between artworks and products can impact customer perceptions
Annet Reij
Master Marketing Management Master Thesis
January 11 2021 Supervisor: dr. J. Berger Faculty of Economics and Business
Management Summary
It has become commonplace in the business environment for companies to use art for advertising purposes. Surprising this is not, as art throughout history, has held connotations with luxury and prestige, which on their turn tend to positively affect consumer perceptions. The phenomenon of the positive impact of the presence of art on consumer perceptions is captured in what is called the “art infusion effect”. Whereas various aspects of the art infusion effect had been researched over the years, one aspect that had been left untouched is the influence of symbolism in art on customer perceptions. This study aimed to investigate whether congruency between the thoughts and feelings, as evoked by a work of art containing symbolism, and those evoked by a product, are positively related to product evaluation, as mediated by brand affect. Moreover, it aimed to investigate whether this effect, if present, was moderated by the processing style of an individual (holistic opposed to analytical).
Data was collected through means of a digital survey, in which respondents were presented images of products combined with artworks of various degrees of congruency. They were then asked to indicate their perceptions of the (fictional) brands and products. Lastly, they were asked the questions of the analysis-holism scale (AHS). Data analysis was then conducted through partially-least squares analysis.
The results showed that there was no proof for an interaction effect between processing style and degree of congruency. Moreover, there were no significant relationships between middle level congruency; and brand affect and product evolution. However, high degree of congruency was shown to be significantly related to brand affect, which on its turn was significantly related to product evaluation. Full mediation could be assumed. An alternative analysis showed that the perceived degree of congruency was also significantly related to brand affect, which on its turn was significantly related to product evaluation. Here partial mediation could be assumed.
Preface
Before you lies the thesis “The Influence of Artistic Symbolism on Product Evaluations. A deeper look at the art infusion effect: How congruency between artworks and products can impact customer perceptions”. It was written between September 2020 and January 2021 as the final graduation requirement of the MSc Marketing Management program at the University of Groningen.
Conducting this research has been a struggle as much as it has been a joy. While much time has been spent breaking my head over what I then presumed to be “the one puzzle piece out of place” (which then turned out to be more than one), or untangling myself from ideas going in all directions, I am grateful for the learning process. Moreover, it delights me to have been able to integrate one of my personal interests, that being art, into this research.
Table of Contents
Introduction 5
Theoretical Framework 8
Literature Review 8
The Art Infusion Effect 8
Symbolism in Art 9
Congruence and Incongruence 10
Processing Styles 11
Brand Affect & Product Evaluation 12
Conceptual Model 12
Hypotheses 13
Research Design 15
Field Design 15
Measures 16
Statistical Method Used 16
Results Sample Representativeness 18
Measurement Model 18
Degree of Congruency (IV) 19
Reflective Scales 22
Structural Model 25
Additional Analysis 28
Conclusion 31
Recommendations 32
Limitations and Directions for Future Research 32
References 34
Appendices Appendix A 37
Appendix B 41
Introduction
“Art is for everybody”, were the words once spoken by the late American artist Keith Haring (1958-1990). And it appears, albeit with a different rationale, marketeers would agree. After all, it is commonplace in the contemporary business reality for companies to be engaged in arts. While often associated with the luxury industry (e.g. the 2017 Louis Vuitton x Jeff Koons collection or the Montblanc Patron of the Arts line), the utilization of art by businesses can be seen in mass production companies as well (e.g. the Uniqlo and New York Museum of Modern Art (MoMa) partnership or the 2014 collaboration between coffee brand Douwe Egberts and the Rijksmuseum). Not only are works of art made part of the product being sold, depictions of artworks are also displayed in advertisements of products and services completely unrelated (e.g. a video advertisement by The Wonderful for its bagged pistachios with the Venus de Milo).
The utilization of art by companies does not come surprisingly, given the fact that art through history has been known to be able to evoke many thoughts and emotions from its observer (Balter, 2009). Moreover, art has also been associated with high culture and thereby, with concepts such as luxury and prestige; which on their turn can be related to product evaluation and purchase intention. In 2008 Hagtvedt and Patrick examined the phenomenon in which the ‘presence of art’ is positively related with ‘product evaluation’ through the mediation of ‘perceptions of luxury’. That phenomenon has since then been known by the term ‘art infusion’.
As mentioned by Hagtvedt and Patrick (2008), one aspect of the art infusion effect worth examining, and that is still a grey area (according to a theoretical examination conducted through a snowballing search method on Google Scholar as of September 2020), is the impact of the symbolism of the artwork used. After all, symbolism in art might evoke thoughts and feelings in consumers that affects their perception of the product. Moreover, while Hagtvedt and Patrick argue that art infusion might occur independent from the content of the art, even when the artwork present has negative content, the extent to which this statement holds is unclear. Could it be possible that the presence of art containing symbolism strengthens the art infusion effect, compared to the presence of art without such meaning? Or more precisely, would a congruence between the thought and feelings a piece of art evokes, and the thought and feelings associated with a certain product be more positively related to brand affect (and by extend, product evaluation) than in the case of no congruence or even an incongruence? Moreover, as symbolism in art is related to association and allegorical thinking, will the relationship between congruency and brand affect be moderated by a consumer’s processing style (holistic vs analytical) (Nisbett, Peng, Choi, & Norenzayan, 2001)? After all, symbolism is a rather implicit way of conveying thoughts and/or feelings, which therefore, can go either go unnoticed, be found ambiguous or be obvious to someone, all depending on the individual’s style of thinking.
By examining the aforementioned, researchers could gain more insight in the scope of the art infusion effect and its underlying mechanism. Marketeers on the other hand could gain viable insights on conditions under which they would have to choose what type of artwork to market their products.
In this thesis, the following research question will be answered:
Will the presence of congruency between a product and a work of art, containing symbolism, used in its marketing, affect product evaluation, beyond art infusion as caused by an unrelated work of art? And if so, is this relationship moderated by the processing style of
consumers?
Theoretical Framework
Literature Review
The Art Infusion Effect
How to define ‘art’ has been the subject of debate since the early days of civilization. One definition would be “the making of objects, images, music, etc. that are beautiful or that express feelings” (Cambridge Dictionary, 2020). Although this would then likely lead into a discussion on what is to be considered ‘beautiful’ or whether every expression of feelings can truly be considered art. However, when taking a consumer-perspective, as is here the case, the definition of art that is upheld is one similar to that of Dewey (1934), in which a work of art comes to be when there is an interaction between the observer’s experience and the object that generates substances towards a sense of fulfillment. In other words: art is what is regarded as such in the eye of the beholder. Moreover, this research will limit itself to the visual fine arts opposed to other art forms such as music and dance.
Symbolism in Art
have this innate desire to mimic the movements made by figures on the artwork, (in the case of styles such as cubism, the movements one assumed the artist to have made) which then results in the evocation of emotions (Freedberg, 2011). Furthermore, symbolism can be expressed through the use of color. Color after all, can carry in itself a meaning, both subjective (e.g. based on cultural associations (Wegman & Said, 2011), but also in a more objective manner (e.g. warm and cool tones (Eiseman, 2006). Generally, though it can be stated that color can be considered an “implicit affective cue” influences humans in terms of affect, cognition and behavior (Friedman & Förster, 2010). To put it shortly, through different media and in different forms, humans have been able to evoke feelings, cement social relations, communicate messages and convey ideas and feelings regarding intangible concepts, by using symbolic expression.
Congruence and Incongruence
Congruence refers to whether or not something fits the context. According to Heckler and Childers (1992) congruency is built out of two components: relevancy and expectancy. The former refers to the extent to which a stimulus (in the context of art infusion, the piece of art used in marketing) is directly related to and can easily be understood in the context of the product being advertised. Furthermore, expectancy refers to the extent to which the information being carried through the stimulus fits into pre-determined structures in the mind. Incongruence, naturally, occurs when a stimulus is not perceived as relevant or expected at all.
Previous research with music as the stimulus has shown that congruency of an advertising stimulus is positively related to positive affective reaction regarding the product (Galan 2009). A similar conclusion was reached in a research on ad-context congruency in advertisements on online blogs (Segev, Wang & Feranandes, 2014) and in research on congruency of environmental issues and product categories (Shin & Ki, 2018).
expectations and pre-determined structures regarding the product categories, nor the artworks were examined.
When presented with a product, or an image of, consumers are likely to possess precognitions regarding the object (unless it is far beyond what they have ever encountered before) and therefore have certain recollections of and associations with the product. Similarly, when exposed to a work of art, thought and ideas will be evoked within the consumer. Thus, depending on whether the product and the symbolism in the work of art can be considered in one another’s context, we can speak of congruency, or possibly, a lack thereof.
Processing Styles
Brand Affect & Product Evaluation
Brand affect is “a brand’s potential to elicit a positive emotional response in the average consumer as a result of its use” (Chaudhuri & Holbrook, 2001: 82). While Hagtvedt and Patrick (2008) found ‘perceived luxury’ to mediate the art infusion effect, Estes et al. (2018) found that by including ‘brand affect’ as a mediator, the mediating role of perceived luxury was diminished. This can be explained, as luxury perceptions themselves, are comprised of antecedent factors, one being emotion, which is captured within brand affect. Moreover, Sung & Kim (2010) found that excitement and sophistication are highly related to the brand affect. Sophistication on its turn, is heavily associated with luxury.
In relation to the dependent variable ‘product evaluation’, research has shown that emotion can affect a plethora of consumer behaviors (Holbrook & Batra, 1987). Therefore, brand affect has been established as a mediator for product evaluation.
Thus, to summarize the relationship between brand affect and product evaluation; the presence of art can elicit emotions/affect, which, when the work of art is used in a marketing context, have the potential of creating brand affect and thereby improving product evaluation.
Conceptual Model
The following conceptual model was created (Model 1).
Model 1: Conceptual Model
Product evaluation Degree of congruency
between symbolism in work of art and product
Hypotheses
Since the concept of art infusion was coined in 2008, the topic has been researched across different art styles, different ways of integrating it into marketing, as well as with different moderators. Among all those researches, one phenomenon always stands: the fact that the use of art in marketing has a positive spillover effect of general favorable perception, positively related to product evaluation. After all, there appears to be some inherent relationship between the presence of art and allusions to luxury and class, which in its turn leads to a higher brand affect, which mediates the relationship between the presence of art and product evaluation (Estes et al., 2011). Moreover, as stated before, perceived congruency appears to be related to a positive reaction towards a product (Galan, 2009; Segev, Wang & Fernandes, 2014; Shin & Ki, 2018). Based on the positive relationship between both the presence of art and congruency individually; and brand effect and product evaluation, the following hypotheses were formed.
H1: The degree of congruency between symbolism in a work of art and the product type, is
positively related to product evaluation.
H2: The degree of congruency between symbolism in a work of art and the product type, is
positively related to brand affect.
when no art work is present, independent from the content of the artwork; the product evaluations (through the mediator ‘brand affect’) of participants with an analytical processing style are still expected to benefit from the presence of art. However, given that, as aforementioned, the congruency between the product and the symbolism in the artwork might go unnoticed by the analytical thinker, the hypothesized added value of congruency will likely not be present. Based on those considerations, the following hypothesis was formed.
H3: A holistic processing style has a positive effect on the relationship between congruency
and brand affect.
Previous research has found brand affect to be a mediator of several consumer behaviors, with one of them being product perceptions, but also price evaluations (Holbrook & Batra, 1987). As this relationship has been considered well-established among a plethora of consumer behaviors and perceptions and no reasons to argue this have been presented by this research, the following hypothesis was adopted.
Research Design Field Design
Participants were gathered by distributing the online questionnaire through social media and instant messaging within the personal sphere. The survey was made available in English, Japanese and Dutch (the latter two having been created through the backward translation process).Participants were first guaranteed that their answers would be processed anonymously and were then asked to perform consent. If participants agreed to the terms and conditions, they were shown the following message: “To investigate consumer attitudes towards new brands
that might be introduced into the market, some (preliminary/draft) advertisements for these new brands were created. On the following page, you will be presented with an ad for a new brand, and you will be asked to evaluate several aspects of the brand. Once completed, the process will be repeated with two more advertisements for new brands. Please take a good look at the following ad:” Participants were randomly assigned to one of three sets of images. They were
all shown advertisements for chocolate, a pen and toothpaste; however, across the three image sets the artworks combined with the product differed. However, every participant was still shown an image across all congruency conditions (low, middle, high). This method was chosen as to prevent the survey from becoming too repetitive, as well as to prevent participants from feeling manipulated (as would likely be the case if participants were asked to rate the same product type thrice). Then they would continue to the questions relating to ‘brand affect’ and ‘product evaluation. Following that, manipulation-checks were performed. Participants were asked to indicate the degree to which they considered the image used in the ad to fit the product and also whether they considered the image to be an artwork (in terms of the manner in which it was created). As the art infusion effect can only take place when an artwork is also considered to be art in the eye of the beholder. Afterwards, they were shown the following message: “You
have nearly completed the survey. The next, and last, section contains various statements regarding your worldview. Please indicate to what extent you agree with each statement.” After
Measures
The independent variable was the level of congruency between artwork and product category (ad image: low, middle, high congruency; between-participants). As symbolism is inherently present (or not) in a work of art; however, will likely not be picked up on by all viewers (depending on their processing style), congruency was determined and appointed to the ad images based on objective criteria. The ad images as presented to the participants can be found in Appendix A.
The moderater, processing style, was measured (on a 1-7 Likert-scale (very strongly disagree to very strongly agree), by using the 24 item Analysis-Holism Scale (AHS) as developed by Choi, Koo & Choi (2007).
The mediator, brand effect, was measured on a 1-7 Likert-scale based on brand affect items (I would feel good if I use this brand; This brand makes me happy; This brand gives me pleasure) from Chaudhuri & Holbrook (2001).
The dependent variable, product evaluation, was measured similarly as in Hagtvedt and Patrick (2008, 2011), by asking participants to evaluate the product on five semantic differences on a 1-7 Likert-scale (dislike/like, bad/good, negative/positive, unfavorable/favorable and unpleasant/pleasant).
An overview of all the items for each construct can be found in Appendix B.
Statistical Method Used
and thus, relying solely on reliability analysis through the Cronbach’s alpha, would always lead to a high outcome due to the high number of items. Moreover, the variables for ‘brand affect’ and ‘product evaluation’ were created by combining the 3 and 5 scale items (which were all reflective to the construct opposed to formative) respectively. To confirm whether the reflective constructs were reliable, as well as valid, composite reliability, Cronbach’s Alpha, item validity, within-method convergent validity and discriminant validity were analyzed following Leroi-Werelds et al. (2014).
Results
Sample Representativeness
The survey was completed and submitted by 94 respondents, aged between 18 and 53 (M=24.43). 47 of the respondents identified as male, 45 as female, 1 as other and 1 person did not specify. Participants from various countries in Asia, Europe, North America and South America were represented in the sample. Moreover, 72 percent of all participants are currently students.
As the aim of the study was to gain more insight in the scope of the art infusion effect as a whole, and not in regards to e.g., age or educational level, the population of interest could be defined as the general population. Thus, although people from various age groups and countries were represented in the sample, the sample cannot truly be deemed as representative. After all, the sample size, will sufficient to run the statistical analyses with, was rather small and was not reflective of the general population. Moreover, this did not only pertain to variables reflective of the general demographics (i.e., those that were not the main focus of the research), but also to a passive variable that was essential to the study. After all, after closer examination, the data showed that range of the variable used to measure processing style (AHS) measured at 2.67, with a minimum of 3.42 and a maximum of 6.08 (Table 1). Given the fact that the variable was measured on a 1-7 scale, representing the spectrum between analytical processing on one hand (1) and holistic processing on the other (7), very analytical thinkers were underrepresented in the sample.
Table 1: Descriptive statistics ‘AHS’
Measurement Model
Before the analysis of the structural model in the following section, this section will lay out the constructs used in the analysis and their operationalization.
Variable N Range Minimum Maximum Mean
Degree of Congruency (IV)
As stated in the previous chapter, participants of the survey were each presented with a set of three images, each of which was representative of a different degree of congruency (low, middle and high). As the variable representing the degree of congruency, between the symbolism in a work of art and product type, was not used in research before, no assumptions regarding validity could be made based on previous research. For that reason, as stated in the previous chapter, objective criteria were used to select artworks and combine them with products. After all, doing so would increase the likelihood of the degrees of congruency of the images being perceived as intended. It should be considered however, that this would be no guarantee.
While the run of pre-test could have confirmed whether all artworks selected were considered to be art and, whether the levels of congruency appointed to the combinations of artworks and products were deemed to match, the circumstances did not enable this. After all, the pool research participants the researcher had access to was rather limited in number, which would then cause a problem later on, as those who would have participated in a pre-test would later on not be allowed to participate in the main survey, as they likely would already have too much insight in the research to give unbiased answers. Instead, as mentioned before, manipulation checks were included in the actual survey.
objective criteria were used, and the researcher tried to refrain from using artworks with symbolism and allegories of which meanings would widely vary amongst cultures, or using very obscure/unrecognizable art styles, complete objectivity could never be guaranteed, nor subjectivity completely mitigated.
Table 2: Descriptive statistics on the ‘degree to which the images were considered to be art’
Table 3: Descriptive statistics on the ‘perceived degree of congruency’
Category N Range Minimum Maximum Mean Std. Deviation
Congruency_Choco_1_1 99 6 1 7 2.55 1.486 Congruency_Choco2_1 99 5 1 6 3.33 1.813 Congruency_Choco3_1 84 6 1 7 3.64 1.767 Congruency_Pen1_1 84 5 1 6 3.32 1.805 Congruency_Pen2_1 99 6 1 7 3.85 1.734 Congruency_Pen3_1 99 6 1 7 4.55 1.567 Congruency_TP1_1 99 5 1 6 2.76 1.604 Congruency_TP2_1 84 6 1 7 2.39 1.481 Congruency_TP3_1 99 5 1 6 3.61 1.544
While it was taken into consideration to remove all data from the toothpaste images from the analysis, ultimately it was decided not to do this. Not only would this have heavily reduced the number of cases in the dataset, but moreover, it would have obstructed the design of the research, as then each participant would only provide data on two out of the three degrees of congruency, which then would reduce the insight on the effects of the different treatments for the same individual.
Category N Range Minimum Maximum Mean Std. Deviation
To verify whether the degrees of congruency were significantly different from one another, an analysis of variance was conducted. First of all, to determine whether a parametric or a non-parametric test had to be conducted, a test of normality was conducted using Kolmogorov Smirnov (as the sample size >50). As the level of significance for each category (low, middle and high) were 0.005, 0.001 and 0.003 respectively (table 4), the null hypothesis of normality had to be rejected for all three. Therefore, a non-parametric test had to be conducted.
Table 4: Kolmogorov-Smirnov of ‘degree of congruency’
Degree of Congruency Statistic df Sig.
Low 0.113 94 0.005
Middle .123 94 .001
High .118 94 .003
As the survey measured every participant three times (once for every degree of congruency), the sample was random, the dependent variable (product evaluation) was interval data and the variable was, as just stated, not normally distributed, it was chosen to conduct a Friedman test (Laerd Statistics, 2020). As shown in table 5, the different groups (low, middle and high) were overall statistically different, with a score of 0.003.
Table 5: Friedman test ‘degree of congruency’
However, as the Friedman test does not show which of the groups are statistically different from one another, a post-hoc test, the Wilcoxon Signed Rank test, was conducted. As show in table 6, the groups ‘low’ and ‘middle’ were statistically different to one another (sig. 0.045). Similarly, ‘low’ and ‘high’ were statistically different (sig. 0.002). However, this was not the case for ‘middle’ opposed to ‘high’ as the significance level (0.362) was larger than 0.05.
N 94
Chi-Square 11.540
df 2
Table 6: Wilcoxon Signed Rank test ‘degree of congruency’ Middle - Low High - Low High - Middle Z -2.007 -3.073 -911
Asymp. Sig. (2-tailed) .045 .002 .362
Reflective Scales
As stated before, the variable representing the variable ‘processing style’, henceforth referred to as ‘AHS’, was measured by using the 24-item as developed by Choi, Koo & Choi (2007). Of the 24 items, 6 each were aimed at capturing a different facet of holistic (compared to analytical) thinking. The four categories are causality, attitude towards contradictions, perception of change and locus of attention. Through confirmatory factor analysis (CFA) (with Promax rotation), it was tested whether or not the 24 items indeed loaded to the same factors as described in the research, which was indeed the case. 4 factors were shown to be appropriate as the KMO value passed the benchmark of 0.5 at 0.613 and the Bartlett’s test of sphericity was passed at sig. 0.000 (see Appendix C). The 6 items loading to the same factor were then also put through a reliability analysis. The CA of all 4 factors passed the benchmark of 0.6: 0.767 (causality), 0.692 (contradictions), 0.641 (change) and 0.776 (attention). To confirm whether all 4 factors (thus in fact, 24 items) all indeed measured the same variable (holistic processing style) (as shown in Choi, Koo & Choi, 2007), the CA was also measured for all 24 items, resulting in a score of 0.703 and thus also surpassing the 0.6 benchmark (Hair, et al., 2006).
As stated before, the 24 items of which the AHS scale was comprised, were aimed at capturing 4 different facets of holistic thinking. Therefore, strictly speaking the construct ‘AHS’ would be a formative scale; however, as all the items are interrelated, they were treated as being reflective in nature. Moreover, the items/indicators of both the variables ‘brand affect’ and ‘product evaluation’ were both reflective in nature as well, as high interrelations between the indicators were expected.
Thus, as all the constructs are reflective, following Leroi-Werelds et al. (2014), composite
reliability and the Cronbach’s Alpha were analyzed to determine reliability. Moreover, item validity, within-method convergent validity and discriminant validity were analyzed to
Reliability
Table 7 shows the Cronbach’s alpha and composite reliability scores for all the reflective constructs in the model. All CA and composite reliability values measure at >0.7, meaning that the threshold of both as set by Leroi-Werelds et al. (2014) has been passed for all, with the exception of the composite reliability of ‘AHS’ (0.650). However, as the score is still relatively high, all in all it can be stated the constructs as used in the model were reliable.
Table 7: Reliability of reflective constructs
Cronbach's
Alpha Composite Reliability
AHS 0.708 0.650 Brand Affect 0.921 0.950 Product Evaluation 0.952 0.963 Validity
Table 8 shows the items loadings of all items to their respective constructs. According to Leroi-Werelds et al. (2014) item validity is satisfied when the item loading >0.7. Whereas all items for the constructs ‘brand affect’ and ‘product evaluation’ do satisfy this requirement, this cannot be stated for items of the construct ‘AHS’ (with the exception of PS_causality_5, which scores at 0.711).
Table 8: Item validity of reflective constructs
Construct Item Item Loading
Table 9 shows the average variance extracted (AVE) scores for the reflective constructs. The AVE is a reflection of the extent to which the construct explains the variance in the items of which it is constructed. The threshold of >0.5, according to Leroi-Werelds et al. (2014), that has to be surpassed to be allowed to assume within-method convergent validity, was passed for both ‘brand affect’ and ‘product evaluation’.
Table 9: AVE of reflective constructs
AHS 0.148
Brand Affect 0.864
Product Evaluation 0.838
Lastly, in regards to validity, table 10 shows the heterotrait-monotrait ratio of correlations PS_Change_4 PS_Change_5 PS_Change_6 PS_causality_1 PS_causality_2 PS_causality_3 PS_causality_4 PS_causality_5 PS_causality_6 PS_contradictions_1 PS_contradictions_2 PS_contradictions_3 PS_contradictions_4 PS_contradictions_5 PS_contradictions_6 0.238 0.443 0.196 0.578 0.153 0.681 0.561 0.711 0.587 0.421 0.527 0.430 0.400 0.288 0.256
Brand Affect Brand_Affect_1
(HTMT) scores for all the reflective constructs. According to Hair et al. (2017), discriminant validity can be assumed between two reflective constructs when the HTMT value is <0.90. As all the HTMT values are <0.9, it can be stated that the constructs are significantly unrelated to one another.
Table 10: HTMT of reflective constructs
AHS Brand Affect AHS Brand Affect 0.230 Product Evaluation 0.270 0.832
All with all, it can be stated that both the constructs ‘brand affect’ and ‘product evaluation’ met all the criteria needed to be met to assume reliability and validity. In other words, test results for the constructs were consistent, and the constructs measured what they were intended to measure. However, for the construct ‘AHS’ the results were shown to be less promising. Whereas, as stated before, reliability did not appear an issue based on the CA and composite reliability scores, the same could not be stated regarding the validity of the construct. After all, from the three validity tests, only the criterion to assume discriminant validity was met. Item validity and within-method convergent validity were not met. Presumably, this was caused by the nature of the construct (24 items, representative of 4 factors, representative of 1 scale). However, as the CFA showed that the items did in fact load on the expected factors and the use of one single constructed was supported by previous research (Choi, Koo & Choi (2007), it was decided to proceed with the existing construct.
Structural Model
Model 2: Structural model
Bootstrapping was set at 1000 subsamples and α=0.05.
First of all, the coefficient of determination, R2 was examined. R2 reflects the proportion of variance in the dependent variable that can be explained by the independent variables. Table 11 shows the R2 to have been rather strong at 0.611 for ‘product evaluation’ but rather low for ‘brand affect’ at 0.102; however, as the study took place in the field social sciences (opposed to the natural sciences), for which R2 as low as 0.10 are acceptable, both pass the benchmark.
Table 11: Coefficient of determination (R2)
R Square Brand Affect 0.102 Product Evaluation 0.611
Table 12: Path Coefficients and Hypotheses
Hypothesis Path Path Coefficient P-Values
H1: The degree of
congruency between symbolism in a work of art and the product type, is positively related to product evaluation. Congruency (middle) à Product Evaluation 0.058 0.207 Congruency (high) à Product Evaluation 0.024 0.590 H2: The degree of congruency between symbolism in a work of art and the product type, is positively related to brand affect.
Congruency (middle) à Brand Affect 0.045 0.506 Congruency (high) à Brand Affect 0.170 0.007 H3: A holistic processing style has a positive effect on the relationship between congruency and brand affect. Congruency (middle) x AHS à Brand Affect -0.002 0.975 Congruency (high) x AHS à Brand Affect
-0.008 0.907 H4: Brand affect is positively related to product evaluation. Brand Affectà Product Evaluation 0.779 0.000
Regarding the dummy variable representing the degree of congruency ‘high’, its relationship (XàM in Baron & Kenny, 1986) with ‘brand affect’ was shown to be significant at 0.007 (table 10). The relationship between ‘high’ and ‘product evaluation’ on the other hand, was not significant at 0.590. However, going by Baron & Kenny (1986), this was desired as the inclusion of the mediator would have to render the relationship XàY (in this case ‘high’ and ‘product evaluation’), (partially) insignificant, whereas it had been significant before the involvement of the mediator. This too was the case in this study, as the direct relationship (Xà Y) between the degree of congruency ‘high’ and ‘product evaluation’ was indeed significant at 0.019, before the involvement of the mediator.
Moreover, the relationship between the mediator ‘brand affect’ and the DV ‘product evaluation’ (Mà Y in Baron & Kenny, 1986) was proven significant at 0.000.
In other words, as in Model 2 the relationship XàMàY was proven to be significant, whereas XàY was not anymore, all the four conditions of Baron & Kenny (1986) as mentioned in chapter 3 were satisfied and full mediation could be assumed. Moreover, as the significant path coefficients were all positive, this meant that H4 was accepted; as well as H1 and H2 for the category ‘high’ regarding degrees of congruency.
However, while the relationship between ‘AHS’ and ‘brand affect’ was significant (p-value 0.000) and positive with a standardized path coefficient of 0.280, the moderation effect itself between ‘high level of congruency’ and ‘AHS’ was not significant at 0.907. Therefore, H3 was rejected. This appears to indicate that there is simply no interaction effect between ‘degree of congruency’ and ‘AHS’; however, regarding the concerns raised earlier (in the section on the measurement model) regarding the validity of the construct, the root cause might also lie in the construct of the variable.
Additional Analysis
symbolism, and in extend congruency between artwork and product, was considered to be inherent to the image. However, as shown by the manipulation check, respondents did not always perceive the degree of congruency as intended.
Alternatively, however, instead of examining the relationship between an objectively determined degree of congruency; and brand affect and product evaluation (potentially moderated by processing style), one could analyze the relationship between the perceived degree of congruency; and brand affect and product evaluation. In the latter, processing style is purposefully not included, as that would likely lead to multicollinearity, as the theoretical framework supports the notion that the perception of congruency is already related to an individual’s processing style. Moreover, in the same line of thinking to the original model, the relationship between ‘perceived degree of congruency’ and both brand affect and product evaluation is expected to be positive.
Model 3 shows the structural model of the alternative analysis.
Model 3: Structural model – Alternative Analysis
Bootstrapping was set at 1000 subsamples and α=0.05.
Table 13: Coefficient of determination (R2)
R Square
Brand Affect 0.392
Product Evaluation 0.655
Moreover, as shown in table 14, all relationships were significant at 0.000. Additionally, the path coefficients were all positive, further confirming the hypothesized relationships.
Table 14: Path Coefficients
Paths Path Coefficients P Values
Brand Affect -> Product Evaluation 0.606 0.000
Perceived Degree of Congruency -> Brand Affect 0.626 0.000
Perceived Degree of Congruency -> Product
Evaluation 0.278 0.000
Conclusion
In conclusion, this research was conducted to gain a deeper insight in the art infusion effect, or more specifically, in the way congruency between associations related to a product type and the presence of symbolic meaning in art, when used in advertising, could affect customer perceptions on a brand and product. Moreover, this research aimed to discover whether such a relationship between congruency and brand affect would be impacted by differences in processing styles (analytical vs holistic) among individuals.
After all, art, even beyond the surface level of the literal content depicted, is able to evoke strong emotions from people. However, as people differ in their ability, or urge, to find deeper meaning, associations and relationships in the world around them, such symbolism might not be registered by everybody. Furthermore, as congruency has been found to be positively related to affect towards products, this research aimed to discover whether such significant relationships between congruency between symbolism in art and product type; and processing style, exist.
The research as described in this thesis showed that there indeed appears to be a significant positive relationship between congruency and product evaluation, mediated through brand affect; however, only at high levels of congruency (and not at the middle level). Moreover, processing style did not appear to be a moderator between congruency and brand affect. While the fact that not all hypotheses could be confirmed could indeed indicate that said hypotheses were incorrect, it should be noted that this research has suffered some methodological issues (on which will be elaborated further in the ‘limitations’ section). Nonetheless, this research has resulted in an interesting finding.
Recommendations
Based on the outcome of this study, several recommendations were formed. This study showed that the scope of the art infusion effect has not yet been fully discovered. Research into the art infusion effect related to symbolism and congruency in specific, could be expanded by examining the effects of different types of symbolism on consumers’ perceptions on brands and products. As stated before, the artworks chosen for this study were chosen to be understood by the general population; however, by for example conducting a study in a specific country, the effects of cultural-specific symbols (of which the meaning has to be thought for it to be understood), opposed to generally understood symbolism, on consumer perceptions could be analyzed. Moreover, a distinction could be made between symbolism evoking positive emotions, opposed to negative emotions and how they would affect consumer perspectives differently. Generally, it would be recommended to conduct more extensive research on the effects of symbolism in art on consumer perceptions and gain more understanding of its underlying mechanisms.
Regarding managerial recommendations, as the study supports the notion that the art infusion effect is in fact not content-independent, managers, when deciding to use art in advertisement, should reflect on the connotations a work of art may have beyond what is literary depicted; and how such associations between symbolism in art and the product/service being sold could positively affect consumer perceptions on the brand and its products/services.
Limitations and Directions for Future Research
This study had various limitations. First of all, the sample lacked representativeness in relation to an essential passive variable in the study. Although there were enough data entries to allow for the analyses to be run, as mentioned before the participants did not cover a wide range on the AHS scale. Therefore, the “true” analytical people were not represented in the sample.
For future research it is therefore advised to be very precise with the selection of a construct to measure holism with. Whereas the AHS does provide an extensive inside in several factors of holism, perhaps it is not best suited to be employed in relation to the symbolism in art, as it might even cover facets of holism that are irrelevant to the way symbolism in art is being perceived. Perhaps a more simplified scale/constructs would prove better suited.
Furthermore, because of time constraints and limited resources (in terms of participants that could be gathered), no pre-tests were conducted for the advertisement images, to measure whether or not the intended congruency levels were indeed perceived as such. Instead, manipulation-checks were run in the actual survey. The results of that showed that while participants generally did perceive the chosen artworks to be art; they did not always consider the various degrees of congruency as intended by the researcher. While most images were placed in the same sequence (from low to middle to high), this was not always the case. Moreover, the perceived differences in congruency between them were substantially smaller than intended. Likely, this is one of the reasons why in the analysis of the data, only the advertisements with a high degree of congruency appeared to be significantly related to brand affect and product evaluation, whereas the middle degree of congruency did not.
Therefore, for future research, it is recommended to thoroughly pre-test whether the advertisements that are used for the IV ‘degree of congruency’ are indeed being perceived the same way by participants.
References
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51(6), 1173–1182.
Baumgarth, C. & Wieker, J. (2019). From the classical art to the urban art infusion effect: The effect of street art and graffiti on the consumer evaluation of products. Creativity and
Innovation Management, 29(S1), 116-127.
Cambridge Dictionary (2020). Art. Cambridge Dictionary.
https://dictionary.cambridge.org/dictionary/english/art [Last accessed January 10 2021]. Chaudhuri A, Holbrook M.B. (2001). The chain of effects from brand trust and brand affect to brand performance: The role of brand loyalty. Journal of Marketing, 65(2), 81-93.
Choi, I., Koo, M., & An, J. (2007). Individual differences in analytic versus holistic thinking.
Personality and Social Psychology Bulletin, 33(5), 691–705.
Dewey, J. (1989/1934). “Art as experience”, in Boydston, J. (Ed.), John Dewey: The later works, 1925-1953, Vol. 10, Southern Illinois University Press, Carbondale, IL, 1-400. Eiseman, L. (2006). Color: Messages and meanings: A pantone color resource.
Hand Books Press, Gloucester, MA.
Estes, Z., Brotto, L. & Busacca, B. (2018). The value of art in marketing: An emotion-based model of how artworks in ads improve product evaluations. Journal of Business Research, 85, 396-405.
Freedberg, D. (2011). Memory in art: History and the neuroscience of response. MIT Press, Cambridge, Massachustts, 337-358.
Friedman RS, Förster J. (2010) Implicit affective cues and attentional tuning: An integrative review. Psychology Bulletin, 136(5),875–93.
Galan J-P. (2009). Music and responses to advertising: The effects of musical characteristics, likeability and congruency. Recherche et Applications en Marketing (English Edition), 24(4), 3-22.
Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis (Vol. 6). Pearson Prentice Hall, Upper Saddle River, NJ.
Hagtvedt, H., & Patrick, V. M. (2008). Art infusion: The influence of visual art on the perception and evaluation of consumer products. Journal of Marketing Research, 45(3), 379– 389.
Heckler S.E. & Childers T.L. (1992) The role of expectancy and relevancy in memory of verbal and visual information: What is incongruency? Journal of Consumer Research, 18(4), 475-492.
Herrman, A.R., Omori, K., McNallie, J. & Allen, M. (2013) Evidence of analytic versus holistic thinking in viewing magazine images: Preliminary findings, Qualitative Research
Reports in Communication, 14(1), 53-61.
Holbrook, M. B., & Batra, R. (1987). Assessing the role of emotions as mediators of consumer responses to advertising. Journal of Consumer Research, 14(3), 404–420.
Ishii, K., Miyamoto, Y., Rule, N. O., & Toriyama, R. (2014). Physical objects as vehicles of cultural transmission: Maintaining harmony and uniqueness through colored geometric patterns. Personality and Social Psychology Bulletin, 40, 175-188.
Lacey, S., Hagtvedt, H., Patrick, V.M., Anderson, A., Stilla, R., Deshpande, G., Hu, X., Sato, J.R., Reddy, S., & Sathian, K. (2011). Art for reward's sake: Visual art recruits the ventral striatum. NeuroImage, 55, 420-433.
Laerd Statistics (2020). Friedman test in SPSS Statistics. Laerd Statistics.
https://statistics.laerd.com/spss-tutorials/friedman-test-using-spss-statistics.php [Last accessed January 10 2021].
Leroi-Gourhan, A. Through: Why did prehistoric people draw in caves? Google Arts &
Culture. https://artsandculture.google.com/exhibit/why-did-prehistoric-people-draw-in-the-caves/lQJiKr1gnrB2LQ [Last accessed January 10 2021].
Leroi-Werelds, S., Streukens, S., Brady, M.K., & Swinnen, G. (2014). Assessing the value of commonly used methods for measuring customer value: a multi-setting emperical study.
Journal of the Academic Marketing Science, 42, 430-451.
Logkizidou, M. & Bottomley, P. & Angell, R. (2018). Why Museological merchandise displays enhance luxury product evaluations: An extended art infusion effect. Journal of
Retailing. 95(1,2019), 67-82.
Merriam-Webster (2020). Symbolism. Merriam-Webster. https://www.merriam-webster.com/dictionary/symbolism [Last accessed January 10 2021].
Michael, B. (2009). On the origin of art and symbolism. Science, 323(5915), 709-711. Naletelich, K., & Paswan, A. K. (2018). Art infusion in retailing: The effect of art genres.
Journal of Business Research, 85, 514–522.
Nisbett, R. E., Peng, K., Choi, I., & Norenzayan, A. (2001). Culture and systems of thought: Holistic versus analytic cognition. Psychological Review, 108(2), 291–310.
Peluso, A. M., Pino, G., Amatulli, C., & Guido, G. (2017). Luxury advertising and recognizable artworks: new insights on the “art infusion” effect. European Journal of
Peng, X. & Lai, F. (2012). Using partial least squares in operations management research: A practical guideline and summary of past research. Journal of Operations Management. 30(6), 467-480.
Segev, S., Wang, W., & Fernandes, J. (2014). The effects of ad-context congruency on responses to advertising in blogs: Exploring the role of issue involvement. International
Journal of Advertising, 33(1), 17-36.
Shin, S. & Ki, E. (2018). The effects of congruency of environmental issue and product category and green reputation on consumer responses toward green advertising. Management
Decision, 57(1), 606-620.
Sung, Y.& Kim, J. (2010). Effects of brand personality on brand trust and brand affect.
Psychology and Marketing. 27, 639 -661.
Appendices
Appendix A: Ad images as shown in the survey
Set 1
Chocolate-Low (Andy Warhol-Fish, 1983)
Toothpaste-High (Georgios Jakobides)
Set 2
Pen-High (Sir Alma-Tadema- A Roman Scribe Writing Dispatches)
Set 3
Chocolate-High (Frank Bernard Dicksee- Romeo and Juliet, 1884)
Toothpaste-Middle (Katagawa Utamoro- The Window of Asahiya, 1796)
Appendix B: Items per construct
Construct Item Description
AHS
(Choi, Koo & Choi, 2007)
PS_Attention_1 PS_Attention_2 PS_Attention_3 PS_Attention_4 PS_Attention_5 PS_Attention_6 PS_Change_1 PS_Change_2 PS_Change_3 PS_Change_4
A1. Everything in the universe is somehow related to each other. A2. Nothing is unrelated. A3. Everything in the world is intertwined in a causal relationship. A4. Even a small change in any element of the universe can lead to significant alterations in other elements.
A5. Any phenomenon has numerous numbers of causes, although some of the causes are not known.
A6. Any phenomenon entails a numerous number of consequences, although some of them may not be known.
Ch1. It is more desirable to take the middle ground than go to extremes. Ch2. When disagreement exists among people, they should search for ways to compromise and embrace everyone’s opinions.
Ch3. It is more important to find a point of compromise than to debate who is right/wrong, when one’s opinions conflict with other’s opinions.
PS_Change_5 PS_Change_6 PS_causality_1 PS_causality_2 PS_causality_3 PS_causality_4 PS_causality_5 PS_causality_6 PS_contradictions_1 PS_contradictions_2 PS_contradictions_3 PS_contradictions_4 PS_contradictions_5 PS_contradictions_6
rather than in discord, with others of different opinions than one’s own. Ch5. Choosing a middle ground in an argument should be avoided. Ch6. We should avoid going to extremes.
Ca1. Every phenomenon in the world moves in predictable directions. Ca2. A person who is currently living a successful life will continue to stay successful.
Ca3. . An individual who is currently honest will stay honest in the future. Ca4. . If an event is moving toward a certain direction, it will continue to move toward that direction.
Ca5. . Current situations can change at any time.
Ca6. . Future events are predictable based on present situations.
Co1. The whole, rather than its parts, should be considered in order to understand a phenomenon. Co2. It is more important to pay attention to the whole than its parts. Co3. The whole is greater than the sum of its parts.
Co4. It is more important to pay attention to the whole context rather than the details.
Co5. It is not possible to understand the parts without considering the whole picture.
Co6. We should consider the situation a person is faced with, as well as his/her personality, in order to understand one’s behavior.
(*highlighted items are reverse-coded) Brand Affect
(Chaudhuri & Holbrook, 2001)
Brand_Affect_1 Brand_Affect_2 Brand_Affect_3
1.I would feel good if I use this brand 2.This brand makes me happy 3.This brand gives me pleasure Product Evaluation
Appendix C: Confirmatory Factor Analysis construct ‘AHS’
KMO and Bartlett’s Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.
0.613
Bartlett's Test of Sphericity Approx. Chi-Square 2576. 165
df 276
27/01/2021
The Influence of Artistic Symbolism
on Product Evaluations
A deeper look at the art infusion effect: How congruency between artworks and products can impact customer perceptions
Thesis Defense – January 27 2021
Annet Reij
Supervisor: dr. J. Berger Master Marketing Management Faculty of Economics and Business University of Groningen 1
Table of contents
Introduction • Research Background • Research Question Theoretical Framework Research Design Results 01 02 03 04 Conclusion • Main Concepts • Conceptual Model &Hypotheses
• Field Design • Measures
• Statistical Methods Used
• Sample
• Measurement Model • Structural Model • Additional Analysis 05
• Limitations & Future Research • Recommendations
27/01/2021
Introduction
• Art used in marketing
o Luxury and mass production o Products and advertising • Art evokes thoughts and emotions • Associations with high culture, luxury and
prestige à positive spillover
àArt infusion effect
3
‘presence of art’ is positively related with ‘product evaluation’ through the mediation of ‘perceptions
of luxury’
(Hagtvedt & Patrick, 2008)
27/01/2021
Research Background
• Recent findings
• Gap in literature: possible impact of symbolism à context dependency • Implicit information given by the artwork
o Evocation of thoughts/feelings à purchase intention
o Congruency between perceptions/associations art and product? • Symbolism goes beyond what is depicted
àmoderation by processing style?
• More insight in underlying mechanisms art infusion effect • Marketeers could become more specific in what art to select
5
Research Question
Will the presence of congruency between a product and a work of art, containing symbolism, used in its marketing, affect product evaluation, beyond art infusion as caused by an unrelated work of art? And if so, is this relationship moderated
by the processing style of consumers?
27/01/2021
Theoretical Framework
The Art Infusion Effect
• Definition of Art
o Art is what is considered as such by the beholder (Dewey, 1934)
• Presence of art is positively related to product evaluation (Hagtvedt & Patrick, 2008)
o Even when art is negatively perceived à content independent
• Art images activate rewarding processes in the brain (Lacey et al, 2011)
7
Main Concepts
Symbolism and Content Dependent Influence of Art
• Content dependency likely limited (Hagtvedt & Patrick, 2008)
• Content explored in the context of concretizing the use of art (Hagtvedt & Patrick, 2011)
o Symbolism still unexplored
• Allegorical/metaphorical meaning & associations
o Through e.g. symbols, figures, movements, colours
o Interaction between art and the observer through implicit information
àSymbolic expression to convey ideas and evoke feelings
27/01/2021
Congruence and Incongruence
• Does something fit the context?
o Relevancy and expectancy (Heckler and Childers, 1992)
• Congruent advertising positively related to product perception (Galan, 2009; Segev, Wang &
Fernandes, 2014)
• Congruency between symbolism in art and product yet unexplored
Processing Styles (Nisbett, Peng, Choi & Norenzayan, 2001)
• Analytical vs Holistic
àSymbolism: relationships, implicitly
àCongruency: artwork and product as one entity
9
Brand Affect & Product Evaluation
• Brand affect is “a brand’s potential to elicit a positive emotional response in the average
consumer as a result of its use” (Chaudhuri & Holbrook, 2001:82)
• Emotion can affect a plethora of consumer behaviours among which ‘product evaluation’
(Holbrook & Batra, 1987)
27/01/2021
Conceptual Model & Hypotheses
Brand Effect Processing Style (+) (holistic opposed to analytical) Degree of congruency between symbolism in work of art and product
type
Product evaluation
H1: The degree of congruency between symbolism in a work of art and the product type, is positively related to product evaluation.
H2: The degree of congruency between symbolism in a work of art and the product
type, is positively related to brand affect. H3: A holistic processing
style has a positive effect on the relationship between congruency and brand affect.
H4: Brand affect is positively related to product evaluation.
11
Research Design
Field Design• Online questionnaire
• Shared in the personal sphere • No age/regional restrictions
• Manipulation checks (perceived congruency & art) Measures
• Congruency: Objective criteria
• Processing style: AHS (Choi, Koo & Choi, 2007) (1-7 Likert scale)
• Brand Affect: Statements (Chaudhuri & Holbrook, 2001) (1-7 Likert scale)
• Product Evaluation: Semantic scales (Hagtvedt & Patrick, 2008; 2012) (1-7 Likert scale) Statistical Methods Used
• SmartPLS
27/01/2021
13
Results
Measurement Model
Degree of Congruency (IV)
• All images considered art
• Order of images generally as indented (exc. toothpaste
middle)
• Range not as indented
• Friedman test à categories significantly different (exc.
middle-high) Reflective Scales Sample Age Gender male female other not disclosed
18 53 M:24 Cronbach’s Alpha Composite Reliability
27/01/2021
Structural Model
R2(coefficient of determination) • BA: 0.102
• PE: 0.611
Path Coefficients & Hypotheses
Hypothesis Path P-value H accepted
H1 MiddleàPE 0.207 ✗
HighàPE 0.590 ✓ (full mediation)
H2 MiddleàBA 0.506 ✗ HighàBA 0.007 ✓ H3 Middle x AHS 0.975 ✗ High x AHS 0.907 ✗ H4 BAàPE 0.000 ✓ 15
Additional Analysis
R2(coefficient of determination) • BA: 0.392 • PE: 0.655Path P-Value H accepted
Brand Affect -> Product Evaluation 0.000 ✓
Perceived Degree of Congruency -> Brand
Affect 0.000 ✓
Perceived Degree of Congruency -> Product
Evaluation 0.000 (complementary✓
mediation)
27/01/2021
Conclusion
• Positive relationship between degree of congruency between
symbolism in work of art (high); and product type
• No evidence for moderation through processing style
àArt infusion effect not context independent
Limitations & Future Research
• Sample did not contain highly analytical participants • AHS construct reliable, but not valid
o Measure not suitable? Specific facets of holism? • No pre-tests were run on images
o Perceived congruency did not cover intended range • Qualitative Research
17
Recommendations
• Further research on the art infusion effect o Different types of symbolism
o Cultural differences
o Positive vs negative emotions evoked through symbolism in art
o Select artworks for marketing purposes with care o Art infusion context dependent
o Connotations beyond literal depiction
o Congruency positively affects product evaluation
27/01/2021
CREDITS: This presentation template was created by Slidesgo, including icons by Flaticon, infographics & images by Freepik and
illustrations by Storyset
