The Influence of the Basket Composition
on Anticipated Embarrassment:
An investigation of the product prominence hypothesis
and the self-expression hypothesis as possible
explanations why products in addition to an
embarrassing product reduce embarrassment when
purchasing an embarrassing product
Author: Claudia Frey
Faculty of Economics and Business
Master Thesis – MSc Marketing Management
The Influence of the Basket Composition
on Anticipated Embarrassment:
An investigation of the product prominence hypothesis
and the self-expression hypothesis as possible
explanations why products in addition to an
embarrassing product reduce embarrassment when
purchasing an embarrassing product
Author: Claudia Frey Leopoldweg 1a CH-6210 Sursee Tel.: +41 79 848 49 33 frey.claudia@bluewin.ch Student Number: S3123065 Supervisor:
Dr. Martijn Keizer, Faculty of Economics and Business
ABSTRACT
Embarrassment is a widely a “widely occurring emotion” (Dahl, Manchanda, & Argo, 2001, p. 473) and especially can play a crucial role in consumer purchase context for embarrassing products like condoms or certain medication. Because embarrassment is a negative feeling people try to avoid it (Edelmann, 1987; Miller, 1987) and therefore use several coping strategies to reduce or minimize embarrassment (Blair & Roese, 2013; Nichols, Raska, & Flint, 2015). One such coping strategy is to buy products in addition to the embarrassing product in order to reduce embarrassment. The main goal of the present thesis was to find evidence for the self-expression hypothesis which tries to ex-plain why this coping strategy reduces embarrassment: The self-expression hypothesis proposes that products in addition to an embarrassing product lower embarrassment (i.e. the threat to the public identity) because of their identity and self-expression function that enhances the self-expressiveness of the whole basket and therefore reduces or coun-terbalances the threat to the public identity coming from the embarrassing product (i.e. lowers embarrassment). The self-expression hypothesis is a refinement of the basket
balance hypothesis from Blair & Roese (2013) who only investigated the
counterbal-ancing aspect of additional products with a very specific experiment. Thus, the basket balance hypothesis so far cannot explain for all additional products and purchase situa-tions why embarrassment decreases when products are purchased in addition to the em-barrassing product. With the aim to propose a refinement of the basket balance
hypoth-esis (i.e. the self-expression hypothhypoth-esis), it was also an aim of this thhypoth-esis to confirm the
results of Blair & Roese (2013) which claim that the product prominence hypothesis is not valid. The product prominence hypothesis states that additional products reduce embarrassment because they would make the embarrassing product less salient during purchase; meaning that the person thinks others would pay less attention to the embar-rassing product (Blair & Roese, 2013).
Hence, the main research question of this thesis investigated whether the identity and self-expression function of the whole basket (and not attention paid by others to the barrassing product) mediates the impact of the basket composition on anticipated em-barrassment.
products and also more specifically when buying brands as additional products that dif-fer in their identity and self-expression function. Additionally, the personality trait lic self-consciousness, which is an indicator for how much one cares about his/her pub-lic identity, was proposed as (direct and indirect) moderator of the impact of the basket composition on embarrassment through self-expressiveness of the basket.
ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to my supervisor Dr. Martijn Keizer. He introduced me into the interesting research area of embarrassment and gave me full freedom to develop a research question that I would like to investigate within this area,
yet providing very helpful guidance and comments throughout this first and important stage that builds now the basis for this thesis. I would also like to thank him for his al-ways highly valuable inputs and constructive remarks which helped me to improve my
thesis in every stage.
I would like to give my warmest thanks to my beloved sister who supported me not only throughout this master thesis but also throughout my study time and my whole life with
her unconditional love, support, and her fabulous sense of humour. Whenever I could not see the wood for the trees, you helped me with your advice and a good dose of
pragmatism. I will always be grateful!
My heartfelt thanks go to my dear parents who are there for me in every stage of my life, providing me with their continuous encouragement, advice, and support. Without
you, this accomplishment would not have been possible, thank you!
Furthermore, I would like to extend a big thank you to my friends here in Groningen and in Switzerland for their continuous motivation and all the fun moments we had
dur-ing assignments and outside of university. I learned a lot from you and am happy to have you as my friends.
Lastly, I would also like to express my appreciation to all the participants who shared their precious time and carefully filled in my surveys. Without you, this master thesis
CONTENTS
ABBREVIATIONS ... VIII FIGURES ... IX TABLES ... XI
1. Introduction ... 1
2. Literature Review and Hypothesis Development ... 6
2.1 Conceptualization of Embarrassment ... 6
2.1.1 Social Evaluation Theory ... 7
2.2 Hypothesis Development ... 8 2.2.1 Hypotheses 1a and 1b ... 8 2.2.2 Hypotheses 2a and 2b ... 13 2.2.3 Hypothesis 3 ... 15 2.2.4 Conceptual Model ... 17 3. Method ... 19
3.1 Design and Participants ... 19
3.2 Materials ... 20 3.2.1 Pre-Tests ... 20 3.2.1.1 First Pre-Test ... 20 3.2.1.2 Second Pre-Test ... 27 3.2.2 Independent Variable ... 29 3.2.3 Dependent Variable ... 29 3.2.4 Mediator Variable ... 30 3.2.5 Moderator Variable ... 31 3.2.6 Control Variables ... 31 3.3 Survey Design ... 33 3.4 Analysis Plan ... 35 4. Results ... 38 4.1 Hypotheses 1a and 1b ... 38 4.2 Hypotheses 2a and 2b ... 42 4.3 Hypothesis 3 ... 43
4.4 Analyses with Control Variables ... 46
5. Discussion... 50
5.1 Hypotheses 1a and 1b ... 50
5.2 Hypotheses 2a and 2b ... 54
5.3 Hypothesis 3 ... 55
5.4 Hypotheses with Control Variables and Supplementary Analyses ... 56
5.5 Theoretical Contribution ... 58
5.6 Managerial Implications ... 59
5.7 Limitations and Directions for Future Research ... 60
6. Conclusion ... 64
REFERENCES ... 65
IMAGE REFERENCES... 71
ABBREVIATIONS
PSC Public self-consciousness NBC National brand condition
CC Control condition (embarrassing product alone) SBC Store brand condition
M Mean
FIGURES
Figure 1. Explanation of the cause of embarrassment according to the Social Evaluation
Theory ... 7
Figure 2: Reasoning for the Self-Expression Hypothesis ... 12
Figure 3: Conceptual Model of proposed mediation in hypothesis 1b ... 13
Figure 4: Conceptual Model of proposed mediation in hypothesis 2b ... 15
Figure 5: Conceptual model and hypotheses ... 18
Figure 6: Product picture of the hemorrhoidal cream selected for the second pre-test and the main survey (Image: Walmart Stores (2017)) ... 22
Figure 7: Product pictures used for the national brand condition (Images: Jumbo Supermarkten (2017a)) ... 27
Figure 8: Product pictures used for the store brand (Jumbo) condition (Images: Jumbo Supermarkten (2017b)) ... 27
Figure 9: Chosen products for the three conditions of the main survey (Images: Walmart Stores (2017), Jumbo Supermarkten (2017a, b)) ... 29
Figure 10: Anticipated embarrassment and self-expression for the participants in the two basket compositions ... 39
Figure 11: Statistical diagram of the mediation analysis: Basket composition -> Self-expression -> Anticipated embarrassment ... 40
Figure 12: Statistical diagram of the mediation analysis: Basket composition -> Attention -> Anticipated embarrassment ... 41
Figure 13: Anticipated embarrassment and self-expression for the participants in the three conditions ... 43
Figure 14: Visualization of the conditional indirect effect in the moderated mediation 45 Figure 15: Statistical diagram of the mediation analysis: Basket composition -> Basket Liking -> Anticipated embarrassment ... 48
FIGURES APPENDIX Figure A. 1: Nationality and time living in NL for participants in the store brand condition ... 114
TABLES
Table 1: Explanation Basket Balance Hypothesis and Product Prominence Hypothesis . 9
TABLES APPENDIX
Table A. 1: Overview of data set preparation procedure ... 112
Table A. 2: Overview demographics ... 116
Table A. 3: Overview of assumption tests ... 118
Table A. 4: Comparison of ANOVA with Kruskal-Wallis ... 119
Table A. 5: Model coefficients for the mediation analysis with self-expression as mediator ... 122
Table A. 6: Model coefficients for the mediation analysis with attention as mediator 122 Table A. 7: Model coefficients for the mediation analysis with self-expressiveness as mediator ... 123
Table A. 8: Model coefficients for the moderated mediation analysis with self-expressiveness as mediator and PSC as moderator ... 124
Table A. 9: Conditional indirect effect(s) of the basket composition on anticipated embarrassment at the mean and plus/minus one SD from the mean of PSC ... 125
Table A. 10: Model coefficients for the mediation analysis with self-expression as mediator and the control variables ... 126
Table A. 11: Model coefficients for the mediation analysis with attention as mediator and the control variables ... 126
Table A. 12: Model coefficients for the moderated mediation analysis with self-expressiveness as mediator, PSC as moderator, and the control variables ... 127
Table A. 13: Model coefficients for the regression analysis with all control variables (except price as it did not have an influence on anticipated embarrassment) ... 128
Table A. 14: Model coeff. for the regression analyses of three models showing that the effect of the basket composition turns insignificant when inserting basket liking in the model, but not when inserting gender in the model ... 128
Table A. 15: Model coefficients for the mediation analysis with liking as mediator ... 129
1. Introduction
Embarrassment is a “widely occurring emotion” (Dahl et al., 2001, p. 473) and, based on the Social Evaluation Theory, can be defined as “[…] the negative, self-conscious feeling that arises when an event [is perceived to threaten] one’s [public] identity” (Miller, 1995; Puntoni, de Hooge, & Verbeke, 2015, p. 72).1 Public identity is the con-ception which others have of who or what another person is (Cheek & Briggs, 1982; Modigliani, 1968) and includes “overt externally observable aspects such as physical appearance” (Edelmann, 1985, p. 223) as well as “the set of traits and characteristics portrayed to others in a given situation“ (Blair & Roese, 2013, p. 677). 2 It is important to notice that it is not the public identity itself that is the cause for embarrassment, but it is the self-perception of the public identity, called the subjective-public-identity (Modigliani, 1968). Manifold circumstances can lead to a perceived threat to one’s pub-lic identity and thus to embarrassment. Especially in the consumer context embarrass-ment can occur during purchase, which is the focus of this study, but also during usage, and/or disposal of certain products (Krishna, Herd, & Aydinoğlu, 2015; Lau-Gesk & Drolet, 2008). Products that can lead to embarrassment during purchase – in this work called embarrassing products – are for example condoms, douches, adult magazines, adult undergarments, feminine hygiene products, or medication (Kumar, 2008; Walsh, Albrecht, Hofacker, Grant, & Takahashi, 2016). 3
People might avoid buying such embarrassing products or might avoid making use of certain services that could elicit embarrassment because embarrassment is a negative feeling with an uncomfortable psychological state, and people therefore try to avoid it (Edelmann, 1987; Miller, 1987). However, not buying certain required embarrassing products or services can have negative consequences (Krishna et al., 2015; Lau-Gesk & Drolet, 2008; McCambridge & Consedine, 2014; Puntoni et al., 2015): People might avoid health care or certain preventive products (e.g. condoms) what can seriously
1 Embarrassment should be differentiated from shame “even though common English usage does no
al-ways distinguish between” those two (Walsh et al., 2016, p. 1139). However, they are “considered to be distinct psychological phenomena” (Walsh et al., 2016, p. 1139). Modigliani (1966, p. 10) states that “[…] one is primarily ashamed of oneself, while one is primarily embarrassed about one’s presented self.” This is consistent with the definition of embarrassment used in this work.
2 Please not that some literature use the term “social identity” instead of “public identity” (e.g. Miller
(1987), Sabini et al. (2000)). However, these two terms are used interchangeably. To be consistent, only the term “public identity” will be used throughout this work.
threaten people’s health. Another implication is that sales for companies can be nega-tively affected because of people not purchasing certain products that elicit embarrass-ment (Krishna et al., 2015). Interestingly, embarrassembarrass-ment in a consumer purchasing con-text has received sparse attention so far (Krishna et al., 2015, p. 474) and thus, only lit-tle guidance exists on how consumers can reduce embarrassment during the purchase of embarrassing products. For this reason and because of the above mentioned negative implications of embarrassment and its unpleasant feeling, it is important to investigate further how consumers can reduce embarrassment when purchasing an embarrassing product. Hence, the problem statement for the present thesis is:
How can consumers prevent or reduce embarrassment when purchasing an embarrass-ing product?
The sparse existing research (e.g. Blair & Roese, 2013; Brackett, 2004; Dahl et al., 2001; Nichols, Raska, & Flint, 2015) on the topic of embarrassment in a purchasing context shows that when people already anticipate embarrassment, they use a range of coping behaviours to reduce or minimize embarrassment during the purchase of an em-barrassing product. These coping behaviours, or coping strategies, “protect and, if nec-essary, restore the desired public identity” (Blair & Roese, 2013, p. 677) during pur-chase and therefore reduce embarrassment. According to the Social Evaluation Theory this means that coping strategies help to reduce or minimize the perceived threat to the public identity (i.e. embarrassment) caused by the embarrassing product(s). Such coping strategies are for example shopping in stores with only few people present to reduce the likelihood to be observed during purchase, avoiding asking store employees for help, waiting until nobody is at the store aisle when taking the product out of the shelf, or purchasing additional non-embarrassing products4 (Blair & Roese, 2013; Brackett, 2004; Nichols et al., 2015).
The present work will focus on the last named coping strategy, i.e. to buy additional products to reduce the perceived threat to the public identity (i.e. lower embarrassment) when purchasing an embarrassing product. This thesis will focus on this coping strategy because the reason why this coping strategy reduces the perceived threat to the public identity is until now insufficiently investigated. Traditionally it is stated that additional
4 If in the following the following the term “additional products/s” is used, it always means “additional
product(s) would make the embarrassing product less salient, i.e. the person supposes that other people will be less likely to see the embarrassing product, what would de-crease embarrassment (Blair & Roese, 2013; Brackett, 2004; Nichols et al., 2015). Thus, the perceived threat to the public identity (i.e. embarrassment) would be reduced “because attention will be spread across multiple products rather than focused solely on the embarrassing product” (Blair & Roese, 2013, p. 676). This explanation is called the
product prominence hypothesis (Blair & Roese, 2013). However, Blair & Roese' (2013,
p. 676) recent results contradict the traditional product prominence hypothesis because they showed that additional purchases reduce the perceived threat to the public identity (i.e. embarrassment) during the purchase of an embarrassing product not because of reduced salience of the embarrassing product, but because , and only if, the additional products are “perceived to counterbalance […] the undesired identity communicated“ by the embarrassing product. Thus, it is the counterbalance perception of the person itself that mediates the impact of additional products on embarrassment (Blair & Roese, 2013): The more he/she perceives the product to counterbalance the undesired identity of the embarrassing product, the less he/she will be embarrassed. This approach is called the basket balance hypothesis (Blair & Roese, 2013).
However, because Blair & Roese (2013) came to the basket balance hypothesis with a rather specific experiment5, the basket balance hypothesis needs further support and refinement that specifies clearly how additional products are able to counterbalance the undesired identity during purchase (i.e. lower the perceived threat to the public identi-ty). For this reason and because it is important to know for retailers and brand managers why additional products can lower embarrassment, it is the main goal of this thesis to develop a refinement of the basket balance hypothesis that can explain in all circum-stances and for all products why products purchased in addition to an embarrassing product lower embarrassment. As a refinement of the basket balance hypothesis, the present work proposes the expression hypothesis as possible explanation. The
self-expression hypothesis suggests that additional non-embarrassing products can
counter-balance the embarrassing product because the identity and self-expression function of the additional products enhances the identity of the whole basket and therefore reduces
5 The experiment involved two directly counterbalancing products, i.e. the Complete Idiot’s Guide
(asso-ciated with unintelligent persons) and Rubik’s Cube (asso(asso-ciated with intelligent persons), see 2.2.1
the perceived threat to the public identity, i.e. embarrassment. With the proposition of the self-expression hypothesis, it is assumed that the product prominence hypothesis is not valid like Blair & Roese (2013) already claimed. However, to be sure about another possible explanation why additional products lower embarrassment, the product
promi-nence hypothesis will also be investigated in this study. Therefore, the main research
question is:
When purchasing products in addition to an embarrassing product, is it the identity and self-expression function of the additional products that lowers the perceived threat to the public identity, i.e. embarrassment – and not the reduced salience of the embarrass-ing product?
Based on the main research question, two sub-questions will be investigated. First, it will be investigated whether the self-expression hypothesis also holds in a more specific context. More specifically, it will be investigated whether national brands with their stronger brand personality than store brands (Beldona & Wysong, 2007) and with their self-expressive function (Cătălin & Andreea, 2014) are better able than store brands to counterbalance the perceived threat to the public identity coming from the embarrassing product (i.e. lower embarrassment). Second, it will be investigated whether the person-ality trait public self-consciousness moderates the impact of the basket composition (i.e.
embarrassing product alone or embarrassing product with additional products) and its
identity and self-expression function on embarrassment. Public self-consciousness is an indicator for how much one cares about his/her public identity. Thus, people high in public self-consciousness put high value on how others see them (i.e. put high value on their public identity) (Edelmann, 1985). Consequently the higher a person is in public self-consciousness the more value he/she should put on the identity and self-expression function of the basket what should raise the impact of the basket composition on embar-rassment.
Moreover, even though the extent to which participants thought others would pay to attention the embarrassing product during purchase was not influenced by the basket composition; it had a strong influence on anticipated embarrassment.
The results of the research at hand have high empirical value as they add to the literature about embarrassment and contribute to the existing but sparse knowledge about why and how consumers can reduce their embarrassment when purchasing an embarrassing product. The results are especially useful because the “shopping basket composition is fully controllable by the consumer once he or she is in the store” (Blair & Roese, 2013, p. 688). Additionally, the results can not only give guidance for consumers, but also for marketers. They could give a good advice for potential bundling strategies for marketers (Blair & Roese, 2013). Ultimately, the application of the results can contribute to con-sumer well-being, concon-sumers might feel less embarrassed when buying an embarrassing product and might buy products they would have otherwise be too embarrassed to buy. Consequently this can positively affect the sales of the respective companies.
2. Literature Review and Hypothesis Development
Because embarrassment is the central construct in the present study, the following sec-tion introduces the conceptualizasec-tion of embarrassment and the Social Evaluasec-tion
Theo-ry which is used to explain the causes of embarrassment in this thesis. The latter is
im-portant as the present work and the developed refinement of the basket balance
hypoth-esis (i.e. the self-expression hypothhypoth-esis) is based on this explanation.
2.1 Conceptualization of Embarrassment
Embarrassment belongs to the self-conscious emotions which “play a central role in motivating and regulating people’s thoughts, feelings and behaviours” (Tracy & Robins, 2007, p. 190). It can be seen as a form of social anxiety (Edelmann, 1987) because it is the feeling resulting from the perception that oneself is being evaluated inappropriately by other people (Miller, 1996). People experiencing embarrassment feel foolish, inept, awkward, and uncomfortable and it is not uncommon that they blush (Miller, 1996). As response, they explain themselves, make a joke, try to ignore the predicament or try to escape the situation (Miller, 1996).
The literature about embarrassment dates back to Goffman (1956) and goes further to other major works from Modigliani (1968), Edelmann (1987), Parrott, Sabini, & Silver (1988) and Miller (1996). Since Goffman, different explanations for the causes of barrassment have been proposed. The many different circumstances that can cause em-barrassment like acting clumsy, committing a faux-pas, being the centre of attention, or having to buy certain products have contributed to the fact that today, generally three theories for explaining the causes of embarrassment do exist (Kumar, 2008; Tangney, 1999) 6. Thus, the theories can mainly be used in different circumstances to explain the emerged embarrassment, but the three theories can also interact with each other (Edelmann, 1987; Miller, 1995; Sabini, Withers & Sherblom, 2008; Withers & Vernon, 2006).
In this work, I am particularly interested in embarrassment caused by the purchase of certain products like condoms. Therefore, I will focus on the theory that is most suitable
6 See Miller (1996) for a detailed review and comparison of the three accounts: Situational Self-esteem
Theory (embarrassment caused by a loss of situational self-esteem), Dramaturgic Model (embarrassment
for explaining the emergence of embarrassment in these situations. The theory that is most suitable in explaining embarrassment that may occur when having to buy embar-rassing products like condoms is the Social Evaluation Theory. 7 This theory will be
introduced in the following. 2.1.1 Social Evaluation Theory
The core of this theory is the “concern for others’ opinions” (Miller, 1995, p. 324). Em-barrassment results “when unwanted public events create acute concern for how one is being evaluated by others” (Miller, 1995, p. 323). Thus, one perceives his public identi-ty, the conception which others have of who or what the person is (Cheek & Briggs, 1982; Modigliani, 1968), to be threatened (Miller, 1995). This self-perception of the public identity, is called subjective-public-identity (Modigliani, 1968). Thus, a certain self-perception, or more specifically, a certain subjective-public identity is the cause for embarrassment, and not the perception that others have about the person’s identity, see
Figure 1. 8
Figure 1. Explanation of the cause of embarrassment according to the Social
Evaluation Theory
However, according to the conceptualization of Miller (1995), besides the specific event also two other conditions must be fulfilled for embarrassment to occur: (1) One must be able to understand others’ judgments. Therefore, “mature” embarrassment does not oc-cur until the age of 11 years (Miller, 1996). (2) Besides the ability to comprehend oth-ers’ judgments about oneself, one must also care about these judgments. A person who
7 The Social Evaluation Theory is generally also the favorite theory among researchers (Miller, 1995, 1996; Withers & Vernon, 2006) because it is able to provide an explanation for almost all events that people might find embarrassing (Miller, 1996).
8 Please note that some researchers, e.g. Miller (1995) and Kumar (2008), talk about the desired public
simply doesn’t care about others evaluations should not be embarrassed. Thus, embar-rassment occurs in any situation, in which one’s desired identity is perceived to be threatened (Withers & Vernon, 2006), and in which the person to the same time cares about (and understands) his/her public identity.
The following definition of embarrassment summarizes the aspects mentioned above and will be used in this work:
“Embarrassment is the negative, self-conscious feeling that arises when an event [is perceived to threaten] one’s [public] identity.” (Miller, 1995; Puntoni et al., 2015,
p. 72)
With the Social Evaluation Theory, it can be explained very well why for example hav-ing to buy condoms, or woman’s hygiene products can lead to embarrassment (Kumar, 2008): The purchase of such products leads to embarrassment because one perceives the purchase of those products as not consistent with his/her own identity and therefore he perceives the purchase of such products as a threat to his public identity, the conception which others have of a person’s identity (Modigliani, 1968). This leads consequently to the feeling of embarrassment concerning the Social Evaluation Theory (Miller, 1995). Lastly, it is important to mention that the Social Evaluation Theory in existing research sometimes gets mixed with a theory that explains embarrassment as a loss of situational self-esteem (Modigliani, 1968)9. However, as research failed to find a relationship be-tween embarrassment and situational-self-esteem (e.g. Modigliani, 1968; Parrott et al., 1988), the Social Evaluation Theory should be treated in its original form as previously described.
2.2 Hypothesis Development
In the following, based on the presented theory about embarrassment and based on addi-tional theory, five hypotheses will be developed in the following.
2.2.1 Hypotheses 1a and 1b
The fact that purchasing non-embarrassing products in addition to an embarrassing product can lower anticipated embarrassment during purchase was already repeatedly shown (Blair & Roese, 2013; Brackett, 2004; Nichols et al., 2015). However, as shown
in chapter 1. Introduction, the recently developed basket balance hypothesis by Blair & Roese (2013) contradicts the traditional product prominence hypothesis as explanation why additional non-embarrassing products help to reduce the perceived threat to the public identity (i.e. embarrassment) when purchasing an embarrassing product (Blair & Roese, 2013) (see Table 1) 10. Also Gestalt theory which states that “impressions of in-dividuals form not as discrete collections of associated knowledge but rather as holistic integrations (i.e., a gestalt) in which observed behaviors and traits interact dynamically” (Blair & Roese, 2013, p. 678) does not support the product prominence hypothesis as the product prominence hypothesis claims that embarrassment is only determined by the embarrassing product and its salience (and not by the whole basket like the product
prominence hypothesis suggests).
Table 1: Explanation Basket Balance Hypothesis and Product Prominence
Hypothesis
However, to definitively rule out the product prominence hypothesis as explanation for reduced embarrassment when purchasing additional products, the basket balance
hy-pothesis needs further support and refinement. This also because Blair & Roese (2013)
derived this hypothesis from an experiment where they only measured the effect on an-ticipated embarrassment of purchases that directly counterbalance the embarrassing product, i.e. “an embarrassing book entitled The Complete Idiot’s Guide to Improving
Your IQ” and two products associated with intelligence counterbalancing the inference
that the person is dumb (Blair & Roese, 2013). But in the real world, a person might rarely purchase counterbalancing products (e.g. unintelligent “The Complete Idiot’s
Guide” – intelligent Rubik’s cube, unhealthy eater Pizza – healthy eater Meal Salad) or
for most embarrassing products no directly counterbalancing product is available (e.g. condoms, medications). However, existing research already showed that various addi-tional to the embarrassing product unrelated products reduce embarrassment during
10 To the author’s knowledge, Blair & Roese (2013) are the first ones who claim that the traditional
prod-uct prominence hypothesis is not valid and who suggest instead the basket balance hypothesis to explain
why purchasing non-embarrassing products in addition to an embarrassing product can lower the per-ceived threat to the public identity (i.e. embarrassment).
Basket Balance Hypothesis1 Product Prominence Hypothesis1
Explanation for the fact that additional products reduce the perceived threat to the public identity (i.e. lower embarrassment) when purchasing an embarrassing product:
Additional products are “perceived to counterbalance […] the undesired identity communicated” (Blair & Roese, 2013, p. 676) by the embarrassing product.
Additional products reduce the salience of the embarrassing product.
chase (e.g. Blair & Roese, 2013; Brackett, 2004; Nichols et al., 2015). Therefore, the idea of the basket balance hypothesis should be verified in a broader context and with additional products that are not directly related to the embarrassing product. This is the main goal of this thesis (and goes hand in hand with an attempt to definitively reject the
product prominence hypothesis).
Remember, the main idea of the basket balance hypothesis is that additional products lower the perceived threat to the public identity (i.e. embarrassment) when purchasing an embarrassing product because they are “perceived to counterbalance […] the unde-sired identity” (Blair & Roese, 2013, p. 676) communicated by the embarrassing prod-uct. Thus, the basket balance hypothesis explains the resulting embarrassment as a func-tion of the identity of the whole basket which is signalled or made up by all the products in the basket (Blair & Roese, 2013). This implies that when purchasing embarrassing products like condoms or woman’s hygiene, a solution to minimize or decrease the per-ceived threat to one’s public identity is to purchase additional products because they would make up for the embarrassing product in that they will be additional (positive) indicators to form the impression of that person (i.e. public-identity) besides the nega-tive indication from the embarrassing product11. This implies that products must have a certain identity and self-expression function which is also confirmed by Gao, Wheeler, & Shiv (2009, p. 31): „Consumption of products can convey the possession of specific self-characteristics.“ That people make use of this identity and self-expression function of consumption is nowadays well-known: “How people make decisions around con-sumption are influenced by their perceptions of themselves and how they would like to be seen” (Szmigin & Piacentini, 2015, p. 233). This is because “people have a pervasive desire to be viewed in a positive light” and therefore regularly engage in impression management (Argo, Dahl, & Manchanda, 2005, p. 208).
With this reasoning, the idea of the basket balance hypothesis, that additional products can counterbalance the perceived threat to the public identity coming from the rassing product, could be refined with saying that they can counterbalance the embar-rassing product because of their identity and self-expression function enhancing the identity of the whole basket. This explanation would also hold for the reduced
11 Remember that embarrassment is coming from a subjective perception of how others perceive the
rassment found in the study of Blair & Roese (2013): A person who buys the Complete
Idiot’s Guide possibly cannot or does not want to identify with this book, and especially
does not want to show this symbolic meaning to the public. The additional products associated with intelligence enable the person to express other (desired) characteristics of him/herself and thus, embarrassment is lower.12 Another example is that a man order-ing a “ladies’ cut steak” with a pint of the manliest beer on the menu will feel less threatened regarding his identity than when he would only order the “ladies’ cut steak” (Blair & Roese, 2013). Thus, the pint of the manliest beer enhances the identity and self-expression function of the whole basket and therefore reduces the perceived threat to the public identity respectively embarrassment. However, not all products will have the same identity and self-expression function13, and this will also not be the same for every person. If the man would have ordered the “ladies’ cut steak” with a coke for ex-ample, he might feel not so much less threatened regarding his identity than when he additionally orders the pint of the manliest beer with which he can identify or express his (desired)identity better. Therefore, the more someone can identify and express him-self with the whole basket, the less he should feel embarrassed. This identity and him- self-expression function should be higher for a basket containing an embarrassing product and additional non-embarrassing products because the embarrassing product should have a low identification and self-expression function and the additional products add positively to the basket identity and can make up for the negative characteristics sig-nalled by the embarrassing product. Thus they can compensate to a certain extent the perceived threat to the public identity coming from the embarrassing product, and thus embarrassment should be lower for the basket with additional products (see Figure 2).
12 It is assumed here that if a person can identify with a product, he/she also wants to show/express this
characteristic to others.
13 This could also explain why the additional counterbalancing product in the study of Nichols et al.
Figure 2: Reasoning for the Self-Expression Hypothesis
This means that the reduction of purchase embarrassment when purchasing non-embarrassing products in addition to an non-embarrassing product should be a function of the extent to which one can identify and express his/her (desired) identity with the whole basket. Or differently formulated: The identity and self-expression function of the basket should mediate the impact of the basket composition on anticipated embar-rassment (see the conceptual model of the mediation in Figure 3). This theory as expla-nation why additional products lower anticipated embarrassment is in the following named self-expression hypothesis. This proposition goes hand in hand with the sugges-tion, that the attention that people think others would pay to the embarrassing product during purchase does not mediate the impact of the basket composition on anticipated embarrassment. Because the expected attention that others would pay to the embarrass-ing product should not be a mediator, attention14 is intentionally not included in the conceptual model. However, it is included in the hypothesis because it will also be test-ed as a mtest-ediator between the basket composition and anticipattest-ed embarrassment in or-der to definitively rule out the product prominence hypothesis as alternative explanation for reduced embarrassment when purchasing products in addition to the embarrassing product.
14 Please note that term “attention” relating to the product prominence hypothesis means throughout the
Figure 3: Conceptual Model of proposed mediation in hypothesis 1b15
Concluding, like previous research (Blair & Roese, 2013; Brackett, 2004; Nichols et al., 2015) already did, this thesis proposes that purchasing additional non-embarrassing products in addition to an embarrassing product lowers anticipated16 embarrassment. Hypothesis 1b therefore is:
H1a: A basket containing an embarrassing product and additional other non-embarrassing products leads to lower anticipated embarrassment than a basket containing only an embarrassing product.
As a possible reason for that, hypothesis 1b proposes the self-expression hypothesis:
H1b: The identity and self-expression function of the basket, but not the attention of others to the embarrassing product, mediates the impact of additional purchases on anticipated embarrassment.
2.2.2 Hypotheses 2a and 2b
Based on the main idea of the basket balance hypothesis that additional products are able to counterbalance the perceived threat to the public identity coming from the em-barrassing product and based on the self-expression hypothesis, developed in the previ-ous section, that this is due to the enhanced identity and self-expression function of the whole basket, in the following a refinement of the coping strategy in which people buy products in addition to the embarrassing product in order to lower embarrassment is propsed.
15 Please note that if self-expression is stated, also the identity function is meant. However, for a better
reading fluency, the term identity (function) will mostly be omitted.
16 The term “anticipated” is used because the hypotheses will be tested with an online questionnaire and
The refinement proposes to buy specific brands in order to enhance the identity and self-expression function of the basket even more because it is commonly known that brands can have a symbolic or identity function and are means of self-expression and are also chosen for these reasons (e.g. Cătălin & Andreea, 2014; Chernev, Hamilton, & Gal, 2011; Gao et al., 2009; Walsh et al., 2016). Cătălin & Andreea (2014, p. 103) provide convincing evidence that “[…] customers rely on brands with a desirable brand identity to express their own identity and they tend to prefer brands that are convergent with their identity […].“ The self-expression function of brands can especially be seen in conspicuous consumption where persons acquire specific brands mainly for showing social status or wealth (Chernev et al., 2011), for example when purchasing green more expensive brands to signal status (Griskevicius, Tybur, & Van den Bergh, 2010). Thus, specific brands could help to signal a specific public identity (i.e. identity and self-expression function) and therefore could especially strongly decrease the weight of the perceived threat of the embarrassing product to the public identity. This would result in even more decreased embarrassment. Also Nichols et al. (2015, p. 51) state: “[…] if the self-view is “shaken” by needing to buy something that causes embarrassment, one may purchase certain products or brands that [especially] help reinforce their public identi-ty.” However, some brands might be more suitable for an identity and self-expression function and others less. Therefore, in the following section it will be shown which brands could be used for the proposed refinement to purchase specific brands as addi-tional products in order to lower embarrassment when purchasing an embarrassing product.
able to make one’s public identity more desirable and therefore are able to stronger counterbalance the perceived threat to the public identity coming from the embarrassing product and thus, weaken the felt embarrassment more. Consequently, a basket with no additional products should lead to the smallest identity and self-expression function and thus to the highest embarrassment. Hypothesis 2a therefore is:
H2a: A basket containing an embarrassing product and additional products that are store brands leads to higher anticipated embarrassment than a basket containing an embarrassing product and additional products that are national brands, but to lower anticipated embarrassment than a basket containing only an embarrassing product.
As a reason for this, like hypothesis 1b, hypothesis 2b also proposes the self-expression hypothesis as an explanation and states that the impact of the basket compositions on anticipated embarrassment is mediated by the identity and self-expression function of the whole basket (see Figure 4). Hypothesis 2a therefore is:
H2b: The identity and self-expression function of the basket mediates the impact of the basket on anticipated embarrassment.
Figure 4: Conceptual Model of proposed mediation in hypothesis 2b
2.2.3 Hypothesis 3
In addition to the previously proposed hypotheses, the thesis proposes that a moderator influences the relationship of the basket composition on anticipated embarrassment through self-expression. The proposed moderator is public self-consciousness. In the following, the reasons for this and the derived hypothesis will be presented.
the ability of public self-consciousness is required in order for embarrassment to occur (Miller, 1995). This is also the reason why embarrassment is called a self-conscious emotion, and does not belong to the basic emotions like fear or anger do (Tracy & Robins, 2004). Public self-consciousness is one major dimension of self-consciousness (Fenigstein, Scheier, & Buss, 1975). Self-consciousness is a trait (and thus varies be-tween persons) and describes the “consistent tendency of persons to direct attention in-ward or outin-ward” (Fenigstein et al., 1975, p. 522). Public self-consciousness describes the “outward-component” of self-consciousness and is associated with the concern about one’s public identity (Blair & Roese, 2013; Tracy & Robins, 2004), or differently said, a “general awareness of the self as a social object that has an effect on others“ (Fenigstein et al., 1975, p. 523). Private self-consciousness describes the “inward-component” and is a general tendency to attend „to one's inner thoughts and feelings“(Fenigstein et al., 1975, p. 523). However, as previously stated, for embar-rassment to occur, especially the ability of public self-consciousness is required (Miller, 1995) because embarrassment was defined as “the negative, self-conscious feeling that arises when an event [is perceived to threaten] one’s [public] identity.” (Miller, 1995; Puntoni et al., 2015, p. 72).
self-consciousness moderated the effects of brand label [(national brand vs bargain brand products)] on brand preference“ (Bushman, 1993, pp. 859–860).
Concluding, the trait public self-consciousness, which is not influenced by a certain situation but is a general characteristic of a person, will moderate the impact of the bas-ket’s identity-and self-expression function on anticipated embarrassment. Because pub-lic self-consciousness is a personality trait, it should affect as hypothesized, the indirect effect of the basket composition on anticipated embarrassment through the mediator (i.e. self-expression), as well as the direct effect of the basket composition on anticipat-ed embarrassment. For the conceptual model see Figure 5, chapter 2.2.4 Conceptual
Model. Hypothesis 3 therefore is:
H3: Public self-consciousness moderates the impact of the basket on embarrassment directly and indirectly through the basket’s identity-and self-expression function.
2.2.4 Conceptual Model
Figure 5 below shows the full conceptual model based on the hypotheses developed in
Figure 5: Conceptual model and hypotheses17
17 Please note that the conditions are marked in certain tables in the appendix as following: Control
3. Method
For testing the proposed relationships as shown in the conceptual model in Figure 5 above, an experimental study using an online survey was conducted.
In the following, information will be given about the design of the study, the partici-pants, the used materials (i.e. scales), the exact procedure of the study, and the methods of analysis for the hypotheses.
3.1 Design and Participants
Based on the conceptual model and in order to obtain the basket composition variants 1 and 2 (see Figure 5) which are required for testing the hypotheses, a 1 (embarrassing product) x 3 (national brand products, store brand products, no additional products)
Be-tween Subjects Design was used. This design leaded to three conditions for the
inde-pendent variable, i.e. the basket composition:
1. Control condition: Embarrassing product alone
2. National brand condition: Embarrassing product + National brand products 3. Store brand condition: Embarrassing product + Store brand products
The participants were randomly assigned to the three conditions in order to ensure that differences across participants were evenly distributed across the conditions (Aronson, Wilson, & Brewer, 1998). Participants were privately recruited through Facebook and with sending private messages containing the link to the survey. They were told before-hand, that they have the chance to win EUR 30 if they indicate their e-mail address at the end of the survey18. Because Dutch store brands were included in the survey (see
3.2.1.1 First Pre-Test), a requirement for the selection of persons to which the survey
was sent was that the persons are currently living, or have lived, in the Netherlands (Dutch or international people) and thus are most probably familiar with the Dutch store brand used in the survey.
173 persons took part in the online survey. However, 18 respondents had to be deleted from the data set because of several criteria used during the procedure of preparing the data set for the analyses (see appendix E). Finally, 155 participants were left in the data set for the analyses, 49 in the control condition, 53 in the national brand condition, and
18 The winner was draw by chance on the 22nd May 2017. He was informed and also the other participants
53 in the store brand condition. Of the participants, 96 were female (61.9%) and 59 were male (38.1%). The average age was 25 years (SD = 5.755) with the youngest per-son being 18 years and the oldest being 53 years. 103 participants were Dutch (66.5%) and 52 were not Dutch (33.5%). 94 participants were students, 26 were studying and working and 35 were full-time working19.
3.2 Materials
As seen in the previous section, the basket composition (i.e. independent variable) is obtained in combining an embarrassing product with either no additional product, na-tional brand products, or store brand products. Therefore, an embarrassing product as well as national brands and a store brand that meet specific criteria had to be selected for the main survey. Thus, in the following, it will be shown how the specific embar-rassing product and the national brands and store brand have been selected based on two pre-tests. After presenting the results of these two pre-tests, for each variable (independ-ent, depend(independ-ent, mediator, moderator, and control variables) the used materials will be presented.
3.2.1 Pre-Tests 3.2.1.1 First Pre-Test
The first pre-test had the aim to investigate which embarrassing product, which national brands and which store brand should be chosen for the main survey. The pre-test was given to students in the Plaza at the University of Groningen and to some people of my dancing association. Participants were asked to fill out a short paper or online question-naire. The two versions had been made to make it as pleasant as possible for the partici-pants to fill out the survey. The questions were in both versions the same. In total, 40 persons (66.7% female with one person failing to report gender, average age M= 24, SD = 4.101) filled out the questionnaire; 26 the paper questionnaire and 14 the online ques-tionnaire20. The questionnaire (see appendix A) had two main parts:
19 A table with all the collected demographic information overall and for each condition can be found in
appendix F.
1st Part: Embarrassing product
For the experimental design it was crucial that the embarrassing product used in the main survey really evokes a feeling of (anticipated) embarrassment. Previous research (Blair & Roese, 2013; Dahl et al., 2001; Kumar, 2008; Lau-Gesk & Drolet, 2008; Nichols et al., 2015) commonly found the following eight products to be embarrassing: condoms, anti-gas-medicine (against flatulence), hemorrhoidal cream, adult magazine, cream against foot fungus, anti-odour foot spray, personal lubricant, and anti-diarrheal medicine.21 Therefore, these eight products have been included in the pre-test. In the pre-test only the product-names were shown in order to prevent a possible influence of the product-design on anticipated embarrassment.
For each product, participants were asked to rate how embarrassed they would feel buy-ing the product at a local store (1 = not at all embarrassed, 7 = extremely embarrassed, and in case people didn’t know what the product is “I don’t know what this is”) (over all products: M= 3.67, SD= 1.380). This question is slightly adapted from one of the 3-items of the 7-point embarrassment scale of Dahl et al. (2001) which was also used in the main survey. Only the item of the scale which asks directly for the felt (anticipated) embarrassment (and not all three items) was used in order to make the pre-test as short as possible for participants. This procedure is justified with the high alpha (α = .88) that Dahl et al. (2001) indicated for the embarrassment scale of these three items. This means that the internal consistency reliability for the scale is high (Malhotra, 2010), i.e. the reliability is high that the three items measure very well the same factor or construct (Voerman, 2016), and thus it is appropriate to only use one of the items.
For getting an even clearer picture of which is the most embarrassing product, partici-pants had to rank the eight products from 1 to 8 according to how embarrassed they would feel buying each of the products at a local store, with 1 for the most embarrassing and 8 for the least embarrassing product.
The results showed that among participants the hemorrhoidal cream was perceived as the most embarrassing product to purchase at a local store (M= 4.71, SD = 1.883), fol-lowed by adult magazine (M = 4.59, SD = 2.245) and anti-gas medicine (M = 4.05, SD = 2.026). Also the agreement between the participants was highest, which can be seen in
21 Products that are specifically for woman (e.g. woman’s hygiene) or that are tailored to men and women
the lowest standard deviation. Also in the ranking, the hemorrhoidal cream was seen as the most embarrassing product (M = 3.03, SD = 1.905 with 1 = most embarrassing and 8 = least embarrassing), followed by the adult magazine (M = 3.46, SD = 2.522) and anti-gas medicine (M = 3.64, SD = 1.755). However, a paired samples t-test showed that embarrassment for the hemorrhoidal cream was not significantly different from the adult magazine, t(33)= 0.796, p= .432, but it did differ significantly from the anti-gas medicine, t(32)= 2.248, p= .032*. Even though the difference between the hemorrhoidal cream and the adult magazine was not significant, the hemorrhoidal cream was selected because only 11.8% indicated to not be embarrassed at all when purchasing hemorrhoi-dal cream which was the lowest value across all eight products and it was very im-portant for the main survey to take a product that really evokes embarrassment. For adult magazine this value was higher with 20.5% of all participants indicating to not be embarrassed at all. Moreover, in contrast to the adult magazine which is a hedonic product, hemorrhoidal cream is a utilitarian product. Thus, health circumstances (i.e. hemorrhoids) might require buying the product. Also a gender-effect did not seem to exist for the hemorrhoidal cream.22 This was congruent with the findings of Nichols et al. (2015) who found no gender difference for hemorrhoidal cream.
The product picture of hemorrhoidal cream below in Figure 6 was selected for the second pre-test (see 3.2.1.2 Second Pre-Test) and the main survey (see 3.2.2
Independent Variable). A criterion for the selected product package was that the product
purpose is clearly visible on the package. Additionally, the logo was erased in order to prevent a possible influence of the brand.
Figure 6: Product picture of the hemorrhoidal cream selected for the second pre-test and the main survey (Image: Walmart Stores (2017))
22 Female: M= 4.41 (SD= 2.197), Male: M= 5.25 (SD= 0.965). Because the homogeneity of variance was
2nd Part: Additional products in the basket - National brands and store brands
For the national brand and the store brand products, groceries served as the additional products because Nichols et al. (2015) found that most commonly people buy groceries as additional products to mask the embarrassing product. It was decided to take 2 addi-tional products (and not more) for each condition – naaddi-tional brand condition and store brand condition – in order to better control for possible interactions that could exist be-tween the different products in the basket. Taking only two additional products is sup-ported by Nichols et al. (2015, p. 47) who state that it is possible “that people feel that only one or two extra items are needed to [reduce embarrassment]“. Moreover, existing research mostly only uses one or two additional products for evaluating the effect of additional products in laboratory experiments (e.g. Blair & Roese, 2013). Thus, choos-ing two products as additional products is reasonable. The two products in each of the two brand conditions had to fulfil specific criteria regarding the brand(s):
For the store brand, the requirement was that it has a lower identity and self-expression function than the national brand but nevertheless that it should not have a strong nega-tive image and thus should not signal an undesired identity. This was important in order to prevent an additional factor that could increase anticipated embarrassment additional-ly to the embarrassment caused by the embarrassing product. Rather people should view the store brand as being as neutral as possible (no positive or negative image or feelings for the brand). Like that, it was better possible in the main study to interpret the effect of the national brand products (vs store brand products) on anticipated embarrassment. Jumbo, Albert Heijn, and Coop were included in the pre-test for the store brands be-cause those are one of the biggest retailers in the Netherlands (Pinckaers & Phillips, 2016) and should be most commonly known in Groningen as the survey was mainly distributed to people living in Groningen. Plus and Dirk would also belong to biggest retailers in the Netherlands (Pinckaers & Phillips, 2016) but they have no (Dirk) or less (Plus) stores in Groningen than Coop. Moreover, Lidl and Aldi also belonging to the biggest retailers in the Netherlands (Pinckaers & Phillips, 2016) were not used for the pre-test because according to Zielke (2014, p. 329), shopping at discount stores is often related to shame and guilt because “discounters might violate several standards, espe-cially when customers suspect that low prices result from non-moral retailer behaviour or quality compromises”.23. Also Willems & Swinnen (2011) mention that still today, being seen at discount stores is for some people a taboo. Thus, discount stores could increase anticipated embarrassment in addition to the embarrassing product. Therefore, Lidl and Aldi were excluded from this study. The extent to which Jumbo, Albert Heijn, and Coop were perceived as neutral (no positive or negative image) and with a lower identity and self-expression function than the national brands was investigated in the second part of the first pre-test:
In the questionnaire, for each brand (store brands and national brands) one item of the inner self subscale and one item of the social self subscale of the self-expressiveness of
brands scale from Carroll & Ahuvia (2006) plus additionally 1 self-constructed item
23 Even though (Zielke, 2014) relates to shame, his explanation of shame is consistent with the concept of
was used. 24 The inner self subscale measures the degree to which “the brand reflects the inner self” and the social self subscale measures the degree to which “the brand enhanc-es the social self” (Carroll & Ahuvia, 2006, p. 83). Only taking one item of each of these scales ensured that the questionnaire will stay short and the high alpha (α= .94) measured by Carroll & Ahuvia (2006) of the scale represents that all 8 items of the orig-inal scale describe very good a single underlying construct what justified to shorten the scale.25 The item “This brand has a positive impact on what others think of me.” was used from the social self subscale and the item “This brand symbolizes the kind of per-son I really am inside” from the inner self subscale. The additional item was: “When putting this brand on the cash desk, I think that others who see me buying this brand will have a good impression of me.” This was integrated in order to have another item which reflects the extent to which the brand can enhance the social self but in a way which is closer related to the described scenario. Like this, more certainty can be gained that the brand really can be a way - or not - to enhance the public identity. The three items were measured on a five-point Likert scale as in Carroll & Ahuvia (2006) and were anchored at 1 = Strongly agree, 7 = Strongly Disagree, and additionally “I don’t know this brand”. As a check and to get an even clearer picture, the pre-test asked par-ticipants to rank the brands on a scale based on the degree to which they can express their personality and identify themselves with the brands; from the brand with which they can identify and express themselves least to the one with which they can identify and express themselves most.
For both tasks, only the brand logos (and not products of the brands) were shown be-cause when showing a certain product of the brand the evaluation of the brand could be influenced by the specific product.
Because the chosen brands should also not be disliked by participants what could addi-tionally lead to the feeling of embarrassment, also the following question was included: “What feelings do you have towards this brand?” The question had to be answered on a 7-point scale (1 = dislike very much, 4 = neither like nor dislike, 7 = Like very much, additional option “I don’t know this brand”).
24 Please see 3.2.4 Mediator Variable for the reason why both, the inner self subscale and the social self
subscale were used.
The results of the pre-test showed that it was most appropriate to select Bertolli and Magnum as national brands and Jumbo as store brand for the main survey: When the three question items showing the expressiveness of a brand were averaged to one varia-ble (α= .891, M= 3.55, SD= 0.894) 26, Bertolli (M = 4.45, SD = 1.239) was rated high-est, followed by Albert Heijn (M = 3.94, SD= 1.065) and Magnum (M = 3.84, SD = 1.173) 27. It seems that Albert Heijn has a very strong and good brand image which serves also to express ones identity. It was therefore not possible to use Albert Heijn for the main study. However, both Jumbo (M = 3.61, SD= 1.208) and Coop (M = 3.40, SD= 1.059) showed a lower average of the degree of expressiveness than Bertolli and Magnum. Because for Coop 26.7% indicated to dislike Coop (ranging on the scale from 2 to 3 28) what could probably lead to the feeling of embarrassment, Coop could not be taken for the main study and Jumbo was selected as the store brand (5.6% indicated to dislike Jumbo and this only on the scale point 3 which is one below “neither like nor dislike”).
Regarding the specific products for the selected brands, basilicum tomato sauce was chosen as the product for Bertolli and Jumbo, and Ice Cream Mini Mix as the product for Magnum and Jumbo (see Figure 7 and Figure 8). These products were chosen be-cause the ice cream packages respectively the tomato sauce packages of the national and store brand were very similar (including the claims on the packages) to each other. This should prevent an influence of the package design on anticipated embarrassment in the main survey.
However, the given results have to be interpreted with caution because effect sizes (in-ferred) were rather small and significance-tests could not be performed due to the small sample sizes because only half of the participants received a certain national brand to make the pre-test shorter. Therefore, in the second pre-test a question to test the results
26 Cronbach’s alpha for the two items that reflect the public expressiveness of brands (item of social self
subscale and self-constructed item) was α = .961 (M= 3.74, SD= 0.913). With this two items, Bertolli (M = 4.59, SD = 1.307) and Magnum (M = 4.24, SD = 1.200) showed the highest means overall, followed by Albert Heijn (M = 4.12, SD = 1.063). Jumbo showed a mean of M = 3.71 (SD = 1.233) and Coop one of M = 3.62 (SD = 1.142).
27 It has to be indicated, that generally, the ratings for the expressiveness were rather low, around the
middle scale point 4 which says “neither agree nor disagree”. However, it was assumed that for the pre-sent study it is mainly important that the brands show a difference in these ratings and not how high the values themselves are.
result from this pre-test – that Bertolli and Magnum have a higher expressiveness than Jumbo – was included.
Figure 7: Product pictures used for the national brand condition (Images: Jumbo Supermarkten (2017a))
Figure 8: Product pictures used for the store brand (Jumbo) condition (Imag-es: Jumbo Supermarkten (2017b))
3.2.1.2 Second Pre-Test
The second pre-test was sent as an online questionnaire to some friends here in Gro-ningen. 9 participants (56% female, average age 24.33 (SD= 1.581)) filled out the ques-tionnaire (see appendix B).
To gain certainty about the correctness of the first pre-test results regarding the choice of brands made for the main survey, the following question was included in the second pre-test: “Please choose the set of products with which you can better identify yourself and better express your personality:” One set showed Bertolli tomato sauce and
Mag-num Mini Mix ice cream, the other set showed Jumbo tomato sauce and Jumbo Mini Mix ice cream. 78% (7 out of 9) chose the set with the national brands Bertolli and
Additionally, because the whole basket should also not signal an additionally more sev-er undesired identity due to possible complementary or intsev-eraction effects between the products in the basket what could flaw the effects of the additional products on antici-pated embarrassment, this was tested in the second pre-test. Four sets of products were shown to the participants, one consisting of the relevant products intended for the main survey: hemorrhoidal cream, tomato sauce and Mini Mix ice cream. All sets were shown as Jumbo products in order to prevent an influence of the brand (except the hem-orrhoidal cream had no logo). For each set, participants had to answer the following two questions taken and adapted from Blair & Roese (2013): “How related are these three products?” (1 = not related at all, 7 = very closely related) and “How likely is it that others would think that all three or two of the three products are being purchased for the same reason?” (1 = very unlikely, 7 = very likely).
The results over the two questions (α= .965, M= 3.84, SD= 0.995) revealed that the hemorrhoidal cream, the tomato sauce and the Mini Mix ice cream did not seem to complement each other (M = 1.69, SD = 0.753). Moreover, they showed the lowest av-erage rating over all four product sets which means that participants perceived the ucts in this set as least related to each other; thus the perceived relatedness of the prod-ucts within the other three product sets was higher. Thus, it can be said pretty certainly that hemorrhoidal cream, tomato sauce and Mini Mix ice cream are not perceived to be related or complementary. 29
In summary, Figure 9 shows the products that were used for the 3 conditions of the main survey.
29 Note: Due to the small sample size it was not possible to conduct a significance test for the difference
Figure 9: Chosen products for the three conditions of the main survey (Images: Walmart Stores (2017), Jumbo Supermarkten (2017a, b))
National brand condition:
Hemorrhoidal Cream (not branded) + Bertolli tomato sauce+ Magnum Mini Mix ice cream
Store brand condition:
Hemorrhoidal Cream (not branded) + Jumbo tomato sauce+ Jumbo Mini Mix ice cream
Control condition:
Hemorrhoidal Cream (not branded)
3.2.2 Independent Variable
In the survey design, the independent variable, namely the basket composition, was ma-nipulated with the instructions for the participants and with the products and brands selected in the pre-test (see 3.2.1 Pre-Tests). Consequently, as shown in section 3.1
De-sign and Participants, the basket composition (variants 1 and 2, see Figure 5) was
ob-tained with three conditions of the survey to which participants were randomly as-signed: Embarrassing product alone (control condition), embarrassing product +
na-tional brands (nana-tional brand condition), embarrassing product + store brands (store brand condition) (see Figure 9).
3.2.3 Dependent Variable
