LYFT – A Feminine
Alternative for Snus
MASTER THESIS WITHIN: Business Administration NUMBER OF CREDITS: 15
PROGRAMME OF STUDY: International Marketing AUTHOR: Cilla Palmroos & Samuel Bengtsson JÖNKÖPING May 2021
An Exploratory Study of Consumer Perceptions of
Gendered Brands
Acknowledgements
The researchers would like to acknowledge and show gratitude towards everyone who took part in the research process.
Firstly, we want to express our gratitude to all participants of the study who provided us with valuable knowledge and insights and enabled us to fulfil the purpose of the study.
Secondly, we would like to show our appreciation to our tutor Brian McCauley for his guidance throughout the research process. With his feedback and expertise, we gained useful insights and ideas for the research topic.
Lastly, we want to thank our peers in the seminar group for providing insights and continuously helping us develop the thesis.
Cilla Palmroos
Samuel Bengtsson
i
Master Thesis in Business Administration
Title: LYFT – A Feminine Alternative for Snus. An exploratory Study of Consumer Perceptions of Gendered Brands
Authors: Cilla Palmroos & Samuel Bengtsson Tutor: Brian McCauley
Date: 2021-05-24
Key terms: Brand Gender, Brand Love, Gender Congruence, Snus
Abstract
Background: Brand gender is the set of human personality traits associated with masculinity
and femininity applicable and relevant to brands. Brand gender perceptions can be affected through a range of different marketing techniques, to attract specific customers, and positively influence their consumer-brand relationships and their brand equity. Similar to past marketing techniques used by tobacco companies who targeted women with cigarettes using slogans such as “Torches of Freedom” and the message that smoking makes you slim, marketers are now trying to change consumer perceptions of snus, a tobacco product originating from Sweden, where a drastic increase in usage among women can be seen.
Problem: Previous literature on the topic of brand gender has mainly focused on products and
product categories that are highly connected to consumers’ self-presentation. Current literature also expresses the need for future research to be done using specific brands and product categories. Researchers have also expressed the need for examination of brand gender in contexts of highly feminine cultures.
Purpose: The purpose of this study is to determine if LYFT’s marketing portrays a clear brand
gender positioning and to see how consumers perceive gendered branding. The aim is to expand existing knowledge on brand love, brand gender congruence and brand-gender positioning, using a product category that is less connected to consumers’ self-presentation. To fulfil the purpose of this research, a content analysis of LYFT’s Instagram was conducted, followed by an analysis of regular users of snus to identify their perceptions and relationships with LYFT.
Method: This study has followed a mixed-method approach, where a quantitative content
ii
interviews of regular users of snus in the age range of 22-28 years in Sweden who are highly familiar with LYFT. The interviews were analysed following interpretative phenomenological analysis (IPA).
Results: The interview participants perceived LYFT’s typical customer, their communication,
as well as their product attributes and benefits to be feminine, which was consistent with the results of the content analysis. Congruent with previous research, this study found that the typical user of the brand had the most significant effect on brand gender perceptions and that a strong brand gender positioning can enhance brand love. However, the study also found that female individuals were less likely to consume masculine products or product categories than what previous studies have claimed. This study is useful for marketing practitioners wanting to enhance their customer-brand relationships through clear brand gendering and shows the importance of utilizing marketing techniques such as social media to affect consumer perceptions of brand gender.
iii
Table of Contents
1.
Introduction ... 1
1.1 Background ... 1 1.2 Problem Discussion ... 3 1.3 Purpose ... 4 1.4 Research Question ... 4 1.5 LYFT Description ... 5 1.6 Definitions ... 52.
Literature Review ... 7
2.1 Branding ... 7 2.2 Brand Gender ... 82.2.1 Six Dimensions of Brand Gender ... 9
2.3 Underlying Factors that Affect Brand Gender ... 10
2.4 What does Brand Gender Affect? ... 12
2.5 Self-congruence and Gender Congruence ... 13
2.6 Self-expressive Brands ... 15
2.7 Brand Love ... 16
2.8 Femininity and Masculinity in Sweden ... 18
2.9 Smoking, Snus and Gender Differences... 19
3.
Methodology and Method ... 21
3.1 Methodology ... 21 3.1.1 Research Philosophy ... 21 3.1.2 Research Approach... 21 3.1.3 Research Design ... 22 3.1.4 Research Strategy ... 22 3.2 Method... 23 3.2.1 Data Collection ... 23 3.2.1.1 Secondary Data ... 23 3.2.1.2 Primary Data ... 24 3.2.1.3 Sampling Approach ... 25 3.2.2 Content Analysis ... 25 3.2.3 Semi-structured Interviews... 26 3.2.3.1 Interview Questions ... 27 3.3 Data Analysis ... 27 3.3.1 Content Analysis ... 27
3.3.2 Analysing Semi-structured Interviews ... 27
3.4 Ethics ... 29
3.4.1 Anonymity and Confidentiality ... 30
3.4.2 Credibility ... 30
4.
Empirical Findings ... 31
4.1 Content Analysis ... 31 4.2 Interview Findings ... 33 4.2.1 Social Influence ... 33 4.2.1.1 Influence on Usage ... 33 4.2.1.2 Influence on Behaviour ... 344.2.2 Snus Culture and Stigma ... 35
iv
4.2.4 Perception of Product ... 39
4.2.5 LYFT Brand Perceptions... 40
4.2.6 Perceived Typical Customer ... 40
4.2.7 Perceptions of the Typical Tobacco Snus Customer ... 42
4.2.8 Brand Femininity ... 42
4.2.9 Brand Love ... 44
5.
Analysis ... 45
5.1 Snus, Masculinity and Culture ... 45
5.2 Perception of LYFT’s Marketing ... 46
5.3 Perception of Product ... 47
5.4 Brand Love ... 48
5.5 Typical Customer and Gender Congruence ... 50
5.6 Brand Femininity ... 51
6.
Conclusion ... 52
7.
Discussion ... 53
7.1 Theoretical Contribution ... 53 7.2 Managerial Implications ... 54 7.3 Limitations... 54 7.4 Future Research ... 548.
Reference list ... 56
v
Figures
Figure 1 ... 31 Figure 2 ... 33
Tables
Table 1 Participant Overview ... 24 Table 2 Table of superordinte and emergent themes ... 28 Table 3 Traits ... 32
1
1. Introduction
______________________________________________________________________
This section starts by introducing the reader with a background of the topic, followed by the problem discussion and purpose, leading to the research question. It will further provide a description of LYFT and lastly, a list of relevant definitions is provided.
_____________________________________________________________________
1.1 Background
Gendered brand marketing has been a popular technique used by marketers for many decades (Ulrich & Tissier-Desbordes, 2018). During the 19th century, attractive women were featured in advertisements for cigarettes, however, their role was to appeal to male customers (Amos & Haglund, 2000). During World War One, women needed to take on more typical male occupations, so they also started to wear trousers, play sports and smoke (Waldron, 1991). The tobacco companies saw their opportunity to market cigarettes to females by communicating how you can keep yourself slim when smoking, instead of eating sweets, with the Lucky Strike campaign “Reach for a Lucky instead of a sweet” (Amos & Haglund, 2000). The campaign was developed by Edward Bernays, also known as the father of public relations (Amos &Haglund, 2000). In 1929, Bernays and American Tobacco hired women to smoke their “torches of freedom” as they participated in the Easter parade protesting against women’s inequality (Brandt, 1996). The “Torches of freedom” expression was then widely used by tobacco companies for several years and smoking became socially acceptable and even socially desirable for women (Amos & Haglund, 2000). Today, we can see a similar shift with the range of discussions about an increasing number of women that are using snus, especially tobacco-free nicotine pouches (Larsson, 2020; Tottmar, 2020; Yttergren, 2020). Some argue that the marketing techniques used by these companies are targeting a younger audience and are most appealing for women (Borglund, 2019; Yttergren, 2020). While snus has predominantly been found to be associated with masculinity and male-dominated professions (Edvardsson et al., 2012).
Brands have been trying to depict a feminine or masculine personality in order to attract customers for many decades (Ulrich & Tissier-Desbordes, 2018). Brands such as Harley
2
Davidson have targeted male consumers through a range of different marketing techniques, where they depict themselves as a masculine brand (Roper et al., 2013). Meanwhile, brands such as L’Oreal have for decades targeted women, trying to portray a feminine brand personality (Ulrich & Tissier-Desbordes, 2018). Research has found that a strong brand gender identity can be achieved through a range of different means such as brand design elements, which would be logo shape, brand name, type font and colour (Lieven et al., 2015). Lieven et al. (2015) found that for instance brand masculinity perceptions are enhanced by angular, bold logo shapes, while round, slender shapes increase femininity perceptions. A well-positioned marketing mix has also been found to have a significant impact on consumer’s brand gender perceptions (Ulrich et al., 2011). Research has found that having a strong brand gender identity, feminine or masculine, has a direct effect on brand love, brand equity and brand loyalty of the targeted consumers (Alreck, 1994; Bajac et al., 2018; Machado et al., 2019; Vacas de Carvalho et al., 2020). Past research has also established that there is a positive connection between brand gender congruence and brand loyalty (Bajac et al., 2018; Grohmann, 2009; Machado et al., 2019; Vacas de Carvalho et al., 2020). Bajac et al. (2018) found that there is a positive relationship between an individual’s gender identity and their loyalty and general attitude towards a brand that resembles a similar gender identity. An important finding by Bajac et al. (2018) was also that congruence between consumers’ gender identity and the brand gender resulted in brand loyalty, while congruence between biological sex and brand gender did not.
Recent research has focused almost exclusively on brands and products that are highly linked to self-presentation (Azar, 2015; Azar et al., 2018; Ulrich & Tissier-Desbordes, 2018). Furthermore, research has been analyzing the relationship between brand gender positioning, brand gender congruence and brand loyalty quantitatively, but there is an admitted gap in the underlying sources of the relationship (Vacas de Carvalho et al., 2020).
3
1.2 Problem Discussion
Reviewing the literature, it becomes apparent that research investigating brand gender and gender congruence has predominantly been focused on high involvement products and brands that are deeply connected with consumers self-presentation (Azar, 2015; Azar et al., 2018; Ulrich & Tissier-Desbordes, 2018) and the effect that brand gender image has on factors such as brand loyalty, brand equity, and brand love (Alreck, 1994; Machado et al., 2019; Vacas de Carvalho et al., 2020). Van Tilburg et al. (2015) studied the effect of brand gender on positive affective and behavioural responses, using product categories that are highly linked to self-presentation. However, the authors recognize that there is a need to check whether these results apply in the context of product categories that are less connected to the owners’ self-presentation and gender. Moreover, Azar et al. (2018) suggest that further research within brand gender perceptions is needed using symbolic product categories that are less status-linked.
Vacas de Carvalho et al. (2020) studied the link between brand gender, brand loyalty, and brand love. The study was done on social media and found that a clear brand gender identity will enhance loyalty towards a brand. Since they were the first to investigate this relationship in the context of social media, further research is needed to strengthen the findings. Moreover, the authors suggest that future research could be done on specific gendered brands or product categories, to provide a more realistic assessment. Further, research has established that there is a connection between brand gender congruence and brand loyalty (Grohmann, 2009; Vacas de Carvalho et al., 2020; Machado et al., 2019). However, there is limited research on the underlying sources of this relationship (Vacas de Carvalho et al., 2020).
There is evidence that a clear brand gender positioning has a significant impact on brand love (Machado et al., 2019). Where brand love and other consumer-brand relationships have been found to result in brand loyalty, positive WOM communication as well as willingness to pay (Bairrada et al., 2018; Batra et al., 2012; Malär et al., 2011). However, since research within brand love is still scarce, there is a need for further investigation within the subject (Bairrada et al., 2018). Lastly, recent research has suggested further investigation on brand gender in other cultural contexts (Azar et al., 2018; Machado et
4
al., 2019), especially highly feminine cultures such as Sweden (Vacas de Carvalho et al., 2020).
1.3 Purpose
The purpose of this study is to determine if LYFT’s marketing portrays a clear brand gender positioning and to see how consumers perceive gendered branding. We explore this in the context of LYFT, a modern tobacco-free nicotine pouch within the snus product category. This research aims to expand existing knowledge on brand love, brand gender congruence and brand-gender positioning, using a product category that is generally less connected to consumers self-presentation. To fulfil the purpose of this research, firstly a content analysis of LYFT will be conducted to determine their brand gender positioning. Followed by an analysis of regular users of snus in order to identify their perceptions and how these influence their relationship with LYFT.
1.4 Research Question
Given the lack of research examining consumer perceptions of brand gender, and how these perceptions influence consumers-brand relationships with brands, especially in the context of brands that are less connected to consumers self-presentation. The first question is:
RQ1: Does LYFT’s marketing portray a clear brand gender positioning?
And was asked in order to determine LYFT’s brand gender positioning. And the second research question:
RQ2: How do snus users perceive gendered branding?
Investigates consumers' perceptions of LYFT’s brand gender positioning, as well as their attitudes towards the brand, and how the two influence their relationship with the brand.
5
1.5 LYFT Description
LYFT’s nicotine pouches, produced by British American Tobacco Sweden AB, are tobacco-free nicotine pouches that one places under the upper lip and can be used as a substitute for smoking, and other types of nicotine products (Golyft, n.d.). The product has many similarities to the more well-known Swedish snus, which is oral smokeless tobacco that resembles a miniature teabag that is also placed under the upper lip (Clarke et al., 2019). Snus is a traditional Swedish product with a very long history of use that dates back several hundred years (Digard et al., 2009). Due to Snus cultural significance in Sweden, many use it as a substitute for smoking (Digard et al., 2009). Where the rate of death amongst males due to smoking in Sweden is the lowest in the world, presumably due to Snus serving as a less harmful substitute (Foulds et al., 2003). Research has found that snus occurs more often amongst Swedish men, in comparison to women in general (Digard et al., 2009). An obvious difference, however, when comparing traditional snus with LYFT is that LYFT’s product is entirely tobacco-free (Golyft, n.d.). Apart from a health aspect, the absence of tobacco in LYFT’s products also gives them the ability to be more flexible in their marketing towards consumers, as the Swedish tobacco laws do not apply to their product (FOHM, 2021). Some of LYFT’s media channels where they advertise their products are their website (Golyft, n.d.), Instagram page (LYFT, n.d.), and in-store advertising through third-party retailers. Through these channels they target potential customers with stylishly designed pictures, using well-known social media personalities as spokespeople for their products.
1.6 Definitions Brand Gender
“The set of human personality traits associated with masculinity and femininity
applicable and relevant to brands” (Grohmann, 2009, p. 106).
Brand Love
“the degree of passionate emotional attachment a satisfied consumer has for a particular trade name” (Caroll & Ahuvia, 2006, p.81).
6
Brand equity
Brand equity refers to consumer perceptions and the global value associated with a brand, which stems from the brand name as well as physical aspects of the brand (Lassar et al., 1995). Brand equity is also correlated to competition and can positively influence the financial performance of a firm (Lassar et al., 1995).
Self-expressive Brands
“the customers' perception of the degree to which the specific brand enhances one's social self and/or reflects one's inner self” (Carrol & Ahuvia, 2006, p. 82).
Social influence
Social influence occurs when individuals alter their thoughts, behaviour, and physical state to meet demands within their social sphere (McKenna, Joinson, Reips & Postmes, 2007).
Gender identity
Gender identity is defined as the extent to which an individual identifies with masculine or feminine personality traits (Deaux, 1985).
Brand Self-Congruence
Self-congruity occurs when the brand/product-user image matches the consumer’s perceived self-image (Sirgy, 1982).
Brand Gender Congruence
Gender congruence occurs when the consumer perceives that his/her gender identity matches the gender identity of the brand (Grohmann, 2009).
Instagram is a social media platform that allows users to upload pictures and short video clips on their individual pages. It is one of the leading social media tools and is used by both individuals and companies (Robinson, 2020).
7
2. Literature Review
______________________________________________________________________
This part provides relevant literature on branding, brand gender, brand love, self-congruence, self-expressive brands, and the Swedish culture. This is to present clarity and a deeper insight into the relevant theories of the topic.
_____________________________________________________________________
2.1 Branding
The terms brand and branding started as expressions for a simple entity with limited application, primarily created, interpreted, and controlled by its creator (Bastos & Levy, 2012). The term entered marketing in 1922 simply as an expression for “brand name” (Stern, 2006). Today branding is used as an expression for how businesses and organizations portray and differentiate themselves using their name, symbol, design and many other aspects of the business (Rowley, 2004). Branding is now recognized as an essential part of any business that can be used as a powerful tool to gain a competitive advantage (Rooney, 1995). Where choosing the right name, the right advertising, and effectively applying the relevant techniques are a few of the necessary factors needed to create a successful branding strategy (Bastos & Levy, 2012). Further, with the rise of social media and the rapid digitalization of modern society, branding has become increasingly more demanding (Kohli et al., 2015; Lipiäinen & Karjaluoto, 2015). Therefore, the success of businesses now more than ever lies in their capability to quickly adapt and change strategies in response to consumer preferences (Kohli et al., 2015). Consumers are now acquiring those brands that allow them to meet their deeper needs for social, economic and environmental justice, and therefore go beyond just the product or the company and base their choices on associations and emotional benefits (Gómez-Suárez et al., 2017). Moreover, branding in the digital age requires both strong internal communication as well as strong external communication (Lipiäinen & Karjaluoto, 2015). The brand must be marketed to external stakeholders such as suppliers, customers, investors, media, and society consistently, and internal stakeholders are seen as both a target of the brand as well as a key component of it (Lipiäinen & Karjaluoto, 2015).
8
2.2 Brand Gender
Brand personality remains a very important aspect for marketers, as it is one of many factors that can help differentiate a brand and enable them to gain a competitive advantage (Keller et al., 2002). A widely used and discussed theory around the topic of brand personality is Aaker’s five-factor brand personality scale, which includes a range of different dimensions that help form the personality of a brand (Aaker, 1997), and among those are the masculinity and femininity traits. However, Grohmann (2009) argued that the model does not adequately capture the gender factor of brand personality and therefore developed a scale to measure gender dimensions of brand personality supplementing Aaker’s five dimensions of brand personality. Where two independent concepts were found, namely: masculine brand personality (MBP) and feminine brand personality (FBP) (Grohmann, 2009). Both MBP and FBP are defined consistent with the definition of brand personality (Aaker, 1997), namely as “a set of human personality traits associated with masculinity and femininity applicable and relevant to brands” (Grohmann, 2009, p. 106), where MBP is characterized using traits such as Adventurous, aggressive, and brave, and FBP is characterized using traits such as fragile, graceful, and sensitive. The MBP/FBP scale complements and can be used in concurrence with the original five dimensions of Aaker’s model (Grohmann, 2009). Furthermore, it was found that the MBP/FBP scale was more appropriate in measuring gender dimensions of male/female brand personalities than human personality scales. Moreover, the argument behind adding these two additional dimensions used for measuring the gender of a brand, is that masculine and feminine personality traits are heavily used when examining human personality and is therefore readily accessible to people (Grohmann, 2009). This, therefore, suggests that consumers also associate masculine and feminine personality traits with brands.
Grohmann’s (2009) research suggests that there are four different groups which brands can be divided into in terms of brand gender, namely: (1) high masculine/low feminine, (2) low masculine/high feminine, (3) high feminine/high masculine, and lastly (4) low feminine/low masculine, where brand gender appears to influence brand attitude, purchase intentions and word-of-mouth communication among consumers. Moreover, the results of Grohmann’s study suggests that femininity and masculinity are negatively correlated, meaning that when masculinity is high, femininity is often low, and vice versa. Due to this, the study found that there is a clear sparsity in androgynous brands, brands
9
high in both femininity and masculinity, while there are several cases of brands low in both femininity and masculinity (Grohmann, 2009).
2.2.1 Six Dimensions of Brand Gender
In 2011, a study was done by Ulrich et al., where the results were found to be congruent with those of Grohmann’s study in 2009 in terms of the four types of brand genders, and the lack of androgynous brands. Additionally, Ulrich et al. (2011), found support for the importance of brand gender as a distinct concept from brand personality as portrayed by Aaker (1997), as well as masculine/feminine brand personality by Grohmann (2009). More specifically, Ulrich et al. (2011) found that brand gender is a multidimensional concept that includes six different dimensions; (1) Gender of the main brand user, (2) Gender attributes of communication, (3) Gendered brand personality, (4) Gender of the brand name, (5) Gender attributes and benefits of the product, and lastly (6) Gendered attributes of the logo. These six dimensions of brand gender appeared through exploratory in-depth interviews, where the participants were asked to express all brand associations that came to mind in regards to a range of different brands.
The first dimension, gender of the main brand user, appeared to be the main dimension since its impact on consumer perceptions was very significant. The subjects of the study identified the gender of a brand by its main user, independently of the gender of the product category that the brand operated in (Ulrich et al., 2011). The second dimension, gendered brand personality referred to the personality traits of the brand that might indicate a specific gender. These are for example seductive, tender, cheerful, and soft, which would indicate femininity, or assertive, competitive, and aggressive, which would indicate masculinity (Ulrich et al., 2011). Moreover, the third dimension, Gender attributes of the communication, was found to be a commonality between a majority of the subjects to be something that affected their perception of a brand’s gender (Ulrich et al., 2011). By communication attributes, the most significant attributes were the spokesperson associated with the brand, the advertising baseline, or a gendered symbol associated with the brand (Ulrich et al., 2011). Both the fourth dimension (Gender of the brand name), and the fifth dimension (Gender attributes of the logo), seemed to have less of an impact on brand gender perceptions among the subjects. Where the name of the brand might be perceived feminine if the pronunciation of the name is soft and round, in
10
contrast to sharp with hard consonants (Ulrich et al., 2011). However, the authors acknowledge that there might be differences in different languages. In terms of gender attributes of the logo, colour and shape had a certain impact on gender perceptions, where red and black might indicate masculinity. Meanwhile, the shape had more to do with the associations the subjects could do, where a rose was considered feminine, and swords for example were considered masculine (Ulrich et al., 2011). Lastly, the gender attributes and benefits of the product refer to the shape and colour of both the packaging as well as the product itself. In terms of packaging, they refer to a product offered by the brand Nickel, where the packaging looked like a petrol can with abrupt shapes, indicating masculinity (Ulrich et al., 2011).
2.3 Underlying Factors that Affect Brand Gender
Research investigating the factors affecting the perceived brand gender has found that it is often determined based on the gender of the product promoter (Debevec & Iyer, 1986; Ulrich et al., 2011) and the perceptions of the product’s typical consumer (Keller, 1993; Neale et al., 2016; Ulrich et al., 2011). Furthermore, different brand design elements such as logo shape, brand name, type font and colour frequently influence brand gender perceptions (Lieven et al., 2015; Ulrich et al., 2011). Precisely, brand masculinity perceptions are enhanced by angular, bold logo shapes, while round, slender shapes increase femininity perceptions (Lieven et al., 2015). Consonants in the brand name also influence consumers’ brand perceptions, more specifically, in the English language, stops (k and t) enhance brand masculinity, whereas fricatives (f and s) enhance brand femininity (Guevremont & Grohmann, 2015).
In addition to examining brand gender, research about product gender has also shown some interest (Fugate & Phillips, 2010; Milner & Fodness, 1996; Van Tilburg et al., 2015). Research on person perception explains that gender is one of the first things you acknowledge when meeting someone new (Dion et al., 1972) and you assume the gender based on the person’s physical appearance (Deaux & Lewis, 1984). Thus, appearance is argued to also indicate product gender (Van Tilburg et al., 2015). Van Tilburg et al. (2015) studied how the aesthetics of a product influence consumers’ perception of its gender. The authors found that heavier products with bulky proportion, an angular shape, and
11
darker colours enhanced the product perception of a product’s masculinity, whereas lighter products with slim proportions, round shape, curvy lines and bright colours enhanced the perception of the product femininity. Like brand gender, a clear product gender has been found to increase positive consumer perceptions and behaviour (Van Tillburg et al., 2015).
The idea that colours, surfaces, patterns, and shapes influence gender perceptions were discussed by Alreck (1994), where lighter colours, pliable surfaces and natural or blended patterns were found to have a strong connection to femininity. On the contrary, darker colours, rough surfaces and geometrical surfaces were found to have a strong connection to masculinity (Alreck, 1994). Moreover, straight, and sharp angular shapes; larger sizes, as well as heavier weights were also found to have masculine connotations, where the opposite, smooth, and round shapes; smaller sizes, and lighter weights were connected to femininity (Berlyne, 1977; Alreck, 1994). All of these factors are vital not only in product appearance but the container, packaging as well as the general promotion of the brand, which can include a range of different marketing tools where these factors can be utilized to portray a certain brand gender (Alreck, 1994).
Furthermore, that colour has a strong correlation with femininity and masculinity has been found by both Lieven et.al (2015) and Kirkham (1996), where both studies established that darker colours increase brand masculinity perceptions, and lighter colours increase brand femininity perceptions. Another finding on colour differences was that there is a difference in preference regarding coolness and warmness between men and women, where women tended to prefer warm colours and men tended to prefer cool colours (Singh, 2006; Iijima et al., 2001).
Lemm et al. (2005) found that an important primer for gender associations in pictures used in brand communication is the specific items included. Where a baseball mitt included in a picture will have strong connections to masculinity, perhaps because it is an item more widely used and owned by males, or perhaps due to a secondary feature that is associated with masculinity (Lemm et al., 2005). In contrast, an oven mitt could then be associated with femininity for the same reasons (Lemm et al., 2005).
12
2.4 What does Brand Gender Affect?
Grohmann’s research in brand gendering has been extensively cited in research during the past few years, where the implications of femininity and/or masculinity in a brand on many factors, such as brand equity, brand loyalty and brand preference have been established (Lieven et al., 2014; Lieven et al., 2015; Machado et al., 2019; Vacas de Carvalho et al., 2020).
Lieven et al. (2014), examined the extent to which brand masculinity and femininity are correlated to brand equity and does so by conducting three different studies. The results show that brand gender influences brand equity since consumers often find it easy to categorize sex-typed stimuli even in brands, which would be for instance highly feminine or highly masculine brands (Lieven et al., 2014). The ease of categorization then results in a more positive response to these either highly feminine or highly masculine brands (Lieven et al., 2014). However, the study also found that brands that are low in both femininity and masculinity or high in both femininity and masculinity do not reap the benefits of ease of categorization resulting in them having less overall appeal and associated with lower brand equity ratings (Lieven et al., 2014). However, in a study done by Lieven and Hildebrand (2016), the findings somewhat contradict those of previous research, where androgynous brands are found to have a significantly higher impact on brand equity than previously thought, even so, that androgynous brands generate higher brand equity than either exclusively masculine or feminine brands (Lieven & Hildebrand, 2016). The reason for this is that previous studies have not included androgynous brands to the same extent as exclusively feminine or masculine brands, inhibiting any positioning to be made concerning androgynous brands (Lieven & Hildebrand, 2016). The study also established that highly feminine brands had a more significant impact on brand equity among female consumers, and highly masculine brands had a more significant impact on brand equity among male consumers.
Vacas de Carvalho et al. (2020) examined brand gender and its effect on brand loyalty by studying brand’s Facebook pages, where they found that brand gender has a major indirect effect on brand loyalty, through the mediators of brand love and brand quality. Moreover, they found that a strong brand gender positioning, in general, has a positive effect on brand loyalty, however, brand gender has a stronger effect on brand loyalty when the brand is highly masculine, compared to highly feminine. They did not examine
13
one specific product category, and instead included a multitude of product categories. Therefore, the researchers realize that there is a need for more extensive research using a specific product category and a specific brand in order to provide a more realistic appraisal of brand gender’s influence on brand love, brand quality and brand loyalty.
2.5 Self-congruence and Gender Congruence
Sex is one of the most used characteristics for demographic categorization and is often used interchangeably with gender to indicate if an individual is biologically male or female (Fugate & Phillips, 2010). Gender identity, however, is not provided to us at birth but rather performed by us through situated, symbolic social interactions (Butler, 1988; West & Zimmerman, 1987) and is defined as the extent to which an individual identifies with masculine or feminine personality traits (Deaux, 1985). Thus, femininity and masculinity can be argued not to be biological traits but personality traits (Fugate & Phillips, 2010). We construct our gender identity in social interactions by adjusting our behaviour to correspond to the normative conceptions of masculinity and femininity that stand in our culture (Gherardi, 1995). Within a culture, there are different forms of masculinity and femininity present (Carrigan et al., 1985) and the almost hegemonic standard that emerges leads to stereotypes, which is defined by Ellemers (2018) as assumptions about members of specific social groups. Stereotypes can often lead to people overemphasizing distinctions between groups and underestimating the differences within them, because of the perception that a particular feature characterizes membership of the group (Ellemers, 2018). Gender role stereotypes are often difficult to change once established (Ulrich, 2013). These stereotypes about gender roles affect consumer behaviour, influencing the evaluation and use of products and brands (Paek et al., 2011). However, the tolerance towards individuals engaging in roles conventionally connected to the opposite sex has increased throughout the years and also led to shifts in the typical gender identifications of both sexes (Fugate & Phillips, 2010).
Consumption of products and brands is part of this performance of gender since gender identity explains differences in consumer behaviour (Fischer & Arnold, 1994; Ulrich & Tissier-Desbordes, 2018), which is in line with the self-congruency theory by Sirgy (1982) that states that individuals favour products and brands that reflect their self-image or identity. Also, the fact that consumers don’t buy products that much for the functional
14
value, but more for the symbolic value of the product that matches their identity (Grubb & Grathwohl, 1967; Levy, 1959; Whitley et al., 2018). Positive self-congruence influences consumer behaviour in several ways, such as purchase intention (Landon, 1974), products choice (Malhotra, 1988), product adoption (Antón et al., 2013), brand loyalty (Kressmann et al., 2006), and emotional brand attachment (Malär et al., 2011). However, Bajac et al., (2018) suggest that the need for congruence might be different for addictive products. The authors explain that congruence may play a significant role to generate initial interest in a product but once addiction occurs, the need for congruence diminishes.
Research shows that both men and women want to know what products and brands are suitable for them in a normative sense and are uncomfortable using products that don’t seem to be made for them (Milner & Fodness, 1996). Individuals create, strengthen and achieve their gender identities through consumption, therefore, our possessions serve as symbolic gender identity markers (Avery, 2012). The desire for self-congruency is evident in both product and brand choices (Cowart et al., 2008). Fugate and Phillips (2010) found that individuals who aspire towards self-congruence with product personality also seek product gender congruence. Moreover, the authors suggest that individuals who don’t perceive themselves agreeing with traditional gender roles or were raised in a home where traditional gender roles were not demonstrated are less likely to seek gender congruence through consumption.
Both men and women prefer brands that express their own gender identity, however, masculine individuals are less likely to buy feminine brands than feminine individuals are to buy masculine brands (Fry, 1971; Fugate & Phillips, 2010; Neale et al., 2016; Worth et al., 1992). For example, a study made by Worth et al. (1992) revealed that the subjects who rated themselves highly masculine preferred products that were described in very masculine terms. In contrast, the subjects who rated themselves as highly feminine preferred products described with highly feminine terms but preferred next most products described in both feminine and masculine terms. Gender identity is a more sufficient dimension for customer segmentation than biological sex and overall, masculine brands succeed better than other gendered brands for attracting consumers with different gender identities (Neale et al., 2016). Similar findings have been revealed when examining
cross-15
gender extensions. Jung and Lee (2006) suggest that cross-gender extensions are more likely to get accepted when it’s a masculine brand that aims to target female consumers than the other way around. Their study also found that women are more positive towards cross-gender extensions than men and, in general, the acceptance is higher for functional products than symbolic ones. Ulrich (2013) also suggests that consumers with a high tendency to stereotype gender roles are more likely to react unfavourably towards brand extensions to the opposite gender, than consumers with a low tendency to stereotype gender roles.
2.6 Self-expressive Brands
Self-expressive brands are defined as “the customers' perception of the degree to which the specific brand enhances one's social self and/or reflects one's inner self” (Carrol & Ahuvia, 2006, p. 82). Self-expressive brands also serve as important symbols of accomplishment, as well as a tool for social integration that allows consumers to express their individuality (Schembri et al., 2010). An example of a typically self-expressive product category would be the clothing industry, where consumption of clothing is often used to express, construct and enhance one׳s self-concept (Millan & Reynolds, 2014). There is a clear lack of research examining the self-expressiveness of snus, perhaps due to its rarity outside of Sweden and other Nordic countries (Salokannel & Ollila, 2021). However, Kim et al. (2017) examined e-cigarette use and behaviour among adult consumers and found that there is a clear functional value to the product, as it is commonly used as a substitute for regular tobacco cigarettes. Where the utilitarian value of e-cigarettes is conferred through the ability to reduce nicotine cravings without having to smoke a tobacco cigarette (Kim et al., 2017). The same argument is disputable for snus, as it is largely used as a substitute for smoking (Lund & Lund, 2014). Although there is a certain self-expressive value to snus, through the tendency for users to be influenced by their social circle (Edvardsson et al., 2012), the self-expressive value is arguably quite low compared to typically self-expressive products.
According to Khalil (2000), there are two types of products regarding how these products confer utility, namely: symbolic and substantive. Where symbolic products are defined as products that give a self-regarding utility to the consumer (Khalil, 2000), which in most
16
cases applies to luxury brands and products such as fashionable clothing or expensive cosmetic products (Piacentini & Mailer, 2004). The reason for consuming symbolic products is not solely due to the functional value of the product, but more so the symbolic meaning created by the consumption of the product, which is used as an outward expression of the consumer’s self-concept and general connection to society (Elliot & Wattanasuwan, 1998). Substantive products on the other hand confer utility primarily based on the function of the product and the monetary value that it gives (Khalil, 2000). Substantive products are therefore products that satisfy conventional needs such as comfort, warmth, or hunger, which could, for example, be food, medicine or other primarily functional products (Khalil, 2000).
2.7 Brand Love
Consumer brand relationships are defined as the psychological bonds that consumers form with brands (Fournier, 1998). Apart from loving people, it has been established that love is also a common consumption-related emotion (Richins, 1997) and we can, therefore, also love brands (Ahuvia, 2005). Fournier (1998) laid the theoretical foundation for consumer-brand relationships, where one of the 6 identified relationship types is love (Rauschnabel & Ahuvia, 2014). Brand love is defined by Carroll and Ahuvia (2006, p. 81) as “The degree of passionate emotional attachment a satisfied consumer has for a particular trade name”. Similar to interpersonal love, brand love includes passion, attachment, positive assessment, positive emotions, and declarations of love for the brand and is often a result of a long-term relationship between the consumer and the brand (Carroll & Ahuvia, 2006).
Batra et al. (2012) constructed through consumer interviews seven core elements of brand love. These include:
1. Self-brand integration
Meaning that consumers’ personal identity matches brand identity, it reflects their desired self-identity and provides meaning to life
17
Including that the consumer has a passionate involvement with the brand, has had frequent interaction with the brand in the past, desires to use it and has invested time, energy and money in it
3. Positive emotional connection
Referring to consumers’ psychological feeling that the brand matches their needs and desires and experiencing a positive emotional connection with the brand
4. Long-term relationship
Meaning that the consumers want the brand to be part of their life for a long time
5. Positive attitude valence
Including satisfaction and having an overall positive evaluation of the brand
6. Attitude Strength
Referring to having strong feelings towards the brand and thinking about it often with certainty and confidence
7. Anticipated separation distress
Includes feeling worried and anxious if the brand ceases to exist
Several authors have examined both the antecedents and outcomes of brand love. Various authors emphasize the importance of perceived quality as a predictor of brand love, however, Rauschnabel and Ahuvia (2014) found that anthropomorphism was significantly more important when explaining consumers’ love for brands. This is in line with Batra et al. (2012) and Bairrada et al. (2018) who stated that perceived quality is important but not a sufficient driver for brand love. Thomson et al. (2005) suggest that a repeated post-consumption satisfaction over time may lead to an emotional bond between the consumer and brand. However, satisfaction is not a sufficient predictor of brand love, since consumers can be equally satisfied with several brands (Roy et al., 2013).
Individual personality traits can also play an important part in the development of brand love (Roy et al., 2013). For instance, self-congruity with a brand has shown to be an important factor for brand love (Albert et al., 2013; Bıçakcıoğlu et al., 2018; Malär et al., 2011; Rauschnabel & Ahuvia, 2014). Carroll and Ahuvia (2006) found that brand love is often higher for products that have symbolic and/or hedonic benefits. More precisely, the authors argue that symbolic brands that offer self-expressive benefits offer the greatest potential to brand love also encouraging desirable post-consumption behaviour. This is
18
also confirmed by Wallace et al. (2014) who found a connection between self-expressive brands and brand love, particularly, that consumers who engage with a self-expressive brand through liking them on Facebook had high levels of brand love. Bairrada et al. (2018) explain that the evidence antecedents of brand love are still scarce, so they tested new dimensions and found that brand uniqueness, brand prestige and perceived value had a direct contribution to brand love.
When it comes to brand love outcomes, several studies have proved that brand love leads to brand loyalty (Bairrada et al., 2018; Bıçakcıoğlu et al., 2018; Carroll & Ahuvia, 2006; Chaudhuri & Holbrook, 2001; Thomson et al., 2005), higher resistance to negative brand information (Batra et al., 2012), positive word of mouth, and willingness to pay a higher price (Bairrada et al., 2018; Albert et al., 2013; Thomson et al., 2005). A recent study also investigated brand coolness and suggested that brand love is an outcome of coolness (Tiwari et al., 2021). Brand coolness has also shown similar outcomes to brand love, such as positive word of mouth and willingness to pay (Warren et al., 2019). However, the dimensions of perceived coolness and the impact of coolness on brand love are still under-researched constructs (Tiwari et al., 2021).
2.8 Femininity and Masculinity in Sweden
Since gender identity is constructed through social interactions adjusted to correspond with the normative behaviour within the culture (Gherardi, 1995), what is perceived as feminine or masculine depends on the country and culture that is observed. Hofstede (1985, 2001) demonstrated that countries are different in terms of masculinity and femininity and the effect of it in different aspects such as gender roles and consumption. Sweden, like the other northern countries, is highly feminine compared to other cultures. One reason why Scandinavia and especially Sweden is ranked as highly feminine may be the history of dividing social and economic responsibilities between both genders (Nelson et al., 2006).
In masculine countries, gender roles are clearly distinguished, whereas in feminine societies gender roles and norms overlap more (Hofstede, 2001; Nelson et al., 2006). Masculine countries are also characterized by having expectations on men to be tough, assertive and focused on material success, but feminine countries allow men to be more
19
feminine, gentle and weak and women are allowed to describe themselves in their own terms, often nurturing but also competitive (Nelson et al., 2006). An example within consumer behaviour is that in highly masculine countries it is seen as important to have a car with high engine power since the car is a part of the extended self, representing masculinity (Hofstede, 2001).
Sweden is often positioned as world-leading when it comes to gender equality (Towns, 2002) and ranks as number five in the Global Gender Gap Index (World Economic Forum, 2021). According to the social role theory (Eagly & Steffen, 1984; Eagly et al., 2000) such results in gender equality when it comes to labour division often also leads to a decrease in gender stereotypes.
2.9 Smoking, Snus and Gender Differences
During the 19th century, attractive women were featured in advertisements for cigarettes, however, their role was to appeal to male customers (Amos & Haglund, 2000). During World War One, there was a change when women needed to take on more male occupations which also led to smoking becoming more common among women (Waldron, 1991). Tobacco companies seized the opportunity to develop new meanings to women smoking to overcome the negative associations and smoking was repositioned as sociable, feminine, stylish, and fashionable (Amos & Haglund, 2000). Edward Bernays and American Tobacco succeeded in making smoking for women socially accepted and even socially desirable through the “torches of freedom” campaign, which has been widely used by tobacco companies for decades (Amos & Haglund, 2000; Brandt, 1996).
Research on gender and cigarette advertisements describe how tobacco companies have specifically targeted women with female cigarette brands, advertisements and promotional campaigns (Amos & Haglund, 2000; Boyd et al., 2003; O'Keefe & Pollay, 1996). The efforts to market to women has led to smoking becoming more common among females and enhanced the association of smoking with appealing attributes such as female liberation, glamour, success and thinness (Amos & Haglund, 2000; Boyd et al., 2003; Ernster, 1985). A study by Carpenter et al. (2005) found that the targeting of women by the tobacco industry goes beyond advertising and marketing. The authors describe that the industry has focused on identifying gender-based differences from motivational
20
factors to product differences to promote and enhance smoking among women. Smoking in Sweden is still today more common among men, and the prevalence of smoking is estimated to decrease from 17,5 percent in 2019 to 15,2 percent by 2025 among men (Statista, 2021a) and from 17, 2 percent among women in 2019 to 14,6 percent by 2025 (Statista 2021b).
Snus use in Sweden has been proven to be a more common occurrence amongst men in comparison to women, where recent data shows that approximately 18% of men aged between 18-84 use snus, in comparison to only 4% in the same age range for women (Swedishmatch, 2018). It has also been found that women tend to choose other variants of Snus, namely miniature-sized alternatives to snus, and rarely use “loose” snus, which is tobacco that is not sealed in a pouch (Digard et al., 2009). Women also tend to consume snus in lower amounts in comparison to men (Digard et al., 2009).
Research examining the gender aspect of Snus from a cultural perspective has concluded that Snus is generally associated with masculinity and traditionally manly behaviour (Edvardsson et al., 2012). Edvardsson et al. (2012) interviewed young adults who associated Snus with traditionally male-dominated professions such as car mechanics and farmers. Helme et al. (2019) found that Snus was highly impacted by gender stereotypes. They argue that the typically masculine personality trait of engaging in something they know can be harmful to demonstrate their masculinity is a significant factor in the use of Snus (Helme et al., 2019). It was also found that men often reacted negatively towards women consuming snus, as it was associated with masculine behaviour (Edvardsson et al., 2012).
21
3. Methodology and Method
______________________________________________________________________
This section starts by presenting the methodology of the thesis, namely the research philosophy, design, and strategy. Followed by the method, which gives a thorough explanation of the data collection, which includes the sampling approach and an overview of the content analysis. Lastly, the process of the data analysis is explained as well as ethical considerations.
______________________________________________________________________
3.1 Methodology
3.1.1 Research Philosophy
Due to the nature of this research, it follows an interpretive philosophy, where the purpose is to understand and analyse consumer perceptions of gendered branding. Interpretivism is based on the argument that individuals and their social worlds cannot be studied in a similar way as physical phenomena since people create meanings (Saunders et al., 2016). One principle of the interpretive approach is that the empirical data generation is a product of a socially constructed interaction, co-produced by both the researchers and the subjects (Walsham, 1995). Where in the case of this study, the co-produced data generation will be a result of the qualitative interviews with the subjects conducted by the researchers. Furthermore, besides contributing to a detailed understanding of contextualized human experience, interpretivism is used to pursue theoretical generalizability, otherwise known as transferability (Gheondea-Eladi, 2014).
Moreover, due to the nature of the research, where it is assumed that the participants of the research may have different perspectives and their own subjective opinions on the matter, the interpretive paradigm is undoubtedly best suited for the situation. The reason behind this is that the epistemological assumption that reality is subjective validates the use of subjective data collected through qualitative methods (Walsham, 1995).
3.1.2 Research Approach
While deduction moves from theory to data and induction from data to theory, abduction includes moving back and forth, essentially combining induction and deduction (Suddaby, 2006). In abduction, the data is used to explore a phenomenon to identify themes and patterns, to generate new or modify existing theory and is often most
22
appropriate for a topic where there is extensive information in one context but limited in the specific context you intend to research (Saunders et al., 2016).
Content analysis can be conducted with an inductive or deductive approach (Elo & Kyngäs, 2008) and some authors even suggest a combination of the two (Lewis et al., 2013). In deductive content analysis, the codes are derived from earlier theory and the categories are determined before data collection (Elo & Kyngäs, 2008; Polit & Beck, 2004), whereas in the inductive content analysis, the codes are derived from the data (Elo & Kyngäs, 2008). The codes for this research were derived from earlier theories of brand gender, thus it followed a deductive content analysis approach. Deductive content analysis is often used when the researchers want to retest existing data in a new context (Catanzaro, 1988), which made it suitable for this research. Therefore, this research primarily follows an inductive approach, with a content analysis done deductively, and therefore resulting in an abductive overall approach (Saunders et al., 2016).
3.1.3 Research Design
In this research, a mixed-method design was chosen. Mixed method research involves combining both quantitative and qualitative methods to collect data (Clark et al., 2008). The fundamental reason for conducting mixed method research is that neither quantitative nor qualitative methods are adequate alone in capturing the topic in full detail (McKim, 2017). The research firstly performed a quantitative analysis of already published data through a content analysis, followed by qualitative interviews to explore perceptions and attitudes. The combination of a quantitative and qualitative approach provides a deeper and broader understanding of the studied phenomenon in comparison to research utilizing only one particular approach (Hurmerinta-Peltomäki & Nummela, 2006).
3.1.4 Research Strategy
This study will include several perspectives from participants about their brand gender perceptions and experiences considering LYFT, resulting in multiple meanings and understandings of social realities. Additionally, this research performed a content analysis, which was analysed quantitatively by dividing the data into defined criteria using theory from prior research.
23
phenomena from a first-person perspective in qualitative research (Smith & Osborn, 2003). The focus of IPA is to highlight the importance of the individuals’ experiences within the whole scope (Eatough & Smith, 2008). IPA is commonly used to investigate personal narratives and experiences. For example, it has been used to investigate women’s experiences concerning barriers to leadership in science and technology (Howe-Walsh & Turnbull, 2016), strategic self-presentation on social networking sites (Uski & Lampinen, 2016), and experiences of individuals supporting adults with intellectual disabilities (Handley & Hutchinson, 2013).
The participants of studies focusing on experiences, provide the researchers with life stories, significant events, and meaningful experiences relevant to the topic (Eatough & Smith, 2008). Which in this particular case is to understand the experiences and perceptions of snus consumers in Sweden. IPA acknowledges that experiences are subjective (Bush et al., 2016). Therefore, the authors recognise potential biases and assumptions during the research process. However, Dahlberg (2006) argues that biases and assumptions are not only negative but instead something that can be utilized to gain more valuable interpretations. Following IPA, the aim is to understand the world of the participants to examine it and be susceptible to new ideas and uncertainty (Eatough & Smith, 2008).
3.2 Method
3.2.1 Data Collection 3.2.1.1 Secondary Data
Secondary data refers to existing data originally collected for another purpose and thereafter reused in another research (Heaton, 2003). In this research, secondary data was used regarding branding, brand gender, brand love, and self-congruence from journal articles and books relevant to the subject. The data was primarily retrieved from the Primo database by Jönköping University Library and Google Scholar. To assist the progress of searching and ensuring to find relevant data the following search parameters were used in the process: Literature types: Books, journal articles, the internet; Keywords: brand gender, brand love, brand gender perceptions, LYFT, self-brand congruence, gender-congruence; Publication languages: English and Swedish. To ensure high-quality and
24
reliability of the used sources, the journal rankings on the academic journal guide by ABS and Scimago were considered, as well as the number of citations of the articles.
3.2.1.2 Primary Data
To correspond with the research strategy, the primary data for this study is both qualitative and quantitative. The quantitative data was gathered through a content analysis of LYFT’s Instagram posts. The purpose of the content analysis was to determine the brand gender that is communicated through the company’s social media. The quantitative data was collected to determine if the consumer perceptions correspond with the brand gender.
The qualitative data was collected through semi-structured interviews to gain comprehensive data that contributes to the purpose of this study. Interviews allow the researchers to gather significant, heterogeneous and comprehensive understandings of the phenomena to answer the research question (Longhurst, 2003). The choice of semi-structured interviews leads to a more open discussion about the topic, which could provide relevant data that might not be directly asked for (Saunders et al., 2016).
25
3.2.1.3 Sampling Approach
When finding the images to be used for the content analysis, it is important that the sample images appropriately represent the available data, and that the images are applicable to the research question (Rose, 2007). For the content analysis, the researchers evaluated the different social media pages of LYFT and discovered that the company was most active and had the most followers on Instagram. Furthermore, although LYFT is also active on Facebook, the images posted there are identical to those posted on Instagram, the only downside is that the number of images on Facebook is much lower, and therefore more difficult to properly analyse using content analysis. The sample size is highly dependable on the variation in the total population, where a low sample size can be representable if the variation is low (Rose, 2007). The total population of posts on the company’s Instagram page was 132 and a sample of 100 pictures was chosen since identical pictures were excluded, as they might skew the results due to the identical pictures being analysed multiple times.
Given the nature of this study, the participants for the interviews were selected through snowball sampling, which is a form of purposive sampling. This sampling technique is helpful when the members of the desired population are difficult to recognize (Saunders et al., 2016). To meet the objectives of this research and to ensure that the sample would contribute meaningful and rich data the individuals needed to fulfil some criteria. The requirements for participants included that they were regular users of snus and highly familiar with the LYFT brand. These criteria were set so that the individuals’ experience and familiarity with the context could provide sufficient data for answering the research question. The first participant was recruited from one of the author’s acquaintances using Facebook messenger and the following participants through the snowballing technique, resulting in eight participants (See Table 1).
3.2.2 Content Analysis
Content analysis is defined as “a method by which selected items of qualitative data are systematically converted to numerical data for analysis” (Collis & Hussey, 2014, p. 166). The content analysis serves well as the validity of findings when combined with other research methods (Krippendorff, 2018). In this study, the content analysis was used to determine the brand gender of LYFT as well as to strengthen the findings of the
semi-26
structured interviews. There are four steps to conducting a content analysis; finding your images, devising your categories for coding, coding the images, and analyzing the results (Rose, 2007). Content analysis is based on counting the frequency of particular visual elements in a well-defined sample of images, followed by an analysis of those frequencies (Rose, 2007). The process of conducting a content analysis is not purely quantitative, since all the stages require decisions about meaning and significance (Rose, 2007). Furthermore, content analysis alone is seldom able to support statements about the significance, effects or interpreted meaning of a domain of representation (Bell, 2001).
Coding involves attaching a set of descriptive labels or categories to the chosen images (Bell, 2001). The codes must be relevant to the research question, be based on relevant theory, and also be clearly defined to be replicable (Krippendorf, 2018). When applying the codes to the images, it is crucial that all images are carefully examined and that all relevant codes are applied to each image (Rose, 2007). The simplest way to count the codes is to produce either absolute or relative frequency counts, relative can for example be a percentage of the total number of images (Rose, 2007). Since the sample in this study was 100 images, the absolute frequency was easily converted to a relative frequency in the form of percentages.
3.2.3 Semi-structured Interviews
Semi-structured interviews were chosen to be able to have a flexible discussion without leaving out the key concepts of the topic. This is in line with the interpretive nature of the study, as semi-structured interviews allow the researchers to achieve a deep understanding of the phenomena (Collis & Hussey, 2014).
Because of the ongoing COVID-19 pandemic, the interviews were entirely conducted through online video calls on Zoom. The interviews were conducted in Swedish, allowing the participants to speak in their mother tongue, thus generating more extensive answers. The interviews were then transcribed in Swedish and the chosen quotes for this thesis were translated into English after analysing. The main consideration of a semi-structured interview is to make the participant feel comfortable (Longhurst, 2003). At the beginning of each interview session, the researchers and the topic were briefly introduced for the
27
participant, followed by some questions about the participant to create a positive and safe environment.
3.2.3.1 Interview Questions
The main purpose of the interview questions was to identify customers’ gender perceptions regarding LYFT, recognize their attitudes towards the brand and explore the relationship between these two. The interview questions were constructed using the seven core elements of brand love by Batra et al. (2012) and the six dimensions of brand gender by Ulrich et al. (2011). These were used to determine what brand gender perceptions the consumers have and if brand love could be recognized. The participants were not asked anything specifically about femininity or masculinity until towards the end of the interview, in order to get the most reliable results of the participant’s perceptions of the brand. The majority of the questions were open-ended questions, which refers to questions that require more developed answers including the participant’s own reflection (Collis & Hussey, 2014). Closed-ended questions that only require a “yes” or “no” answer (Collis & Hussey, 2014) were used when necessary to confirm a mutual understanding.
3.3 Data Analysis
3.3.1 Content Analysis
The codes for the content analysis were derived from a broad range of literature discussing brand gender perceptions and gender cues in images, as previously identified in the literature review (Alreck, 1994; Berlyne, 1977; Iijima et al., 2001; Kirkham, 1996; Lemm et al., 2005; Lieven et al., 2015). This resulted in a total of 16 codes whereof eight were feminine traits and the other eight masculine traits (see Table 3). Each image was analysed individually and applied to the relevant codes manually. Which is suggested by (Newbold et al., 2002) who claim that media content analysis is done more appropriately manually, than using computer software.
An important consideration when analysing and applying the codes is to secure intercoder reliability, which is the extent to which independent coders evaluate the content and reach the same conclusion (Lombard et al., 2010). Intercoder reliability is essential in content analysis for the data and interpretations to be considered valid (Lombard et al., 2010). A
28
common way to measure intercoder reliability is using Cohen’s Kappa test (Cohen, 1960; Lavrakas, 2008; McHugh, 2012), which has predominantly been used in content analysis (Krippendorff, 2004; Stemler, 2000). Cohen’s kappa measures the agreement between two raters, calculating the overall agreement while taking agreement by chance into consideration (Cohen, 1960; Lavrakas, 2008). The value for Kappa can be between 0 and 1, where 0 represents the amount of agreement that can be expected from random chance, and 1 represents perfect agreement between the raters (McHugh, 2012). A kappa value of 0.8 indicates 80% agreement, which is often recommended as the minimum agreement rate (Lavrakas, 2008; McHugh, 2012).
The kappa for this research was calculated in accordance with Stemler (2000). The formula for Cohen’s kappa coefficient is k = Pa-Pc /1-Pc, where Pa is the proportion of units on which the raters agree and Pc is the proportion of units for which agreement is expected by chance (Stemler, 2000). Pa is calculated by all items the raters agreed upon divided by the total number of instances, while Pc is calculated by multiplying the total agreement of rater A with the total agreement of rater B and multiplying the total disagreement of rater A with the total disagreement of rater B and adding those numbers together. Following Cohen’s Kappa formula, the kappa resulted in 0.88, which is acceptable since 0.81-1.00 is considered almost perfect (Landis & Koch, 1977).
3.3.2 Analysing Semi-structured Interviews
The aim of IPA is for the authors to recognize how the participants make sense of their experiences (Pietskiewicz & Smith, 2014). The analysis followed the steps recommended by Pietskiewicz and Smith (2014) for IPA. The process started with conducting and transcribing 4 interviews. The first step concerned individually and thoroughly reading and re-reading the transcripts while taking notes. Both authors’ notes were then discussed and compared carefully, to ensure that all important findings were included. The emerging themes were then identified and evaluated. The third step involved detecting links between the emerging themes and grouping them to produce a descriptive label for the superordinate themes (See Table 2). The themes were then assessed in relevance to the research question. Before conducting the last 4 interviews, the questions were adapted using the superordinate themes, and to detect new ones relevant to the topic. The same analysis process was applied to all 8 interviews, to gain the most accurate results. The last
