towards the advertisement and the purchase of the product”

(1)

“Green advertising: the effects of imagery, lettering and the moderating role of perceived environmental threat on lettering on purchase intentions and attitudes

towards the advertisement and the purchase of the product”

Supervisor:

Dr. J.I.M. De Groot

Co-assessor:

Prof. J. W. Bolderdijk

Author:

Arianna Melchiori

Msc. Marketing Management S3790274 – a.melchiori@student.rug.nl

January 13

^th

, 2020

University of Groningen

Faculty of Economics and Business

(2)

Acknowledgments

I would like to express my sincere gratitude to my supervisor Dr. Judith de Groot for her support, her advice and for the constructive feedback she provided me along the writing of this thesis. Moreover, I would like to thank my second supervisor Prof. J. W. Bolderdijk for taking the time to read and evaluate my thesis.

I wish to sincerely thank my parents and my family for their infinite support, their patience and for accepting my choices. I would like to thank my uncle Roberto for always being by my side and for his encouraging words.

A special thank goes to all the friends that I met during this thrilling academic year. I feel thankful to my

“Marketing Team” friends especially Claudia, Ludovica, Jordi, Lena, Giulia, Alberto and Tommaso. Thanks for the unforgettable memories and experiences that we shared together. Even the time spent in the UB was fun with you.

I would like to thank my friends Laura, Chiara and Veronica for being next to me when I needed them the most and for their precious advice. I would like to thank all of my friends for their support during this year in the bad days and for their enthusiasm in the happy ones.

Next, I would like to thank my patient cousin Marco for editing the pictures of the experimental conditions of this research.

Lastly, I would like to thank my thesis groupmates, in particular Nicole, for their ongoing help, their support and their helpful advice.

Thank you, dank je wel, grazie

(3)

Abstract

In the last decade, the overconsumption of goods has dramatically increased, leading to damages to the environment. The dissemination of information through the use of media and the promotion of green products, has resulted in consumers shifting their purchases of products towards ‘greener’ options. One common way to increase purchase intentions of green products is to adopt advertisements that emphasise the sustainable aspects of the products. To increase the effectiveness of such ‘green’ advertising, we examined in a two by two between-subject experimental design (N= 298) the effect of imagery (nature vs neutral) and lettering (upper- vs lower-case) on the purchase intention of a green product, and, the attitude toward the advertisement and toward the purchase of the product. Additionally, we examined the extent to which perceived environmental threat moderated the effect of lettering on the purchase intention and attitudes toward the advertisement and the purchase of the product. The results showed that an advertisement including a background of nature imagery was not more effective to change one’s purchase intention and attitudes toward the advertisement and the purchase of the product than a neutral background (Hypothesis 1). Furthermore, upper-case lettering was not more effective than lower-case lettering to change one’s purchase intention and attitudes towards advertisement and purchase of the product (Hypothesis 2).

The mean scores between the four experimental conditions showed not a lot of differences in purchase intention and attitudes toward the advertisement, and, a few differences between attitude toward the purchase of the product, implying the main effects were not qualified by an interaction effect. Indeed, the MANCOVA revealed that there was no significant interaction-effect between imagery and lettering on purchase intentions and attitudes (Hypothesis 3). Finally, the results showed that the more consumers perceived an environmental threat, the stronger their intention to purchase the green product and more favourable their attitudes toward the advertisement and the purchase of the green product (no specific hypothesis). However, environmental threat did not moderate the relationship between lettering and purchase intentions and attitude toward the advertisement and the purchase of the product (Hypothesis 4).

The findings and implications of the present results show how advertisers could use imagery and lettering in green marketing and advertising to foster the purchase of green products and to generate favourable attitudes.

Key words: green advertising, nature imagery, upper-case lettering, perceived environmental threat, pro- environmental behaviour

(6)

Introduction

The effects of climate change are negatively affecting nature and, as a consequence, our lives (National Geographic, 2019). In the last ten years, the overconsumption of goods has dramatically increased, leading to the depletion of natural resources and damages to the environment (Chen & Chai, 2010). Countries and governments realize the consequences that pollution has on our planet and thus, they have started working toward reducing the harmful impact of business activities and consumption (Joshi & Raman, 2015).

The diffusion of information about environmental issues through media and spokespersons such as Greta Thunberg has led people to become more concerned and aware about the environment and its importance (Kim, Malek, & Roberts, 2019). This flow of information generated a heightened sense of consumer responsibility (Haanpää, 2007).

The stronger sense of consumer responsibility has resulted in consumers' behaviours and purchasing intentions shifts towards sustainable options. As a consequence, companies are challenged with integrating environmental issues into their strategies (Nidumolu, Prahalad, & Rangaswami, 2009). Consumers are encouraged to change their daily behaviours towards an environmental-conscious way. For example, we are spurred to recycle, to use reusable products such as water bottles, shopping bags and, to take public transportation instead of the car, to protect our environment from global warming. This concern has led to

"sustainable development": the promotion of sustainability and all forms of development that minimize the damaging impact on the environment and society (Joshi & Raman, 2015).

Sustainable development nourishes the concept of green consumption: "an environmentally responsible consumption where consumers consider the environmental impact of purchasing, using, and disposing of various products, or using various green services" (Moisander, 2007). Therefore, the pivotal purpose of green consumption is to reduce consumption's environmental impact by the growth of products and services (Liobikiene & Bernatoniene, 2017). Past literature has observed that consumers can prevent or decrease environmental damage by purchasing green products (Joshi & Rahman, 2015). "Green" products refer to those goods that have a minimal environmental impact during their lifecycle (Joshi & Rahman, 2015).

The present research will, therefore, focus on the purchase of green products. Specifically, this study focuses on bamboo toothbrushes as a green product and, how companies can positively influence green purchasing intentions through advertising.

One popular way to increase purchase intentions of green products for companies is to use advertisements that emphasize the sustainable aspects of their business and products. Since a few years, it is common to find advertisements advocating green appeals and practices (Schmidt & Donsbach, 2012).

These advertisements are referred to as "green advertising", as they present a range of marketing activities that firms adopt to promote the environmentally friendly aspects of their business operations and the eco-

(7)

friendly product features (Leonidou et al., 2011). Furthermore, green advertising aims to show the company's commitment to pro-environmental efforts (Papadas, Avlonitis, & Carrigan, 2017). The primary purpose of green advertising, therefore, is to reduce uncertainty around promotions and to positively influence consumers' purchasing intentions about purchasing a "greener" option (Dangelico & Vocalelli, 2017).

The main objective of this research is to examine the effects of specific advertising elements that can be used together in green advertising to enhance the purchase intentions of green products and additionally, how these elements can positively influence consumers’ attitudes toward the advertisement and the purchase of the advertised product. The present study focuses on intentions and attitudes as they are referred to as good predictors of actual behaviour (Ajzen, 1991; Ajzen, 2005). Therefore, it is pivotal to understand how these concepts can be influenced to encourage the effectiveness of green advertising. More specifically, this study focuses on how green advertisements can use nature imagery (pleasant pictures of natural landscapes) and lettering (words written in capital or regular letters) as visual and verbal elements to promote the intention to purchase green products, and, attitudes toward the advertisement and toward purchasing the product.

To boost the consumption of green products, pictures are frequently used in advertisements because they show strong communication effects that will generate purchase intentions in general (Mitchell et al., 1986), and more specifically green purchase intentions (Hartmann et al., 2010). Past literature argues that nature images (i.e., pictures of pleasant nature landscapes) can be efficiently applied in advertisements because people innately prefer nature experiences compared to urban or neutral counterparts (Cacowski &

Nasar, 2003). Especially concerning the purchase of green products, using nature images highlights environmental product features, and therefore, it can positively influence cognitive processes and affective responses of consumers toward the advertised green products (Hartmann, Apaolaza & Eisend, 2016). Nature imagery is used in advertisements to trigger positive associations between the product advertised and brand and the nature picture itself (Hartmann & Ibañez, 2010). Due to this spill over effect, consumers are expected to exhibit positive attitudes towards the advertisement and the product. Furthermore, they are assumed to display higher intentions to purchase (Hartmann & Ibañez, 2010).

Furthermore, the present study examines also the extent to which lettering influences consumers' intentions to purchase green products and attitudes toward the advertisement and the purchase of the product. According to Berlyne (1974), upper-case lettering can capture one's attention and enhance the importance attached to words because upper-case is reserved for particular purposes (e.g. for bans "DO NOT SMOKE"). Therefore, when exposed to upper-case lettering, people are expected to sense an authoritative image that influences their words and actions, compared to its lower-case counterpart (Bowman et al., 2004;

Cialdini & Goldstein, 2004). As a consequence, we assume that consumers will be more persuaded to follow

(8)

This study intends to contribute to the green advertising area by investigating whether the main effects of imagery (nature vs neutral), lettering (upper- vs lower-case) and their interaction effect increase the purchase intentions of the advertised product and additionally, whether they positively affect the attitudes people hold about the advertisement and the purchase of the product. Namely, we want to examine whether consumers’ intention to purchase the green product increases and whether consumers generate more favourable associations between the advertisement and the product. Therefore, this research seeks to address the following questions:

(1) To what extent are nature images and lettering type effective in promoting consumers' intentions to purchase green products and additionally, in generating favourable attitudes toward the green advertisement and the advertised product?

In addition to the effects caused by the pairing of lettering type and imagery and their single main effects, we intend to examine whether the effect of lettering is stronger (e.g. participants' purchase intentions are increased, attitudes toward the advertisement and the purchase of the product are more favourable) when consumers are threatened by the negative consequences of climate change in case, they do not buy the advertised green product. We focus on this aspect due to the authoritative influence that upper-case (compared to lower-case) lettering holds and its power in arousing fear. Past research has observed that people experience higher levels of fear arousal when exposed to upper-case lettering, and thus, they pay more attention to the message (Bandura et al., 1982). We intend to arouse a sense of perceived environmental threat through the message claim of the advertisement that highlights the negative consequences that can happen when consumers do not buy the advertised green product. Previous research has observed that fear induced by awareness of environmental threats might motivate pro-environmental coping behaviour (e.g. environmental concern, recycling) (Hartmann, Apaolaza, D'Souza et al., 2014). This happens because people want to remove the perceived threat, and the best solution is to adopt coping behaviour (Hartmann, Apaolaza, D'Souza et al., 2014). As a consequence, we expect that the effect of lettering is enhanced when participants perceive a severe environmental threat. We assume that the higher the level of perceived threat on the authoritative effect of upper-case lettering, the higher the probability that consumers act in line with the advertising pro-environmental message. Hence, we will examine the following sub-research question:

(2) To what extent is the perceived threat about environmental pollution important to increase the effectiveness of upper-case lettering on green purchase intentions and additionally on attitudes toward the advertisement and the purchase of the product?

Overall, this research helps to understand better how advertisers could boost the consumption of green products and therefore, how we can prevent the environment from pollution when purchasing. This

(9)

study extends the knowledge acquired by past literature about imagery and lettering type on people's perceptions and subsequent behaviour intentional responses. The present research provides to contribute to the growing literature of green advertising by investigating the effects of imagery, lettering and their combination on purchase intentions of green products and attitudes towards the advertisement and the purchase of green products. Eventually, this study will provide new guidelines on how imagery and lettering can be used together to increase the effectiveness of green advertising in generating stronger intentions to purchase and more favourable attitudes toward the advertisement and the purchase of the product.

Furthermore, the current research examines the influence of perceived environmental threat on the effect of lettering as an additional element to increase the likelihood of purchase of green products and to generate favourable attitudes toward the advertisement and the purchase of the product.

(10)

Literature review

The increasing media’s interest in environmental issues has led people to get more concerned and aware of the environment and its resources (Kim, Malek & Roberts, 2019). Indeed, consumers perceive a growing responsibility in preserving and protecting the environment during their daily activities and behaviours (Haanpää, 2007). This concern gave rise to the concept of sustainable development, the promotion of all forms of development which minimise the damaging impact on the environment and society (Joshi & Raman, 2015).

Sustainable development boosts the concept of green consumption, a form of conscious consumption that aims to reduce consumption’s environmental impact (Liobikiene & Bernatoniene, 2017).

Joshi et al. (2015) argue that consumers are capable of preventing environmental damage by purchasing green products, goods that have a minimal environmental impact during their lifecycle. To increase the consumption of green products, companies invest in green advertisements that emphasise the sustainable sides of the business and products. Indeed, green advertising’s purpose is to reduce uncertainty about promotions and thus, to better influence purchasing intentions (Dangelico & Vocalelli, 2017).

This study aims to analyse how green advertising can boost green consumption to prevent the environment from pollution. More specifically, the present research focuses on the effects of imagery and lettering, as advertising elements, on the purchasing intentions of consumers toward green products and, their attitude toward the advertisement and the purchase of the product. We intend to analyse intentions rather than the actual behaviour as it is the best and most accurate proxy of behaviour (Ajzen, 1991). In support of this view, Ajzen and Fishbein (1970) argue that behavioural intentions are assumed to mediate overt behaviour, and therefore, intentions are precise behavioural estimates. Additionally, we examine attitudes, referred to as “a learned predisposition to respond in a consistently favourable or unfavourable manner in relation to some object” (Szmigin & Piacentini, 2018, p. 213). This definition suggests that attitudes are directly linked to behaviour, leading to a predisposition to act in a certain way. According to the Theory of Planned Behaviour (Ajzen, 2005), attitudes are an accurate predictor of whether or not a specific behaviour will happen. Indeed, behaviour only takes place after the formation of attitudes and intentions has accomplished. Therefore, we also investigate attitudes because they are a good estimator of actual behaviour as they influence the formation of intentions.

2.1 Using green advertising as a way to promote intentions to purchase green products

One common way to increase purchase intentions of green products for companies is to use advertisements that emphasise the sustainable aspects of their business and products. Green advertising relates to a range of marketing activities that firms adopt to promote the environmentally friendly sides of

(11)

their business operations and to show their commitment to environmental issues (Papadas, Avlonitis &

Carrigan, 2017). It is a method used by marketers to position their products as green, sustainable goods in consumers’ minds (Eren-Erdogmus et al., 2016); it allows to disseminate green image and to build environmental awareness among consumers (Grillo et al., 2008). Past literature sustains that green advertising can be considered as a promotional message that catches consumers’ needs and wishes related to the environment (Zinkhan & Carlson, 1995). This form of advertising has spread out rapidly in the last decade across different industries due to the increasing consumers’ attention to the environmental impact of their deeds (Leonidou, Leonidou, Palihawadana & Hultman, 2011). According to the International Trade Center (2017), 88% of American people reckon that it is vital to take care of the planet; 52% think that the government should protect the environment. Tanner et al. (2003) argue that this growing environmental consciousness may lead to favourable attitudes and as a consequence, increased intentions to purchase toward products and brands that are perceived as pro-environmental.

In order to meet consumers’ needs, companies are challenged with integrating environmental issues into their strategies (Nidumolu, Prahalad & Rangaswami, 2009). Research shows that pro-environmental promotions help companies to gain sustainable competitive advantage and to accomplish optimal performance (Leonidou et al., 2011). The primary purpose of green advertising, therefore, is to minimise uncertainty about promotions and better influence consumers’ purchasing decisions of green products

(Dangelico & Vocalelli, 2017).

2.2 Using green images in advertisements to generate purchase intentions and attitudes toward the advertisement and the purchase of the product

Advertisers frequently use pictures as they show strong communication effects that will generate purchase intention (Mitchell et al., 1986). To increase consumers’ purchase intentions and to generate favourable attitudes toward the advertisement and the purchase of the product, the present study focuses on nature imagery, pictures of pleasant nature landscapes.

Nature images are often used in advertisements because people innately prefer nature experiences compared to urban or neutral counterparts (Cacowski & Nasar, 2003). This kind of imagery can positively influence affective responses of consumers (Hartmann, Apaolaza & Eisend, 2016). Hartmann et al. (2016) argue that exposure to pleasant nature imagery leads to positive feelings similar to those experienced in contact with real nature. The higher the congruence between advertising-triggered feelings and authentic- nature experiences, the more positive the recipients’ affective responses (Hartmann, Apaolaza & Eisend, 2016). As a result, the advertisement will be more effective to encourage purchase intentions of green

(12)

products and to generate positive attitudes towards the advertisement and the purchase of the green product (Hartmann et al., 2010).

There are different reasons why using nature images evoke strong and positive emotional responses.

Firstly, nature pictures are extremely fractal: natural objects such as leaves and waves include self-similar patterns regardless of zoom of the brain (Mandelbrot, 1997). Therefore, nature images are easy to process for the human brain (Gigerenzer & Goldstein, 1996). Based on the Fluency Theory, stimuli that are fluent to process are judged as more familiar and thus, they elicit a more positive affective response than disfluent stimuli (Schwarz et al., 2009). As a result, the easier to process, the more likely the pictures induce positive feelings. Secondly, past literature suggests that nature imagery has several positive psychological and physical effects on individuals (Raanaas, Patil & Hartig, 2012). For example, people experience less stress and recover faster from mental and physical illness when exposed to natural pictures or when they look at a natural view rather than urban (Ulrich et al., 1991). Hence, these positive consequences of exposure to nature pictures lead people to prefer natural landscapes and natural images inherently.

Nature imagery is used in advertisements to trigger positive associations between the product advertised and brand and the nature picture itself (Hartmann & Ibañez, 2010). Due to this positive connection, consumers are expected to exhibit positive attitudes toward the advertisement and the product.

Furthermore, consumers are assumed to display higher intentions to purchase a green product (Hartmann

& Ibañez, 2010). Based on what has been discussed, the following hypothesis has been developed:

H1: Consumers exposed to a nature image (in an advertisement) will experience a stronger intention to buy a green product and generate more favourable attitudes toward the advertisement (H1alternative1) and the purchase of the product (H2alternative2) than those who are exposed to a neutral image.

2.3 Using lettering to influence intention to purchase green products and attitudes toward the advertisement and the purchase of the product

A way to enhance the promotional effects of an advertisement is the adoption of lettering (words are written in capital, regular or lower-case letters). According to Berlyne (1974), upper-case lettering can capture one’s attention and enhance the importance attached to the words because upper-case is reserved for particular purposes (e.g. for bans “DO NOT SMOKE”). When exposed to upper-case lettering, people experience higher levels of fear arousal, and thus, they pay more attention to the information of the message (Bandura et al., 1982). We are exposed to lower-case lettering more frequently: Jones and Mewhor (2004) found that the counts of lower-case letters in the New York times were 17 more than upper-case on average.

(13)

Hence, we are more familiar with the sight of lower-case lettering and refer to it as perceived friendliness (Perea & Panadero, 2014).

The attention of upper-case lettering can be related to perceived strength and authority (Mehrabian

& Russell, 1974). Therefore, when exposed to upper-case lettering, people are expected to sense an authoritative image combined with an authoritative tone that influences their words and actions (Bowman et al., 2004; Cialdini & Goldstein, 2004). For this reason, we assume that consumers will be more persuaded to follow the pro-environmental message of the advertisement and thus, to engage in coping behaviour (i.e., to purchase the green product), when the text is written in upper-case lettering than when it is in written in lower-case lettering. In addition to purchase intentions, we chose to observe the effects of lettering on attitudes toward the advertisement and the purchase of the product because according to the Theory of Planned Behaviour (Ajzen, 2005), behaviours are determined by intentions, which are influenced by attitudes. Therefore, by analysing attitudes we can get better insights about the actual behaviour of consumers. In line with the literature above, the following hypothesis has been developed:

H2: Participants exposed to upper-case lettering (in an advertisement) will show a stronger intention to buy a green product and experience more favourable attitudes toward the advertisement (H2alternative1) and the purchase of the product (H2alternative2) due to its authoritative effect than those who are exposed to lower-case lettering.

2.3 Using the interaction effect between nature imagery and lettering type on purchase intentions and attitudes toward the advertisement and the purchase of the product

Past research has shown the positive effects that nature imagery has on the formation of consumers

‘purchase intentions and attitudes (Ulrich at el., 1991; Hartmann et al., 2016; Cacowski & Nasar, 2003;

Raanaas et al., 2012). For example, Ulrich et al. (1991), demonstrated that people exposed to nature images recovered faster from physical and mental illnesses; Hartmann et al. (2016) showed that exposure to nature imagery leads to the formation of positive feelings. Furthermore, past literature suggests that people perceive an authoritative influence when exposed to upper-case lettering because it is less common to see than lower-case lettering (Jones & Mewhor, 2004; Bowman et al., 2004). Hence, people pay more attention to the underlying message of a claim written in upper-case lettering (Bandura et al., 1982). We expect to observe an interaction effect between nature imagery and upper-case lettering because both of them are congruent with and foster the concept of nature protection within the field of green advertising (Hartmann et al., 2016). Past research argues that when advertising elements are congruent, information transfers easily from the advertisement to the product and elicits favourable attitudes because the information is more accessible in memory (Rifon et al., 2004; Song et al., 2019). Therefore, as both the two advertising elements

(14)

go towards the same direction (i.e., protecting nature), we expect that their single effects reinforce each other when mixed together.

We assume that participants exposed to the interaction between nature imagery and upper-case lettering will hold strong purchasing intentions due to the influence of upper-case lettering and favourable attitudes toward the advertisement and the purchase of the product because of the nature imagery effects.

As a conclusion, the following hypothesis has been formulated:

H3: Consumers exposed to nature images and upper-case lettering will show a stronger intention to buy green products and more favourable attitude toward the advertisement (H3alternative1) and the purchase of the product (H3alternative2) than the other experimental conditions.

2.4 Using advertisements to promote green products: the interaction effect between lettering and perceived environmental threat

The media’s interest in environmental issues has led people to become more aware of the environment (Kim et al., 2019), resulting in a growing perceived responsibility in protecting natural resources (Haanpää, 2007). It is pivotal that people understand the consequences of consumption to prevent pollution.

A way to boost green consumption and thus, to minimise environmental damage is to threaten consumers about the consequences of an unconscious consumption. According to Hartmann et al. (2014), fear induced by awareness of environmental threats might motivate pro-environmental coping behaviour (e.g.

environmental concerns, recycling). Kao (2019) defined “ecological fear” as an emotional state in which people make judgements on the risk of climate change and worry becomes the prevalent emotional response.

According to the Protection Motivation Theory (Maddux & Rogers, 1983) people feel the need to remove the perceived threat and the best solution is to adopt coping behaviour, so to pursue the message claimed by the threat (Hartmann, Apaolaza, D’Souza et al., 2014). As a consequence, we expect that people exposed to fear arousing advertising messages are more likely to engage in pro-environmental behaviour such as green purchasing. The higher the level of perceived threat, the higher the probability that consumers follow the advertising pro-environmental message. The efficacy of a threat appeal depends on the level of triggered fear. Past literature has demonstrated that up to an optimal fear level, persuasive effects increase with rising fear (Hartmann et al., 2014).

Based on Davis’ (1995) research, environmental communications focused on threats increased behavioural intentions about natural resource conservation, recycling and green purchasing. Fear motivates coping behaviour (Janis, 1967), and it leads to aversive behavioural reactions seeking an absence of

(15)

punishment (Holbrook & O’Shaughnessy, 1984). Furthermore, fear arousal contributes to the appraisal of the severity of a threat appeal, influencing advertising acceptance and relative behaviour (Liu & Stout, 1987).

Based on what has been mentioned above, we assume that the level of perceived environmental threat moderates the effect of upper-case lettering on intentions to purchase and attitudes toward the advertisement and the purchase of the product. Namely, we expect that perceived environmental threat strengthens the effect of upper-case lettering. Hence, when consumers perceive a severe environmental threat, they are more likely to engage in coping behaviour (e.g. intention to purchase the green product, attitude toward the purchase of the product) when the advertisement message is written in upper-case lettering rather than lower-case lettering because of its authoritative tone (Mehrabian & Russell, 1974).

Specifically, we assume that the combination of the fear induced by the message claim and the fear triggered by the authoritative tone of upper-case lettering results in an “exponential growth in fear”. Namely, we expect that consumers’ worry for the environment increases when they perceive a threat and they see the message claim written in upper-case lettering compared to when they perceive a threat and the message is written in lower-case lettering. As a consequence, consumers would more likely to need to engage in coping behaviours to remove the perceived threat. Therefore, this assumption has been developed:

H4: The effect of upper-case lettering on purchase intentions and on attitudes toward the advertisement (H4alternative1)and the purchase of the product (H4alternative2) is stronger when participants perceive a strong environmental threat; the effect of lower-case lettering on purchase intentions will not be strengthened by perceived environmental threat.

The purpose of the current study is to examine how imagery (nature vs neutral), lettering and the interaction effect between these two variables affect consumers’ green purchasing intentions and attitudes towards the advertisement and the purchase of the product. Additionally, this research studies whether the perceived environmental threat has a positive moderating effect on lettering and hence, whether the effect of upper-case lettering on purchasing intentions is stronger when people feel worried about the negative consequences of pollution on the environment.

2.5 Conceptual model of the hypotheses

The conceptual model has been developed in line with the literature that has been provided in the previous paragraphs and with the present research purposes. The conceptual model is depicted in Figure 1.

(16)

Figure 1: Conceptual model

(17)

Methodology

The present research examined whether the effect of imagery (nature vs neutral), lettering (upper- vs lower-case) and the interaction effect between these two variables have a positive impact on consumer purchase intentions and attitude toward the advertisement and the purchase of the product. Furthermore, this study analysed whether one’s perceived environmental threat of a message influences the relationship between using (upper- versus lower-case) lettering and intentions to purchase the product.

3.1 Sample

The data of this study was gathered along with 348 participants. In line with the concept of convenience sampling (Malhotra, 2010), a questionnaire created with Qualtrics has been shared in the form of a link on social media such as Facebook, LinkedIn, Instagram, WhatsApp and through direct messages to recruit the sufficient number of participants (N= 300) due to a limited time available. Moreover, participants were asked to share the survey questionnaire with their contacts to increase the sample size. Convenience sampling technique was employed because it allows selecting “convenient respondents” that happen to be in the right place at the right time (Malhotra, 2010). Therefore, everyone can participate in the survey including relatives, friends and colleagues (Blumberg, Cooper & Schindler, 2014). This helps us to reach a sufficient number of respondents to increase the reliability of our results.

The research sample consisted of 63.1% of women (n= 188) and 36.9% of men (n= 110). The age of the participants differed from 18 to 66 years old (M= 26.87), which indicates that participants were relatively young. 85.9% of participants came from Italy, 1.3% of participants were Dutch, and 12.8% of participants came from other countries. On average, the educational attainment of respondents was a high-school degree (43%). Additionally, 47% of respondents gain a yearly income of less than €20.000. We excluded the participants that already had a bamboo toothbrush (n= 26) to avoid biased answers because of their pre- existing connections with the product (Mau, Silberer & Constien, 2008). Moreover, we did not include participants who stated they did not regularly use a toothbrush to clean their teeth (n= 5) as they would not be interested in the experiment product. Moreover, we filtered out participants (n=19) younger than 18 years old because they are not fully capable of making thoughtful decisions yet (Cauffman & Steinberg, 1995). As a consequence, the final sample consisted of 298 participants.

To increase the reliability of the outcomes and thus, to increase the probability of rejecting a false null hypothesis (Wilson, Van Voorhis & Morgan, 2007) we decided to achieve a suitable large number of participants. There are two main reasons why we chose to reach a large sample size. First, we have control over the number of sample size (Cohen, 1990; Cohen, 1992). Second, past research suggests that large samples depict more accurately the features of the population from which they are derived (Cronbach,

(18)

Gleser, Nanda, & Rajaratnam, 1972; Marcoulides, 1993) and thus, they are more reliable. As we conducted a factor analysis in our research, we decided to achieve a sample size of at least 300 participants and thus, 75 respondents per each condition, as it is suggested by Tabachnick and Fidell (1996). The data was collected from November 11th to November 17th.

3.2 Research design

This research adopted a 2 (imagery: nature versus neutral) x 2 (lettering: upper-case versus lower- case) between-subject experimental design (see Table 1). Participants were randomly assigned to one of the four experimental conditions to guarantee that the differences in the results were caused by the manipulation and not by external factors. We decided not to include a control group because we focused on the differences among the experimental conditions rather than comparing the outcome to a benchmark.

Based on the conceptual model, we defined the following dependent variable: intention to purchase (main dependent variable of the present study). We analysed intentions rather than the actual behaviour as it is the best and most accurate proxy of behaviour (Ajzen, 1991). In support of this view, Ajzen & Fishbein (1970) argued that behavioural intentions are assumed to mediate overt behaviour, and therefore, intentions are precise behavioural estimates. Additionally, we included attitudes toward the advertisement and purchase of the product as alternative dependent variables because attitudes are good predictors of whether or not a behaviour will take place (Ajzen, 2005). Therefore, including these alternative dependent variables help to get better insights about consumer behaviour in the area of green advertising.

Table 1: 2 x 2 between-subject design (N= 298)

N= 298 Upper-case lettering Lower-case lettering

Nature imagery n= 74 n= 76

Neutral imagery n= 72 n= 75

3.3 Pre-test

Pre-testing was applied to test the questionnaire on a small sample of respondents (N=32) to identify and then exclude potential problems (Malhotra, 2010). For the pre-test to be reliable, its sample group needed to be similar in terms of background characteristics, familiarity with the topic, attitudes and behaviours of interest (Malhotra, 2010). The purpose of the pre-test was to select a nature image considered to be pleasant and congruent to the product and the ad message according to our final sample. These criteria were important because pleasant nature images lead to more favourable attitudes (Hartmann et al., 2010) and congruent elements allow the information flow to become more accessible in memory, resulting in more favourable attitudes toward the target object (Rifon et al., 2004; Song et al., 2009). Therefore, through the pre-test, we assured that the natural image would satisfy these two criteria to increase the effectiveness of

(19)

the green advertisement and thus, to enhance the likelihood of increasing purchase intentions and generating favourable attitudes towards the advertisement and the purchase of the green product.

We contacted via WhatsApp 32 Italian people, and we provided them with a link to fill in the pre-test questionnaire. We chose to contact only Italian people because the test sample would be mainly composed by Italian participants and thus, we needed to customise the manipulation of the nature imagery to the specific cultural preferences (Hartmann et al., 2013). Respondents who did not complete all the survey questions were excluded from the final pre-test sample (n= 5). As a result, the final pre-test sample was composed of 27 participants, of which 66.66% were women. The age of respondents differed from 22 to 28 years old (M = 24.67).

Participants were shown nine different versions of the same advertisement of a bamboo toothbrush at random order (see Appendix I). Each advertisement included both product and text written in regular lettering (i.e., the initial letter of the first word after a dot was written in upper-case, while the other letters in lower-case). The advertisements differed only in the background picture. Based on Hartmann and Apaolaza-Ibáñez’s (2013) experiment, we selected pictures of a rocky desert and the urban landscape of Milan. According to these scholars, people should not prefer and find pleasant these two pictures. The other seven different images of nature included vistas of natural vegetation and landscapes: savannah grassland with trees, dolomites, a waterfall, a stretch of Mediterranean coastline, a lake, a pine forest and a vineyard.

Besides the savannah and rocky desert, all the other landscapes were Italian. Past research argues that people tend to favour landscapes they are familiar with (Hartmann et al., 2010). Therefore, we selected Italian places because the majority of our test’s participants were Italian (85.9%).

We used three items to measure whether participants found the nature image to be pleasant and the overall advertisement congruent with the underlying message of “nature protection”. Namely, “I think the background picture of the advertisement is pleasant”, “I think the bamboo toothbrush fits with the background” and, “I think the message included in the ad fits with the background”. All these items were measured on 7-point Likert scale ranging from (1) strongly disagree to (7) strongly agree. Furthermore, to make sure that the final sample would understand the purpose of the advertising message, we asked participants what the advertisement claim was suggesting.

The pre-test showed that respondents found the most aligned and most pleasant the nature image depicting a waterfall (M= 5.74; SD= 1.03). This result is supported by Hartmann et al. (2010), as they suggest that on average people find more pleasant pictures of water or moist vegetation. Moreover, all of the respondents understood that the advertising message fostered the consumption of green products, specifically the bamboo toothbrush, to reduce the environmental impact of plastic substitutes.

(20)

3.4 Materials

The advertisements for the main study were created with Adobe Photoshop (see Appendix II).

3.4.1.Product: a bamboo toothbrush

We decided to base the green product of the experiment on a bamboo toothbrush. A bamboo toothbrush is a product that everyone can be familiar with due to its daily usage. Furthermore, this product is not expensive, and it entails minimal consumers’ effort before purchase because it does not have a significant impact on their lifestyle (Szmigin & Piacentini, 2018). A bamboo toothbrush is not a significant investment and therefore, consumers are low involved with the purchase. As a consequence, people do not have to overthink whether they would buy the product (Szmigin & Piacentini, 2018).

To enhance the reliability degree of the results, we decided to create a fictitious brand

“Woodoobrush” rather than an existing brand of the bamboo toothbrush to make sure that participants did not have pre-existing connections with the product in their minds. Hence, we avoided pre-existing biases (Mau, Silberer & Constien, 2008).

3.4.2. Congruence among the advertising elements to foster the concept of “nature protection”

Congruence takes place when advertising elements are thematically matched (e.g. the image of a natural landscape or green background and a pro-environmental text). It is the extent to which a component helps to identify the related information displayed by the advertisement (Song et al., 2019). Congruence allows information to transfer easily from the advertisement to the brand and product. As a consequence, the brand and product yield a high recall rate among customers, thereby eliciting favourable attitudes as information is more accessible to the viewer’s memory (Rifon et al., 2004; Song et al., 2019).

In the case of green advertising, the emotions elicited by the advertisement are conditioned by the relationship of the nature image with the message of “nature protection” and the promoted environmental features (Hartmann et al., 2016). This relationship between image and message claim affects consumers’

affective responses (Hartmann et al., 2016). Therefore, the higher the congruence between the advertising elements and the underlying message of nature protection, the more positive consumers’ affective responses and thus, the more effective the advertisement in increasing consumers’ intention to purchase the advertised product and in generating favourable attitudes toward the advertisement. For example, Grimmer et al. (2014) observed in their study that consumers noticed and responded more favourably to the green appeal advertisement than to the cost-saving one, as the latter provided incongruent information with consumers’ pre-existing schema, that is that green products were expensive. As a consequence, we decided to design an advertisement whose elements were thematically matched towards the concept of “nature

(21)

protection”. The purpose was to generate more favourable attitudes toward the overall advertisement and to improve the attentional skills of participants.

3.4.3. Manipulations of the independent variables: imagery and lettering

The independent variable imagery (i.e., type of image for the background) was manipulated by changing the background picture: the “nature condition” displayed a waterfall landscape while the “neutral condition” displayed a green background.

Based on the results of the pre-test, we chose the picture of a waterfall for the nature condition. In support of the pre-test outcome, past research suggests that advertisements including pictures of clear water and lush green vegetation evoke the most favourable responses and thus, lead to more favourable attitudes toward the advertisement (Hartmann et al., 2010). Furthermore, humans prefer scenes of water or moist vegetation than sceneries which are more arid such as deserts (Balling & Falk, 1982; Lyons, 1983; Ulrich, 1981). Past literature argues that we may prefer these kinds of landscapes because they present specific characteristics that were adequate for survival and reproduction than others (Hartmann et al., 2010).

Moreover, pictures of familiar landscapes lead to more favourable emotional responses and enhanced attitudes towards the advertisement (Hartmann et al., 2010). To further increase the familiarity of the landscapes, we chose a typical Italian landscape rather than exotic vistas because the majority of participants were Italian (85.9%). Additionally, the water in the background of the advertisement can be considered to be congruent with the bamboo toothbrush as people need water to brush their teeth. Congruent elements help the transfer of information and thus, they elicit more favourable attitudes toward the target object as information is more accessible in memory (Rifon et al., 2004; Song et al., 2019).

For the neutral condition, we chose green as a background colour because of its underlying connection with nature (Moller et al., 2009) and consequent association with feelings of security and with the outdoors (Kaya & Epps, 2004; Clarke & Costall, 2007). People usually are attracted to colours associated with salient objects that elicit positive reactions (e.g. greens with positively valued grass fields and trees) and repulsed by colours linked to objects that elicit adverse reactions (e.g. browns with rotting food) (Palmer et al., 2010). We chose the green colour because it is aligned with the purpose of the advertisement (i.e., nature protection, green product) and because it is linked with positive natural elements and thus, it can boost the creation of positive attitudes (Palmer et al., 2010). Second, we were interested in analysing the differences between nature imagery and neutral imagery. Specifically, both types of imagery boost the concept of protecting nature (Moller et al., 2009; Hartmann et al., 2016). Therefore, we can investigate whether the specific natural elements included in a background picture are pivotal to generate more favourable attitudes toward the advertisement and product rather than a general green background itself.

(22)

All of the four conditions displayed the same product and the same message (see Appendix II). To better improve the attentional skills of recipients, we decided to position the product (bamboo toothbrush) on the left and the text on the right of the advertisement, according to the Hemispheric Lateralization Theory (Fennis et al., 2016). This theory suggests that the right hemisphere of the human brain is specialised in holistic and impressionistic processing, while the left one is specialised in data-driven and feature analysis (Fennis et al., 2016). According to Fennis et al. (2016), the location in the visual field determines the hemisphere where the information is processed. Therefore, placement on the right activates the left hemisphere while placement on the left activates the right brain hemisphere. As a consequence, a pictorial advertisement benefits the most from being placed on the left side as it will activate more the right hemisphere.

The manipulation of the second variable (lettering) was based on the study of Song et al. (2019). We manipulated the variable on two levels: lower-case and upper-case. To avoid any potential confounding effects, the total length of the claim was the same, either when the text was written in lower- or upper-case (see Appendix II). We chose Franklin Gothic font because this typeface is widely used in advertising (Song &

Luximon, 2019) and thus, it helps to reduce any possible confounding with pre-existing biases and opinions people hold about green advertising. Moreover, to catch the attention of participants, we set the text in white because this colour created contrast with the background of both the nature and the neutral condition (Song et al., 2019).

One of the purposes of this study was to examine the effect of perceived environmental threat paired with lettering. To trigger threat and concern for the environment and climate change, we adopted a loss message framing, highlighting the negative consequences that can happen to the environment if consumers do not buy the green product advertised. According to past literature, environmental communications focused on threats increase behavioural intentions about natural resources conservation, recycling and green purchasing (Davis et al., 1995). We chose not to manipulate perceived threat but to measure it after participants were exposed to the advertisement because we wanted to observe whether it had an influence on the lettering effect.

Our experiment was based on a low-involving product. According to Andreu et al. (2015), when consumers are not highly involved in a purchase, they do not have the effort to scrutinise the advertisement elements. Instead, they rely on peripherical cues such as images and heuristics. As a consequence, in these cases, it is best to adopt an emotional advertising appeal. That is, the advertisement needs to influence consumer’s feelings rather than informing. Therefore, the message claim did not include any statistical data about pollution, and it was not supported by a credible source (Wood et al., 1985). We did not include strong quality arguments so the message could fit with the emotional appeal of the advertisement.

(23)

3.5 Procedure

We created the survey questionnaire with Qualtrics, and we translated it in English and in Italian to reach more respondents (participants could choose the survey language at its beginning). At first, participants read an introduction cover in which it was explained the purpose of the survey. Namely, the research goal was to investigate how consumers perceive an advertisement of a green product.

To refine the sample number and thus, to avoid biased answers, participants had to answer whether they already had a bamboo toothbrush. If they already had it, the survey would end. Moreover, they were asked whether they regularly use a toothbrush to clean their teeth. If they answered “no”, the survey would end. To all the respondents that were not excluded from the survey after the first two questions, one of the four experimental conditions was shown (see Appendix II).

After the participants were being exposed to the advertisement (i.e., experimental condition), we measured the dependent variables. Participants were asked about their intentions to purchase the bamboo toothbrush and questions about their attitudes toward the advertisement and the purchase of the green product. Next, they had to answer to what extent they perceived the advertising message to be environmentally threatening. After this, we tested whether participants paid attention when observing the advertisement by asking three questions.

Finally, participants were asked about their demographical information including age, gender, country of origin, education attainment and yearly income to obtain a detailed understanding of the sample characteristics. At the end of the survey, participants were thanked for their participation and asked to fill in their email in case they wanted to have a chance to win the €20 Amazon gift card.

3.6 Measures

Participants had to fill in questions regarding their intention to purchase the product, their attitudes towards the advertisement and the purchase of the product, and whether they felt threatened for the environment.

Purchase intention of a bamboo toothbrush: four items were used to measure the intention to purchase the green product, adapted from Ajzen (1991). Namely “I intend to buy a bamboo toothbrush when I need to purchase a new toothbrush”, “I am willing to buy a bamboo toothbrush in the future”, “I would consider buying the advertised bamboo toothbrush” and “I would purchase the advertised “green” version of a toothbrush rather than the non-recyclable substitutes in the future”. All these items were measured on 7-point Likert scale ranging from (1) strongly disagree to (7) strongly agree.

Attitude toward the advertisement: this variable was measured along with four items on a 7-point semantic differential scale, adapted from Andreu et al. (2015) and Decrop (2007). Namely, “To what extent

(24)

do you believe the advertisement is”: Bad (1) – Good (7), Unpleasant (1) – Pleasant (7), Unfavorable (1) – Favorable (7), Believable (1) – Unbelievable (7).

Attitude toward the purchase of the product: this variable was measured along with four items on a 7-point semantic differential scale, adapted from Andreu et al. (2015) and Decrop (2007). Namely, “How would you describe the bamboo toothbrush?”: Bad (1) – Good (7), Unpleasant (1) – Pleasant (7), Unfavorable (1) – Favorable (7), Believable (1) – Unbelievable (7).

Perceived environmental threat: one item was used to measure the perceived environmental threat aroused from the advertisement message, adapted from Hartmann et al. (2010). Namely, “To what extent do you perceive the advertising message to be environmentally threatening?”, measured on a scale ranging from (1) not at all, (2) slightly, (3) somewhat, (4) very much to (5) extremely.

3.6.1 Manipulation check of the independent variables

We applied manipulation checks to ensure that the effects resulting from the experimental conditions were due to the manipulation of the independent variables (i.e., nature imagery versus neutral imagery and lower-case versus upper-case lettering) and not because of other external factors and confounding variables. Moreover, we applied the manipulation check of the nature imagery to assess whether participants paid enough attention to the questions and thus, to check whether they understood the advertisement.

Type of imagery: one item was used to measure the effectiveness of the manipulation of the nature image. Namely, “To what extent do you associate the background of the advertisement with an image including natural elements?”, measured on a 7-point Likert scale ranging from (1) strongly disagree to (7) strongly agree.

Type of lettering: two items were used to measure the effectiveness of the manipulation of upper- case lettering. Namely, “I think that the use of letter type capitalisation (upper- or lower-case) in the ad is aligned with the message it entailed” and, “I think that the use of letter type capitalisation (upper- or lower- case) in the ad is threatening”, both measured on a 7-point Likert scale ranging from (1) strongly disagree to (7) strongly agree.

3.7 Data Analysis

The data of the survey were analyzed through SPSS 23, a statistical software designed by IBM. We used descriptive statistics to report means and standard deviations of relevant variables. We conducted factor analysis, specifically the Principal Component Analysis, to summarize the variation in the dataset and

(25)

to show congruence among the variables. Then, to study the reliability of the constructs, we examined the Cronbach’s Alpha. Furthermore, we conducted a two-way MANCOVA to study the effects of the hypotheses 1, 2 and 3. Finally, we adopted PROCESS model 1 of Hayes (2009) to analyse the effect of perceived environmental threat on lettering. For all the statistical tests that have been conducted, we applied a confidence level of 95% and thus, an alpha level of .05.

3.7.1 Manipulation checks

We conducted Independent Sample t-tests to understand whether participants perceived the manipulations (i.e., imagery and lettering). The two independent sample t-tests were performed to check whether the manipulations were successful. The t-tests compared the means of the groups between (1) the nature imagery condition and neutral condition, and, (2) the upper-case lettering condition and lower-case lettering condition. The purpose of these tests was to determine if a statistically significant difference was present between the samples.

3.7.2 Hypothesis testing

To test hypotheses 1, 2 and the interaction effect of hypothesis 3, a two-way MANCOVA was conducted as the variable “perceived environmental threat” was just measured and not manipulated. We chose to include perceived environmental threat as a covariate because although we could not control it, the linear regression analysis revealed that this confounding variable had a significant result for each dependent variable (Purchase intention: F(1, 297)= 10.639, p< .001; Attitude towards the advertisement: F(1,297)=

10.443, p< .001; Attitude towards the purchase of the product: F(1,297)= 9.941, p< .002). Therefore, perceived environmental threat could significantly predict each dependent variable, and as a consequence, it reduced the total variance explained by the other independent variables (i.e., imagery and lettering). The results of the linear regression are supported by Hartmann et al. (2014), as they suggest that fear induced by awareness of environmental threat may boost pro-environmental copying behaviour. Therefore, the higher the level of consumers’ perceived environmental threat, the more likely consumers engage in pro- environmental behaviour such as green purchasing. The two-way MANCOVA examines whether participants of the nature imagery condition show more or less favourable intention to purchase the green product and attitudes towards the advertisement and the purchase of the product than participants of the neutral imagery condition (hypothesis 1). Moreover, the two-way MANCOVA indicates whether respondents of the upper-case lettering condition show more or less favourable intention to purchase the green product and attitudes towards the advertisement and the purchase of the product than participants of the lower-case lettering condition (H2). This test is also applied to verify the interaction effect between the two independent variables (i.e., imagery and lettering) (H3). Finally, the PROCESS model number 1 (i.e., simple moderation) of Hayes (2009) was employed to test whether the control variable “perceived environmental threat”

(26)

influenced the effect of the lettering variable on the intention to purchase and attitudes towards the advertisement and the purchase of the product (H4).

3.7.3 Parametric tests assumptions

To check whether it was possible to perform the Independent Sample t-tests and the two-way MANCOVA and thus, to get reliable results, the data needed to meet six assumptions of the MANCOVA analysis.

The first assumption states that the scale of measurement needs to be continuous or ordinal (Statistics Solutions, 2019). All of the three dependent variables that we analyzed were measured on a 7- point Likert scale; therefore, this assumption was met.

Second, the data needs to be collected from a representative, randomly selected population, and each observation needs to be independent of one another (Statistics Solutions, 2019). We met this condition as participants were randomly assigned without replacement to the four experimental conditions.

The third assumption assumes that the values of the dependent variables are normally distributed, resulting in the plot of a bell-shaped curve (Statistics Solutions, 2019). This assumption is supported by the fourth condition, which states that the sample needs to be relatively large. The normality assumption was checked through a Shapiro-Wilk test as the sample was smaller than 2000. The test and the visual inspection of the histograms and Q-Q plots (see Appendix III) revealed that the dependent variables were not normally distributed as the p-values were significant for all the four groups (nature image p<.01; neutral image p<.01;

upper-case lettering p<.01; lower-case lettering p<.01). Even though the normality assumption was violated, past research argues that the two-way MANCOVA remains a robust statistical procedure for the violation of normal distribution (Blanca et al., 2017). Therefore, we kept conducting this analysis.

Furthermore, there should not be any significant outlier among values (e.g. point that significantly differs from the other observations). We met this assumption as we did not detect any outlier in the dataset through the use of Boxplots when the multiplier of IQR was 3. Nevertheless, three observations differed from the average at a multiplier of 1.5 IQR’s for the “Purchase intention” dependent variable (see Appendix III). As the 1.5 IQR’s is not always a good estimator of outliers (Hoaglin & Iglewicz, 1987), we chose to keep these observations in the dataset.

Next, we checked whether we met the last assumption of homogeneity of variance. This condition is achieved when the standard deviations of samples are approximately equal. For the Independent sample t- test by looking at Levene’s test, we could understand that this assumption was not met for the imagery manipulation (manipulation check question as DV and imagery as IV (1= nature; 0= neutral). Levene’s test showed p<.05, and thus, the manipulation check question groups do not have equal variances. For the lettering type manipulation check, the two manipulation check questions were used as DV and the lettering type group as IV (1= upper-case; 0= lower-case). Based on Levene’s test results, we could see that for both

(27)

manipulation check questions p>.05 (specifically, p= .819; p= .083). Therefore, there were equal variances between the two groups, and thus, the homogeneity assumption was met for lettering. In line with the assumptions that have been mentioned and the results achieved, as the homogeneity assumption was not fully met, Welch’s test would be used for the nature imagery manipulation check. For the MANCOVA analysis, the Levene’s test was performed to check whether the variance was homogeneous for the combination of the groups of the independent variables (i.e., imagery and lettering). The analysis revealed there was homogeneity of variance as F(3, 294)= 0.253; p= .859. Therefore, the assumption of homogeneity of variance for MANCOVA was satisfied.

Finally, we checked whether there was homogeneity of regression slopes for the MANCOVA analysis.

Namely, whether the relationship between the covariate and each dependent variable was equal in each group of the independent variables (Statistics Solutions, 2019). For all the three dependent variables (i.e., intention to purchase, attitude toward the advertisement, attitude toward the purchase of the product) the analysis revealed no significant interaction effects between the covariate “perceived environmental threat”

and the independent variables imagery and lettering as p-values were lower than 0.05. Precisely, for purchase intentions: lettering and environmental threat p=.138, imagery and environmental threat p=.958;

for attitude toward the advertisement: lettering and environmental threat p=.581, imagery and environmental threat p=.973; for attitude toward the purchase of the product: lettering and environmental threat p=.123, imagery and environmental threat p=.297.

towards the advertisement and the purchase of the product”

“Green advertising: the effects of imagery, lettering and the moderating role of perceived environmental threat on lettering on purchase intentions and attitudes