1
Measuring the effectiveness of scarcity and
message framing in the context of banner
advertising
Adrian Jasinski 11735910 June 21, 2018
Master Thesis MSc Business Administration Digital Business track
University of Amsterdam
2
Statement of Originality
This document is written by Student Adrian Jasinski who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
TABLE OF CONTENT
ABSTRACT ... 6 1. Introduction ... 7 2. Literature Review ... 11 2.1 Scarcity ... 112.1.1. Psychological background - commodity and reactance theories... 11
2.1.2. Scarcity and message framing - types overview ... 12
2.1.3. Scarcity – demand- and supply-generated... 14
2.1.4. Scarcity – time limitation ... 15
2.1.5. Message framing ... 18
2.1.6. Scarcity – mediating effect of competition ... 18
2.1.7. Scarcity – online ... 19
2.2. Banner ads – overview ... 19
2.3. Indicating the gap and formulating hypotheses ... 21
3. Methodology ... 28 3.1 Method... 28 3.2 Design ... 29 3.3 Measures ... 31 3.4 Pre-test ... 32 3.5 Procedure ... 33 4. Results ... 37
4.1 Sample and data collection ... 37
4.2 Preparing data ... 39
4.3 Reliability and validity analysis ... 39
4.3 Descriptives and Correlations ... 40
4.4 Manipulation checks... 43 4.5 Hypotheses testing ... 47 4.5.1 Testing hypothesis 1 ... 47 4.5.2 Testing hypotheses 2,4,6 ... 48 4.5.3 Testing hypotheses 3,5,7 ... 51 4.5.4 Summary ... 54 5. Discussion ... 56
5.1 Limitations and additional insight ... 57
5.2 Theoretical implications ... 60
5.3 Managerial implications ... 62
4
6. Conclusions ... 65
References ... 66
Appendix A – survey transcript from Qualtrics ... 72
Appendix B –Summary Table of descriptive statistics and bivariate tests ... 78
Appendix C – factor analyses tables ... 79
Appendix D – banner ads ... 80
Appendix E – websites with banner ads ... 85
TABLE OF FIGURES
Figure 2.3.1 Research model 1 ... 27Figure 2.3.2 Research model 2 ... 27
Table 3.2.1 –Summary of conditions ... 30
Table 3.5.1- Measurement items and Constructs ... 34
Table 4.1.1 – Sample demographics ... 38
Figure 4.1.2 – Sample population pyramid ... 38
Figure 4.3.1 – Reasons for not clicking... 41
Table 4.2.1- Spearman correlation matrix ... 42
Table 4.4.1 - Results of t-test and Descriptive Statistics for manipulation checks of perceived scarcity... 43
Table 4.4.2- Average perceived scarcity scores per condition ... 44
Table 4.4.3 - Results of Chi-square Test and Descriptive Statistics for scarcity type by condition ... 45
Table 4.4.4 -Results of Chi-square Test and Descriptive Statistics framing type by condition ... 46
Table 4.4.5 - Results of t-test and Descriptive Statistics for manipulation checks of perceived time limitation ... 47
Table 4.5.1.1 - ANOVA for the scarcity conditions on click intentions ... 48
Table 4.5.1.2 - ANOVA for the scarcity conditions on purchase intentions ... 48
Table 4.5.2.1 Two-way ANOVA for type of scarcity and time limitation on click intentions 49 Table 4.5.2.2 Two-way ANOVA for type of scarcity and time limitation on purchase intentions ... 49
Table 4.5.2.3 – Summary table of PROCESS model 7 moderated mediation of type of scarcity on click intensions. ... 50
5
Table 4.5.2.4 – Summary table of PROCESS model 7 moderated mediation of type of scarcity on purchase intensions... 51 Table 4.5.3.1 Two-way ANOVA for type of framing and time limitation on click intentions 52 Table 4.5.3.2 Two-way ANOVA for type of framing and time limitation on purchase
intentions ... 52 Table 4.5.3.3 – Summary table of PROCESS model 7 moderated mediation of type of
framing on click intensions. ... 53 Table 4.5.3.4 – Summary table of PROCESS model 7 moderated mediation of type of
framing on purchase intensions. ... 54 Table 4.5.4.1 – Summary of hypotheses tested ... 54
6
ABSTRACT
Scarcity appeals are common persuasion tactic used by marketers to create the feeling of urgency and increase sales. Since companies are allocating more and more of their marketing budgets in banner advertising every year, the objective of this research was to investigate whether scarcity appeals, and message framing increase click intentions and purchase intentions in context of banner advertisements. It was anticipated that this happens because people feel competition and they are acting in aversion to expected loss that they might encounter. Moreover, it was hypothesized that this effect is stronger when time limitation is present. Thus, experiment was designed. 352 respondents acquired through online platform - MTurk took part in this experiment. The results, surprisingly, were in contrary to what was expected and suggested that consumers do not pay attention to online banner advertising with scarcity appealand that they have a negative attitude toward it. These finding were in accordance with the banner blindness concept. Hence, marketing managers should carefully consider using banner advertising as one of the online channels, because most of the advertisements even if displayed on the screen are simply disregarded by consumers.
Keywords: scarcity, framing, quantity limited scarcity, time limited scarcity, banner advertising, banner blindness.
7
1. Introduction
Many airline companies tell their customers that only 3 seats are left at their flight, the travel agents stress that a lot of people already booked their holidays, and fashion companies announce that their new collections are limited. Especially, we see more and more banners ads with similar claims and scarcity appeals in behavioral retargeting situation i.e. when online advertising is targeted to consumers based on their previous Internet actions. Why? Because, as commodity theory states, the restricted availability of products increases their values (Brock, 1968). This is why commodity theory was applied to marketing as concept of scarcity (Lynn, 1991).
Scarcity is used in marketing as a persuasion tactic. By creating the sense of urgency marketers want to convince people to make a purchase. They know that exposing people to scarcity appeals may trigger their emotional reactance, decrease their critical judgement and make them feel that value of the promoted product is higher (Cialdini, 1993), or it may put a threat on them, and increase their attention to the product (Brehm, 1966). In either way, their purchase intentions usually increase.
Nevertheless, the scarcity is not applied to marketing communication in just one standard form. There are many different types of scarcity. Verhallen (1982) make a distinction between scarcity generated by high demand or generated by limited supply. Both of those types of scarcity were later categorized by Gierl and Huettl (2010) as Limited Quantity Scarcity. These researchers also described Time Limited Scarcity as scarcity generated by imposing limited time to buy. These different types of scarcity are used to increase the effectiveness of marketing among different channels ranging from billboards, through tv adverting to most popular right now – online advertising.
The recent years have been a time of rapid digital transformation in the marketing landscape. Remarkably, marketers followed the prevailing direction of change in attitudes
8
towards planning their advertising spending. Companies worldwide shifted their media budgets from traditional channels like television, radio or print to online channels. This is why more and more companies started using scarcity to drive conversions online.
The highly visible presence of scarcity appeals in advertising led to appearance of considerable amount of literature. Researchers examined, among the others, the moderating effects of product type (Ku, Kuo, Yang, & Chung, 2013), motivational orientations (Ku, Kuo, & Kuo, 2012), and message framing (Roy & Sharma, 2015). None of the existing publications, however, has as yet examined whether using scarcity appeals, so common in offline marketing, could improve performance of online advertising, especially banner advertising.
Banner advertising is among the most widely used channels of online advertising, playing an important role in online marketing strategies. It is expected to retain this position, with estimated spending on ads reaching almost 48 billion USD in 2018 in the United States of America alone (eMarketer 2017). However, taking into consideration the finding that only 39% of US internet users trust banner ads while making a purchase decision whereas 82% trust print ads and 80% trust TV ads (MarketingSherpa, 2016), it is intriguing to examine the impact of different combinations of scarcity appeals and see whether their influence on click intentions and online purchases is actually positive or not.
Although digital marketing is a new field not only for practitioners but also for academics, the number or research papers on this topic has been steadily increasing over the past few years. However, to date, the concept of scarcity in academic literature has not been applied to the banner advertising. There is still a need for addressing the effects of the presence of scarcity appeals on the effectiveness of banner advertisements measured by click-through rate and online purchase intention.
9
Thus, the central question of this thesis is: Can using different types of scarcity and message framing on banner advertisements increase consumers’ purchase and click intentions?
First, it is examined whether just the presence of scarcity on banner advertising can result in higher click intentions and purchase intentions. Secondly, it is studied which type of scarcity, demand-generated or supply-generated works better. Regardless of the progress in investigating effects of these two types of scarcity, the results of studies conducted so far (Gierl and Huettl 2010, Ku, Kuo, & Kuo, 2012, Ku, Kuo, Yang, & Chung, 2013) cannot be easily generalized to the banner advertising. Further, this study explores the influence of message framing placed on banner advertisements on their effectiveness. Message framing is a technique of expressing information as either positive, which indicates a gain, or negative, which indicates a loss (Tversky and Kahneman, 1981). Although many banner advertisements on the Internet display claims such as “Do not lose your chance and click here!”, effects of framing used in banner advertising are yet unknown. It is especially worth investigating since the results of previous studies have proved inconclusive (Ganzach, Karsahi, 1995; Yoo , 2011). Next, following the claim of Cialdini (2008) that people want an item when it scarce and they want it even more when they are in competition for it, this study inspects if competition can mediate the relationship among scarcity, message framing, click intentions and purchase intentions. Previous studies proved the importance of competition in the pursuit of limited resources (Aggarwal et al. 2011) but never in the banner advertising context. Lastly, taking into account the fact that time pressure might generate anxiety which in turns stimulate consumers to act faster (Maule, Hockey, and Bdzola, 2000), it is analyzed if signaling time limitation increases feeling of competition.
Considering high amounts of money that companies spend on banner advertising this research provides imperative insights for managers and shows how different banner
10
advertisements affect consumers online behavior, how consumers perceive banner advertising and what managers may do to improve their campaign results.
This paper is divided into five sections. It starts with a review of existing literature on scarcity and banner advertising. The literature review ends with indication of research gap and presentation of hypotheses and conceptual models. The subsequent section describes the research methodology. In the next chapter, results are introduced. Following, the discussion section presents implications, both theoretical and managerial, limitations and directions for future research. Lastly, in the final section conclusions are drawn
11
2. Literature Review
This section explains theoretical background of this research starting at the psychological theories behind the concept of scarcity and finishing on short overview of academic literature about banner advertising. Finally, the research gap, hypotheses and conceptual models are presented.
2.1 Scarcity
The marketing science adopted scarcity as a concept derived from well-known commodity theory. This theory implies that a certain product might be regarded as of a higher value if the availability of this product is limited (Brock, 1968). In the meta-analysis, which aimed at applying scarcity in a marketing setting, Lynn (1991) concluded that scarcity indeed increased the value of products that consumers find useful. Lee and Seidle (2012) proved that scarcity appeals might be used in marketing to effectively increase purchases. In their study they exposed participants to two advertisements of the same wristwatch – one included scarcity appeal and the other one didn’t. They found that participants in scare condition showed the willingness to pay for the wristband even twice as high as participants in not scare condition proving that scarcity affects consumers’ purchase intentions.
2.1.1. Psychological background - commodity and reactance theories
In addressing the question of how scarcity actually influences consumers’ behavior, marketing researchers considered several psychological mechanisms behind scarcity appeals and found two contradicting explanations. Cialdini (1993) explains that when people see scarcity appeals, which inform that certain product is becoming less accessible, the emotional reaction is triggered. This, in consequence, decreases people’s ability to process information and therefore, people perceived product’s value becomes higher. It leads, then, to consumers’
12
automatic responses. To put it succinctly, he argues that scarcity reduces consumers’ cognitive abilities.
The second, wider spread but opposing explanation relates to reactance theory (Brehm, 1966). This theory indicates that people possess certain behavioral freedoms. If under some circumstances people perceive a threat that these freedoms might be reduced, as a result, they encounter psychological reactance. This is the state in which cognitive abilities increase - people are motivated to pay more attention to information and act in order to protect their behavioral freedoms. The product scarcity is recognized as an impersonal source of threat to freedom (Clee and Wicklund 1980). The upshot of this is that when product scarcity is present, people recognize the threat, they cognitive abilities increase, and they pay more attention to this product. This, in turn, may make the product more attractive.
2.1.2. Scarcity and message framing - types overview
Although this is true that scarcity induces similar psychological processes among people by either decreasing or increasing their cognitive abilities, it does not necessarily follow that it always works to the same extent. There exist, after all, different methods to communicate scarcity to consumers.
Various approaches have been proposed to categorize scarcity appeals (Verhallen 1982, Cialdini 2008, Roy and Sharma 2015). Verhallen (1982) distinguished between demand-generated and supply-demand-generated scarcity. The demand-demand-generated scarcity is created when the consumers display high demand for the product. The example of this type of scarcity appeal might take a form of a sentence – “In high demand - 34 rooms booked in the last 24 hours” which is displayed on the website of Booking.com, an e-commerce hospitality company (http://www.booking.com). The second type, supply-generated scarcity, is created when the
13
supply of particular product is restricted. As an illustration, online fashion company Zalando (https://www.zalando.com) promotes some products using phrases such as “Limited edition – leather jacket. Only 2 left.” However, by focusing on supply and demand for products, Verhallen overlooked the broader perspective that scarcity does not necessarily have to be highlighted only by the limitation in quantity.
Cialdini (2008) presented a different classification of scarcity appeals. He drew a distinction between a limited-time scarcity (LTS), which refers to products that are available to obtain within a specific time frame, and a limited-quantity scarcity (LQS), which refers to products that are available to obtain in specific predetermined amounts. Limited-time scarcity might be communicated with an appeal which indicates time left to take advantage of the offer. To take a case in point, on Friday 9th March Ryanair announced on its website a “Seat Sale” with 500 000 seats up to 25 percent off and used a Limited Time Scarcity by adding information – ‘Book by midnight Sunday 11th March.’ Limited-quantity scarcity might be communicated in the same way as previously mentioned demand-generated or supply-generated scarcities because these both types represent it.
The last classification reviewed in this study is characterized by Roy and Sharma (2015) and it refers to the message framing. The concept of message framing was brought to consumer behavior by Simonsons and Nowils (200). It suggests that consumers react differently whether they are exposed to message implying a gain or a loss. Process of “framing” means choosing a way of presenting the information either as positive one or negative one. The concept was adapted from psychology. It was first described in the seminal paper of Tversky and Kahneman (1981) who explain that people are apt to be risk averse if they expect a gain. On the other hand, people are apt to be risk-taking if they expect a loss. It is in a sense similar to scarcity because framing message as a loss might suggest the similar threat as presenting limitation by scarcity appeal. A given message triggers a subsequent psychological reaction. The concept of framing
14
is already applied in marketing communication. One of the clear illustration might be the Coca-Cola’s summer 2017 campaign slogan “Win a dream holiday!” ( http://www.coca-cola.co.uk/stories/win-a-dream-holiday) which is a great example of gain-framed appeal. Whereas using the opposite “Don’t lose your chance for dream holiday” would be an example of loss-framed appeal.
2.1.3. Scarcity – demand- and supply-generated
These different types of scarcity were investigated in more details by researchers. Many attempts have been made (Gierl and Huettl 2010, Ku, Kuo, & Kuo, 2012, Ku, Kuo, Yang, & Chung, 2013, Aguirre-Rodriguez, 2013) with the purpose of understanding how differences in communication of demand or supply-generated scarcity might influence consumers. Gierl and Huettl (2010) studied the effect that presence or absence of scarcity appeals has on consumers’ product evaluation whether product is aimed to be a conspicuous consumption good or not. They wanted to examine if different type of scarcity (demand-generated or supply-generated) might moderate this relationship. They referred to the idea of conspicuous consumption formulated by Veblen (1899) as the consumption through which some people want to impress others while displaying their wealth. Eventually, Gierl and Huettl concluded that supply-generated scarcity is indeed better for promoting conspicuous consumption goods because products which are limited in quantity help people to differentiate from the others whereas demand-generated scarcity makes people associate themselves with other consumers who already purchased the product. Hence, demand-generated scarcity might suggest losing a chance to display one’s social status and differentiate from friends whereas supply-generated scarcity indicates this chance.
The motivation, which consumers possess, determines whether demand-generated or supply-generated scarcity is more effective (Ku, Kuo, & Kuo, 2012). Researchers referred to
15
the distinction between promotional-oriented consumers and prevention-oriented consumers derived from Regulatory Focus Theory (Higgins, 1997) and checked which type of scarcity worked better for the above-mentioned groups. They agreed with Mourali et al.(2007) that prevention-oriented consumers tend to avoid taking too much risk while promotion-focused consumers tend to aim at maximizing benefits. Hence Ku, Kuo, and Kuo made very realistic assumptions that prevention-oriented group would be significantly influenced by demand-generated scarcity than supply-demand-generated scarcity because demand-demand-generated indicates less risk, since it high demands suggests that product was purchased by many people before. The promotion-oriented group for contrary was expected to be under stronger impact of supply-generated scarcity because quantity limitation of the product could make product to be seen as more beneficial. The results of the conducted studies indeed supported their expectations and proved that type of motivation moderates the effect of demand-generated or supply generated scarcity on purchase intention.
The more recent study (Ku, Kuo, Yang, and Chung 2013) shows that also the product type might have a moderating effect on relationship between scarcity type and purchase intentions. Researchers demonstrated that consumers who want to acquire a utilitarian product should be exposed to demand-generated scarcity messages, while consumers who want to acquire a hedonic product should be exposed to supply-generated scarcity messages in order to gain better results. This finding is of a great importance in context of travel industry because as Juergen Gnoth (1997) described “holiday tourism is a hedonic activity”. Thus, supply-generated scarcity messages should be more effective than demand-generated scarcity messages in travel industry context.
2.1.4. Scarcity – time limitation
16
literature. Much work on the potential of time-limited scarcity has been carried out (Brock & Brannon 1992, Howard and Kerin, 2006, Aggarwal, et al. 2011, Chang and Chen, 2015 Mou and Shin, 2018). Brock and Brannon (1992) argued that scarcity, which initiates the process of evaluative scrutiny among consumers, might be enhanced by a time limitation.
Mattson (1982) demonstrated that presence of time pressure could significantly influence consumers’ behavior. Maule, Hockey, and Bdzola (2000) supported this finding. They conducted an experiment and reflected that people who were exposed to time pressure became more energetic while making their decisions. They also suggested that the reason for this was the anxiety provoked by a deadline, which increased the awareness of involvements needed in solving a task.
However, there is a slight discordance in the literature on whether time scarcity has such a strong positive influence on marketing effectiveness. Some researchers stay skeptical about significant effects of time-limited scarcity while others advocate for it. For instance, researchers examined and compared time limited versus quantity limited scarcity and discovered that quantity limited scarcity had a higher influence on purchase intention than time-limited scarcity (Aggarwal, et al. 2011). In their study, they used a hedonic product i.e. watch. Another research compared impact of quantity-limited scarcity and time limited scarcity for both conspicuous and nonconspicuous products. Again, the superiority of quantity-limited scarcity over time limited scarcity was stressed. (Jang, at al., 2015). In addition, using time-limited scarcity appeals is confirmed to be a good practice only when the product is of high quality. This is because low-quality products presented with time-limited scarcity appeals are negatively evaluated by consumers (Shen, 2013). Furthermore, more recent evidence even suggests that time-limited scarcity might have negative effects on consumers as they try to devaluate presented products and find better offers somewhere else (Hmurovic, Goldsmith, & Lamberton, 2016).
17
Many other researchers contend, on the other hand, that time-limited scarcity might be indeed used to improve consumers’ responses (Howard and Kerin, 2006, Godinho et al. 2016, Mou & Shin 2018). For example, Howard and Kerin (2006) proved that scarcity messages that include time limitation strengthen the effect that reference pricing has on increasing store shopping intentions. More consumers were willing to buy a product when the advertisement included a comparison of the new and old price and time-limited scarcity than when time limited scarcity was absent. In the lights of studies, which have been carried in recent years, time-limited scarcity gained credits due to improving consumers’ responses in an online marketplace. (Godinho et al. 2016, Mou & Shin 2018).
In the experiment, researchers exposed participants to the website resembling an e-commerce platform and gave them a chance to choose a camera (Godinho et al. 2016). In one condition participants were given an unlimited amount of time to make their decision while in the other one they were given only 3 minutes. The study showed that when time pressure is present in the online marketplace, the likelihood that consumers defer their purchase decision declines.
Furthermore, findings of Mou and Shin (2018) signal that time-limited scarcity is of the great importance in the online environment. According to the study they conducted, if website shows offers of products with time-limited scarcity, these offers are more likely to gather consumers’ attention online. These results seem to be strong because the study is very well-grounded – the authors supported the experiment with an eye-tracking software to check if participants paid attention to the messages as well as reduced all the elements which might associate products or website with any similar brands or online platforms. That is why, the expectation is that time scarcity signals might increase the effect of scarcity messages on banner effectiveness.
18
2.1.5. Message framing
A number of marketing studies sought to examine in more detail previously mentioned message framing. The field experiment, built upon the knowledge of Tversky and Kahneman (1981), confirms that framing a message as a gain or as a loss might change consumers’ behaviors in different ways (Ganzach, Karsahi 1995). In the experiments, researchers presented a gain- or loss-framed message to inactive clients of the credit card company and observed their card usage for the next two months. They discovered that the number of individuals who started using a card again after receiving loss massage was more than twice as high as the number of individuals who received gain massage.
However, when the similar approach was investigated to measure the effectiveness of online search ads, the message presented as positive was more influential than the negative ones. Yoo (2011) argues that search ads with positively framed messages generate higher click-through rates than search ads with negatively framed messages. So, gain-framed messages might increase the attention and intention to click to banners more than loss-framed messages.
Roy and Sharma (2015) analyzed the interaction between scarcity type (demand or supply), message framing (gain or loss) and need for uniqueness, which they referred to as behavior of some consumers who mark themselves as unique through consumption. They argue that message framing has an important impact on advertising. The gain framing should be used to emphasize social benefits for consumers with a low need for uniqueness whilst the loss framing should be used to address consumers with high need of uniqueness and enhance their tendency to make an unconventional purchase decision.
2.1.6. Scarcity – mediating effect of competition
19
both scarcity and competition. He suggests that not only people want an item more when it becomes scarce but that they want it most when they see that there is a competition for it. Researchers defined consumer competition as “the act of a consumer’s striving against one or more consumers for the purpose of achieving a desirable economic or psychological reward” (Aggarwal, et al. 2011, p. 20). They proved that consumer competition is mediating the relationship between scarcity and purchase intentions. When consumers feel that they have to compete with others for a scare product then the influence of scarcity on purchase intention is stronger, because the pressure increases.
2.1.7. Scarcity – online
Although scarcity messages might have some significance for online marketing, the research on it is scarce. Only very recent studies investigated this topic. The effectiveness of scarcity in the context of viral marketing was examined, and it was shown that the presence of scarcity appeal indeed has a positive effect on referral likelihood (Koch & Benlian, 2015). Nevertheless, when displaying scarcity online, it is important to show only products that are still in stock, as showing rare products which are out of stock on the website makes consumers more dissatisfied (Peinkofer, Esper, & Howlett, 2016).
2.2. Banner ads – overview
A growing body of literature has studied the impact of display advertising. For instance, experts examined what the effect of exposure to banner advertising could be on customers’ internet purchasing patterns (Manchanda, Dubé, Goh, & Chintagunta, 2006). This study differentiated among factors such as the number of websites, the number of pages and the number of exposures. The researchers drew the conclusion that these factors positively affected purchase probabilities, whilst one factor, the number of unique creatives, had a negative
20
influence. Nevertheless, this study might seem outdated since it was conducted before the introduction of concepts of omnichannel marketing and digital marketing attribution that are nowadays considered crucial to assessing the contribution of display on customers’ online purchases.
Lambrecht and Tucker (2013) examined how customized content of banner advertising affects online shopping behavior. Internet users collect cookies while browsing. These might be used to customize the context of the banner advertisement that would be displayed to them. Researchers studied generic and dynamic advertisements and concluded that, contrary to what had been expected, generic advertisements worked better than corresponding dynamic advertisements. However, this was true only at the first stages of browsing. Later, when customers already formed their preferences or redefined them, dynamic advertisement could also become effective.
Lately, it was additionally proven that the fit between the context of a banner advertisement and the online behavior of consumers are not the only important factors. Auschaitrakul and Mukherjee (2017) studied whether banner advertisements are more effective when placed on social websites or commercial websites. Researchers referred to fit-fluency theory stating that existing fit between some consumers’ characteristics and some message being sent might increase ease of processing. Thereby this fit could influence consumers’ judgements and final evaluations of products. (Hong & Sternthal, 2010; Kim, Rao & Lee, 2009) Specifically, they demonstrated that banner advertisements placed on commercial websites would have a stronger impact on consumers’ attitudes towards brands, as well as on attitudes towards the advertisement itself, than banner advertisements placed on social websites. This is because, in accordance with the fit-fluency theory, consumers felt stronger ease of processing when banner advertisement with the objective of selling matched the commercial websites with the same objective.
21
However, there is so far no research that examined if using scarcity communication on banner ads improves their effectiveness. This might be surprising concerning the fact that the latest studies highlight the importance of banner advertisements, which were for a long time underestimated (Li & Kannan, 2014; Papadimitriou, et al. , 2011). In recent years there has been a considerable interest in measuring the relationship among different online marketing channels as well as their actual contribution to conversions, which has shown a renewed, higher importance of display advertising. In their investigation into the connection between display advertising (advertising in the form of banners or rich media on different websites) and search advertising (advertising in the form of text ads within search engines such as Google or Yandex), Papadimitriou, et al. (2011) proved that the augmentation of the number of search queries is one of the indirect effects of display advertising. Thus, they showed that display advertising allows to move prospects to another medium and retain contact with a brand. Moreover, a study using binary logit tested customers’ interactions between display advertising and search engine advertising and confirmed that these interactions have a significant influence on click probabilities, emphasizing again the impact of display advertising (Nottorf, 2014). Additionally, more recent studies demonstrated that display ads enhance not only clicks, but also conversions referred to search engine advertising (Kireyev, Pauwels, & Gupta, 2016). This means that even though last-click attribution (which is still commonly used by marketers) does not assign many credits for conversions to display advertising, this medium does help to improve results of search advertising. Hence stopping display campaigns might negatively affect digital marketing results.
2.3. Indicating the gap and formulating hypotheses
Many previous studies have indicated that scarcity influences consumer behavior and can indeed increase purchase intention (Godinho et al. 2016, Mou & Shin 2018). It is not clear,
22
however, that the same applies if scarcity is presented on banner advertisements and if the same effect can be generalized to click intentions. However, the increasingly popular technique of behavioral retargeting uses banner advertisements with scarcity appeals a lot. Investigating the effectiveness of scarcity appeals on banners advertisements is very relevant and timely topic which is not researched sufficiently. This is why this study aims at examining whether using scarcity appeals on banner advertisements has a real influence on consumer behavior. Moreover, since banner advertising and scarcity are used extensively by companies such as airlines, travel agencies and hospitality, scarcity will be examined in the travel industry setting.
Scarcity strategies are adapted by many companies to increase product appearance with higher perception of exclusiveness, rarity, or being one-of-a-kind (Kaptein and Duplinsky. 2013). Desire for scarce products increases with scarcity appeals because the ownership of such products produces feelings of personal distinctiveness or uniqueness and even feeling of victory. Moreover, scarcity of the product signals its popularity, since low level in stocks usually indicate that many people have bought the product. In the airlines, travel agencies and hospitality industries, scarcity appeals on banner ads attract the attention of consumers, produce urgency to react, and increase the perceived value of products and opportunities. Mou and Shin (2018) with their eye tracking study indicated that, products which are showed with time-limited scarcity are more likely to gather consumers’ attention online. Scarcity on banner ads might increase purchase and click intentions in an online advertising context (Yoo, 2011). That is why I formulated the first hypothesis as follows:
H1 Scarcity on banner ads positively influences purchase intentions and click intentions.
23
literature, there appears to be no study that would combine demand-generated and supply-generated scarcity in terms of banner advertising. The Internet environment plays a key role. Most Internet users take a peripheral route, which is characterized by low involvement and less motivation to analyze and process the content of messages, they usually take heuristic approach to process the message and they judge it simply by accepting or rejecting it (Bajaj et al. 2006; Haans, Raassens, & Hout, 2013; Park et al. 2007). Hence, they are not representing prevention-oriented, since they do not deeply analyze information, but rather group but rather promotion-oriented. This is why, in alignment with previous studies, it is expected that they would be influenced in a stronger manner by supply-generated scarcity than by demand-generated scarcity (Ku, Kuo, & Kuo, 2012). Finally, since supply-generated scarcity works better for hedonic products (Ku, Kuo, Yang, and Chung 2013) and tourism is regarded as hedonic (Juergen Gnoth, 1997) I formulate the second hypothesis that:
H2 Purchase intentions and click intentions of ads with supply-generated scarcity are higher than purchase intentions and click intentions of ads with demand-generated scarcity.
A large body of literature argues that using loss framing in marketing communication strategies provides better results (Roy and Sharma, 2015; Tversky and Kahneman, 1981). Nevertheless, most of these studies examined the effects of framing in offline environments. Again, the difference between the offline and online environment is a crucial aspect. When the effects of message framing were examined in online environments, the results were the opposite. It was demonstrated that using gain framing works better in an online setting (Yoo 2011). This is again because most Internet users follow the peripheral path and process messages heuristically (Bajaj et al. 2006; Haans, Raassens, & Hout, 2013; Park et al. 2007). It was found that when a heuristic approach is taken, the gain framing tends to be more effective and this is because the gain is simply perceived as more attractive. This perception is
24
subsequently transferred from the context to the message itself (Maheswaran & Meyers-Levy 1990; Veer & Pervan 2008). The positive message in the form of a gain then becomes more persuasive (Sternthal & Craig 1974). Since this study investigates banner advertising, which is a part of advertising in the Internet, I formulate the third hypothesis as follows:
H3 Purchase intentions and click intentions of ads with gain-framing scarcity are higher than purchase intentions and click intentions of ads with loss-generated scarcity.
Purchase intentions also increase when consumers feel the pressure of competition, because consumers might feel that there are other buyers who can choose to buy the product and thereby make it unavailable (Aggarwal, et al. 2011). This perception might increase consumers’ willingness to be the first to buy the product before others do. The reason to act faster is that people tend to value products more, if these products might become subjects to competition (Worchel, Lee, and Adewole, 1975). Possible unavailability of the product, which might occur as a consequence of competition in the marketplace, might be considered as another source of threat to freedom which subsequently triggers a reaction (Clee and Wicklund 1980). Since it is proved that competition mediates the influence of scarcity appeals, which are present on printed advertisements, on purchase intentions, it is therefore expected that the same happens when scarcity appeals are presented on online banner advertisements. Moreover, it is expected that because consumers feel the competition of others to buy the product, they would like to click on the advertisements faster to gain more insights about the offer. Hence, I formulate hypothesis three and four as follows:
H4 Competition mediates the relationship between a type of scarcity (demand/supply-generated) and purchase intentions and click intentions.
25
Message framing, similarly to scarcity appeals, is sending an information to consumer. It can be stressed as either a positive or a negative one. These two ways affect consumers’ behavior differently (Ganzach, Karsahi 1995). However, both types of framing trigger a reaction – consumers either want to act to get the expected benefit or want to take more risk and act to prevent from expected loss. It is likely, that people experience this urge to respond to the message because they feel the threat that if they do not, others might react to the information first. Again, the incentive to act would be faster in light of competition (Worchel, Lee, and Adewole, 1975). Therefore, next hypothesis is formulated:
H5 Competition mediates the relationship between a type of framing (loss/gain) and purchase intentions and click intentions.
Lastly, this mediating effect of competition should be even higher when time limitation is present. This is because the time pressure significantly influences consumers behavior (Mattson 1982). The anxiety which is a result of time pressure increases engagement, so consumers react faster (Maule, Hockey, and Bdzola, 2000). Studies confirmed that people are paying more attention to the offers when time limitation is present on them and that time limitation might be improving their responses (Howard and Kerin, 2006). This is especially true in the online environment (Mou and Shin, 2018), probably because people following a peripheral route do not feel the need to pay too much attention until they notice time limitation. It is then anticipated that time limitation increases consumers’ attention to both scarcity and framing. Consequently, consumers want to make a smart decision by reacting faster, earlier than other consumers, and this increases their feeling of competition. Therefore, I formulate the last two hypotheses as follow:
26
H6 Competition mediates the relationship between a type of scarcity (demand/supply-generated) and purchase intentions and click intentions when time scarcity is present
H7 Competition mediates the relationship between a type of framing (loss/gain) and purchase intentions and click intentions when time scarcity is present
27
Figure 2.3.1 Research model 1
28
3. Methodology
This chapter introduces how the focal research question of this study was answered. It gives a detailed account of how the research was carried out, what method was used and how the hypotheses were verified.
3.1 Method
The main objective of this study is to analyze the effect of scarcity appeals used in banner advertising on consumers’ click intentions and purchase intentions. Researchers have used a variety of methods to evaluate impact of scarcity on consumers behavior including large field experiments or laboratory experiments with eye-tracking software. Each method has its advantages and drawbacks, and none can be regarded as unflawed.
Measuring the effectiveness of banner advertising is not easy. Marketers use more advanced techniques such as statistical attribution models. These can include variables like exposure, frequency, and recency (Google 2017). Nevertheless, these techniques usually show the impact of overall customer journey. It means that touchpoints which in fact do not significantly increase consumers’ purchase intentions but only appear on customer journey are still granted credits. This way, estimating real impact of each contact with a banner advertisement on consumers’ purchase intentions becomes complex. Furthermore, many consumers click on banner ads by mistake. This is especially true for mobile banner ads, which are clicked accidentally in around 60% cases (Media Post). Thus, it is important to measure if scarcity increases real intentions to click on the advertisements.
Hence, in order to answer the research question, a between-groups online experiment has been conducted. This particular method was chosen because it is one of the most feasible approaches. Moreover, previous literature about scarcity messages (Gierl and Huettl 2010; Ku,
29
Kuo, Yang, & Chung, 2013; Mou and Shin, 2018) most of the times used experiments. This method enables analysis of the influence of scarcity messages on consumer behavior in the face of many limitations to conduct the research in real settings. Additionally, the data from experiments might more accurately estimate if scarcity used on banner advertisements increases purchase and click intentions when all the other factors (such as different advertisements which consumers might encounter) are excluded.
3.2 Design
To determine whether type of scarcity or message framing has bigger influence on click intentions and purchase intentions, this experiment was designed as between-groups with 9 different conditions. The minimum required sample size per condition with accordance to Central Limit Theorem was 30. Nevertheless, this study reached approximately 40 participants per condition. There were 4 quantity scarcity conditions: 2x (supply-generated vs. demand-generated) x2 (time limitation vs. no time limitation), 4 framing conditions: 2x (gain-framed vs. loss-framed) x2 (time limitation vs. no time limitation), and a control condition where there was no scarcity present on the advertisements. The online experiment was performed using Qualtrics software. Participants were recruited via Amazon Mechanical Turk (MTurk) and paid for their participation.
Manipulations: The type of scarcity was manipulated across the banner advertisements by modifying their copies. In an effort to keep scarcity realistic, the appeals similar to those present on popular travel websites were used. To signal demand-generated scarcity the phrase “High demand – 40 rooms already booked!” was used, and to signal supply-generated scarcity the phrase “Just 2 rooms left” was used.
30
modifying the copy of banner advertisements. The appeal was framed as gain which indicated possible benefit to be obtained, or as loss, which indicated a threat to miss a chance. To manipulate gain condition the phrase “Just 2 rooms left! Win this chance!” was used. To simulate loss condition, the message “Just 2 rooms left! Don’t miss this chance!” was displayed.
The construct of time scarcity used was derived from previous studies and described as time limitation which a buyer faces to purchase a product (Chang & Chen, 2015, Mou & Shin 2018). To manipulate for the presence of time scarcity the banner advertisement was changed so that it contains additional information “Only today”. In the condition without time scarcity, this information did not appear on the banner. The banners are attached in Appendix A.
Table 3.2.1 –Summary of conditions
Number Scarcity type Time scarcity Condition presented to participants 1 Supply-generated Absent Just 2 rooms left! Book now!
2 Supply-generated Present Just 2 rooms left! Only today! Book now!
3 Demand-generated Absent High demand – 40 rooms already booked!
4 Demand-generated Present High demand – 40 rooms already booked! Only today! Book now
5 Gain-framed Absent Just 2 rooms left! Win this chance! Book now
6 Gain-framed Present Just 2 rooms left! Win this chance! Only today! Book now
31
chance! Book now
8 Loss-framed Present Just 2 rooms left! Don’t miss this chance! Only today! Book now
9 Control condition (no scarcity) Book now
3.3 Measures
The click intentions and purchase intentions, the dependent variables, were measured using the average of two seven-point Likert-scale items. The items used for click intentions were “I will probably click on this advertisement to see the offer” and “I am willing to click on this advertisement”. Purchase intentions were measures with items “I am likely to book a room at this hotel” and “I am willing to purchase my stay at this hotel” (1 = Strongly disagree and 9 = Strongly agree). The Cronbach’s α were respectively α(click)= 0.86 and α(purchase)= 0.91. Two consumer competition items were borrowed from the study of Aggarwal, Jun and Huh (2011) customized to match the context of this experiment. The average of two seven-point Likert-type scale items was utilized: “I may lose the opportunity to book my room if others book it first” and “There is a lot of competition from others for booking this room”. The Cronbach’s α was equal to 0.66
As a manipulation check variable, I measured perceived scarcity. Participants were instructed to respond to a statement adopted from a study by Eisend (2008), “How available do you think Product X is”, on a seven-point scale anchored by “rather inadequate” and “rather adequate”. Then participants were further asked to identify the stated cause of scarcity or framing – depending on the condition they were assigned to. They were either asked “How would you describe scarcity?” with possible answers suggesting demand-generated or supply generated scarcity, or they were asked “Which sentence do you agree with most?” with possible
32
answers “By clicking on this advertisement I may gain the opportunity to book a room at this hotel” and “By not clicking on this advertisement I may lose a chance to book a room at this hotel.” As a second control variable I used perceived time limitation, which was measured by the seven-point Likert-scale item “Please, rate to what extent do you agree with the following statement: There is a time limitation to book a room.” (1= Strongly disagree, 7= Strongly agree).
3.4 Pre-test
I first conducted a pre-test with 23 students who voluntarily agree to participate. All nine conditions were tested. I check whether click intentions for participants in scarcity conditions (M=3.83, SD=1.92) where indeed higher than click intentions for participants in control condition where there was no scarcity (M=1.33, SD=0.58). There was a significant difference between scarcity and non-scarcity conditions when it comes to click intentions (t (10.06) = 4.78, p = 0.001.) The average click intentions score for participants in scarcity conditions was 2.49 points higher than the average click intentions score for participants in control condition. However, the perceived scarcity was not measured correctly because the scale used for the item was not understood clearly. After qualitative feedback from participants it was determined that scale from “rather inadequate” and “rather adequate” corresponding to question - How available do you think rooms in this hotel are?” was found wordy and hard to understand. This is why scale was adjusted and changed to “rather unavailable” and “rather available”. Moreover, initially and in the pre-test, I used banners with claims: “Just 5 rooms left!” to manipulate scarcity. However, very important comment was registered many times that participants did not find amount of 5 as scarce. It was suggested that the amount of 1-3 rooms would be perceived as truly scarce, Thus, all banner ads were adjusted, and table presented before in this study already includes proper claims. Additionally, some small changes were
33
applied to make survey more user friendly (e.g. font on the blog was increased in order to make text easier to read).
3.5 Procedure
Participants were firstly asked to give consent to participate in the study. Then, they were instructed that they were taking part in the study about travel destination preferences. They were asked to imagine that they were going to Santorini in 2 months and that they still had not booked a hotel. Next, they were asked to read a blog about Santorini. The blog imitated tourist website of http://www.visitgreece.gr/en/greek_islands/cyclades/santorini. It was chosen in order to comply with fit--fluency theory (Hong & Sternthal, 2010; Kim, Rao & Lee, 2009). The touristic blog was expected to be a good match for hotel advertisement and to increase processing fluency.
Participants were randomly assigned to one of the 9 conditions. On the blog website that they were reading there was a banner advertisement placed in the middle of the text. There were nine different banners manipulated according to the conditions. To mitigate any confounding effects of using a popular brand name, I did not place any logo of recognizable travel company on the banners. However, to make the banner ad realistic I used a name (Canaves Oia Epitome) and photos of a real hotel. Nevertheless, again to mitigate any confounding effects I used the hotel which could not be recognized by any of the participants since it was not opened or advertised at the time of conducting the experiment. In summary, each of the banner ads consisted of a photography of a hotel on top, the name of the hotel in the middle, and copy with the call to action (“Book now”) on the bottom. They differ by two elements. Firstly, by the copy which could also include scarcity/framing appeal (demand- or supply generated and loss-framed or gain-framed). Secondly, by the additional graphic element presenting message “Only today” in condition where time limitation was present. The other
34
elements of the banner advertisements such as name of the hotel, photography used, color of the background, fonts used, color of the button and claim “Book now” remain the same in all the conditions.
To improve effectiveness of the experiment the size of banner advertisements was 336x280 (large rectangle) which is considered by Google as most successful ad size (Google AdSense Help, 2018). Furthermore, I chose to use static ads with no interactive elements. This was because according to the experiment scenario participants were at the beginning of the browsing process, (particularly no accommodation consideration was presented to them) and finding of Lambrecht and Tucker (2013) confirmed that static banners were found more effective for Internet users at the first stages of browsing.
After participants spent time on reading the blog they were asked 2 or 3 attention questions. Firstly, if Santorini was inactive or active volcano (where active was a correct answer). Secondly, if they had noticed the banner advertisement on the blog and if they had checked answer “yes” they were subsequently and lastly asked to choose what was presented on the banner advertisements (a restaurant, a hotel or a cruise offer).
Then once again the banner advertisements which were located on the blog were presented participants. They were given some time to observe them and evaluate by filling the rest of the survey (Appendix A).
Table 3.5.1- Measurement items and Constructs
Construct/variable Source Scale Scale type
Click Intentions - Q1 I will probably click on this advertisement to see the offer Q2 I am willing to click on this advertisement 7 points likert scale (mean on items used as a metric)
35
Purchase Intentions
- Q1 I am likely to
book a room at this hotel Q2 I am willing to purchase my stay at this hotel 7 points likert scale (mean on items used as a metric)
Competition Aggarwal, Jun and Huh (2011)
Q1 I may lose the opportunity to book my room if others book it first
Q2 There is a lot of competition from others for booking this room. 7 points likert scale (mean on items used as a metric) Independent
variables: variable type
Type of scarcity (demand vs. supply) categorical Type of framing (gain vs. loss) categorical Manipulation check questions Item variable type Type of scarcity How would you describe scarcity?
0 - There seems to be a limited amount of rooms available
1- The rooms in this hotel seem to be highly demanded.
categorical
Type of framing Which sentence do you agree with most? 0- By clicking on this advertisement I may gain the opportunity to book a room at this hotel.
1- By not clicking on this advertisement I may lose a chance to book a room at this hotel.
categorical
Time limitation There is a time limitation to book a room. metric(1=strongly disagree, 7=strongly agree) Attention check questions
Is Santorini an inactive or active volcano?
1- Active 2- Inactive
36
Have you noticed a banner advertisement on the blog?
1- Yes 2- No
What did the banner advertisement promote?
1- A cruise 2- A hotel 3- A restaurant
Control variables: Source Scale Scale type Perceived scarcity Eisend (2008) How available do you
think the rooms at this hotel are? metric (1=rather unavailable, 7=rather available) It was counter-indicated in analysis (1= rather available, 7= rather unavailable) Demographic variables: type Age metric Gender categorical
37
4. Results
In this chapter statistical analyses will be introduced. Multi-item scales’ reliability and validity will be tested with Cronbach’s alphas and factor analysis, respectively. Manipulation checks will be conducted with independent sample t-tests to examine there are significant differences in scarcity perception among scarcity conditions compared to control condition. To test whether proper identification of type of scarcity and type of framing were conducted by consumers Chi-square test will be analyzed. For hypothesis testing, we will use ANOVA analysis to determine if group means of click intentions and purchase intentions differ between scenarios. The analysis will be conducted with SPSS and PROCESS macro (Hayes, 2017). Mediating effect and moderated meditating effect will be conducted by running PROCESS model 7 four times.
4.1 Sample and data collection
Participants for this study were recruited through Amazon Mechanical Turk (MTurk). Mturk was proven as a reliable tool to collect data efficiently for research purposes (Buhrmester, Kwang & Gosling, 2011). This way of data collection was chosen for two main reasons. First of all, users of MTurk are familiarized with the Internet since they are working in the Internet. Thus, it was assumed that they had already seen banner advertisements before and would react naturally when exposed to another banner advertisement in the experiment. Secondly, because the experiment was conducted in English and MTurk allowed for data collection from English native speakers.
The total number of 410 people took part in the experiment and 352 finished it. Participants were paid small amount of money for participating. The sample was represented in 66.7% by males, 33%. by females and 0.3% by people who chose not to state their gender.
38
The average age of participants was 31 years. Sample was represented only by consumers older than 18 years old.
Table 4.1.1 – Sample demographics
Gender Frequency Percent
Female 116 33%
Male 235 66.7%
N/A 1 0.3%
Total 352 100%
Figure 4.1.2 – Sample population pyramid
50 40 30 20 10 0 10 20 30 40 18-25 26-33 34-41 42-49 50-57 58-65 66-73 Age Male Female
39
4.2 Preparing data
I started the analysis by cleaning the data. Since after participants read the blog they were given another chance to take a look at the banner advertisement I deleted only answers of respondents who:
Firstly, failed the attention check on one of two possible steps: 1) They answered that Santorini was inactive volcano.
2) They answered that they had noticed banner advertisement on the website but later chose the wrong type of the presented offer (either a restaurant or a cruise)
Secondly, checked the same scores on all Likert-type scales.
This process led to deleting 7 responses and left 345 responses to analyze further.
4.3 Reliability and validity analysis
Reliability analysis was conducted using Cronbach’s α. As previously mentioned the Cronbach’s α for two items measuring click intention was at acceptable level (α(click)= 0.86). Thus, I used both items to compute new variable “click intention ratio” which was a mean score of two items. The Cronbach’s α for two items measuring purchase intention was also at acceptable level (α(purchase)= 0.91). Again, I used both items to compute new variable “purchase intention ratio” which was a mean score of two items. Lastly, the score of Cronbach’s α for two items measuring perceived competition was equal to 0.66 which was again considered as acceptable. I calculated the mean of two item scores and used it as a “perceived competition ratio” for further analysis.
To test validity a principal axis factor analyses were conducted. In first analysis for purchase intentions and competition (mediator) the Kaiser-Meyer-Olkin measure verified the sample adequacy for the analysis KMO= 0.667. Bartlett’s test of sphericity (2 (6) = 627.84, p
40
= .000) indicated that correlations between items were sufficiently large for PAF, competition and purchase intention were valid constructs, and confirmed discriminant validity. The table (Appendix C) shows that items load on valid factors.
Second analysis for click intentions and competition (mediator) gave similar results- the Kaiser-Meyer-Olkin measure verified the sample adequacy for the analysis KMO= 0.718. Bartlett’s test of sphericity (2 (6) = 530.75, p = .000) once more indicated that correlations between items were sufficiently large for PAF, competition and click intention were valid constructs, and confirmed discriminant validity. The table (Appendix C) shows that items load on valid factors.
4.3 Descriptives and Correlations
A Spearman's rank-order correlation was run to determine the relationship between variables in the model. There was a positive correlation between competition and click intentions which was statistically significant (rs = .587, p = .000). Statistically significant, positive correlation was also found between competition and purchase intentions (rs = .581, p = .000) and perceived time limitation (rs = .533, p = .000). There was also a negative correlation between competition and perceived scarcity which was statistically significant (rs = .000, p = -.205). Positive correlation which was statistically significant was also found between click intentions and purchase intentions (rs = .686, p = .000), perceived time limitation and purchase intentions (rs = .422, p = .000).
Preliminary bivariate tests, descriptive statistics are indicated in Appendix B. Figure below (Figure 4.3.1) presents reasons why people would not click on the banner advertisement.
41
Figure 4.3.1 – Reasons for not clicking
0% 5% 10% 15% 20% 25% 30% 35% 40% I think it is a spam.
I do not trust it. I am worried about tracking. This banner advertisement is not relevant I am worried about getting a virus.
Why wouldn't you click on this banner advertisement? (Please,
choose the main reason)
42
Table 4.2.1- Spearman correlation matrix
competition click intentions purchase intentions perceived scarcity perceived time limitation scarcity type time limitation framing type competition Correlation Coefficient 1.000 .587** .581** -.204** .533** -0.051 -0.046 0.130 Sig. (2-tailed) 0.000 0.000 0.000 0.000 0.533 0.428 0.106 N 345 345 345 345 345 149 306 157 click intentions Correlation Coefficient .587** 1.000 .686** -.358** .422** -0.037 -0.053 0.027 Sig. (2-tailed) 0.000 0.000 0.000 0.000 0.656 0.354 0.735 N 345 345 345 345 345 149 306 157 purchase intentions Correlation Coefficient .581** .686** 1.000 -.505** .422** 0.005 0.028 -0.024 Sig. (2-tailed) 0.000 0.000 0.000 0.000 0.953 0.629 0.763 N 345 345 345 345 345 149 306 157 perceived scarcity Correlation Coefficient -.204** -.358** -.505** 1.000 -.295** -0.125 0.037 -0.020 Sig. (2-tailed) 0.000 0.000 0.000 0.000 0.130 0.520 0.807 N 345 345 345 345 345 149 306 157 perceived time limitation Correlation Coefficient .533** .422** .422** -.295** 1.000 -0.032 -0.061 0.022 Sig. (2-tailed) 0.000 0.000 0.000 0.000 0.699 0.288 0.788 N 345 345 345 345 345 149 306 157 scarcity type Correlation Coefficient -0.051 -0.037 0.005 -0.125 -0.032 1.000 -0.019 Sig. (2-tailed) 0.533 0.656 0.953 0.130 0.699 0.822 N 149 149 149 149 149 149 149 0 time limitation Correlation Coefficient -0.046 -0.053 0.028 0.037 -0.061 -0.019 1.000 -0.019 Sig. (2-tailed) 0.428 0.354 0.629 0.520 0.288 0.822 0.809 N 306 306 306 306 306 149 306 157 framing type Correlation Coefficient 0.130 0.027 -0.024 -0.020 0.022 -0.019 1.000 Sig. (2-tailed) 0.106 0.735 0.763 0.807 0.788 0.809 N 157 157 157 157 157 0 157 157
**. Correlation is significant at the 0.01 level (2-tailed). Scarcity type (0=supply, 1=demand)
Time limitation (0= absent, 1=present) Framing type (0=loss, 1=gain)
43
4.4 Manipulation checks
Eventually, manipulation check for scarcity was formulated with a question - “How available do you think the rooms at this hotel are?” with adjusted item scale 1 = “Rather unavailable” to 7 = “Rather available”. However, for the purpose of analysis the scale was counter-indicated to obtain perceived scarcity score (1 = “Rather available” and 7=” Rather unavailable”). Then, a series of independent samples t-tests was conducted to examine whether participants in any of manipulation conditions perceived rooms as less available than participants in control condition where no scarcity message.
Results for first condition show that the mean scarcity scores differed t (67.221) = 2.34, p = 0.022. The average scarcity score for respondents exposed to supply-generated scarcity (M=3.56, SD=1.86) was 0.92 points higher than for respondents exposed to non-scarcity condition (M=2.64, SD=1.50). However, this was the only manipulation check that gave the positive results.
Table 4.4.1 - Results of t-test and Descriptive Statistics for manipulation checks of perceived scarcity Condition 95% CI for Mean Difference Supply-generated and no time limitation scarcity No Scarcity M SD n M SD n t df Perceived scarcity 3.56 1.86 36 2.64 1.50 39 .91, .39 2.37* 73 Note: * p < .05.
Unfortunately, the rest of manipulation checks appeared not working as intended. When tested, the mean scarcity scores of supply-generated scarcity condition with no time limitation and non-scarcity condition did not differ t (78) = 1.684, p = 0.096. The same applies for all other conditions. The mean scarcity scores of demand generated scarcity condition and non-scarcity condition did not differ t (72) = 0.592, p = 0.556, the mean non-scarcity scores of
demand-44
generated scarcity condition with time limitation and non-scarcity condition did not differ t(74) = 0.575, p = 0.567.
In the framing conditions, manipulations checked were found again as invalid. The mean scarcity scores of gain-framed condition and non-scarcity condition did not differ t (77) = 1.018, p = 0.312. The mean scarcity scores of gain-framed condition with time limitation and non-scarcity condition did not differ t (77) = 1.663, p = 0.100. The mean scarcity scores of loss-framed condition and non-scarcity condition did not differ t (74) = 1.023, p = 0.310. Eventually, the mean scarcity scores of loss-framed condition with time limitation and non-scarcity condition did not differ t (77) = 1.907, p = 0.060. This makes the research questionable.
The table (Table 4.4.2) shows the average perceived scarcity scores per conditions. It is demonstrated that even though results did not reach statistical significance, the scarcity conditions reached higher perceived scarcity scores than non-scarcity condition (M=2.6410).
Table 4.4.2- Average perceived scarcity scores per condition
Subsequently, to verify if participants correctly identified demand-generated or supply generated scarcity they were asked the question: “How would you describe scarcity?” with two possible answers – “There seems to be a limited amount of rooms available.” regarded as
3.5556 3.2927 2.8571 2.8108 3.0000 3.2750 3.0000 3.3500 2.6410 0.0000 1.0000 2.0000 3.0000 4.0000 1 Supply with no time limitation 2 Supply with time limitation 3 Demand with no time limitation 4 Demand with time limitation 5 Gain with no time limitation 6 Gain with time limitation 7 Loss with no time limitation 8 Loss with time limitation 9 No scarcity
