The impact of perceived sleep quality on consumer's exploratory behaviour

(1)

1

Master Thesis

‘The Impact of Perceived Sleep Quality on Consumer’s Exploratory

Behaviour’

Submitted by:

Laura-Katharina Swienty (11376511)

June 23

rd

_{2017 | Final version}

University of Amsterdam

Faculty of Economics and Business

MSc Business Administration – Marketing Track

Under the supervision of:

(2)

2

STATEMENT OF ORIGINALITY

This document is written by Laura-Katharina Swienty who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is 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)

3

ABSTRACT

Bad sleep quality can have a devastating effect on a person’s cognitive functions, e.g. stimuli and information processing. However, research on sleep quality effects has not yet been expanded to all areas of consumer behaviour. The Optimum Stimulation Level (OSL) theory states that the processing of stimuli is a way to reach optimum stimulation. In order to be exposed to new stimuli consumers engage in exploratory behaviour, e.g. seeking for new product information or purchasing innovat ive products. Several studies have investigated the effect of OSL on exploratory behaviour. This research takes a first approach to combine theory on perceived sleep quality, OSL and exploratory behaviour. It examines possible effects of perceived sleep quality on exploratory behaviour and the mediating role of OSL.

An online between-subjects experiment was conducted in Germany and the Netherlands with a total of 168 participants. Participants were randomly assigned to a Good or Bad Sleep Quality condition receiving false feedback about their sleep quality. Afterwards, exploratory behaviour was measured by having participants rate common and fictitious innovative products across four different product categories

Findings of this study suggest that there is no influence of sleep quality perception on either OSL or a consumer’s preference for stimulating products. Furthermore, OSL is not mediating the relationship between perceived sleep quality and exploratory behaviour. However, the influence of OSL on exploratory behaviour was confirmed and has implications for marketing. Marketers need to emphasize the innovativeness of new product offers in order to appeal to consumers high in OSL.

(4)

4

LIST OF FIGURES

Figure 1: Conceptual Framework ... 11

Figure 2: Conceptual Framework including hypotheses ... 19

Figure 3: Sleep Survey questions; sleep experiences ... 22

Figure 4: Sleep Survey Q uestions; external influences ... 22

Figure 5: Speedometer 77% for Good Sleep Quality Condition ... 23

Figure 6: Speedometer 30% for Bad Sleep Quality Condition ... 23

Figure 7: Example product presentation and questions ... 26

Figure 8: Slider for Preferred Level of Excitement ... 28

LIST OF TABLES

Table 1: ANCOVA for Perceived Sleep Quality on Preferred Level of Excitement ... 39

Table 2: ANCOVA for Perceived Sleep Quality on Excitement Difference Score ... 41

Table 3: ANCOVA for Perceived Sleep Quality on Stimulation Preference Score ... 43

Table 4: PROCESS Results for Excitement Difference Score as Mediator ... 45

(7)

7

1 INTRODUCTION

In times of a fast-moving consumption environment, consumers are facing an unlimited choice of products, providing them with great product variety and new experiences (Smith & Agrawal, 2000). New product releases and brand extensions try to lure the consumer to switching brands and changing their old shopping habits (Wan & Dresner, 2015). Consumers seem less likely to stay loyal to a brand given the increased product stimulation their experience every day.

Stimulation plays a key role in the search for new product variations. According to Steenkamp and Baumgartner (1992), people have an individual optimum level of stimulation (OSL) which they seek to reach through either receiving or refusing stimuli in their environment, like unfamiliar product offers. Research has shown that exploratory behaviour is the result of the discrepancy between a consumer’s preferred level of stimulation and his actual level of stimulation (Steenkamp & Baumgartner, 1992). Hence, exploratory behaviour describes an individual’s desire for something unknown and can be affected by several external and internal factors (McAlister & Pessemier, 1982).

One possible factor affecting consumers’ variety seeking behaviour might be the consumer’s sleeping pattern due to bad sleep during the previous night. According to the 2008 ‘Sleep in America’ poll, a survey conducted by the NSF (National Science Foundation), a person’s average sleep duration on workdays amounts to 6.7 between 7.4 hours on non-workdays. However, survey participants described a need for 7 to 8 hours sleep per night as being the personal optimum (Swanson et al., 2011). Furthermore, 49% of all participants reported that they experience a non-refreshing sleep a few nights per week, while 42% stated they would experience frequent awakenings during the sleep period (Swanson et al., 2011). Those findings show, that especially consumers who work full time

(8)

8 experience at least occasionally restrictions to their optimum sleep quantity or quality resulting in sleepiness throughout the next day.

Research has established that bad sleep quality can have a devastating effect on a person’s cognitive functions, inter alia information processing, reaction time or memory (Ratcliff & Van Dongen, 2009). Nevertheless, research on sleep quality has not yet been expanded to all areas of consumer behaviour. Although it is known that a poor sleep quality limits the consumers’ capacity to absorb stimuli, research has not been conducted on how exploratory behaviour might be influenced by the perception of a good or bad sleep quality.

This work aims at closing this research gap. The research objective shall be as follows:

Identifying the influence of consumers’ perceived sleep quality on their exploratory behaviour, and pointing out the role the optimum stimulation level might play in this relationship.

Findings on this relationship can have great implications for marketers. Advertisers need to explic it ly address a consumer’s need and preference for stimulation in order to successfully promote new products (Hirschman & Wallendorf, 1980). When a consumer has a high need for stimulation and wants to satisfy this need with the purchase of a new or unfamiliar product, marketers have to emphasize the product’s innovativeness in promotional messages (Hirschman & Wallendorf, 1980). In addition to that, retailers need to be up to date with newest trends and need to offer a great product variety (Wan & Dresner, 2015). This will not only positively affect demand and sales but also bind the consumer (Smith & Agrawal, 2000). If sleep quality perception is a factor influencing consumers’ interest in product variety than it is crucial for marketers to find out how to influence this perception and how to use it for the promotion of their products.

(9)

9 The work is structured in four sections. First, the existing literature on Sleep Quality, Exploratory Behaviour and OSL will be reviewed. Furthermore, the theoretical framework will explain the research model and hypothesis. In the second part, the methodology of the research will be explained. This includes the setting, the explanation of variables and methods used in the research. The third section will show the results of the analysis. To conclude, the last section will discuss the main conclusions of the results leading to managerial implications that can be derived from this researc h as well as main research limitations.

2 LITERATURE REVIEW

2.1 Research gap

Research on cognition and sleep deprivation has found severe cognitive dysfunctions due to the loss of sleep. People who suffer from sleep loss have a reduced ability to cope with minor as well major cognitive tasks, such as innovative thinking (Harrison & Horne, 1999). Thus, research on mood states found that a poor sleep quality correlates with a decrease in positive affect and satisfaction in life (Pilcher, Ginter, & Sadowsky, 1997). Furthermore, tired people are more easily distracted through their surroundings and are less likely to focus on external stimuli like advertisements, promotions or product information (Ratcliff & Van Dongen, 2009). Sleep deficit must not be the only reason for reduced cognitive functioning. Findings show that the mere perception of bad sleep quality is sufficient for individuals to perform worse on cognitive tests (Draganich & Erdal, 2014). Research has not yet examined the influence of sleep quality perception on a consumer’s exploratory behaviour. One may assume however, that the perception of bad sleep quality not only affects cognitive functioning but also the consumer’s exploratory behaviour.

(10)

10 The concept of the OSL states that a person which falls below its optimum level of external stimulation explicitly engages in exploratory behaviour in order to increase level of stimulat io n (Baumgartner & Steenkamp, 1996). One reason for a low level of stimulation might be the consumer’s tiredness. This may suggest, when a consumer feels tired he might want to experience some kind of thrill by exploring new product alternatives.

Little research has been done on combining those different streams of research. In fact, both streams of research, in part, provide contradictory statements. Optimum Stimulation Level Theory (OSLT) relies on consumers’ boredom or sleepiness to activate their exploratory behaviour (Hirschman, 1980; McAlister & Pessemier, 1982) A consumer who is tired and high in his OSL would engage in exploratory behaviour in order to increase his level of stimulation. In contrast to that, research on sleep quality suggests that a sleepy consumer has a decreased ability to absorb and process stimuli and therefore will not engage in exploratory behaviour (e.g. seeking out for new product informatio n).

This research seeks to establish which stream of research proves right and investigates possible effects of perceived sleep quality on exploratory behaviour. It will also focus on the underlying process which mediates this relationship. The research aims at answering the following questions:

How does perceived sleep quality affect a consumer’s exploratory behaviour? How does

perceived sleep quality affect a consumer’s level of stimulation? Does the optimum stimulation level play a mediating role in the effect of perceived sleep quality on a

consumer’s exploratory behaviour?

The conceptual framework developed in this thesis can be seen in Figure 1. The variables and relationship between the three variables will be explained in detail in the following sections. For this

(11)

11 purpose, existing literature on exploratory behaviour in form of variety seeking and innovat ive behaviour as well as the underlying OSL concept will be reviewed. Furthermore, the current state of the research of the effect of sleep quality on cognitive functioning and its impact on consumer decision making will be examined. Finally, hypotheses will be developed to show how these variables might be related to each other and what the consequences of this relationship might be.

Figure 1: Conceptual Framework

2.2 Sleep Quality Effects

Sleep is a natural recreational human state of mind and body. During sleep, the body can restore its energy resources (Oswald, 1980), repair cell tissue and regulate the metabolic system and adaptive immune functions (Rasch & Born, 2013). However, the main function of sleep is the consolidat io n of memory. It helps to integrate new labile memory traces into pre-existing knowledge structures (McGaugh, 2000; Rasch & Born, 2013). Furthermore, sleep helps to filter information and erases information obtained through stimulation overload or bizarre associations (Rasch & Born, 2013).

Research found that an average sleep duration of 7.5 hours is an optimum for most people (Ferrara & De Gennaro, 2001). A reduction in quality or quantity of sleep can lead to an increased desire for sleep. This can be understood as a physiological need state comparable to hunger or thirst (Mullins,

(12)

12 Cortina, Drake, & Dalal, 2014). Consequences are physiological indispositions and the feeling of sleepiness that lead to a high decrease in performance and cognitive functions (Mullins et al., 2014; Pilcher et al., 1997).

Feeling sleepy is a condition known to every person. Two groups of people seem to struggle with sleepiness particularly, college students and full-time workers. Full time employees work an average 7 hours per day (Basner et al., 2007), while spending an average of 6.68 hours per night for sleeping (Barnes, Wagner, & Ghumman, 2012). 70% of college students stated that they slept less than the recommended 8 hours of sleep, resulting in daytime sleepiness (Hershner & Chervin, 2014). Both full-time workers and college students experience a constant time-based conflict where they have to divide their remaining daytime between family and friends (Barnes et al., 2012), grocery shopping or recreational activities, like the use of media (Basner & Dinges, 2009). They are voluntarily restricting their amount of sleep and sleep patterns (Pilcher et al., 1997) in order to have more leisure time. Both groups take sleepiness and thus lowered performance into account (Hershner & Chervin, 2014; Mullins et al., 2014).

Irregular sleep patterns have an important influence on a person’s sleep quality. Sleep quality is a complex phenomenon without a clear definition (Krystal & Edinger, 2008). It can be understood as a collection of quantitative and subjective measures that lead to a description of a person’s sleep habits (Krystal & Edinger, 2008). The most common and widely employed measure, the Pittsburgh Sleep Quality Index, retrospectively measures respondents’ sleep duration, sleep latency, sleep disturbances and also includes external influences, e.g. medication (Buysse, Reynolds, Monk, Berman, & Kupfer, 1989). Research on sleep quality found that not only sleep duration, but also sleep efficiency has a huge impact on a person’s cognitive functioning (Blackwell et al., 2006). Disturbances during sleep lead to a decreased sleep efficiency, regardless of the total amount of sleep

(13)

13 hours. The same goes with sleep latency, the amount of time between lying down and falling asleep (Blackwell et al., 2006).

Disturbances as well as a reduced amount of sleep can lead to severe sleep deprivation with strong cognitive consequences. Research found that sleep deprived people suffer from an extremely reduced ability to extract information from stimuli and find it harder to make simple as well as more complex decisions (Harrison & Horne, 2000; Ratcliff & Van Dongen, 2009). Sleep deprivation limits the work of the human’s prefrontal cortex (PFC). The PFC is the brain region responsible for directing, focusing and sustaining attention to a task by isolating all kinds of distraction. Furthermore, the PFC enables innovative and flexible thinking as well as making decisions under novelty and complexit y (Harrison & Horne, 2000).

Interestingly, not only sleep deprivation can affect cognitive functions, but also the mere perception a person has about his sleep matters (Draganich & Erdal, 2014). Draganich and Erdal (2014) found in their study on the ‘placebo sleep effect on cognitive functioning’, that a person’s perception of his sleep is sufficient to determine physiological limits. For the purpose of the study, they tested this effect on 164 undergraduate students. One group was told they have a below average sleep quality which may lead to negative outcomes, e.g. reduced cognitive functioning. Results show, that participants, who were told to have had bad sleep, performed worse on a following test. Those results occurred even when the participants indicated beforehand, that they perceive their sleep quality as good and sufficient (Draganich & Erdal, 2014).

While those findings on sleep effects were mostly tested in a non-consumption setting, it can be assumed that the inhibiting effect of bad sleep quality has also an impact on product choice, decision making and seeking for variety or innovation. Given that consumers who suffer from a bad sleep

(14)

14 quality are not able to perceive and process all given stimuli in their shopping environment, it seems less likely that they will seek for novel or unfamiliar products which require an increased informat io n processing effort. Therefore, the following hypotheses are proposed:

Hypothesis 1: Perceived Bad Sleep Quality lowers a consumer’s level of stimulation. Hypothesis 2: Perceived Bad (Good) Sleep Quality reduces (enhances) a consumer’s

preference for stimulating products.

2.3 Exploratory Behaviour

In the absence of external stimulation, consumers tend to engage in exploratory behaviour (Steenkamp & Baumgartner, 1992). Such behaviour is motivated by the prospect of product variations or new consumption experiences (Baumgartner & Steenkamp, 1996). It may be attributed to mostly intrinsic motivations, e.g. the desire for something unfamiliar, the utility in alternating familiar brands and the consumer’s inner OSL.

2.3.1 Variety Seeking Behaviour

Variety seeking describes the consumer’s desire to switch between familiar product alternatives (McAlister & Pessemier, 1982). More precisely, it describes the behaviour of a consumer who is bored with or tired of a current product choice and explicitly seeks more stimulation in his shopping behaviour (Hirschman, 1980; McAlister & Pessemier, 1982). This behaviour is driven by external and internal forces such as a consumer’s multiple needs or changes in the choice problem (McAlister & Pessemier, 1982). These internal and external forces are referred to as changes to the consumptio n situation. In this case a new product may fit better the consumer’s adjusted needs or altered taste (McAlister & Pessemier, 1982).

(15)

15 Furthermore, interpersonal motivations also play a role in behaviour change. Individuals have an inner desire for group affiliation or individual identity (McAlister & Pessemier, 1982). On the one hand, some consumers try to stand out from their peers through their product choices. They explic it ly engage in varied behaviour by purchasing varied products to reach a higher level of external stimulation. On the other hand, some consumers are seeking social conformity and tend to imitate their peers’ product choice behaviour. In this case, variety seeking behaviour can be explained by the wish to keep up with the peers’ changing behaviour. The underlying goal is to reach some sense of affiliation (McAlister & Pessemier, 1982).

The most important motivation for variety seeking behaviour however is generated by the consumer’s inner desire for external stimulation, also referred to as intrapersonal motives. When the consumer’s level of stimulation falls below its optimum, the consumer tends to engage in variety seeking behaviour to obtain external stimuli (Steenkamp & Baumgartner, 1992).

Research revealed three distinct factors of intrapersonal motives that explain consumer’s engageme nt in variety seeking behaviour (McAlister & Pessemier, 1982; Raju, 1980). First, a consumer’s level of stimulation can be increased by switching between familiar product alternatives (Faison, 1977; McAlister & Pessemier, 1982). The consumer regularly changes between already known product alternatives to escape boredom with the current product choice. It is important to acknowledge, that, in this case, the consumer does not desire an innovative or unfamiliar product but simply a change of pace (Faison, 1977). According to Raju (1980), this alternation between familiar products involves only little risk, since the information needed to make a product choice is already known and processed. The consumer can, thus, rely on previously gained product information as well as own product experiences, which especially reduces the risk of making a dissatisfying choice, while still providing some stimulation (McAlister & Pessemier, 1982).

(16)

16 Second, some consumers engage in variety seeking behaviour to acquire new product informat io n. According to Keon (1980), the more consumers switch between brands, the more they get confused about the true value of products they did not purchase. He argues that repeated purchase of a particular product (X) causes consumers to forget about the performance (good or poor) of another product (Y). This confusion, in turn, leads to an increased probability that consumers will purchase product Y again (Keon, 1980). Based on this theory, consumers switch between products to refresh their memories about products that they have not recently bought. (McAlister & Pessemier, 1982). Besides, the acquisition of product information also facilitates consumers’ need for risk reduction for future product choice decisions (Raju, 1980).

Last, the desire for an unfamiliar product also may hint at exploratory behaviour. Consumers purchase new and innovative products to increase their level of stimulation. Details on the so called innovat ive behaviour will be presented in the next paragraph.

2.3.2 Innovative Behaviour

Leading researchers have produced several definitions of innovative behaviour (Hirschman, 1980; Midgley & Dowling, 1978; Steenkamp & Baumgartner, 1992). Innovative behaviour can be defined as the early adoption of a new product (Burns, 2007). It involves the consumer’s risk-taking tendency (Raju, 1980) as well as the search for new solutions to consumption problems (Hirschman, 1980; Steenkamp & Baumgartner, 1992). The two most prevailing conceptualizations of innovat ive behaviour differentiate between innate innovativeness (Midgley & Dowling, 1978) and inherent novelty seeking (Hirschman, 1980). While Midgley and Dowling (1978) find that innovative ness only comprises the desire to adopt to a new product, Hirschman (1980) argues that a consumer’s desire to acquire information about an innovative product (called novelty seeking) is indistinguishab le

(17)

17 from a consumer’s willingness to adopt it. In fact, she sees novelty seeking as antecedent for innovativeness (Hirschman, 1980).

Novelty seeking can be understood as the striving for an alternative new product (Hirschman, 1980). It animates the consumer to search for novel information and involves a relatively high degree of risk tolerance (Hirschman, 1980; Raju, 1980). More precisely, it bears the risk of making a dissatisfying product choice, as well as running a possible financial loss, in case the product does not live up to its promise. Furthermore, it requires the consumer not only to absorb and process all given stimuli and information that are present in his environment, but also to engage in explicit behaviour to gain more additional product information (Hirschman, 1980). According to Hirschman (1980), a consumer’s search for novel information goes back to the need to store potentially useful information that might solve future consumption problems as a means of self-preservation. Raju (1980) elaborates that the reasons for acquiring new information may also depend on the stimulation level a consumer wants to reach. Consumers with a high stimulation ideal may search for additional information due to their genuine desire for new experiences, while consumers with a lower stimulation ideal will seek for additional information to reduce potential risk (Raju, 1980).

Innovativeness means the extent to which a consumer relatively earlier than his peers tries to adopt an innovative product (Midgley & Dowling, 1978; Rogers & Shoemaker, 1971). Midgley and Dowling (Midgley & Dowling, 1978) state that consumers with a high degree of innate innovativeness make an innovation decision regardless of other consumers’ communicated experiences. In the view of most researchers, innovativeness only comprises the adoption of a newly introduced product prior to information about product experiences from a consumer’s social environment (Burns, 2007; Rogers & Shoemaker, 1971). Midgley and Dowling (1978) find that consumers differ in their reliance on other people’s assistance while making new product decisions.

(18)

18 Consumers which do not seek for their peer’s innovation experiences are susceptible to adopt early to innovative products. They are willing to take the risk of not foreseeing potential consequences, like financial losses or dissatisfaction with the product (Manning, Bearden, & Madden, 1995; Midgley & Dowling, 1978).

2.3.3 Optimum Stimulation Level Theory

As stated above, OSL is one of the main drivers of consumer’s exploratory behaviour. Past research has shown that consumers are highly influenced by stimuli in their environment and that they have a certain need for stimulation (Steenkamp & Baumgartner, 1992). The OSL varies across individ ua ls and is presented as an inverted U-shaped curve. Research shows that intermediate levels of stimulation, as opposed to high and low OSL, are commonly most preferred (Steenkamp & Baumgartner, 1992). Given the findings of Sleep Quality effects on a consumer’s ability to absorb and process environmental stimuli, the following hypothesis regarding the role of OSL is proposed:

Hypothesis 3a: A consumer’s Optimum Stimulation Level mediates the relationship between

Perceived Sleep Quality and Exploratory Behaviour.

Furthermore, according to the concept, it is important to make a distinction between the consumer’s actual stimulation level and his OSL (Sharma, Sivakumaran, & Marshall, 2010; Steenkamp & Baumgartner, 1992; Steenkamp, Baumgartner, & Van der Wulp, 1996). Only when there is a discrepancy between the actual and the optimum level of stimulation, consumers will attempt to increase or decrease stimulation (Steenkamp & Baumgartner, 1992; Steenkamp et al., 1996).

Hypothesis 3b: The difference between Current and Preferred level of Excitement mediates

(19)

19 Figure 2 shows how the variables affect each other according to the hypotheses stated above.

Figure 2: Conceptual Framework including hypotheses

3 METHODOLOGY

3.1 Setting

Participants of all ages were invited to the experiment since literature on sleep deprivation and sleep quality suggests that sleep quality effects and exploratory behaviour are relevant across all age groups. Invitations were sent to participants via mail and social media platforms, including Facebook and LinkedIn. To ensure a high reachability, a link to the experiment was further posted in forums on social media sites. The experiment was only provided in the English language to avoid a langua ge -bias between several different languages. Furthermore, the English language seemed to be a good commonly known language in the target countries.

The experiment was conducted online. The online setting was preferred over a lab environment to facilitate a high reachability in two different countries, the Netherlands and Germany. An online experiment, compared to a lab experiment, entails some great advantages. First, participants have

(20)

20 24-hours access to the experiment, which increases participants’ comfort (Dandurand, Shultz, & Onishi, 2008) and allows to target a wide range of audiences (through mailing list and social media networks). Furthermore, online experiments make it possible to easily obtain large sample sizes, that would otherwise be very time-consuming and expensive in a lab experiment (Birnbaum, 2004). Second, an online experiment is more time efficient, since experimental procedures can be automated and less time is needed for managing the experiment (Dandurand et al., 2008).

3.2 Data and Measures

This section describes the theoretical concept of the study and sets forth the independent, dependent and mediating variable. In the following, the three variables will be defined and the chosen measurement and operationalization will be explained.

3.2.1 Independent Variable

Perceived Sleep Quality is the independent variable in this research. Perceived Sleep Quality is defined as the consumer’s perception of his sleep quality, which can be either good or bad. This definition is based on the finding that changes in a person’s mindset towards perceived sleep quality can have physiological effects. Consequently, a person can overcome physiological limits by psychological means (Draganich & Erdal, 2014). Draganich et al. (2014) show how perceived sleep quality can be successfully manipulated. They operationalised perceived sleep quality as follows. First, they asked participants how deeply they had slept during the last night on a scale from 1-10, with 10 stating a very deep sleep (Draganich & Erdal, 2014). After that, participants were randomly assigned to two different conditions. In each condition, participants were given a 5-min lesson on sleep quality and its impact on cognitive functioning. After that participants were connected to a BIOPAC equipment that measured their pulse, heart rate and brainwave frequency. Participants then watched the computer calculating their sleep quality. Depending on the condition they were in (below or above average sleep quality), participants received random feedback about their sleep quality

(21)

21 (Draganich & Erdal, 2014). The results show that the given ‘false’ feedback has significant impact on consumers’ cognitive functioning. Therefore, the method is suitable to manipulate perceived sleep quality (Draganich & Erdal, 2014).

Nevertheless, the measurement of the sleep quality with the BIOPAC equipment is not feasible in this study, due to the online nature of the experiment as well as monetary constraints. Hence, the manipulation in this study will be operationalised through a questionnaire with following false feedback. The questionnaire was inspired by the Pittsburgh Sleep Quality Index (PSQI) by Buysse and colleagues (Buysse et al., 1989). It comprised questions regarding subjective sleep quality, sleep duration, habitual sleep efficiency and sleep disturbances. First, participants had to answer personal questions about their sleep habits, including average sleep hours, average time needed to fall asleep and average wake ups during night. Second, participants were asked to state how often they have experienced the proposed experiences (see Figure 3). These statements had four different response options: ‘not during the past month’, ‘less than once a week’, ‘once or twice a week’ or ‘three times or more a week’.

Third, another set of questions was included in the questionnaire to ask participants about their sleep quality, i.a. ‘Do you use a mattress that is tailored to your body?’ Response options were ‘yes’ or ‘no’ (see all questions under Figure 4).

(22)

22 Figure 3: Sleep Survey questions; sleep experiences

Figure 4: Sleep Survey Questions; external influences

In the final part of the ‘false feedback survey’, participants were asked to give some personal information, including gender, age, body weight and height as well as occupation. Here, participants had to indicate a number (e.g. Age= 24).

After finishing the ‘false feedback survey’, participants were randomly assigned to one of two conditions, following a between-subject research design. They were either assigned to the Good or Bad Sleep Quality condition.

(23)

23 In order to pretend a real processing of the sleep quality survey data, a gear wheel was shown to the participants. After a few seconds, participants were given false feedback about their sleep quality, depending on the condition they were assigned to. In condition 1, participants were shown the following sleep quality result:

Figure 5: Speedometer 77% for Good Sleep Quality Condition

In condition 2, people were shown a speedometer chart suggesting a below average sleep quality:

Figure 6: Speedometer 30% for Bad Sleep Quality Condition

3.2.2 Dependent Variable

The dependent variable is ‘exploratory behaviour’. Literature offers multiple definitions of exploratory behaviour that all have, more or less, the same core message. Consumers engage in

(24)

24 exploratory behaviour in order to reach an optimum level of external stimulation (Steenkamp & Baumgartner, 1992). Steenkamp and Baumgartner (1992) distinguish between three different forms of exploratory behaviour: curiosity- motivated behaviour, variety seeking and risk taking. Other forms of exploratory behaviour are information seeking and novelty seeking (Hirschman, 1980; Joachimsthaler & Lastovicka, 1984). This study will solely focus on the novelty seeking aspects of exploratory behaviour, the interest for novel and more stimulating products. Most previous studies have measured exploratory behaviour by using a scale. Important examples are the Sensation Seeking Scale (Zuckerman, Eysenck, & Eysenck, 1978), the Change Seeker Index (Garlington & Shimota, 1964) and Arousal Seeking Tendency (Kohn, Hunt, & Hoffman, 1982; Pearson, 1971). These scales measure the results of sets of up to 96 questions participants have to answer. They include questions about general tendencies to engage in specific behaviour, e.g. ‘I would like to try parachute jumping’ (Zuckerman, Kolin, Price, & Zoob, 1964), but do not observe actual behaviour.

A set of up to 96 questions is not very practical in an online experiment, given that participants are more likely to withdraw from online experiments. Furthermore, this research wants to focus on product choices rather than general tendencies concerning stimulating activities. Therefore, exploratory behaviour was measured by letting participants rate 4 product categories, with 18 products in total. Each product category reflects different degrees of innovativeness. The product category ‘Body Lotion’ was exclusively shown to female participants, while male participants exclusively were shown product variations for shaving creams. Product variations of ice cream and chips were shown to all participants.

The assessment of the product preferences was structured as follows. First, participants were shown at least one ‘common’ product, available in every grocery store and produced by a known brand (e.g. Nivea and Häagen Dazs). This familiar product was described in a short sentence e.g. ‘The Classic

(25)

25 Nivea Cream’ to ensure familiarity. Following the familiar product choice, three to four imaginary increasingly stimulating product variations were shown to the participants. Those variations were either flavour variations (e.g. Häagen Dazs ice cream with mango-champagne flavour) or entirely new creations (e.g. Nivea Shaving Capsules), which are not available in the stores. These products were subtitled with more exciting product descriptions, e.g. ‘Häagen Dazs newest composition of mango and champagne bubbles that explode in your mouth and release liquid, tingling champagne.’, to explain the product and provide more excitement.

After the presentation of the product and a short product description, the participants’ preference for products was assessed by four questions. Participants had to indicate how much they like the product, how likely they are to buy or try the product, whether they are interested in more information about the product and how they would rate its degree of innovativeness. Answers were measured on a 7-point scale, with 1 indicating the highest and 7 the lowest preference for the product. An example of the product presentation and the questionnaire can be seen in Figure 7.

The first product category ‘Shaving Creams’ included 4 different products. The first product was the regular ‘Nivea Men shaving foam’ which is widely known and available in every grocery store. This product was chosen since it was meant to not substantially raise a consumer’s interest for the product. A participant who chooses this product is not seeking for further stimulation, he rather sticks with a familiar product. Secondly, participants were shown fictious variations of the standard product or new developments which are not available on the market. Participants who rated high on these options expressed their desire for the unfamiliar. Other product categories were ‘Ice Cream’, ‘Body Lotions’ and ‘Chips. The different products can be seen in the survey overview in the Appendix.

(26)

26 Figure 7: Example product presentation and questions

3.2.3 Mediator Variable

The mediating variable determines the relation between the independent and the dependent variable (Baron & Kenny, 1986). Whereas a moderating variable explains when a certain effect will occur, the mediating variable specifies how or why the variables are affecting each other (Baron & Kenny, 1986). As outlined before, OSL is one prerequisite for consumers’ exploratory behaviour, as a lowered optimum stimulation level entices consumers to engage in exploratory behaviour to achieve a higher level of stimulation (Steenkamp & Baumgartner, 1992). Therefore, OSL might function as mediator in the relation between the independent variable ‘perceived sleep quality’ and the dependent variable ‘exploratory behaviour’.

(27)

27 The Change Seeker Index (Garlington & Shimota, 1964) is regarded as the most suitable measurement for OSL (Steenkamp & Baumgartner, 1992). This index is composed of a one-dimensional scale of 95 items. It measures a person’s need for variation in his stimulation input in order for this person to maintain ideal functioning (Garlington & Shimota, 1964). However, this scale is not very practical due to its length encouraging Steenkamp and Baumgartner to develop a 7-item scale to measure OSL (Steenkamp & Baumgartner, 1995). The scale includes statements like ‘When things get boring, I like to find some new and unfamiliar experience’ or ‘I like a job that offers change, variety, and travel, even if it involves some danger’ (Steenkamp & Baumgartner, 1995). This form of OSL measurement includes aspects of exploratory behaviour as it is also meant to measure a consumer’s willingness to engage in exploratory behaviour. Key statements like ‘I like to continue doing the same old things rather than trying new and different things’ (Steenkamp & Baumgart ner, 1995) give a hint at the purpose of the test – to establish what exploratory tendencies participants have.

Another way of measuring level of stimulation was introduced by Bradley and Lang (1999). They tested affective norms for English words on a visual arousal scale. For this purpose, they used mannequins which displayed five different arousal states with calm and excitement as two extremes (see mannequins used in questionnaire; Figure 8) (Bradley & Lang, 1999).

For the purpose of this study, such visual scale-measurement seems to be the more appropriate approach. The mannequins do not hint at exploratory behaviour or product preferences. Therefore, OSL will be measured visually, by letting participants select their preferred and current level of stimulation via a slider, ranging from 0 to 100. Above the slider, five different figures showed different levels of excitement, from feeling ‘sleepy’ (equivalent to 0) to ‘bursting from exciteme nt’ (equivalent to 100; see Figure 4 for ‘preferred’ level of excitement).

(28)

28 Figure 8:Slider for Preferred Level of Excitement

This visualisation has two advantages. First, participants were not given hints for the follow ing product choice section and therefore, their choice is not being influenced by previously asked questions about their tendencies to engage in exploratory behaviour or preferences for stimulat ing products. Second, the slider also offered the opportunity to see an immediate effect of the false sleep quality feedback. This allowed to establish whether their current level of stimulation matched the false feedback they were given in the good or bad sleep quality condition.

3.2.4 Control Variables

In order to eliminate other influencing factors on the effect of perceived sleepiness on exploratory behaviour, control variables were included in the experiment. Literature suggests that gender has an important influence on both exploratory behaviour and OSL. Research found a significant difference between men and women regarding decision-making and exploratory behaviour. Some studies

(29)

29 suggest that women tend to be more impulsive buyers, while men rely on product information and tend to make product choices less prompt (Mitchell & Walsh, 2004). A study by Surajit et al. (2009) found that men are willing to take more risks and more easily adopt to innovative behaviour as women. Other researchers found that sensation seeking is higher in males than females and that these differences might also be for different product choices (Zuckerman, 1979). Males tend to try new brands more often and purchase new financial products, whereas females tend to be more interested in groceries (Tuteja, 2017). Based on these findings, gender was included as covariate for exploratory behaviour as well as OSL.

Second, participants’ age can also have significant influence on a person’s exploratory behaviour tendencies and level of stimulation. Previous studies found a curvilinear relationship between age and exploratory activity. Exploration tendencies increase to a certain level of age and then fall off with an increase in age (Kish & Busse, 1968). Middle aged people score highest on OSL (Kish & Busse, 1968), while older people are more likely to explore new products by shopping and informat io n seeking (Urbany, Dickson, & Kalapurakal, 1996). This age group engages extremely in informat io n seeking and they regularly experience information overload (Mitchell & Walsh, 2004). Younger people of generation Y tend to engage more in information seeking about new technology, are keener on branded products and switch brands more often (Tuteja, 2017). In conclusion, the variable age was chosen as covariate for both OSL and exploratory behaviour.

Third, a person’s occupation was found to be significantly correlated with OSL by Raju (1980). Employed people are higher on OSL than unemployed people (Raju, 1980). Although several studies found a correlation between education and exploratory behaviour (Dastidar & Datta, 2009; Urbany et al., 1996), no research has further examined a relationship between occupation status and

(30)

30 exploratory behaviour. Therefore, the variable Occupation will only be used as covariate for OSL, not for exploratory behaviour.

3.3 Method

For the analysis of the collected data the data analysis tool SPSS was used. In order to establish whether the manipulation was successful and to test the hypotheses made in this research, three main methods were applied. First, ‘independent-samples t-tests’ were conducted to compare the means of outcome variables in the good and bad sleep quality conditions. The independent-samples t-test is a simple and easy-to-apply method that examines differences between two groups (Field, 2009). It tests the null hypothesis that the means of two groups significantly differ from each other, suggesting that the assignment to different groups has an influence on the outcome variable (Pallant, 2013). The hypothesis will be tested on a 95%-significance interval, meaning that a p-value equal to or below .05 suggests that the null hypothesis can be accepted. A value above .05 instead rejects the null hypothesis and means that there is no difference between the means of the two tested groups (Pallant, 2013).

The second method applied in this research is an analysis of covariance (ANCOVA). ANCOVA not only tests the difference of means between two or more conditions (similar to the independent-samples t-test), it also takes into account the variability of other variables which are called covariates (Field, 2009). These covariates are not of main interest in the research but might influence the relationship between the independent and dependent variable. Providing a constant covariate, ANCOVA delivers adjusted means for the outcome variable, eliminating any influence of covariates on the means between treatment groups (Field, 2009).

(31)

31 The third method concerns the mediation effect hypothesized in this research. A mediation is a causal model. This means that the mediator is influencing the outcome variable and not the other way around. Testing mediation can reveal the underlying mechanism through which the causal variables affect the outcome. Several methods can test mediation, including the causal steps approach by Baron and Kenny (1986). This approach involves the estimation of each of the paths in the model and tests mediation by seeing if statistical criteria are met (Hayes, 2009). This approach has been subject to criticism as it is arguably the least likely method to actually detect the effect of mediation. A reason for this is that the indirect effect of the independent variable X on the dependent variable Y through the mediator M, is only inferred logically and not quantitatively tested (Hayes, 2009). This is why Hayes (2013) introduces the analysis tool PROCESS that can statistically test direct and indirect effects. The direct effect of X on Y is explained by c’. C’ states how much two cases which differ by one unit on X are estimated to differ on Y, without the effect of M on Y. The indirect effect a1b1 is the product of the effect of X on M (a1) and the effect of M on Y controlling for X (b1). The total effect c of X on Y is the sum of the direct and indirect effect.

4 RESULTS

4.1 Descriptive and Frequencies Analysis

This section summarizes and discusses the variables used in this research. Hereby, descriptive and frequency statistics will be examined in detail.

In total, 277 people participated in the online survey. However, only the data of 168 participants could be used in the analysis due to incomplete or suspicious responses. 61 people did not complete the survey and were immediately deleted from the data. 30 participants who needed + 3*SD of the

(32)

32 average response time (M=13.25;SD=8.68) were excluded from the data due to data quality. Another 18 participants were excluded from the data due to suspicious response patterns, where all answers followed the same answer scheme and casted doubt about their reliability.

The age of the respondents ranged from 15 to 58 years. The mean age of all participants (N=168) was 30.06 years (SD=11.008), whereby 58% of the respondents were between 22 and 26 years old. 55 % of the participants were female (N=93) and 45% were male (N=75). The occupation scale shows that the majority of participants in this research were students (42.3%), followed by employees (35.7%), working students (11.9%), self-employed workers (7.1%), and unemployed persons (3%).

In the manipulation section of the experiment, all participants (N=168) were randomly assigned to one of two conditions, Good Sleep Quality or Bad Sleep Quality. 88 participants were assigned to Condition 1 (Good Sleep Quality) and 80 were assigned to Condition 2 (Bad Sleep Quality). After participants were shown the good or bad sleep quality result, they rated their Current Level of Excitement on a scale from 0 to 100, with 0 meaning ‘fully relaxed’ and 100 meaning ‘extremely excited’. Descriptive analysis shows that participants who were assigned to the Good Sleep Quality condition had an average Current Level of Excitement of 44.5 (SD=20.407), while participants in the Bad Sleep Quality condition had an average Current Level of Excitement of 50.01 (SD=21.172). The average Current Excitement Level of the entire sample was 47.13 (SD=20.90). Q-Plot analysis shows that both Current and Preferred Level of Excitement are not normally distributed. The Kolmogorov-Smirnov as well as the Shapiro-Wilk test both provide values below 95% confidence interval. The p-value is below .05 meaning that the null hypothesis can be rejected. Thus, the data is not normally distributed.

(33)

33 In the next step of the experiment, the dependent variable Exploratory Behaviour was tested by showing participants four different product categories, with four to five varying product alternatives each. The two product categories Chips and Ice Cream were shown to all participants (N=168). The remaining two categories were divided by gender. 93 female participants were shown the Body Lotion category, whereas 75 male participants were shown the Shaving Cream product category.

4.2 Preparatory Analyses

This paragraph provides a manipulation check for the independent variable Perceived Sleep Quality and the dependent variable Exploratory Behaviour to ensure that the hypotheses can be tested with the materials used.

4.2.1 Manipulation Check – Sleep Quality and Current Level of Excitement

This section analyses whether the manipulation in this research was successful and thereby answering Hypothesis 1, whether perceived sleep quality has an effect on participants’ current level of stimulation.

Before conducting an independent-samples t-test, it is important to see whether homogeneity of variance can be assumed, as this has severe consequences for the interpretation of the independent -samples t-test results (Field, 2009). To test the homogeneity of the variance, Levene’s test will be used. It tests the null hypothesis that the difference between the variances in the different groups is zero, meaning that the variances are homogenous (Levene, 1960). The test was conducted for Current Level of Excitement, Last Night Sleep and Tired Now as dependent variables and Good Sleep Quality and Bad Sleep Quality as factors. Levene’s test for homogeneity of variances was found to be violated for the present analysis of the Current Level of Excitement, F(1, 166)= .677, p= .412. For Last Night Sleep, variances were also equal for Bad and Good Sleep Quality conditions, F(1, 166)= 3.327, p= .070. Owing to this violated assumption of homogeneity, the results of the independensamples

(34)

t-34 tests for both Current Level of Excitement and Last Night Sleep will be computed while assuming equal variances. In contrast, Levene’s test of homogeneity for Tired Now was not violated, F(1,166)=5.844, p=.017. This result means that equal variances cannot be assumed in the follow ing analysis.

First, an independent-samples t-test was conducted to compare participant’s current level of excitement in the good sleep quality and bad sleep quality conditions. There was not a significa nt difference in the scores for Good Sleep Quality (M=44.5;SD=20.407) and Bad Sleep Quality (M=50.01;SD=21.172); t(166)=-1.718, p=.088. The results suggest the manipulation of the participant’s sleep quality has a marginally significant effect on their current level of stimulation. In both conditions, people experience a relatively similar current level of excitement, with a tendency to differ. This result rejects the assumption, made in hypothesis 1, that perceived sleep quality affects a person’s current level of excitement.

Second, an independent-samples t-test was conducted to compare how participants rated their last night’s sleep in condition 1 and 2. There was not a significant difference in the scores for Good Sleep Quality (M=3.00;SD=1.654) and Bad Sleep Quality (M=3.19;SD=1.519); t(166)= -.763, p= .447. The result shows that participants in the two conditions did not rate their last night’s sleep differently after the manipulation. The assignment to either condition 1 or 2 has no effect on how people perceive their last night’s sleep. Therefore, it can be assumed, that sleep quality manipulation does not influence people`s perception about themselves and their actual sleep quality.

A third independent-samples t-test compared how tired participants felt right after the assignment to condition 1 or 2. Also for Tired Now, no significant difference in the scores for Good Sleep Quality (M=3.59;SD=1.609) and Bad Sleep Quality (M=3.83;SD=1.465) was found; t(165,999)=-.987,

(35)

35 p=.325. It seems, that the assignment to either condition has no influence on participants’ sleepiness right after the manipulation and further suggests that a good or bad sleep quality perception does not manipulate how people feel in a particular moment.

4.2.2 Manipulation Check - Exploratory Items

In order to measure the dependent variable ‘exploratory behaviour’, different kinds of products in one product category need to be chosen. The selection of the products is based on their degree of innovativeness perceived by the participants in this research. In this case, the variable Innovation is of high interest, since it depicts a starting point for choosing the products per category. Paired sample t-tests were conducted to compare the degree of innovativeness of all products in each product category.

In the shaving cream category, there was a significant difference in the scores for classic shaving cream (M=5.57;SD=1.327) and shaving capsules (M=3.52;SD=1.884); t(74)=7.575, p=.00. The results suggest that classic shaving cream and shaving capsules have a different degree of innovativeness. To be more specific, shaving capsules are perceived as more innovative than a classic shaving cream. Although the remaining two products were all also significantly different from classic shaving cream, they were perceived less innovative by the participants. Furthermore, to test whether there might be a third product that is significantly different from the former two products, two more paired sample t-tests were conducted. The first test showed that there was no significant difference between shaving capsules (M=3.52;SD=1.884) and shaving mask (M=3.77;SD=1.805); t(74)= 1.249, p=.22. The second test did not yield a significant difference between shaving capsules (M=3.52;SD=1.884) and anti-stubbles shaving cream (M=3.77;SD=1.977); and t(74)=-1.184, p=.24. The results suggest, that shaving capsules, shaving mask and anti-stubbles shaving cream are

(36)

36 perceived as similar innovative by the participants. Therefore, only scales concerning the two products classic shaving cream and shaving capsules will be used in the following analyses.

Paired sample t-tests were conducted for the ice cream category. There was significant difference between vanilla ice cream (M=5.32;SD=1.505) and champagne ice cream (M=3.44;SD=1.687); t(167)=11.450, p=.00. This result suggests that champagne ice cream is perceived as more innovat ive than vanilla ice cream. The same result showed paired sample t-tests conducted for vanilla ice cream and the remaining three products beer ice cream, beetroot ice cream and coffee ice cream. They were all perceived as more innovative than vanilla ice cream. Nevertheless, the mean of champagne ice cream (M=3.44) was lower than the means of the other products, suggesting that participants perceived champagne ice cream as more innovative than beer, beetroot and coffee ice cream. Paired sample t-tests support this suggestion. There was no significant difference between champagne ice cream (M=3.44;SD=1.687) and coffee ice cream (M=3.66;SD=1.524); t(167)=-1.718, p=.09; champagne ice cream (M=3.44;SD=1.687) and beer ice cream (M= 3.64;SD=1.900); t(167)= -1.376, p=.17; champagne ice cream (M=3.44;SD=1.687) and beetroot ice cream (M= 3.60;SD=1.854); t(167)=-1.077, p=.28. The analysis shows that for the product category ice cream, only two products, vanilla ice cream and champagne ice cream, significantly differ in terms of their degree of innovativeness and therefore, only scales belonging to these two products will be used for further analyses.

In the chips product category, three different products were chosen according to the results of paired sample t-tests. There was a significant difference between salt chips (M=5.90;SD=1.307) and pepper chips (M=5.41;SD=1.522); t(167)=4.358, p=.00. The result suggests that participants perceived pepper chips as more innovative than salt chips. Another paired sample t-test yielded the follow ing result: there was a significant difference between pepper chips (M=5.41;SD=1.522) and chicken chips

(37)

37 (M=4.11;SD=2.039); t(167)= -6.702, p=.00. This suggests that chicken chips are perceived as more innovative than pepper chips. The same result was yielded for the mean comparison of pepper chips and pickle chips. Nevertheless, there was no significant difference between chicken chips (M=4.11;SD=2.039) and pickle chips (M=4.21;SD=1.649); t(167)=0.798, p=.43. This suggests that chicken chips and pickle chips have the same perceived degree of innovativeness. Therefore, only salt, pepper and chicken chips will be included in the following analyses, since pickle chips do not represent an additional degree of innovativeness.

Finally, paired sample t-tests were conducted for the products in the body lotion product category. There was a significant difference between classic body cream (M=5.56;SD=1.514) and ginger body lotion (M=3.46;SD=1.691); t(92)=9.577, p=.00. The results suggest that the classic body cream is perceived as less innovative than the ginger body lotion. A second paired sample t-test showed a significant difference between ginger body lotion (M=3.46;SD=1.691) and body oil pearls (M=2.89;SD=1.387); t(92)=3.857, p=.00. Ginger body lotion and body oil pearls seem to differ in terms of innovativeness. To be more concrete, the result suggests that body oil pearls are perceived as more innovative than ginger body lotion. There was no significant difference between classic body cream (M=5.56;SD=1.514) and body milk (M=5.32;SD=1.568); t(92)=-1.674, p=.10. Body milk and the classic body cream to not differ in terms of innovativeness, both are seen as less innovat ive. Furthermore, there was no significant difference between and body oil pearls (M=2.89;SD=1.387) and thermal water spray (M=3.03;SD=1.433), t(92)=-1.059, p=.29. Also, body oil pearls and thermal water spray do not differ in their degree of innovativeness and are both perceived as relative ly innovative. These results led to the selection of three different products, classic body cream, ginger body lotion and body oil pearls, which all have different levels of innovativeness and therefore will be used for further analysis.

(38)

38

4.3 The Effect of Sleep Quality on Preferred Level of Excitement and the

Excitement Difference Score

4.3.1 Preferred Level of Excitement as DV

An independent samples t-test was performed to test, whether perceived sleep quality affects a person’s preferred level of excitement. The result of this test also shows that no significant difference in the scores for Good Sleep Quality (M=58.03;SD=22.558) and Bad Sleep Quality (M=61.05;SD=23.477) were found; t(166)=-.849, p=40. Equal variances were assumed, F(1,166)=.32, p=.57. This suggests that perceived sleep quality has no influence on a person’s preferred level of excitement and rejects the hypothesis 1 made in this research.

In order to rule out other possible influences of covariates on a person’s preferred level of exciteme nt, an ANCOVA was performed. As already described, the variables Age, Gender and Occupation were included in the analysis, since literature suggests significant correlations between these variables and the outcome variable OSL (Raju, 1980; Zuckerman & Donohew, 2015).

Before performing the main analysis, two assumptions need to be fulfilled: (1) that the covariate and the treatment effect are independent of each other and (2) that regression slopes are homogeneous. A one-way ANOVA showed that Age did not significantly differ between condition 1 and 2; F(1,166)=.004, p=.95. There is no difference in age between participants in the Good or Bad Sleep Quality condition and thus, age is a suitable covariate. Further analysis revealed that the means of Gender were not significantly different in the Good and Bad Sleep Quality condition, F(1,166)=.05, p=.83. Furthermore, Occupation was also not found to be significantly differing between the two conditions, F(1,166)=.81, p=.37. A test for homogeneity of regression slopes showed that all three variables Age, Gender and Occupation have a linear relationship with the independent variable

(39)

39 Preferred Level of Stimulation across both treatment groups (Good and Bad Sleep Quality), all with p>.05. The results suggest that all three variables Age, Gender and Occupation are suitable as covariates in an ANCOVA.

As can be seen in Table 1, an ANCOVA revealed that there was no significant effect of Sleep Quality on Preferred Level of Excitement after controlling for the effects of Age, F(1,163)=.05, p=.82; Gender, F(1,163)=.13, p=.72; Occupation, F(1,163)=.49, p=.49. To be more concrete, whether a person believes his sleep quality is good or bad has no influence on the optimum level of stimulat io n he prefers.

Table 1: ANCOVA for Perceived Sleep Quality on Preferred Level of Excitement

Source Sum of Squares df Mean Square F Sig. Gender 69.34 1 .82 .13 .72 Occupation 259.66 1 259.66 .49 .49 Age 28.51 1 28.51 .05 .82 Condition 440.73 1 440.73 .82 .37 Error 87176.64 163 534.83 DV: Preferred Excitement

4.3.2 Excitement Difference Score as DV

To give an even more clear picture of the potential effect of sleep quality, the analys is was repeated for the preferred excitement by comparing current excitement levels and desired excitement levels. An Excitement Difference Score was calculated to see whether participants prefer a rather high or low level of stimulation. To assign a value to this difference, a gain score was calculated, subtracting Preferred Level of Excitement from Current Level of Excitement. A negative score indicates that a participant’s current level of stimulation is lower than his preferred level of stimulation. In contrast, a positive score means that a person has a higher current level of stimulation than he usually feels

(40)

40 comfortable with. Participants’ excitement difference score had a mean of -12.34 (SD=21.293). It seems that, on average, participants prefer a slight raise in stimulation as opposed to relaxation.

Taking Excitement Difference Score as dependent variable, an independent-samples t-test showed that there was no significant difference between the scores for Good Sleep Quality (M=-13.52;SD=20.63) and Bad Sleep Quality (M=-11.04;SD=22.06); t(166)=-.76, p=.45, assuming equal variances F(1,166)=1.30, p=.26. Participants’ assignment to either condition 1 or 2 did not yield differences in the Excitement Difference Score. To be more clear, it can be assumed that in both Good and Bad Sleep Quality condition people have similar differences between current and preferred level of excitement, so that they were not manipulated in this research.

This result was further tested in an ANCOVA to rule out the covariates Age, Gender and Occupation as possible influencing factors in this relation between Sleep Quality and Excitement Differe nce Score. Testing for homogeneity of regression slopes, results showed that all three variables have a linear relationship to the outcome variable Excitement Difference Score, with p>.05. An ANCOVA (see Table 2) revealed that there was no significant effect of Sleep Quality on Excitement Differe nce Score after controlling for the effects of Age, F(1,163)=2.12, p=.15; Gender, F(1,163)=.02, p=.90; Occupation, F(1,163)=.04, p=.84. This result further shows that there is no effect of the two sleep quality conditions on the difference between current and preferred level of excitement, rejecting hypothesis 1.

The impact of perceived sleep quality on consumer's exploratory behaviour

Master Thesis