The effect of risk reduction methods when purchasing an insurance

The effect of risk reduction methods

when purchasing an insurance

Studentname: Froukje Canter Cremers

Studentnumber: 6377920


Supervisor: Prof. Dr. Ed Peelen


Second supervisor: Drs. ing. A.C.J. Meulemans

Master of Business Studies

Abstract

Marketers are interested in understanding what causes individuals to perceive risks when they purchase goods and services so that they can develop relevant strategies to reduce those risks and increase purchase probabilities. It is suggested that the charateristics of service and mainly the intangibility of services greatly increase the degree of perceived risk in the purchase phase of a proposition. A review of the literature shows that little research has been done considered the relationship between perceived risk and purchase intention for services and insurances in particular and the effect risk reduction methods can have on this relationship. The present study reports on the differenent elements of perceived risk and their relationship to purchase intentions, and test the relationship of seven risk reduction methods in the purchase of a contents insurance. The results confirm a strong impact of all the risk dimensions on purchase intentions. Perceived Financial- and Psychological risk are the best predictors of doubt with the purchase decision.The research didn’t support the hypotheses that insurance tend to be evaluated risky when purchasing but the effect of perceived risk on purchase intention is strongly supported.

Table of Content

1. INTRODUCTION

5

2. LITERATURE REVIEW

8

2.1 Perceived risk 8

2.1.1 Types of perceived risk 8

2.1.2. Perceived risk and online environment 9

2.1.4 Services 12

2.1.5 Services and Perceived risk: 14

2.1.6 Intangibility and perceived risk 15

2.2 Purchase intentions 15

2.2.1

2.3 Risk relievers 17

2.3.1 Brand experience 17

2.3.2. Customer reviews 17

2.3.2. Comparison websites 18

2.3.3. Money back garantuee 19

2.3.4.Service cues 20

2.3.5. Most expensive insurance coverage – expensive model 20

2.3.6. Offline expert (insurance brooker) 20

3. METHODOLOGY

22

4. RESULTS

29

4.2 Reliability of the perceived risk scales 31

4.3 Percieved risk and experienced doubt 31

4.4 Measures, experienced risk, and purchase intention 32

4.5 Hypotheses result: 37

5.1 DISCUSSION

39

5.1.1 Perceived risk and purchase intentions 39

5.1.2 Risk reduction methods and their effects on perceived risk and purchase intentions 39

5.2 MANAGERIAL IMPLICATIONS

41

5.3 LIMITATIONS AND FUTURE RESEARCH

43

5.4 CONCLUSIONS

44

6. REFERENCES

45

ANNEX 1: THE RESEARCH QUESTIONAIRE

51

1. Introduction

When customers buy a new product they always have to deal with a certain amount of risk. Any choice situation involves two aspects of risk: uncertainty of the outcome and uncertainty about consequences. To minimize this risk they experience in the pre-purchase stage, consumers can rely on various risk-reduction methods. That literature traditionally ignored the role of perceived risk at the purchase phase of the consumer buying process.

There are many risk reduction methods, for instance a customer can seek information from formal and informal sources, make use of money back guarantees or free gifts with the purchase or free trial purchases (Akaah and Korgaonkar, 1988). Studies by Roselius (1971) and Derbaix (1983) have shown that some risk relievers may receive a higher or lower preference among consumers if associated with certain methods of purchase or with certain types of products.

According to Cox (1967) customers perceive risk in most store purchase decisions. Given the fact that a consumer cannot always be certain that all of his or her buying goals will be achieved, risk is perceived in most purchase decisions (Cox, 1967). Research has shown that consumers perceived financial, product performance, psychological, physical, social, and time risks when making purchases (e.g. Jacoby and Kaplan, 1972; Peter and Tarpey, 1975; Garner, 1986; Mitchell, 1992; Schiffman and Kanuk, 1994).

Further, according to Akaah and Korgaonkar (1998) customers associate a higher level of risk with non store purchase than store purchase decisions. Unlike offline retail stores, the inability to interact with a salesperson and the merchandise, as well as the reliance on electronic payment methods increases percei ed ris ith re ard to on ine shoppin asa o et al., 2007).

In the insurance industry this is a relevant subject as you can see in figure 1. The market of purchasing insurances from a broker is shrinking and the direct market (internet) is raising.

Figure 1: Purchase channels insurances 2005-2013

Source: Internal numbers of Product Management Nationale Nederlanden 2013

In the marketing literature, research has shown that the use of certain risk reduction strategies such as brand reputation, product trial, and warranty are successful in reducing the risk perception of consumers (e.g. Roselius, 1971; Shimp and Bearden, 1982; Innis and Unnava, 1991; Boulding and Kirmani, 1993).

But besides the growing purchases of insurances on the internet the insurance industry as is, is also perceived more riskier then other industries. Insurances are part of the service industry. Since 1960 services has been seen as a subdicipline of goods. The main reason for this seperate treatment is that services differ from goods by a few characteristics, reviewed by different researches. Frisk, Brow and Bitner (1993) concluded that four features are exhaustive for the case that services marketing is a field distinct from goods marketing. These four features are intangibility, inseparability, hetrogeneity and persishability.

But throughout the years researches began to doubt the differences between goods and services. Gummesson (2007) stated that there is no such thing as service marketing vs goods marketing. Gummesson (2007) states that when you take any service and there will be a goods element and take any good and there are service elements involved. Furthermore he explaines that customers do not buy goods but that they buy something that they perceive to be of value for them. The terms goods and services can still be used and should not be abandoned but they should be used with more finesse.

Intangibility and the other characteristics of IHIP can still be used for the value proposition but they do not distinguish between goods and services. Because intangibility is an important characteristic of how customers experience insurances and because it is further exacerbated by the raise of purchases on the internet we will use intangibility as a dependent variable of perceived risk of insurances.

The question if risk reductionmethods are effective to for perceived risk when purchasing an insurance and the effect perceived risk has on purchase intention for insurances have not been addressed yet. Most of the existing literature focuses on the advantages and disadvantages of risk reduction methods but very little research has been done about customers perceived risk in when purchasing insurance and the effect it has on the purchase intention.

Therefore in this paper we will address the issue of perceived risk in the insurance industry and the potential effect on purchase intention. We will determine the risk internet shoppers of insurance products perceive and their responses to the different risk reduction methods available and the effect on their purchase intentions.

This study is interested in the effectiveness of some of these risk relievers in the internet shopping insurance context.

This leads to the following research question: what is the relationship between perceived risk and purchase intention and what is the relationship of the chosen risk reduction methods with both perceived risk and purchase intention.

The structure of this thesis is as follows. Chapter two contains the literature review were we will shed light on the existing literature on perceived risk, purchase intention, risk reduction methods and services/insurances. In chapter 3 the methodology and data are of this research will be describedand hypotheses and the research design are incorporated. Chapter four contains the results of the different analyses and results found for our hypotheses. The research ends with the discussion of our research results managerial implications, limitations and the conclusion. .

2. Literature review

2.1 Perceived risk

When customers buy a new product they always have to deal with a certain amount of risk. In consumer behavior literature the central problem is choice. The risk element is present because before making a choice to purchase the article or service, the customer has specific buying goals and the outcome of a choice can only be known in the future, so the customer has to deal with a certain amount of risk/uncentaincy.

In the 1960’s Bauer first introduced the concept of percei ed ris to the mar etin iterature. Bauer introduced the idea that consumer behavior can be considered as an instance of taking and risk reducing behavior. Bauer (1960) argued that consumer behavior is risk-taking behavior because actions of a consumer may produce unanticipated consequences, some of which may be unpleasant. Cunningham (1967) conceptualized perceived risk in terms of uncertainty and consequences. Uncertainty is associated with a situation for which there are various possible outcomes. Consequences are the results of the various outcomes. Other researchers have conceptualized perceived risk as a function of two different components: the probability of loss occurring (uncertainty) and the importance of the loss (consequences) (Bauer, 1960; Bettman, 1972; Bettman, 1973; Mitchell, 1999).

According to Cox and Rich (1964) percieved risk refers to the nature and amount of risk perceived by a consumer in contemplating a particular purchase decision.

O er the years percei ed ris is most defined as ris in terms of the consumer’s perceptions of the uncertainty and adverse consequences of buying a product (or service) (Bauer, 1960; Bettman, 1972; Bettman, 1973; Mitchell, 1999).

2.1.1 Types of perceived risk

Regardless of different operational definitions of perceived risk, it is generally thought to have multiple dimensions (Bettman, 1973; Cunningham, 1967; Jacoby & Kaplan, 1972; Kaplan, Syzbillo, & Jacoby, 1974). Jacoby and Kaplan (1972) conceptualized five different types of risk (financial, performance, psychological, physical, and social risk). In other research, Roselius (1971) identified time risk as well as the five types of risk identified by Jacoby and

Kaplan. More recent research of Mitche and Harris 2005) su est that shoppers’ moti es are linked to four main perceived risk dimensions: time, financial, psychosocial and physical. Based on earlier research most common found types are financial, performance, psychological, physical, social and time risk and the operational definitions for these six types of risk are descriped as follows by Bauer (1960), Jacoby and Kaplan (1974), Murray and Schlacter (1990), Chang (2008) and Boksbergen et al. (2007):

(a) Financial risk refers to the probability that purchase results in loss of money or other resources,

(b) Performance risk refers to the probability that a product purchased results in failure to function as expected,

(c) Social risk refers to the probability that a product purchased results in disapproval by family or friends,

(d) Psychological risk refers to the probability that a product purchased results in inconsistency with self-image,

(e) Physical risk refers to the probability that a product purchased results in personal injury, and

(f) Time risk refers to the probability that a purchase results in loss of time to buy or retain the product

Researchers have found that the six types of risk explain significant variation in overall perceived risk (Stem, Lamb, & MacLachlan, 1977; Stone & Gronhaug, 1993).

Jacoby and Kaplan (1974) did a cross-validation of their previous research to their multiple dimensions of perceived risk from 1972, to make sure the findings were the same two years later. They found simular results in this cross validation reserach so we can estimate that financial, time, performance, psychological and social risk are the five types of risk possiblyexperienced by the customer. Besides the cross validation of Jacoby and Kaplan (1974) these types of risk have been selected because they were the most prominent and widely discussed in the perceived risk literature.

2.1.2. Perceived risk and online environment

The majority of research on perceived risk is focused on the traditional purchasing situations (in stores). But with advances in modern technology, the Internet population has increased year by year globally. For young customers who consider convenience and speed as prerequisites, online shopping has become a new type of consumption. Online shopping

differs from shopping in stores. Internet shopping technologies are essentially self-service technologies that offer the benefits of round-the-clock convenience, ubiquitous availability, time and money savings, and a reduction in the anxiety caused by judgmental service representatives (Bitner, 2001; Meuter et al., 2000). Consumers have realized the benefits of online shopping, such as saving time and energy, convenience, competitive pricing, broader selection, and greater access to information (Verhoef & Langerak, 2001). When internet was upcomin in the 80’s and 90’s some would argue there are disadvantages to internet shopping such as computer system complications, computer phobia, and loss of pleasure and social interaction (George, 1987). Internet shopping and perceived risk has been studied generally. Researchers (Vijayasarathy and Jones, 2000) found perceived risk to be a significant factor affecting internet consumer behavior. Liebermann and Stashevsky (2002) and Forsythe and Shi (2003) provide evidence to support a relationship between perceived risk and frequency of use. Forsythe and Shi (2003) and Soopramanien (2011) state that customers perceive risk when purchasing online but the research of Soopramanien (2011) postulate that online shopping experience has a direct effect as well as an indirect effect on the intention to use online shopping and the involved perceived risk when using.

But also before the wide diffusion of internet non-retail shopping and risk has been studied. The most early researchers have found consumers to associate more risk with home shopping (e.g., telephone, catalog, and door-to-door sales) than with store shopping (i.e., department stores) (e.g., Cox & Rich, 1964; Spence, Engel, & Blackwell, 1970). For example, Cox and Rich (1964) found greater perceived overall risk for telephone shopping as compared with store shopping. Spence et al. (1970) found that consumers perceived mail order as more risky than store shopping or than directly purchasing from a salesperson. Festervand et al. (1986) found that financial, performance, and time risks were perceived to be greater for catalog shopping than store shopping.

This past research outcomes are still be seen nowaydays when examining online customers. Consumers are sometimes unwilling to shop online because limited Web trust hampers their judgment of retailer honesty compared to face-to-face interaction (Reichheld & Schefter, 2000) But also more recent studies of (Boksbergen et al. 2007 and Chang 2008) refer to financial, product performance, social, psychological, physical and time risk when customers make transactions online.

The research of Hoffman et al. (1999) estimate that only the four types of perceived risk described below are experienced in an online purchase.

Financial risk is experienced by customers by apparent sense of insecurity regarding online credit card usage stems primarily from a concern about financial risk. Consumers’ unwillingness to provide their credit card information over the Web has been cited as a major obstacle to online purchases (Maignan and Lukas, 1997). Many consumers believe that it is too easy to have a credit card stolen online (Caswell, 2000). Even in 2014 creditcardcompanies and websited warn us for the risk of online transactions (e.g. newsdaily, mastercard and visa)

Product performance risk is experienced online because of the ability to judge service quality online may be limited by barriers to touching, feeling and trying the product or service, inaccurate product colors and insufficient information on quality attributes relevant to the consumer resulting in increased product performance risk. For insurances this is slightly different because of the intangibility of an insurance offline as well as online. For insurances performance is more is the cues and missing human contact when purchasing online which may have influence on perceived performance risk. In the specific example of insurrances the customer is due to the service characteristics which are mentioned above, not sure how a product might perform. Above that the customer is faced with a difficult value trade-offs, for example the price of their insurance versus the coverage the customer gets in return.

Psychological risk may occur because the Internet is often perceived as likely to violate users privacy, a major concern of many Internet users (Maignan and Lukas, 1997; Jacobs, 1997; Benassi, 1999). The feeling of lack of control over the access others may have to their personal information during the online navigation process is a psychological risk that prevents many consumers from providing information to Web providers in exchange for access to information offered onsite (Jacobs, 1997; Hoffmam et al., 1999).

Time/convenience risk may refer to the loss of time and inconvenience incurred due to difficulty of navigation and/ or submitting order, finding appropriate Web sites, or delays receiving products. Insurance is a low interest product and customers see it as difficult, when websites are also difficult customer scan feel it as time/convience risk.

But in a service environment perceived risk also plays an other role as to the recent trend to more self-service technology. Not many research have been done thoughout the effects in the insurance industy but there is in the airline industry which can be compared with insurances because both are part of the service sector. What is seen there is that the internet airline reservation places a significant burden and responsibility on the consumer. The consumer is responsible for searching multiple carriers for fares, comparing prices, and proper booking (Law and Leung, 2000). Mistakes are the sole blame of the consumer who

has very limited recourse for correcting errors which influences the perceived risk negatively. The same trend can be seen in the insurance industry where the customers need to fill in their own wanted coverages and total insured amount of assets and the customers can change their own coverages day by day (examples are shown in figure 2).

Figure 2: Possibilities to change coverages day to day.

Moreover, in order to develop a trust environment to facilitate online transactions and ser ices, it is important to understand consumers’ percei ed ris and therefore their purchase intentions in regard to the use of online shopping.

2.1.4 Services

This research focuses on services marketing and in particular on the marketing of insurances. But what are services and what are their specific charateristics. The last 50-years a lot of different researches tried to create service marketing as a sub discipline of product – manufacturing based markting. According to Lovelock & Wirtz (2011) it is hard to define services because they offer a vast array of different and often complex activities. But the main reason for this seperate treatment s that service marketing differs from goods marketing by a few caracteristics, reviewed by different researches.

These four features are intangibility, inseparability, hetrogeneity and persishability. Also referend to as IHIP. These four characteristics are explained below.

Intangibility: According to the Oxford dictionaries the definition of intangible is: unable to be touched, not having physical presence, difficult or impossible to define or understand, vague and abstract, (of an asset or benefit) not constituting or represented by a physical object and of a value not precisely measurable.

Inseperability: Most services ara characterized by simultanious production and consumption. In comparison with goods who first are produces then sold and then consumed (Regan, 1963). The customer must be present during the production of many services, like haircuts and traveling by airplaine. This is also partly valid for insurrances whereas the customer in most cases is “ forced into intimate contact ith the production process” hen reportin an insurance claim for instance (Carmen and Langeard, 1980).

Hetrogeinity involves the potential of high variability in the service performance. According to Zeithaml et all (1985) this is specifically a problem in a labor intensice service environment. Many different employees can be in contact with the service customer which can lead to different customer experiences.

Perishability concerns the fact that most services cannot be saved (Thomas 1978). For instance to much demand on an insurrance customer service when their has been a storm or too little demand outside the holiday seasons on travel insurrances. Capacity not used cannot be reclaimed later.

Figure 3: Characteristics of services

Authors Intangibility Inseparability Hetrogeneity Perishability

Zeithaml et all 1985 X X x

Lovelock 1983 X X x x

Parasuraman & Varadarajan 1988 X X x x

George & Barksdale 1974 X X

Thomas 1978 X

But over the last years more researches argue the diversification between goods and services. It started with an article on service-dominant logic by Vargo and Lusch (2004), that immediately gave rise to an international discussion of what a service perspective on business can offer marketing in general.

In the following year, a survey was published of the observations of a number of leading international scholars in the service field on service and service marketing (Edvardsson et al., 2005). The key finding was that service was indeed more generally considered to be a perspective than merely an activity. Service is a perspective on value creation rather than a cate ory of mar et offerin s” Ed ardsson et a ., 2005).

In one of their original propositions about a service-dominant logic, Vargo and Lusch (2004, 2008) viewed customers as producers, but later changed this view into customers as co-creators of value. Applying a service logic means that the firm is not restricted to making value propositions only, but also gets opportunities through the value co-creation possibilities during interactions with the customers, to actively and directly participate in value fulfillment for its customers.

According to recent research of Smith (2012) it is arguable that this new insight may not be the case in every service context; customer involvement may vary dramatically across services, for insurances for instance customer involvement is very low.

2.1.5 Services and Perceived risk:

The available research adresses the role of risk in a services environment both conceptually (Zeithalm, 1985) and empirically (Murray and Schlacter, 1990) with theory and evidence found that services are tend to be evaluated and perceived riskier than goods. In every purchase there is a degree of risk involved but services tend to be percieved as riskier to purchase than goods (Murray and Schlacter 1990; Zeithalm 1981, Mitchell and Greatorex 1993) . The characteristics of services, hetrogeinity, perishability, inseparability and intangibility lower consumer confidence and increase percieved risk this effect mainly appears by the degree of uncertainty during the decisionmaking (Mitchell 1999). Murray and Schlacter (1990) identified a greater overall perceived risk associated with services than with goods. Murray and Schlater (1990) also find evidence that customers find services more riskier than goods accross serveral types of risk namely; Psychological risk, Physical risk, Social risk.

The reserach of Mitchell and Greatorex (1993) suggest that the theory of perceived risk plays a greater role in explaining the behaviour of customers of services than behaviour of customers of goods.

Although the impact of perceived risk on the consumer buying process for services is less studied than for goods, the effect of perceived risk is believed to have a greater effect on the consumer for services (Guseman, 1981; Murray, 1991; Murray and Schlater, 1990). Services are generally intangible, non-standardized, usually sold without guarantees, and need to be experienced before they can be assessed. But the research of George et all (1985) where the experienced perceived risk for four goods are compared to four services showed no justification for a different risk percention between services and goods. Most recent research by Sun (2014) in the service sector (hotel) shows that there is still evidence for a higher perceived risk for services because of the intangibility. So for this research we wil keep on

using intangibility and based on the previous research we assume that perceived risk is experienced high for services, in particular insurances.

2.1.6 Intangibility and perceived risk

McDougall (1987) describes typical intangibility as a lack of physical evidence. More recent research states that intangibility exists out of three dimensions, namely; Physical intangibility, generality and mental intangibility (Laroche et all 2001). The study of Laroche et all (2003) shows with their results that intangibility is in fact a three dimension construct in the overall model and across goods and services. Laroche et all (2001) explained these three dimensions as follows.

The physical dimension is the extent to which a good cannot be touched or seen, in fact it is inaccessible to the senses and lacks a physical presence.

The generality dimension represents the difficulty customers experience in defining and describing a particular service .

At last mental intangibility dimension stands for that a service can be physically tangible, but difficu t to rasp menta y. This ast dimension appears especia y if the customer doesn’ t have experience with the good.

The effect of intangibility of services is that it makes it more difficult to evaluate a service which can lead to an increase of the perceived risk when purchasing (Mitchell and Greatorex, 1993).

Insurances

Insurance is an experience product. The insurance product can not be touched or hold and the performance of the product performance can only be evaluated after a first claim. Intangibility makes also insurances difficult, sometimes impossible, to evaluate before purchase and, in many instances, after purchase and use. Complaining about an unsatisfactory insurance, which only exists while it is being performed, is much more difficult and problematic. The complainant has limited physical evidence and memory generally serves as the only source of evidence. This increases the risk when purchasing (Mitchell and Greatorex, 1993).

2.2

Purchase intentions or perceived doubt when purchasing

Purchase intentions can be explained as the intention of the customer to purchase the item of service. (Mitchell, 1999). Consumer intentions have long interested marketing researchers. Intention is used in many researches as an explanatory or dependent measure in different marketing models.

Several studies have proved that intention is a good and precise way to predict behavior. Juster (1964) for instance shows in his research that the purchase rate was much higher among intenders than among the nonintenders. On the strenght of these early findings many researchers used and measured the purchase intention. And this still holds in 2014 because it is still a much used measurement in scientific research (i.e. Hartmann 2012, Kim and Zhang 2014). Coherent with purchase intention is perceived doubt. Because purchase intention is coherent to experiencing doubt when purchasing (Mitchell, 1999)

2.2.1 Perceived risk and purchase i ntentions

Lots of research exists in which perceived risk is recognized as a fundamental concept in consumer behavior (Cox 1964 , Roselius, 1971). The theory used for this thesis focus on the original work of Bauer (1960) which was carried on by others, the perceived risk strategy. The perceived risk strategy generally assumes that consumers act to minimize any expected negative utility associated with the possible purchase.

Howard and Sheth (1969) proposed that one of the determinants of purchase intention is confidence, which is the inverse of perceived risk. Bennett and Harrell (1975) suggested that confidence might play an important role in predicting intentions to purchase and experienced doubt. Confidence about the brand is positively related to intention. This suggests that lower perceived risk may be related to higher purchase intention, which is endorsed by research of others (Dowling 1994, Kim et all 2008, Mitchell 1999). In Internet shopping, Vijayasarathy and Jones (2000) found that consumers perceived risk was an important factor that influenced intention to shop online. Shoppers confidence in judging quality of products or in making decisions to purchase products reduce perceived risk, as consumers develop shopping experience from the Internet (Yoh, Damhorst, Sapp, & Lazniak, 2003).

Little research have been done in understanding the impact of this higher perceived risk on the purchase decision and possible information need and/or promotional tools to lower the perceived risk for the customer.

2.3

Risk relievers

Consumers find themselves trying to evaluate virtually indistinguishable service alternatives and providers. Therefore to make a purchase decision, the consumer will learn to rely on dependable methods for dealing with the risk. In an attempt to make the best possible decision, the buyer will rely on risk relievers, i.e. devices or actions used to allay perceived risk.

The following risk relievers are selected for this research. These risk relievers were derived from previous research of Derbaix (1983), Guseman (1981) and Mitchell and Greatorex (1993) but only the relievers particullary fitted and commonly used for insurances were selected.

1. Brand experience 2. Customer reviews

3. Experts opinion or comparison websites 4. Money back garantuee

5. Service Cues 6. Most expensive

7. Experts opinion offline (insurance brooker)

2.3.1 Brand experience

We expect that the usefulness of the risk relievers will vary between different customers. Nevertheless, studies of both products and services have found that brand loyalty/experience (using the same brand/supplier as before) is the most useful risk reliever [Roselius, 1971; Derbaix, 1983; Guseman, 1981]. It may lead to less time risk; because of the experience the customer does not feel the need to look up a lot of information and process the information. Brand experience may also lead to less performance risk because the customer already has a good experience with another insurance at the same company, therefore it may also lead to less financial risk and psychological risk.

2.3.2. Customer reviews

Consumers who perceive high risks tend to engage in more information-seeking activities (Dowling, 1986; Mitchell & Boustani, 1994). Murray (1991) also states that it is in general logical that the greater the degree of perceived risk is in the purchase phase, the greater also the need for the consumer to seek information about the product is. According to an article in

the Dutch Media shown in figure 4, customers already have great trust in customer reviews when it comes to holiday reviews..

After searching customer reviews customers need to process the reviews. Engel, Kollat, and Blackwell (1973) emphasized the importance of perceived risks in the external search and alternative evaluation stages of decision making. If the given information fails to reduce perceived risks, consumers may reject the purchase. However, if the given information helps to moderate perceived risks it will have a positive effect on purchase intention. Because comparing the different customer reviews and insurance websites could take some time, we estimate that customer re ie s on’t ha e effect on percei ed time ris but i ha e effect on psychological risk, product performance risk and financial risk because of the experience of others.

Figuur 4: Customers have great trust in holiday reviews

Source: Telegraaf.nl, 25-06-2014

2.3.2. Comparison websites

Earlier studies depicted perceived risk as a significant factor primarily associated with the information search stage of the consumer buying process (Cox, 1967; Dowling and Staelin, 1994; Murray, 1991; Murray and Schlater, 1990). Recently, Forsythe and Shi (2003) raise the specter of a number of barriers, including perceived risk, that preclude the conversion of

internetbrowsers (information search) into internet shoppers (evaluation and purchase). With more and more comparison websites were the customer select the comparison criteria and ets “ independent” ad ice about the best fit of insurance based on the customers o n criteria and prices of the insurrances. Based on the above we may assume that this will lower the perceived risk for time and convenience because the customer saves time by searching the different insurance websites and because of the advise given and found information it will also has effect on the other elements of perceived risk.

2.3.3. Money back garantuee

By offering a money-back guarantee a company promises that any unsatisfied customer can return the item or in case of insurances can ask for their insurance premium back within a certain period (Davis, Gerstner, & Hagerty 1995).

In response to intense competitive forces in business environments, especially during the recent recession, money back garantuees have been widely implemented by retailers (see figure 5 for an example) and manufacturers as a promotional tool to gain consumers attention and positively influence their purchase decisions (Sullivan 2009). Several authors argue that MBGs serve as extrinsic cues of quality (Moorthy & Srinivasan 1995), reduce consumers perceived risk (McWilliams, & Zilberman 2001) and enhance purchase intentions (Davis et al. 1995; Wood 2001).

But previous research in comparing risk reduction methods found money back garantuee the third most favorable risk reduction methods by customer (Toh en Heeren, 1982), Akaah & Korgaonkar (1988) found money back garantuee the first most favorable risk reduction methods. Because insurance is an experience product we estimate that a money back garantuee after the first claim handling procedure will only have effect on financial risk and will have no effect on the other parts of perceived risk. We further think that money back garantuee will have a negative effect on performance risk because the posibble customer can interpret the money back garantuee as a possible sign of an underwhelming future performance. The purchased product must function as expacted in the future so a money back garantuee should not be needed.

Figure 5: Nationale Nederlanden advert with a money back garantuee.

2.3.4.Service cues

According to Chang (2008) consumers thus look for cues from the online environment to make sure that they will not experience doubt when making a purchase online. Lots of insurance companies make use of service cues on their website, the most commonly used are the label customer oriented insurances, your own contents claim handler who visits you at home to help with your contents claim and a friendly helpdesk. Service cues make services more tangible, which may help people with their expectations about the performance in the future.

2.3.5. Most expensive insurance coverage – expensive model

The expensive model is the model were the customer makes the choice to buy the most expensive option among the different providers and contents insurances. The most expensive model is interesting for insurances because extra coverages automatically adds up the premium price per month but also coverages more contents insurance claims. Therefore by choosing the most expensive insurance, the customer may think he/she will have the most completely contents insurance package and therefore may experience a lower performance risk. The research of Roselius (1971) shows that this method has a positive effect on performance risk but the most expensive method is overall the least useful risk reduction method compared to the other methods. But the most expensive model may be highly interesting from a margin perspective.

2.3.6. Offline expert (insurance brooker)

An offline expert is comparable with the literature of the comparison website. We expect only one difference and that is on the time/convenient element of perceived risk. The comparison

website can be reached whenever and where ever. For the insurance brooker there are trading hours and it may involve traveling.

3. Methodology

This chapter provides an explanation on the research design, research method, survey and questionnaire design and respondents. The proposed methodology aims to limit the research systematically, to make this research controllable, valid and reliable.

3.1 Research method

As mentioned in the new value delivery model services and particulary insurances still tend to be evaluated with a high perceived risk because of the intangibility and so it is striking that there is comparatively little attention directed to understanding the impact of the riskier nature of services on the purchase process and the information needs and possible risk reduction methods for the services customers. In this research we will further research the relationship between perceived risk and purchase intentions and the effect on this relationship for different risk reduction methods. Based on the literature review, 10 hypotheses are conducted.

Hypotheses

In order to answer our research question and to test the hypotheses a case study is needed. For this casestudy the purchase of a contents insurance is choosen. With a contents insurance, contents are insured for loss, fire, water damage, storm and with a complementary insurance also for drop damage. This insurance is not compulsory by the government such as a health insurance. The research design of this research can be found in figure 7 on the next page.

The first hypotheses are about the relationship between perceived risk and purchase intentions. Because insurances are an experience product we expect that performance risk will be the most experiences element of perceived risk and therefore also will have the most negative relationship with douct compared to the other element of risk. In general, based on the literature review we expect a negative relationship between perceived risk and purchase intention/experienced doubt when purchasing.

H1: Performance risk is the most experienced element of perceived risk for insurances purchase compared to the other elements of perceived risk.

H1b: Performance risk has the most negative relationship with doubt compared to the other elements of perceived risk.

H2: For purchasing a contents insurance there is a negative relationship between perceived risk and purchase intention.

The risk reduction methods are measured separately and the relationships to perceived risk and purchase intention can therefore be compared to eachother. We expect that the

usefulness of the risk relievers will vary between different customers so the following

hypotheses for the risk reduction methods and their relation to both perveiced risk as well as purchase intentions will be tested:

Brand experience: Studies of both products and services have found that brand

loyalty/experience (using the same brand/supplier as before) is the most useful risk reliever [Roselius, 1971; Derbaix, 1983; Guseman, 1981].

H3: Brand experience has the highest negative relationship with all elements of perceived risk compared to the other risk reduction methods.

Customer reviews: If the given information of customer reviews helps to moderate perceived risks it will have a positive effect on purchase intention. Because comparing the different customer reviews and insurance websites could take some time, we estimate that customer re ie s on’t ha e effect on percei ed time ris but i ha e effect on psycho o ica ris , product performance risk and financial risk because of the experience of others.

H4a: The exploration of customer reviews has a negative relationship with Product performance, Financial and psychological risk

H4b: The exploration of customer reviews has a positive relationship with Time risk

Exploring comparison websites: By exploring comparison websites before the purchase we expect a negative relationship for the perceived risk elements time and convenience because the customer saves time by not having to search the different insurance websites. Because of the advise given and the found information we also expect a negative relationship with the other elements of perceived risk.

H5: Exploring comparison website has a negative relationship with all elements of perceived risk

Money back garantuee: Because insurance is an experience product we estimate that a money back garantuee after the first claim handling procedure will only have effect on

financial risk and will have no effect on the other parts of perceived risk. We further think that money back garantuee will have a positive relationship with performance risk because the posibble customer can interpret the money back garantuee as a possible sign of an

money back garantuee should not be needed.

H6a: A money back garantuee has a negative relationship with financial risk. H6b: A money back garantuee has a positive relationship with the other elements of perceived risk

Service Cues: Service cues make services more tangible, which may help people with their expectations about the performance in the future.

H7a: Service cues have a negative relationship with performance risk.

H7b: Service cues have a positive relationship with the other elements of perceived risk The most expensive model: The research of Roselius (1971) shows that this method has a positive effect on performance risk but the most expensive method is overall the least useful risk reduction method compared to the other methods.

H8: Most expensive model only has a negative relationship with performance perceived risk Insurance brooker: We expect only one difference compared to the comparison websites and that is on the time/convenient element of perceived risk. The comparison website can be reached whenever and where ever. For the insurance brooker there are trading hours and it may involve traveling.

H9: An Insurance broker has positive relationship time risk

H9b: An insurance broker has a negative relationship with the other elements of perceived risk

In general: we expect that the risk reduction methods with negative relationships with the different elements of perceived risk will have a negative relationship with purchase intention based on our H2 were we state that there is a negative relationship to perceived risk and purchase intention. So by lowering the perceived risk with risk relievers, purchase intentions will be higher.

H10: Risk reduction methods with a negative relationship with the perceived risk elements will have a negative relationship with purchase intention.

The research approach is deductive (Saunders, Lewis & Thornhill 2009). Deductive research is a research approach involving the testing of a theoretical proposition by the employment of a research strategy specifically designed for the purpose of its testing. Research projects are undertaken for different purposes. These can be categorized as exploratory, descriptive and explanatory (Saunders, Lewis & Thornhill 2009). This research project is exploratory, it seeks new insights into phenomena, to ask questions, to test hypotheses and to assess the phenomena in a new light (Saunders, Lewis & Thornhill 2009). With this quantitative data

analyze an explanation of the effects risk reduction methods on perceived risk and purchase intentions can be made.

The main research strategies are; experiment, survey, case study, action research, grounded theory, ethnography and archival research (Saunders, Lewis & Thornhill 2009). In order to test the hypotheses this research makes use of survey design. The specific method to gain this information is to send out an online questionnaire. As populations change and technology evolves, electronic surveys (e.g. email or web) will become much more prevelant (Baruch & Holtom 2008; Porter 2004).

Figure 6: The overall research desgin

3.2 Questionnaire design

To collect the quantitative data for this research an online survey is used. Quantitative data can be decided into two distinct groups: Categorical data and numerical. The data for this research is a mix of those two types. The most data of the demographic questions is categorical but the measurement for perceived risk and purchase intention is on a 5 point likert scale.

The survey questionnaire is made on thesistools.nl. This is a dutch website were you can make your survey and distribute it online. The survey can be found in appendix 1 or through the following link. http://www.thesistools.com/web/?id=423597. Saunders et al. (2007) use a questionnaire or survey as a general term to include all techniques of data collection in which each person is asked to respond to the same set of questions in a predetermined order deVaus 2002). In this type of sur ey there isn’t an inter ie er bein present Saunders, Lewis and Thornhill 2007). The questions in this survey will be standardized.

For knowing which independent variable has effect on the dependent variable quantitative data is needed. The self-administrated questionnaire will be made through earlier research, which will guarantee the validity. Before the questionnaire was online, different people from different ages and background tested the structure and comprehensibility of the questionnaire.

The questionnaire is anonymous, because that’s a so a ay to impro e standard mai sur ey rates accordin to the study of Jobber and O’Rei y 1998). The time to fi up the comp ete survey was in between 15 and 20 minutes for every respondent. According to the research of Linsky (1995) the length of the questionnaire has no influence on response rate.

Besides the demographical questions at the beginning of the questionnaires, the questionnaires will consist questions about perceived risk with the last contents insurance purchase and the effect it had on the purchase decision. The four different items of perceived risk were measured in this study: Financial, Time, Performance and Psychological. The scales used for this research of perceived risk were developed by Stons and Gronhaug (1993) and each scale excist out of 3 statements measured on a five point likert scale. This perceived risk measures were derived from previous risk research literature (for example, Jacoby and Kaplan 1972; Peter and Tarpey 1975; Roselius 1971). For this research it was necesarry to slightly modify item statements to accommodate products of a service nature, namely insurances.

The scales demonstrated good reliability and validity and were built on prior work (Stem, Lamb, and MacLachlan 1977). Three items were employed to measure each risk dimension. he scale was anchored at five positions as follows: strongly disagree, disagree, neutral, agree and strongly agree. Every scale closed with a statement on how the above statements had effect on the doubt felt with the purchase. As mentioned in the literature review we use experienced doubt as a measure for the purchase of the past. This measure is choosen because the purchase has been done, therefore asking what the effect on the purchase intention would not measure accurate what the relationship

with perceived risk was in that specific purchase. The question if the customer first experienced doubt when purchasing because of the different elements of perceived risk is a more accurate measurement. Purchase intention will be used for the second part of the research because the case in that part involves a future purchase. Both doubt and purchase intentions are coherent and measure the same effect on perceived risk only the statements fit better for the specific statements and case above. See for the exact statements used for this research appendix 1, question 9,10,11 and 12.

After the first part influence of different risk reduction methods were measured on perceived risk items and purchase intention. Seven risk reduction methods were selected on the basis of their representativeness, applicability to various methods of purchase, and applicability to the contents insurance. These risk relievers were derived from previous research of Derbaix [1983], Guseman [1981] and Mitchell and Greatorex (1993). No problems were reported with the use of the questionnaire method to study the use of these risk relievers. Subjects were given definitions of some of the terms used in the questions. Again perceived risk was measured on the 4 items and on a 5 point Likert scale. The perceived risk were measured on a five point Likert scale reflecting the respondents opinion as to how they agree with each statement for reducing the risk posed in the situation. Every item of perceived risk was brought back to 1 statement per item and the statements were adjusted to met every risk reliever as follows for example brand loyalty experience:

1. Financia ris : Based on my experience ith the brand I’m ess concerned if the contents insurance will be worth the insurance premium.

2. Performance ris : Based on my experience ith the brand I’m ess concerned that the contents insurance eventually will not function the way I expected when purchasing. 3. Psychological risk: Based on my experience with the brand I feel less uncomfortable and anxiety.

4. Time ris : Based on my experience ith the brand I’m ess orried that purchasin the contents insurance will take much of my time and more time then I expected it would take. After the 4 statements about risk perception for the specific risk reduction method purchase intentions was measured with the statement on a 5 point likert scale if based on the four statements above the respondent would think there would be a huge chance that they would buy the home insurance.

The risk reliever methods were presented per method in the questionnaire as follows, the pictures can be found in appendix 1:

You already have a liability insurance and a health insurance with a certain insurance company where you are very content with.

Method 2: Customer reviews

The next picture shows customer reviews on an insurance websites from other customers just like yourself.

Method 3: Online comparison websites

The next three pictures give an impression of an online comparison website. You can choose wich insurance you want to compare and get information about where you have to think about further. You fill in some questions about yourself and your living conditions and after that you receive a top three insurance based on price quality advice. You can further specify to important coverages for your living situation.

Method 4: Free trial period, money back guarantee,

On the next picture you see an advertisement of Nationale Nederlanden which entitle you to 2 months restitute insurance premium on your car insurance when your not totally satisfied.

Method 5: Most expensive insurance

In this case you choose the most expensive contents insurance with the most extensive coverages and additional coverages.

Method 6: Service cues

To make an insurance more tangible, insurance companies add tangible cues in their communication like the label customer oriented insurances, your own contents claim handler who visits you at home to help with your contents claim and a friendly helpdesk.

Method 7: Experts opinion offline (insurance broker)

You decide not to purchase your insurance online but go to an offline insurance broker who give advise about coverages and different insurances.

3.3 Respondents

The survey link is distributed online by convenience sampling. The link to the survey was shared on the startpage of Facebook were Facebook contacts could see it in the first week of July 2014. The link was also distributed by personal messages on Facebook to different social contacts. The sample group participants are between the 18 and 65 and there are no

background limitations. Facebook is used because of the presence of many contacts.

This research made use of one or more follow ups and/or reminders. The use of follow ups or reminders is highly effective when send to initial non responders, such as telephone or registered email (Rose et al., 2007; Linsky 1995; Kanuk and Berensen 1975; Yu & Cooper 1983).

After collecting the data it will be prepared and set for analyses. The analyses consist of a descriptive, reliability, Mauch y’s Test of Sphericity and a MANOVA analyses. The first test will be done to test the data for reliability of the scales used for this research. The last three analyses will follow to test whether there is a relationship between the variables. All analyses will be done by SPSS. After analyzing the data, the results will be explained. With the analyses the research question will be answered.

4. Results

In this chapter the results of this reseach will be evaluated. This chapter exist out of the following content, perceived risk and experienced doubt, measures methods, effects on experienced risk and purchase intention.

4.1 Respondents overview

This research consist out of 127respondents. Existing out of different age groups, educaction backgrond, gender and linving situation. But also out of different online purchase experience and frequency of online purchases and earlier online insurance purchase expierence.

Respondents

Total

127 respondents

Gender

Men

56 respondents

44,10%

Women

71 respondents

55,90%

Age

18 to 25 years

10 respondents

7,80%

25 to 35 years

58 repondents

45,60%

35 to 45 years

20 respondents

15,75%

45 to 55 years

16 resondents

12,60%

55 to 65 years

17 respondents

13,39%

65 or older

6 respondents

4,72%

Online purchase

Never

5 respondents

3,94%

1 time a year

3 respondents

2,36%

Quartely

35 respondents

27,56%

1 time a month

46 respondents

36,22%

More time per month

38 respondents

29,92%

Education

Highschool

9 respondents

7,09%

MBO

18 respondents

14,17%

HBO

50 respondents

39,37%

WO

50 respondents

39,37%

Living situation

At home

2 respondents

1,57%

Single with no kids

28 respondents

22,05%

Single with kids

3 respondents

2,36%

Married or living together without kids 57 respondents

44,88%

Married or living together with kids

37 respondents

29,13%

Online insurance

purchase

Yes

86 respondents

67,72%

No

41 respondents

32,28%

Channel of purchase

Online

45 respondents

35,43%

Offline

40 respondents

31,50%

Telephone

22 respondents

17,32%

By the mail

3 respondents

2,36%

4.2 Reliability of the perceived risk scales

The four scales of perceived risk show a high reliability per scale. Financial risk has a cronbach’s A pha of 0,902, Time ris 0,922, performance ris 0,903 and psycho o ica ris 0,946.

4.3 Percieved risk and experienced doubt

In this paragraph we describe the relationship between four types of perceived risk and experienced doubt. The table below shows the means and standard deviations of the percieved degree of risk art the most recent insurance purchase.

Perceived risk and experienced doubt are measured on 5-point Likert scales. The mean percei ed ris scores are bet een 1.7 and 2.7, both be o the ‘neutra ’ score 3. The same goes for experienced doubt: the means range from 1.6 – 2.1. This indicates that the respondents perceived low risks and experienced little doubt about their recent insurance purchase.

The correlations between the degree of perceived risk and experienced doubt are strong (all p<.001) and positive: the more risk one perceives, the more doubt one experiences.

Table 1: means and correlations of perceived risk and experienced doubt mean perceived risk SD risk mean doubt SD doubt rrisk,doubt Financial risk 2,3 1,1 2,0 1,2 ,846** Psychological risk 1,7 1,0 2,1 1,2 ,833** Time risk 2,1 1,1 1,6 1,0 ,740** Performance risk 2,7 1,2 1,8 1,0 ,633**

The next table demonstrates the p-values of tests for significance, comparing the correlation coefficients above. Perceived performance risk has the weakest relationship with experienced doubt; significantly weaker than the correlation coefficients between experienced doubt, and Perceived Financial- and Psychological risk.

Perceived Time risk has a weaker relationship with experienced doubt than Perceived Financial- and Psychological risk, but it does not significantly differ from the relationship of experienced doubt with Perceived Performance risk.

The strongest predictors of experienced doubt; Perceived Financial- and Psychological risk, do not differ in predictive power of experienced doubt but the differences are very small. .

Table 2: p-values of tests for significance, comparing correlation coefficients Performance risk Psychological risk Time risk

Financial risk <.001 .73 .02

Performance risk <.001 .10

Psychological risk .05

Hypotheses result:

Hypotheses 1a is supported. Performance risk is the most experienced element of perceived risk for insurances purchase compared to the other elements of perceived risk.

Hypotheses 1b is not supported. Performance risk does not have the most positive relationship with doubt compared to the other elements of perceived risk. Although the differences are small Financial- and Psychological risk seem to have a more positive relationship to experiences doubt then performance risk for insurances.

Hypothese 2 is supported the more risk one perceives, the more doubt one experiences.

4.4 Measures, experienced risk, and purchase intention

Seven measures are compared to test their influence on perceived risks and purchase intention:

 Measure 1: Brand Experience

 Measure 2: Customer reviews

 Measure 3: Online comparison websites

 Measure 4: Money back garantuee

 Measure 5: Most expensive insurrance

 Measure 6: Service cues

 Measure 7: Experts opinion offline (insurance broker)

A repeated measures analysis of variance was used to do these tests, with age, gender, frequency of online shopping and purhcase channel as control variables.

Mauchly's Test of Sphericity shows the sphericity assumption is violated for time risk, and, to a lesser degree, for psychological risk and purchase intention (PI). Therefore we inspect the results of the multivariate tests, which is robust agains violations of the sphericity assumption.

Mauchly’s Test of Sphericityb Within Subjects Effect Measure Mauchly’s W Approx. Chi-Square df Sig. Epsilona Greenhouse-Geisser Huynh-Feldt Lower-bound measure financial ,793 22,944 20 ,292 ,933 1,000 ,167 performance ,752 28,092 20 ,107 ,915 1,000 ,167 psychological ,709 33,959 20 ,026 ,894 1,000 ,167 time ,605 49,672 20 ,000 ,863 ,969 ,167 purchase_intention ,698 35,492 20 ,018 ,898 1,000 ,167

Tests the null hypothesis that the error covariance matrix of the orthonormalized transformed dependent variables is proportional to an identity matrix.

a. May be used to adjust the degrees of freedom for the averaged tests of significance. Corrected tests are displayed in the Tests of Within-Subjects Effects table.

b. Design: Intercept + Watisuwleeftijd + Hoevaakwinkeltuonline + Watisuwgeslacht + kanaal Within Subjects Design: measure

The multivariate tests show no significant Between Subjects effects. This indicates that there are no main effects of age, gender, frequency of online shopping and purhcase channel on the degree of percieved risks and purchase intention.

There is a significant Within Subjects effect of Measure, F(30,72)=1.7, p<.05. This indicates that the different measures result in different degrees of percieved risks and purchase intention. There are no significant interaction effects.

Multivariate Testsc

Effect Value F Hypothesis df Error df Sig.

Between Subjects Intercept Pillai's Trace ,710 47,398a _{5,000 97,000 ,000}

Wilks' Lambda ,290 47,398a 5,000 97,000 ,000

Hotelling's Trace 2,443 47,398a 5,000 97,000 ,000 Roy's Largest Root 2,443 47,398a 5,000 97,000 ,000

Watisuwleeftijd Pillai's Trace ,049 1,007a _{5,000 97,000 ,418}

Wilks' Lambda ,951 1,007a _{5,000 97,000 ,418}

Hotelling's Trace ,052 1,007a 5,000 97,000 ,418 Roy's Largest Root ,052 1,007a 5,000 97,000 ,418

Hoevaakwinkeltuonline Pillai's Trace ,019 ,374a 5,000 97,000 ,865

Wilks' Lambda ,981 ,374a _{5,000 97,000 ,865}

Hotelling's Trace ,019 ,374a 5,000 97,000 ,865 Roy's Largest Root ,019 ,374a 5,000 97,000 ,865

Watisuwgeslacht Pillai's Trace ,054 1,106a 5,000 97,000 ,362

Wilks' Lambda ,946 1,106a 5,000 97,000 ,362

Hotelling's Trace ,057 1,106a 5,000 97,000 ,362 Roy's Largest Root ,057 1,106a _{5,000 97,000 ,362}

Kanaal Pillai's Trace ,116 ,798 15,000 297,000 ,680

Wilks' Lambda ,887 ,797 15,000 268,176 ,681

Hotelling's Trace ,125 ,796 15,000 287,000 ,682 Roy's Largest Root ,093 1,840b _{5,000 99,000 ,112}

Within Subjects Measure Pillai's Trace ,415 1,703a 30,000 72,000 ,034

Wilks' Lambda ,585 1,703a 30,000 72,000 ,034

Multivariate Testsc

Effect Value F Hypothesis df Error df Sig.

Roy's Largest Root ,710 1,703a 30,000 72,000 ,034

measure * Watisuwleeftijd Pillai's Trace ,381 1,479a 30,000 72,000 ,090

Wilks' Lambda ,619 1,479a _{30,000 72,000 ,090}

Hotelling's Trace ,616 1,479a 30,000 72,000 ,090 Roy's Largest Root ,616 1,479a 30,000 72,000 ,090 measure * Hoevaakwinkeltuonline Pillai's Trace ,326 1,162a _{30,000 72,000 ,297}

Wilks' Lambda ,674 1,162a _{30,000 72,000 ,297}

Hotelling's Trace ,484 1,162a 30,000 72,000 ,297 Roy's Largest Root ,484 1,162a 30,000 72,000 ,297

measure * Watisuwgeslacht Pillai's Trace ,335 1,210a _{30,000 72,000 ,253}

Wilks' Lambda ,665 1,210a 30,000 72,000 ,253

Hotelling's Trace ,504 1,210a 30,000 72,000 ,253 Roy's Largest Root ,504 1,210a _{30,000 72,000 ,253}

measure * kanaal Pillai's Trace 1,023 1,277 90,000 222,000 ,077

Wilks' Lambda ,279 1,278 90,000 216,358 ,076

Hotelling's Trace 1,630 1,280 90,000 212,000 ,076 Roy's Largest Root ,812 2,002b 30,000 74,000 ,008 a. Exact statistic

b. The statistic is an upper bound on F that yields a lower bound on the significance level. c. Design: Intercept + Watisuwleeftijd + Hoevaakwinkeltuonline + Watisuwgeslacht + kanaal Within Subjects Design: measure

The Univariate Tests table demonstrates a signifcant effect of Measure on all four types of percieved risk, and on purchase intention. Pairwise comparisons of all means of percieved risk and purchase intention per measure are shown in the Annex 2.

Univariate Tests

Source Measure Type III Sum of Squares df Mean Square F Sig.

measure financial Sphericity Assumed 14,499 6 2,417 2,743 ,012

Greenhouse-Geisser 14,499 5,598 2,590 2,743 ,014

Huynh-Feldt 14,499 6,000 2,417 2,743 ,012

Lower-bound 14,499 1,000 14,499 2,743 ,101

performance Sphericity Assumed 17,860 6 2,977 3,424 ,002

Greenhouse-Geisser 17,860 5,489 3,254 3,424 ,003

Huynh-Feldt 17,860 6,000 2,977 3,424 ,002

Lower-bound 17,860 1,000 17,860 3,424 ,067

psychological Sphericity Assumed 13,220 6 2,203 2,497 ,021

Greenhouse-Geisser 13,220 5,367 2,463 2,497 ,027

Huynh-Feldt 13,220 6,000 2,203 2,497 ,021

Lower-bound 13,220 1,000 13,220 2,497 ,117

time Sphericity Assumed 12,275 6 2,046 2,212 ,040

Greenhouse-Geisser 12,275 5,179 2,370 2,212 ,050

Huynh-Feldt 12,275 5,817 2,110 2,212 ,042

Lower-bound 12,275 1,000 12,275 2,212 ,140

purchase_intention Sphericity Assumed 15,005 6 2,501 3,746 ,001

Greenhouse-Geisser 15,005 5,385 2,786 3,746 ,002

Huynh-Feldt 15,005 6,000 2,501 3,746 ,001

Lower-bound 15,005 1,000 15,005 3,746 ,056

Error(measure) financial Sphericity Assumed 533,793 606 ,881

Greenhouse-Geisser 533,793 565,439 ,944

Huynh-Feldt 533,793 606,000 ,881

Lower-bound 533,793 101,000 5,285

performance Sphericity Assumed 526,817 606 ,869