The reward, the experience, or the
rewarding experience?
Identifying customer segments using drivers of (non-)participation in
short-term loyalty programs
A.I.V. Huitema
The reward, the experience, or the
rewarding experience?
Identifying customer segments using drivers of (non-)participation in
short-term loyalty programs
A.I.V. Huitema H.W. Mesdagstraat 18 9718 HH Groningen (06) 12 58 94 41 A.I.V.Huitema@student.rug.nl Student Number: S 1686003
Master Thesis for MSc Marketing University of Groningen Faculty of Economics and Business
Department of Marketing
Supervisors:
prof. dr. T.H.A. Bijmolt J.H. Wiebenga, MSc
II
Preface
This master thesis is the final examination for a Master of Science degree in Marketing, and by handing in this document I complete my study curriculum and discontinue my life as a student in Groningen. Looking back on six years of studying at the University of Groningen, I have completed a bachelor program, a transition program, and two master programs – all without forgetting to enjoy the student life in all its facets. These six years have provided me with a lot of experience, both on scholarly and personal level, which I take along for any future endeavor. Although I feel ready and look forward to start a professional career, I am sure that I will occasionally look back on these years in Groningen with sentiment and nostalgia.
Writing and finishing this thesis has been a process on which I look back with relieve and joy. Keeping an overview over a database with over 800 variables, more than 600 cases, and over half a million data points has been both a challenge and an opportunity to demonstrate my analytical capabilities. I have deliberately attempted to raise the level of this thesis, by applying several types of analyses that were new to me, in addition to some with which I was already familiarized. Although I have regretted this decision numerous times, I have also greatly enjoyed figuring out new techniques and finding significant results by applying them. Being able to independently work on these analyses was very exciting (even though I sometimes needed a push in the right direction), however I must admit that I am also relieved that it is now converted into a more comprehensible format.
Writing this thesis would not have been possible without the help and support of many friends, family and of course my supervisors. First of all, I greatly appreciate the facilities and resources provided by [company], as well as their trust, feedback, and support in conducting this research and writing my thesis. I especially would like to thank [supervisor] and [supervisor] for their time and effort put into this research. In addition, I would like to thank my supervisors from the University of
Groningen, Tammo Bijmolt and Jacob Wiebenga, for their constructive, helpful, and inspiring
feedback. I would also like to thank my friends and family, for their understanding and patience during the course of writing this thesis. In the finalizing stages, I found myself living as a hermit from time to time, prioritizing my research over social appointments. The support I have received from them has been an encouragement to continue working on this thesis, and to finalize it within my intended timeframe.
III
Management summary
The current study is built around identifying segments of customers in terms of attitudinal similarity and participation status in short term loyalty programs (LPs).The purpose of this study is to identify what changes in the design characteristics of LPs potentially have the greatest effect on the participation status of the customer. The problem statement used in this study is: How can
participation rates in short-term loyalty programs be increased using insights on applicable (non-) participation drivers?
An extensive literature review was performed from which a taxonomy of (non-)participation drivers is composed. From this taxonomy, five main themes of (non-)participation drivers were identified. These constructs concern the reward and redemption desirability, experience contentment, collection scheme appeal, communication and design quality, and the program context. These constructs support the theory that participation depends on the expected benefits and costs of participation. However, the obtained factors propose a slimmed down version of the taxonomy, in which more subcategories were presented.
Together with the participation state of the respondent, the clusters that were found were used as active segmentation variables in a latent class cluster analysis. The results indicate that there are four distinctly different segments: neutral opportunists, unattainable pessimists, rational enthusiasts, and regretful switchers. The unattainable pessimists have the lowest participation rate (17%), followed by the neutral opportunists (55%). The rational enthusiasts and regretful switchers are characterized by high participation rates (82% and 83% respectively). Focusing on the low-participating segments, the neutral opportunists and unattainable pessimists are found to be indifferent with regard to their demographics. The differences originate from predisposition attitudes.
1
Table of content
1. Introduction ... 3
1.1 Aim of this paper ... 3
1.2 Theoretical and managerial relevance ... 5
1.3 Thesis outline... 6
2. Literature review ... 7
2.1 Loyalty ... 7
2.2 Loyalty programs ... 7
2.3 Participation in LPs: benefits, costs, and boundary conditions ... 8
2.3.1 Benefits ... 9
2.3.2 Costs ... 11
2.3.3 Boundary conditions... 15
3. Qualitative inquiry: review of focus groups ... 18
3.1 Design and participants ... 18
3.2 Procedure ... 19
3.3 Findings... 19
3.3.1 Benefits ... 19
3.3.2 Boundary conditions... 20
4. Conceptual model ... 23
4.1 Influence of benefits, costs, and boundary conditions on participation ... 23
4.1.1 Benefits ... 23
4.1.2 Costs ... 24
4.1.3 Boundary conditions... 25
4.2 Taxonomy of non-participation drivers ... 25
2
6.1.4 Factor interpretation ... 35
6.2 Latent class cluster analysis ... 37
6.2.1 Clustering procedure ... 37
6.2.2 Profiling clusters ... 38
6.3 Binary logistic regression ... 43
6.3.1 Comparing factors in binary logistic regression ... 44
6.3.2 Comparing clusters and factors in binary logistic regression ... 46
6.3.3 Targeting clusters ... 48
7. Conclusion and discussion ... 51
7.1 Conclusion ... 51
7.1.1 Theoretical questions ... 51
7.1.2 Empirical questions ... 52
7.2 Discussion ... 54
7.3 Managerial implications ... 56
7.4 Limitations and future research directions ... 57
7.5 Epilogue ... 59
Reference ... 60
Appendix A: Overview of questions ... 65
Appendix B: Comparison of questions in different versions ... 70
Appendix C: Rotated component matrix ... 72
3
1. Introduction
Loyalty programs (LPs) are designed to reward loyal behavior and encourage shoppers to patronize retailers with more regularity (Bijmolt et al., 2010). Due to technological advancements in information technology, and an increasing number of firms adopting a customer-centric approach to retailing, the number of LPs offered by businesses has grown vigorously over the last decades (Gable et al., 2008; Leenheer and Bijmolt, 2008; Smith et al., 2004). Nourished by the recent recession, consumer interest in LPs has increased parallel to this development (Burness et al., 2009; Ferguson and Hlavinka, 2009).
The popularity of LPs has also given rise to academic literature on the topic. This research has extensively studied loyalty and LPs in terms of the behavioral (Leenheer et al., 2007; Meyer-Waarden, 2008) and attitudinal responses (Daams et al., 2008; Hallberg, 2004; Henderson et al., 2011), and the psychological mechanisms (Kivetz et al., 2006; Taylor and Neslin, 2005) and design characteristics (Kumar and Reinartz, 2006) affecting the functioning of an LP. Although LPs have been studied comprehensively, the vast majority of this research has focused on long-term LPs.
Short-term LPs have been under examined in the scientific literature, mainly because the focus of many LPs is on developing what is referred to as true loyalty. True loyalty is conceptualized as attitude-based behavior instead of inertial repeated purchase (Odin et al., 2001). The concept was first introduced by Day (1969), who stated that true loyalty requires the consumer to hold a favorable attitude towards the brand in addition to repeatedly purchasing it (in: Bove and Johnson, 2009). As favorable attitudes are driven by continued emphasis on customer satisfaction, it is typically a long-term construct which is mainly developed by continuous LPs (Kim et al., 2008). Short-term LPs usually involve frequency reward programs, offering monetary and firm-related rewards for repeated purchase during a fixed period of time. These programs can be either sole-proprietary or based on partnerships between retailers (Bijmolt et al., 2010). Short-term LPs also vary in terms of timing: some offer immediate rewards (hereafter referred to as promotion programs), while some have delayed rewards using a collection scheme. Although short-term programs are typically considered to be focused on increasing short-term results, repeatedly offering short-term LPs with time breaks can have the same effect on loyalty as long-term programs.
1.1 Aim of this paper
one-4 fourth of enrolled members redeem rewards (Capizzi and Ferguson 2005). The aim of the current paper is to assess possible drivers of (non-)participation and develop an LP evaluation method that explains why potential participants of an LP choose not to participate. Subsequently, this data can be used to identify segments of customers, that hold different attitudes and respond distinctly different to changes in the LP design. The main question of this research is:
How can participation rates in short-term loyalty programs be improved using insights on applicable (non-)participation drivers?
In order to assess this question, first a thorough understanding of the functioning of short-term LPs is required. The basic mechanisms underlying LPs, such as the points-pressure and the rewarded behavior mechanisms are discussed in the first part of the literature review. It is argued here that participation in an LP is based on a trade-off between expected benefits and expected costs of participating. The main question discussed here is:
1. How do short-term loyalty programs function from a consumer perspective?
In the second part of the literature review, the focus is shifted to LP hindrances. From the body of research focusing on long-term LPs, a taxonomy of (non-)participation drivers is proposed and discussed. These drivers are further expanded based on findings derived from focus group discussions. A conceptual model containing the drivers of (non-)participation is presented. The question discussed in this part is:
2. What drivers of non-participation can be identified in relevant scientific literature and in focus group discussions?
Once a taxonomy of (non-)participation drivers is created, a questionnaire is developed in order to measure the applicable (non-)participation drivers in short-term LPs. Applying the questionnaire to measure (non-)participation drivers will provide results for individual loyalty programs. The evaluation method is designed to expose attitudes regarding possibly decisive issues of the LP. In the current study, the evaluation tool will be tested on one such program. Factor analysis is applied on the data obtained to uncover the various drivers of (non-)participations that show high correlations, and thus form underlying dimensions or constructs. Based on this factor analysis, the following question is discussed:
3. What are the main dimensions on which customers evaluate an LP?
5 using passive variables such as demographics, predisposition measurements, and cognitive approaches to the LP. This provides in indication of the relative differences between groups of customers. The question that is intended to be answered in this part of the study is:
4. What segments of customers can be found based on the constructs and participation status?
Finally, binary logistic regression is used to compare segments with regard to the relative importance of the constructs for participation. Compared to a general model, splitting the model across segments will provide insights into the relative differences in segments in terms of influence of the constructs on participation likeliness. The results obtained, in combination with the participation statistics, can be used to identify which segments can be most likely persuaded to participate in subsequent programs, and which design characteristics should be changed to achieve this. This analysis also indicates which elements of the LP should be changed in order to achieve these results. The question discussed here is:
5. Which elements of the LP should be changed in order to increase the participation rate?
1.2 Theoretical and managerial relevance
A multitude of past research has focused on unveiling the motivational drivers of customers to participate in LPs, focusing on the expected benefits and their impact on participation (see for instance: García Gómez et al., 2012; Mimouni-Chaabane and Volle, 2009; Steyn et al., 2010; De Wulf et al., 2003). However, little research has focused on understanding why customers wish not to participate: the drivers of non-participation. Noble and Philips (2004) and Ashley et al. (2011) are among the few authors that discuss the possible motivational drivers of non-participation. This research is focused on long-term LPs and discusses just a limited number of drivers. The approach of the current study is to theoretically and empirically study (non-)participation in short-term LPs by focusing on possible drivers that prevent the customer from engaging in a relationship with the retailer. This standpoint provides the possibility to identify dissimilar issues as opposed to focusing solely on participation drivers. Reversely, positive attitudes with respect to non-participation drivers might also foster participation, thereby becoming a participation driver.
6 Furthermore, by applying binary logistic regression for individual segments, a detailed view on the contribution of various constructs to participation can be found. This increases both the manageability and accountability for managers working with short-term LPs. Only if detailed information on (non-) participation drivers exists, future LP designs can be enhanced, and participation rates can be increased. In turn, this affects the loyalty fostered by the LP and improves the results of the retailer.
1.3 Thesis outline
7
2. Literature review
The scientific literature on loyalty and LPs is extensive (see: Bijmolt et al., 2010 for a recent overview). Drivers of loyalty and reasons for participating in LPs have been thoroughly analyzed and discussed. However, the focus in many research has been on participation drivers, focusing on customer reasons to participate in loyalty programs (García Gómez et al., 2012). Relatively little is known about participation hindrance: why would a customer not participate in an LP? In this chapter a general definition of loyalty and a description of the loyalty programs analyzed in the current study is presented. It is argued that participation is based on a trade-off between the expected benefits and the expected costs of participating. It is proposed that this trade-off forms the core of the decision to participate and that non-participation drivers are caused by a problem with the benefits, the costs, or the boundary conditions. In the final paragraphs of this chapter, the categories of (non-) participation drivers are discussed using current scientific insights.
2.1 Loyalty
Oliver (1999) defines loyalty as “a deeply held commitment to rebuy or repatronize a preferred product/service consistently in the future, thereby causing repetitive same brand or same brand-set purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior”. A distinction is made between two types of loyalty: behavioral loyalty and attitudinal loyalty (Dick and Basu, 1994; Uncles et al., 2003). Behavioral loyalty relates to purchase patterns (Bijmolt et al., 2010), and is the relative weight or frequency of customer purchases (Romaniuk and Nenycz-Thiel, 2013). Attitudinal loyalty refers to loyalty driven by favorable attitudes and high levels of commitment (Bijmolt et al., 2010). Both types of loyalty reinforce each other: attitudinal loyalty resides only in the mind of the consumer and will not contribute to sales without behavioral loyalty, which is the actual act of purchasing. Reversely, behavioral loyalty without attitudinal loyalty is simply the act of repeated purchase, and might be caused by a lack of alternatives rather than a strong commitment to a product or service (Dick and Basu, 1994; Whyte, 2004). Because strictly behavioral loyalty could be the result of peripheral route processing and low cognitive involvement (Petty et al., 1983), it is more vulnerable to competitive offerings.
2.2 Loyalty programs
8 treatment is based on segmentation according to individual customer profitability (Blattberg et al., 2008). It also excludes promotion programs, in which customers immediately receive gifts after reaching a threshold amount in basket size. The rewards in this definition refer to physical rewards with monetary discounts to the manufacturer suggested retail price (MSRP). The short-term duration of the loyalty programs infers reward expiry (Bijmolt et al., 2010), which means that rewards can only be redeemed during the course of the LP.
Reward expiry triggers an important psychological mechanism for short-term LPs: the points-pressure mechanism (Kopalle and Neslin, 2003; Drèze and Nunes, 2009). Together with the rewarded behavior mechanism they form the core psychological mechanisms for the functioning of LPs. Both are directly related to reward redemption (Bijmolt et al., 2010). The points-pressure mechanism encourages customers to increase their purchase volume when close to reward redemption or reward expiry, due to the fear of disqualification (Drèze and Nunes, 2011; Smith and Sparks, 2009). The rewarded behavior mechanism encourages a customer to maintain or increase its spending level after reward redemption, either through increased attitudinal commitment (Philips Melancon et al., 2010), emotional attachment (Haisley and Lowenstein, 2011; Palmatier et al., 2009; Smith and Sparks, 2009), or the behavioral lesson learned (Rothschild and Gaidis, 1981). The key challenge in designing LPs is to stimulate point-pressure mechanisms during the course of the LP and to maintain the rewarded behavior for as long as possible after the LP has ended.
2.3 Participation in LPs: benefits, costs, and boundary conditions
Customers decide to participate in an LP based on an evaluation of the expected benefits and the expected costs of participation (Ashley et al., 2011; De Wulf et al., 2003). The costs of participation refer to what is given by the customer in order to participate, whereas the benefits refer to what is received by the customer (Ashley et al., 2011). The thought of weighing benefits against costs is consistent with exchange or equity theory (Adams, 1965; Housten and Gassenheimer, 1987; Huppertz et al., 1978; Oliver and Swan, 1989). Equity theory postulates that customers in exchange situations such as LPs compare their input (ie. the spending level) against the output (ie. the reward) of the exchange (Huppertz et al., 1978). When inequity exists, there is discrepancy between the perceived benefits and costs. When evaluating the benefits and costs of an LP, customers attempt to make an assessment of the fairness of the LP. Participation is unlikely when the discrepancy is in favor of the retailer.
9 remainder of this chapter, benefits, cost, and boundary conditions applicable to LPs are discussed using previous scientific findings. Within each category, several distinct themes of non-participation drivers are identified. Each theme itself consists of several issues which might drive non-participation. Table 1 presents the outline of this model. Findings of previous research are used as building blocks for the taxonomy of non-participation drivers, which will be extended in the subsequent chapters using insights from focus group discussions.
Category Theme Issues
Benefit Theme 1 Issue 1 Issue 2 Theme 2 Issue 1 Issue 2
Cost Theme 1 Issue 1
Issue 2 Theme 2 Issue 1 Issue 2
Boundary conditions Theme 1 Issue 1 Issue 2 Theme 2 Issue 1 Issue 2
Table 1: Model overview
2.3.1 Benefits
Benefits refer to what is received by the customer by successfully participating in an LP (Ashley et al., 2011). Within the category of benefits, two types of themes can be identified: reward themes and experiential themes.
2.3.1.1 Reward themes
10 One of the most important issues within the reward themes is referred to as unenticing benefits (Noble and Philips, 2004) and concerns the overall appeal of the reward. Unappealing rewards are less aspired by customers, and are a sound reason not to participate in the LP. These issues are referred to as Aspiration issues. Under aspiration issues, a multitude of reasons might be applicable why the reward is unenticing. Consider a reward category such as cutlery: the brand, functionality, design, details, finish, weight, color, and quality could all be sound arguments why the reward is unenticing. These reasons will all be covered by aspiration issues.
Another reward related issue affecting the participation likelihood is referred to as worthlessness of the reward (Stauss et al., 2005) or cash value of the reward (O’Brien and Jones, 1995). If redeeming the reward does not lead to spending less or cost savings, but rather the opposite in the eyes of the customer, participation is unlikely (Mimouni-Chaabane and Volle, 2009). It is important to note here that this is closely related to the problems with the reward scheme: a reward can only be considered worthless if the redemption scheme is too demanding or the additional payment is too high. Issues in which the additional payment is too high, or the received discount on the reward is too low, will be referred to as monetary savings issues.
Hedonic rewards are more preferred as opposed to utilitarian rewards, as hedonic rewards are often seen as a reward for oneself (Smith and Sparks, 2009). However, there is a tension between saving loyalty points for a hedonic self-gift and using the points on impulse to redeem a utilitarian gift (Smith and Sparks, 2009). Utilitarian products are more likely to be obtained as the probability of attainment increases (Smith and Sparks, 2009). A reward can be appealing, but not seen as a real treat to the customer. If the reward is not considered special enough to be redeemed, for instance because of its utilitarian nature, this issue is referred to as a motivation issue.
Stauss et al., (2005) and O’Brien and Jones (1995) also refer to inaccessibility of the reward. Two types of inaccessibility are important: either the right reward is not included in the LP, or the right reward is simply out of stock. The first is referred to as accessibility issues, the latter is referred to as
delivery issues. Accessibility is thus a matter of being able to choose a suitable reward from the
available options, whereas delivery is a matter of having that reward available at the moment of redemption. Accessibility is related to choice, whereas delivery is related to availability.
2.3.1.2 Experiential themes
11 issues: the feeling of exclusivity, the entertainment value of the LP, and the discovery of new products and services by LP participation.
Many LPs enhance the perception of social benefits (Libermann, 1999). These social benefits can include the feeling to be part of an exclusive group of participators, with whom the participant can identify itself and with whom it shares values associated with the retailer (Muniz and O'Guinn, 2001). If the LP does not create a desired feeling of exclusivity, for instance because there is no recognizable group of privileged customers (Mimouni-Chaabane and Volle, 2009), this is referred to as exclusivity
issues.
Mimouni-Chaabane and Volle (2009) also identify the hedonic benefits of having an entertaining LP. Participating in an LP can provide joy to the customer (Hirshman and Holbrook, 1982; Tauber, 1972). LPs can therefor attract customers due to the pleasure associated with collecting and redeeming points (Johnson, 1999). A customer might experience that he or she is enrolled in a game, and derives a feeling of entertainment from participating (Mimouni-Chaabane and Volle, 2009). On the contrary, if the LP lacks these kinds of entertaining features, this might be a reason not to participate. If participating in the LP is not considered fun, this is referred to as entertainment issues.
Finally, some customers like to participate in LPs because it leads to the discovery of new items or services, which they would otherwise not buy or use (Mimouni-Chaabane and Volle, 2009). If the LP invites customers to try new or innovative products, this is an example of exploratory behavior (Arnold and Reynolds, 2003; Baumgartner and Steenkamp, 1996). Customers generally appreciate to use LPs to get acquainted with new products, as it satisfies curiosity. Consequently, if the LP is lacking the experience of something new, this is referred to as exploration issues.
2.3.2 Costs
The costs of participation refer to what is given by the customer in order to participate (Ashley et al., 2011). There can be various themes identified in the literature that refer to the costs of participation. Upkeep themes, reward scheme themes, time themes, and personal loss themes are proposed here as being the four major categories of costs associated with LP participation.
2.3.2.1 Upkeep themes
12 Noble and Philips (2004) refer to the task of updating and maintaining a relationship with the retailer as maintenance issues. In the light of long-term LPs, these issues involve updating account information or preferences. It is not related to time (as with time themes), but rather with effort associated with participation. Examples of maintenance issues can include presenting the collection leaflet at the cashier to receive stamps, or filling out personal information on the collection leaflet prior to reward redemption (Ashley et al., 2011).
Related to the maintenance issues are the perceived costs of actually carrying collection leaflets or loyalty cards around in order to acquire the benefits of patronizing the retailer (Noble and Philips, 2004). Especially in the case of larger leaflets, customers can experience inconvenience when having to bear the leaflet and can consider the space required to keep the leaflet as a cost. Issues pertaining the material unease of the customer with regard to the LP are termed physical issues. Physical issues can also include putting stamps or points in to the leaflet, or uploading points to a loyalty card. Noble and Philips (2004) state that often consumers will disengage from the relationship when these physical tasks are seen as annoying or require too much effort.
Some LPs require effort from the customer in terms of bearing the loyalty card and bringing the card upon patronizing the retailer (Noble and Philips, 2004). When stamps or points are directly put into the leaflet or uploaded on a loyalty card, it can be frustrating when the card is forgotten or lost. The effort involved in remembering the LP and bringing the leaflet to the store is pertained as mental
issues. Noble and Philips (2004) state that “the irritation of having to remember such cards in the
past in order to continue engaging in a relationship with a retailer left many consumers reluctant to initiate a relationship with another retailer using a similar tactic”.
Other authors, such as Ashley et al. (2011) and O’Brien and Jones (1995), have coined similar issues with LPs under the term convenience issues. If the program is not easily accessible, this can be perceived as an inconvenience (Ashley et al., 2011). Furthermore, if the LP requires too much additional effort, this is also a form of inconvenience (O’Brien and Jones, 1995). All other issues related to effort in initiating or maintaining enrolled in an LP which are not related to physical, mental, or maintenance issues are termed convenience issues. For instance, customers that are required to ask for stamps at the cashier might perceive this as an inconvenience.
2.3.2.2 Reward scheme themes
13 the input of the exchange from a customer’s perspective. Based on the literature, four reward scheme issues are discussed.
The first problems that might be applicable with the reward scheme involve hollow benefits
issues (Noble and Philips, 2004), which refer to the reward not being compensating for the effort put
into collecting. This can be the case if the number of stamps or points required (costs) outweighs the expected benefits of the reward. The customer experiences inequity of exchange and consequently decides not to participate (Huppertz et al., 1978).
A second issue with the reward scheme refers to the qualification barrier (Stauss et al., 2005). A qualification barrier is conceded when a customer does not reach the threshold amount in basket size to receive stamps. Ashley et al. (2011) associate this problem with purchase frequency: the customer does not patronize the retailer frequently enough to become eligible for a reward. This issue is also referred to as relevance of the LP (O’Brien and Jones, 1995). A customer can consider the LP irrelevant if she does not apply for stamps or rewards. All issues of the customer related to not being able to fulfil the reward scheme and redeem rewards will together be termed qualification
issues.
A third issue with the redemption scheme refers to the flexibility of the reward scheme. The more easy it is to use the reward scheme, the more likely participation is (Smith and Sparks, 2009).
Flexibility issues thus refer to problems with the choice alternatives within the use of the reward
scheme. Consider for instance an LP in which it is possible to redeem different rewards depending on the number of points collected, or a reward scheme in which the number of stamps collected has influence on the discount granted. As redemption might not occur as things get in the way (Soman, 1998), inflexibility of an LP might cause participation loss.
14 2.3.2.3 Time themes
Time themes refer to the time required to engage, participate, and redeem rewards in the LP (Noble and Philips, 2004). Here, four issues are discussed which can be grouped under time themes.
In order to participate in an LP, it is inevitable to initiate the relationship via a sign-up procedure (Noble and Philips, 2004). In most short-term LPs, this procedure involves picking up a leaflet or loyalty card and start collecting. However, if this procedure takes a lot of time, for instance because a custom card needs to be created or there is many paperwork that needs to be filled out prior to collecting, the procedure can be perceived as a considerable cost in terms of time. Noble and Philips (2004) refer to problems with the initial outlay of time spent on signing up as initiation issues, a construct which is adopted in this study as well.
Accumulation issues refer to the time required to reach a redemption threshold. This varies
between consumers, depending on their purchase patterns and the collection scheme used in the LP. It is also related to expectations on the points-pressure mechanism (Kopalle and Neslin, 2003; Drèze and Nunes, 2009), as consumers that do not expect to reach the reward threshold within the time-frame of the LP have no reason to participate. Additionally, for LPs with longer duration, accumulation issues might also be considered when customers lose interest due to boredom with the LP (Noble and Philips, 2004).
The decision to participate might also be a matter of location of the retailer. Noble and Philips (2004) term problems with the location of the retailer travel issues. It is a rather practical issue, as some customers simply live too far away from the retailer to purchase frequently. However, travel issues are only a problem for participation once it affects the patronization of the retailer. Here, the term patronization issues is adopted to reflect all non-participation drivers related to the time required to patronize the retailer. This can thus encompass a slightly broader definition than travel issues.
15 To prevent overlap between other issues and themes as much as possible, it is important to note here that time themes explicitly relate to costs associated with time loss as a driver of non-participation. This for instance distinguishes accumulation issues from qualification issues: the first refers to not being able to complete the collection scheme in time because the points-pressure is too high, while the latter refers to not being able to collect points and complete the collection scheme at all, even when time would be irrelevant.
2.3.2.4 Personal loss themes
A final cost associated with LPs involves personal loss themes. Personal loss themes (Noble and Philips, 2004) or privacy concerns (Ashley et al., 2011) are all issues referring to giving up something private or personal. Here, a distinction is made between ambush issues and privacy issues.
Ambush issues pertain to the fear of using the provided personal information for subsequent
inquiry or fraudulent actions by the retailer or a third party (Noble and Philips, 2004; Ashley et al., 2011). The fear of being ambushed by participating in an LP is especially vivid for LPs demanding personal information, such as e-mail addresses and telephone numbers, and can be a significant cost associated with an LP (Fournier et al., 1998; Milne & Boza, 1998; Phelps et al., 2000).
Privacy issues involves giving up something private. This can be the loss of anonymity, or the
feeling that ever more data is gathered around the individual customer. The fear of losing something private is also considered a significant cost of LP participation (Fournier et al., 1998; Milne & Boza, 1998; Phelps et al., 2000). If a customer feels the loss of privacy (cost) does not outweigh the benefits of the LP, participation becomes unlikely.
2.3.3 Boundary conditions
Besides the trade-off between benefits and costs grounded in exchange theory (Adams, 1965; Housten and Gassenheimer, 1987; Huppertz et al., 1978; Oliver and Swan, 1989), a number of boundary conditions should be met in order for the customer to consider the LP. Here, three themes are proposed that revolve around boundary conditions for LP participation: relationship themes, communication themes, and environment themes.
2.3.3.1 Relationship themes
Relationship themes include a number of issues that are related to the past, current, and future relationship with the retailer.
16 (Noble and Philips, 2004). The affiliation with the retailer might cause social resistance, and is incompatible with the interests of the customer. If a customer has trouble having any kind of relationship with the retailer, this is referred to as affiliation issues.
LPs usually have very punctual collection and redemption thresholds: reaching a threshold amount in basket size makes a customer eligible for a stamp or point, and reaching a threshold in collected points make the customer eligible for a reward. This tight focus on the economic aspects of the relationship is termed economization (Stauss et al., 2005). When a customer is offended by economization, for instance when he or she does not receive stamps or points when just falling short of reaching the threshold, this can have detrimental effects on the perceived relationship with the retailer. If this is a reason to seize future participation in LPs, this is referred to as economization
issues.
In the same light, Stauss et al. (2005) discuss defocusing issues. Defocusing issues refer to the retailer losing sight of what is its core business, for instance by putting too much emphasis on the LP (Stauss et al., 2005). Defocusing can be especially harmful for customers that are generally not interested in LPs, as these customers are likely to notice defocusing first. Defocusing can make an LP backfire and cause relationship termination instead of retention.
Discrimination issues refer to a favorable treatment of certain customers, and consequently a less
favorable treatment of other customers (Stauss et al., 2005). For instance, customers that do not participate due to insolvency to redeem a reward might feel discriminated as opposed to customers that are eligible for a reward (Fournier et al., 1998). Especially if participation leads to subsequent benefits, such as shorter waiting lines or other preferential treatment, discrimination in LPs can cause harm to the relationship between customer and retailer. Here, a tension becomes apparent between discrimination issues and exclusivity issues: for customers that do participate in the LP an exclusive preferential treatment might be driver of participation, whereas this favorable treatment consequently might harm non-participators due to the discrimination issues.
2.3.3.2 Environment themes
Environment themes pertain non-participation drivers that can be derived from the environment of the current LP. Here it is proposed that there are two important environment themes: overload issues and competitive issues.
17 trivial and useless, instead of valuable and unique” (Fournier et al., 1998). When the number of LPs in which a customer is enrolled increases steadily, eventually the customer will lose overview and becomes insusceptible to new LPs. This is referred to as overload issues.
18
3. Qualitative inquiry: review of focus groups
In the previous chapter a preliminary taxonomy of non-participation drivers for LPs was presented, using an approach that considers the benefits and costs of participation, and several boundary conditions that should be met. This taxonomy was derived from an extensive literature review, drawn from research on both long-term and short-term LPs. In the current chapter, this taxonomy is validated and extended with findings from focus groups.
3.1 Design and participants
The current study uses two sets of focus group discussions. The first set of focus group discussions were organized in Germany and revolved around the communicative features of collection leaflets. Two focus groups were organized in May, 2012, and four focus groups were organized in September, 2012. The focus groups distinguished between two target groups: young women (age 25-39) and elderly women (age 40-60). Each group consisted of six to eight participants. The duration of the focus group discussions was approximately two hours per group. The participants were profiled as average to higher class in purchasing power, with varying backgrounds in terms of family and profession. Both employed women and housewives were solicited. All participants were the main household shopper and were responsible for the acquisition of household goods, appliances, textiles, etc. With regard to LP experience, approximately 70% of the participants were familiar with LPs. Approximately 30% were generally interested in LPs and could be considered potential participants.
The second set of focus group discussions were organized in Lille, France, and mainly revolved around the rewards used in LPs for a large French retailer. Two sessions of each two hours were held. The first session had five participants, the second session seven participants. The target group of these focus groups were women, aged 30-54, with knowledge of an earlier LP organized by a large French retailer. The participants patronized this retailer regularly, and 50% of the participants had participated in an LP organized by this retailer before. All participants were the main grocery shopper in the household.
19 strength or importance of these issues can vary strongly depending on the country. Therefore, caution is advised for generalizing the findings of this study directly to LPs organized in other countries.
3.2 Procedure
The focus group protocol consisted of postulating several topics regarding LPs, which the participants could openly discuss both with the interviewer and with other participants. All focus groups commenced with an introductory round, in which the interviewer first introduced himself and all participants had the opportunity to state personal information such as their name, age, and occupation.
In the German focus group discussions, some recent LPs were discussed in general terms. Among others, questions regarding knowledge of the LP, participation in the LP, and general feedback on the LP were asked. Participants were free to discuss any aspects of LPs, such as in-store communication or their thoughts about the retailer. Because the participants were free to discuss any aspects that came to mind, discussions included all kinds of general remarks with regard to LPs.
In the French focus groups, laddering techniques were used to uncover the preferences, feelings, and core values regarding the rewards. Laddering techniques involve a specific way of asking questions, such that each reported answer is subsequently elaborated on to assess superordinate or core constructs (Neimeyer et al., 2001). After the introduction, participants were asked to freely discuss the rewards in different programs with regard to usage opportunities, quality perception, etc. Both real and fictive rewards were discussed, to unfold what drives reward aspiration. Although the focus in these focus group was on the rewards, participants were able to freely discuss their thoughts with regard to LPs in general.
3.3 Findings
Findings from the focus group discussions both confirmed and extended the taxonomy presented in the previous chapter. For the sake of parsimony, confirmatory findings will not be discussed here. The focus is on those findings that extend the previously discussed model. Additions were found for the category of benefits and boundary conditions.
3.3.1 Benefits
20 3.3.1.1 Reward themes
Within the reward themes, two issues were added based on findings in the focus groups. The first refers to the necessity to collect entire sets of rewards in an LP. A respondent stated that it is unappealing to redeem just part of a set, such as cutlery or cookware, if the set cannot be completed. If multiple items need to be redeemed separately to complete a set, and it is preferred to have a complete set rather than parts, there might be compatibility issues.
The second issue added for reward themes refers to the link between the reward and the retailer, or to the link between the rewards. Respondents in the focus group generally agreed that, to a certain extent, rewards should have a connection with the retailer, for instance because the reward complements the product categories offered by the retailer. If this link with the retailer is missing, the relevance of the reward within the setting of the retailer decreases. Additionally, participants stated that they seek for a relationship between rewards, searching for a common theme which connects the rewards together. When such an overarching theme is missing, this causes participation hindrance. Both these issues are referred to as connection issues.
3.3.1.2 Experiential themes
Within experiential themes, two new issues emerged from the focus groups. Uncovered in the literature, participants in the focus groups frequently mentioned the importance of the thrill of participating in an LP. Especially when redeeming a reward, excitement was generated by the feeling of winning something. In some cases, however, respondents mentioned that LPs were unenticing in terms of leveraging these feelings of excitement upon reward redemption. Respondents stated that reward redemption should lead to a satisfactory feeling, and when this incentive was missing this could be a driver of non-participation. These issues are referred to as gratification issues, as they refer to problems with the extent to which participation leads to a satisfactory experience.
Many respondents stated that they did not participate in previous LPs, because there was something missing that triggered them to truly feel like they were participants in the program. Although it was hard to specify what this element was, it can be considered as something that had to keep them involved with the program, for instance by continuously reminding them about their progress in collecting and the distance from reward redemption. When customers do not experience that they are true participators in the LP, the chance of completion declines. Issues related to the feeling of participating in an LP are referred to as involvement issues.
3.3.2 Boundary conditions
21 the communication of the LP could be identified as a new theme. In the upcoming section, the findings are discussed in-depth.
3.3.2.1 Relationship themes
Relationship themes encompass a set of issues associated with affiliation of the customer to the retailer. Three new issues were added to this theme based on the focus group discussions: history issues, host issues, and security issues.
Multiple participants in the focus groups mentioned previous bad experience with the retailer with regard to an LP. One participant had collected enough points to redeem a reward, but encountered a more attractive competitive offer for the same product elsewhere. Another participant wanted to redeem her points, but the preferred reward was out of stock. These kinds of negative experiences with LPs could lead to future LP resistance. If previous negative experience drive non-participation, this is referred to as history issues.
One of the participants in the German focus groups mentioned that she could not identify the host of the LP. Hosts of LPs can vary; sometimes LPs are hosted by retailers, whereas some larger LPs are hosted by third-parties. Knowledge of the organizing party of the LP was critical for her to consider participation, as knowledge of the host made it easier to infer the quality of the LP. If the owner of the LP is unclear, this could be a driver of non-participation. These issues are referred to as
host issues.
Another important aspect related to the relationship between customer and retailer is that of the expected after-sale services. For the core business of the retailer, the customer usually has an expectation of, or experience with, the amount of service he or she receives from the retailer. However, as many LP rewards are beyond the core business of the retailer, the after-sale service on these rewards can be doubted by the customer. This concern was shared by participants in the focus groups, who stated that they could not assess what would happen when a reward broke down within the warranty period. If questioning the after-sales services of the retailer with regard to the LP reward becomes a driver of non-participation, this is referred to as security issues.
3.3.2.2 Communication themes
22 Some of the participants in the focus groups stated that they noticed the LP too late to consider themselves eligible for a reward. Additionally, some participants stated that they were customers of the retailer, but did not notice the LP at all. In either case, there was a problem with the communication of the LP, such that consumers were not sufficiently informed about the running LP. If the communication fails to create notion with the customer that there is an LP, there are
knowledge issues.
Even when the customer is informed about the LP, the communication may be unenticing or unconvincing. In discussing the leaflet designs during the German focus groups, participants agreed that some leaflets were not persuasive and did not invite the customer to participate. Two rationales were proposed: either the communication was not triggering the customer to pursue reward redemption, or the lay-out was not appealing. The first is pertained as persuasive issues, the latter is referred to as design issues.
Another boundary condition that became apparent during the focus group discussions, especially in the French discussions, was related to interaction with the rewards. In-store displays showing the redeemable rewards entice the customer to start collecting. Additionally, sales-personnel, especially cashiers, should reinforce the customer and actively promote the LP. If either aspect is lacking, the customer will not be activated by the LP and he or she will not participate. If the LP is too passive and interaction is missing, this is referred to as activation issues.
Most of the LP leaflets promote the rewards based on the granted discount on the manufacturer's suggested retail price (MSRP). One of the respondents in the French focus groups stated that she highly doubted the percentage of discount that was advertised. When the communication is perceived to be misleading, for instance because rewards are portrayed better than reality or fictive MSRPs are being mentioned, the integrity of the retailer and the LP can be questioned. This is referred to as integrity issues.
23
4. Conceptual model
In the previous two chapters, relevant scientific literature and focus group discussions with regard to participation in LPs have been discussed. The findings are combined in this chapter to form a preliminary conceptual model of LP participation. This conceptual model is discussed in terms of the expected relationships between themes and participation. Furthermore, a more specific taxonomy of (non-)participation drivers for LPs is presented at the end of this chapter.
4.1 Influence of benefits, costs, and boundary conditions on participation
The conceptual model presented here contains the benefits, costs, and boundary conditions approach derived from equity theory (Adams, 1965; Housten and Gassenheimer, 1987; Huppertz et al., 1978; Oliver and Swan, 1989) and consists of the various themes identified. It is proposed that within each category, a theme can have a positive or negative influence on participation. Figure 1 shows the conceptual model containing the expected relationships between the various themes and participation. Argumentation for these relationships is presented in the upcoming paragraphs.
Figure 1: Conceptual model
4.1.1 Benefits
In general, benefits of LPs are expected to have a positive influence on participation and satisfaction (Omar et al., 2010). Based on the literature review and focus group discussions, two benefit themes could be identified: reward themes and experiential themes. Here, it is proposed that both have a positive influence on participation.
(Non-)
Participation
in Loyalty
Program
Benefits
•Reward themes (+) •Experiential themes (+)Costs
•Upkeep themes (-) •Reward scheme themes (-) •Time themes (-)•Personal loss themes (-)
24 Leenheer et al. (2007) suggest that the perceived economic benefits (rewards) drive positive attitudes towards an LP. Hence, the better the reward is in terms of monetary value, appeal to the customer, discount provided, etc., the more likely participation becomes. It is therefore proposed that there is a positive relationship between rewards and participation in LPs.
The same argumentation can be used for explaining the positive relationship between the expected experience from participating and LP participation. Leenheer et al. (2007) state that non-economic benefits, such as psychological and sociological benefits, drive a positive attitude towards LPs. It is proposed here that better experiences with LPs also encourage participation. Hence, there is a positive relationship between experiential themes and participation.
4.1.2 Costs
The costs of participation refer to what is given by the customer in order to participate (Ashley et al., 2011). In general, the higher the costs are, the more the customer should give up in order to successfully participate in the LP and the less likely participation becomes (Huppertz et al., 1978), if the reward and other characteristics of the program would stay the same. A negative relationship is therefore proposed for all themes within the cost category.
Upkeep themes refer to the effort required by the customer to remain enrolled in the LP (Noble and Philips, 2004). The more effort is required, the less convenient the LP becomes for the customer (Ashley et al., 2011; O’Brien and Jones, 1995). As inconvenience of the LP is a driver of non-participation, the relationship between upkeep themes and participation is proposed to be negative. Reward scheme themes refer to the way the LP is designed in terms of collection and redemption thresholds. If these thresholds increase, it becomes more difficult for customers to qualify for rewards, use the collection scheme in their favour, and yield the benefits of participation. Again, the proposed relationship between the reward scheme and participation is negative.
Time themes refer to the time it requires to engage, participate, and redeem rewards in the LP (Noble and Philips, 2004). Time can be considered a cost, as it is something that is given up by the customer (Monga and Saini, 2009). The more time is required to initiate, accumulate, or patronize the retailer, the less beneficial the LP is for the customer. Therefore, a negative relationship between time themes and participation is proposed.
25 retailer. Once more, a negative relationship is predicted between personal loss themes and participation.
4.1.3 Boundary conditions
Boundary conditions involve a number of conditions that should be met in order for the customer to consider the LP. Boundary conditions can have both a positive and a negative influence on participation, but differ from benefits and costs because they are not directly related to the design of the program. Rather, the boundary conditions are related to circumstances in the surrounding of the LP, such as the general relationship with the retailer, the way the LP is communicated, and the environment in which the LP is organized.
Relationship themes pertain to the relationship between the customer and the retailer. Generally speaking, the better the relationship with the retailer, the more likely participation in the LP becomes (Ha and Stoel, 2008). A positive relationship between relationship themes and LP participation is expected. However, issues with the relationship might be detrimental for LP participation.
Communication themes refer to the amount and quality of the communication effort made in promoting the LP. Consistent with advertising theory, as the number of communication efforts increase, the customers’ valuations of the LP increase as well (Sahuguet, 2011). A positive relationship between communication themes and LP participation is predicted. However, as with relationship themes, issues regarding the communication might cause participation hindrance.
Finally, environment themes refer to other LPs in the surrounding of the LP under investigation. Especially in saturated markets, competition between LPs or simply an overload of different LPs might cause resistance (Liu and Yang, 2009). It is proposed here that as the number of environmental aspects increases, the likelihood of LP participation decreases. Hence, a negative relationship is expected.
4.2 Taxonomy of non-participation drivers
26 non-participation, for instance when the values of these variables are perceived as negative. Reversely, negative issues can be perceived as positive as well. Here, the integral taxonomy of (non-) participation drivers used as discussed in the previous chapters is presented. Next, it is discussed how this taxonomy is transformed to a questionnaire to measure (non-)participation drivers and how this questionnaire was used in evaluating a recent LP organized in Mexico.
Benefits Reward themes Accessibility issues A1 Aspiration issues A2 Monetary savings issues A3 Compatibility issues A4 Motivation issues A5 Delivery issues A6 Connection issues A7 Experiential themes Exclusivity issues B8 Entertainment issues B9 Exploration issues B10 Gratification issues B11 Involvement issues B12 Costs Upkeep themes Maintenance issues C13 Physical issues C14 Mental issues C15 Convenience issues C16 Reward scheme themes Hollow benefits issues D17 Qualification issues D18 Flexibility issues D19 Redemption issues D20 Time themes Initiation issues E21 Accumulation issues E22 Patronization issues E23 Choice issues E24 Personal loss themes Ambush issues F25 Privacy issues F26 Boundary Conditions Relationship themes Affiliation issues G27 Defocusing issues G28 History issues G29 Economization issues G30 Discrimination issues G31 Host issues G32 Security issues G33 Communication themes Knowledge issues H34 Persuasive issues H35 Design issues H36 Activation issues H37 Integrity issues H38 Consistency issues H39 Environment themes Overload issues I40
Competitive issues I41
27
5. Research methodology
In this chapter, the evaluation tool for short-term LPs is presented and discussed. First, the design of the study is discussed. Next, the various materials used in the questionnaire are presented. This chapter ends with a description of the participants and procedure used in collecting data.
5.1 Design
The design of the current study is a between-subjects exploratory correlational study, in which the respondents are clustered according to their participation status and scores on five factors which are derived from the 41 items in the taxonomy.
5.2 Materials
A current post-program evaluation tool was adjusted and extended to better fit the issues identified in the literature and presented in the taxonomy of non-participation in paragraph 4.7. This was done by identifying which questions in the standard questionnaire showed similarities to the constructs in the taxonomy. If necessary, these questions were adjusted to better fit the issues under investigation. For those constructs that were not present in the questionnaire yet, and that could not be easily adjusted, new questions were added. These questions were integrated in parts with similar themes, or in new question batteries with a common theme. In designing both new and adjusted questions, guidelines for questionnaire development were followed.
28 regard to the interpretation of the measurement and are formed to a single measure per question, regardless of the response format employed.
5.3 Participants
Participants were selected from the market study database of GMI. GMI provides access to an online panel of research respondents, and ensures that the participating respondents are suitable for this research. GMI compensates participants, depending on the length and content of a questionnaire. Fully completed questionnaires reimburse a respondent’s balance at GMI. This balance can be used to redeem vouchers once a threshold amount is reached. This ensures proper motivation for the respondent to finalize the questionnaire.
Questionnaire version Active participators (N) Non-participators (N) 1 – 5-point SD / 5-point AD N = 71 Mean Age = 39.31
(σ = 11.506)
N = 75 Mean Age = 44.25 (σ = 12.700) Female = 56.3% Female = 60%
2 – 5-point AD / 5-point SD N = 71 Mean Age = 42.13 (σ = 12.209)
N = 91 Mean Age = 42.01 (σ = 11.420) Female = 56.3% Female = 57.1%
3 – 7-point SD / 7-point AD N = 98 Mean Age = 39.77 (σ = 12.794)
N = 63 Mean Age = 41.24 (σ = 12.115) % Female = 48% Female = 60.3%
4 – 6-point AD / 7-point SD N = 61 Mean Age = 39.20 (σ = 12.329)
N = 87 Mean Age = 40.85 (σ = 13.499) Female = 57.4% Female = 60.9%
Table 3: Distribution of participants
29 approximately half of the respondents per questionnaire actively participated in the LP. The distribution of the respondents across the questionnaires is summarized in table 3. Analysis of variance on the distribution of respondents indicates no significant differences with respect to the age (F = 1.418; p = 0.195) and gender (F = 0.621; p = 0.738) of the respondents between versions, indicating that the randomization was successful.
5.4 Procedure
Respondents were invited by an e-mail with a hyperlink to the questionnaire. Contingency questions were used to verify that the respondents met certain criteria, which includes being the main grocery shopper in the household, shopped at the specific retailer (Commercial Mexicana) in the last three months, owned the loyalty card (Monedero Naranja) of this retailer, and was aware of the organized LP. If the respondent suffices these criteria, access is granted to the main questionnaire. Approximately 2% of the respondents that were invited to participate in this survey met these criteria. Respondents that did not meet these criteria were denied access to the main part of the questionnaire and thanked for their willingness to participate. Qualified respondents were randomly assigned to each of the four manipulations of the main questionnaire and were able to start the survey immediately.
31
6. Results
This chapter contains the results of the various analyses conducted on the data obtained from the questionnaire. First, the amount of variables in the taxonomy are reduced to a number of underlying dimensions using factor analysis. Next, an attempt is made to segment the respondents based on their factor scores, using latent class cluster analysis. The clusters obtained are profiled based on demographic variables, predisposition measurements, and cognitive approach to the program. Finally, binary logistic regression is performed using the factor scores, both on a general level and with the clusters as covariates. This latter analysis is used to identify the most targetable segments, and indicate which attributes of the program could be changed in order to increase the participation rate.
6.1 Factor analysis
Factor analysis is used to reduce the number of measured variables to a more workable amount by combining variables with shared variance. First, a description is given on how the data is prepared for factor analysis. This is required, since the raw data obtained from the questionnaire contained multiple response scales and differences between positively and negatively keyed questions. Next, the criteria and procedures to obtain factors are described, which leads to an initial factor solution. Subsequently, the internal consistency of the factor is checked using Cronbach’s Alpha and Pearson correlation statistics. This subsection ends with a description of the factors, based on the variables that load well on these factors.
6.1.1 Data preparation
32 for instance because participants got a preferential treatment or belonged to a group of exclusive customers (Muniz and O'Guinn, 2001). In this case, the data shows that exclusivity is actually negatively correlated to participation, which leads to the assumption that exclusivity might cause resilience towards the program. In order to cope with these negative correlations, these variables have been reversed coded such that higher values indicate more positive evaluations. In the discussion, a clarification for this counter-intuitive finding is presented and discussed. After data preparation, all variables from the taxonomy are now positively keyed which means that higher scores indicate more positive attitudes. The data is now ready for factor analysis.
6.1.2 Obtaining factors
Factor analysis uses shared variance between variables to compose factors. The method applied here is principal component factor analysis, which also takes into account the unique variance of a variable in addition to the shared variance. The Kaiser-Meyer-Olkin measure of sampling adequacy (KMO = 0.918) and Bartlett’s test of sphericity (9371.296; p = 0.000) verify that factor analysis is appropriate for this data.
Factor Eigenvalue ( ) % of variance Cumulative 1 10.424 25.425 25.425 2 3.097 7.554 32.980 3 2.005 4.891 37.870 4 1.732 4.225 42.095 5 1.528 3.728 45.823 6 1.186 2.894 48.717 7 1.124 2.742 51.459 8 1.043 2.543 54.002 9 1.007 2.457 56.459 10 0.969 2.362 58.821 11 0.963 2.348 61.169 12 0.879 2.143 63.312
Table 4: Variance explained per factor
33 important for deciding on the final number of factors. Table 4 contains an overview of the Eigenvalues and variance explained by different factor solutions, in which the factor solutions that meet the criteria have been shaded.
34 Figure 2: Scree plot of factor solutions
To interpret the factors, the rotated factor loadings are used. Varimax rotation with Kaiser normalization is used to maximize the sum of the variances of the loadings, whereas minimizing the loadings of variables on other factors. A deliberate decision was made for varimax rotation, to avoid too much correlation or multicollinearity between separate factors. Appendix C contains the rotated component matrix. To ease interpretation, all coefficients below 0.40 are suppressed. The initial solution has factors containing four to ten variables each, which is a relatively balanced distribution of variables across factors. It appears that some variables have high loadings on multiple factors, such as variable B9 (entertainment issues). The decision to which factor these variables should be assigned is based on reliability analysis and interpretation, which is discussed in subsequent paragraphs.
6.1.3 Internal consistency
Two methods to assess the internal consistency of the factors are applied. For factors composed out of three or more variables, Cronbach’s Alpha (α) is estimated. Cronbach’s Alpha is a coefficient of internal consistency which is similar to a correlation coefficient and estimates an alpha level on the interval (-1, 1). The closer the value is to 1, the better the consistency within the input variables. Negative values can be obtained in case of negative correlation, or indicate reverse coded variables. Generally, any level of α obtained that meets the criterion (α > 0.7) is considered suitable for combination in factors. By assessing the effect of removing an item from the factor on the obtained α, it is possible to improve the internal consistency of the factors. For variables which initially loaded on two factors, both the interpretation and the α level determine the factor to which the variable is assigned. Variables that reduce the internal consistency and that hinder straightforward interpretation are removed from the analysis. For factors that contain less than three variables,
