An agency based view on
pay satisfaction and
commitment
2
An agency based view on pay satisfaction
and commitment
Maikel Hidding
Master thesis marketing management
Rijksuniversiteit Groningen
Faculty of Economics and Business
3
Management summary
This research looks at a new way for small and medium sized enterprises to advertise. A company places digital screens to show advertisements in stores of other entrepreneurs. Because one company hires another to work for him and perform parts of its marketing program, and will pay him for his work, an agency dilemma occurs. Ensuring that the agent is motivated and committed is important to make sure that the principal gets the results it desires. This research focuses on compensation and the contract that is used between the principal and the agent.
The analysis shows that agents have a preference for fixed payments. The more fixed payments an agent receives, the higher its pay satisfaction is. If a principal does decide to use variable pay it should take into account that agents will not like the increased risk and pay satisfaction will go down when the degree of variable payments goes up.
One moderator of the relationship between type of payment and pay satisfaction is also proposed. Store size is suggested to have a moderating effect because a larger store with more income might be willing to take more risks. Bigger stores are thus hypothesised to show higher satisfaction for variable pay but this is not supported by the data.
4
Preface
With this thesis I hope to finalize my student life, which started years ago at the Hanze. Not really sure about the exact direction I wanted to take I choose a broad study in business administration. This turned out to be the perfect choice for me and led me to take the pre master year here at the RUG and ultimately to me writing this thesis. I cannot say it has been easy writing this thesis but it has been an interesting and useful experience. It took a few changes along the way but I am happy with the end product. I know I could not have done this without the support of a few people, they know who they are.
December 2013, Hoogeveen
5
Table of contents
Table of contents ... 5 1. Introduction ... 6 1.2 Commitment ... 9 1.3 Pay satisfaction ... 9 2. Theoretical framework ...11 2.1 Agency theory ...11 2.1.1 Moral hazard ...12 2.1.2 Adverse selection ...13 2.1.3 Other Problems ...14 2.2 Commitment ...15 2.3 Pay satisfaction ...162.4 Pay satisfaction & commitment ...17
2.5 Solutions for the agency problem ...18
2.6 Compensation ...19 2.6.1 Contract ...20 2.7 Store size ...22 3. Problem statement ...24 3.1 Conceptual model ...24 4. Methodology ...25 4.1 Research design ...25 4.2 Scenarios ...26 4.3 Questionnaire design ...27 4.3.1 Pay satisfaction ...27 4.3.2 Organizational commitment ...29 4.3.3 Store size ...30 4.4 Analysis method ...30 4.5 Hypothesis testing ...31 4.5.1 Hypothesis one ...31
4.5.2 Hypothesis two and three ...32
4.6 Sample size ...33 5. Results ...35 5.1 Commitment ...35 5.2 Pay satisfaction ...36 5.3 Hypotheses testing ...36 5.3.1 Hypothesis one ...36
5.3.2 Hypothesis two and three ...36
6. Conclusion and discussion ...38
6.1 Recommendations ...40
7. Limitations and further research ...41
Literature ...42
Appendix 1: questionnaire ...54
Appendix 2: Cronbach’s alpha commitment ...56
Appendix 3: VIF analysis pay satisfaction ...57
Appendix 4: Regression for commitment ...58
6
1. Introduction
Companies advertise their products or services to persuade customers to buy the products or use the services of the company. This way the company looks to increase its earnings by attracting new customers or by selling more products to their existing customer base.
Research has found that advertising can lead to a direct increase in sales (Bruce et al 2012). Another aspect is that advertising influences consumers’ cognition, affect and experience and these three drive sales as well (Vakratsas & Ambler 1999). And finally, advertising has also been found to increase word of mouth among consumers (Keller & Fay 2009).
So advertising can be very important, however advertising can be done in many different forms using many different media. A company can also decide between different levels to advertise on, ranging from a local level to national or even international level. To reach the largest audience, a company has to advertise on a (inter)national level.
However, advertising can be expensive especially when advertising on a (inter)national level. Running an advertisement on national television, or advertising in a well read newspaper is too expensive for small and medium enterprises, which means they have to find other ways of advertising, to be able to get national recognition.
Customers still tend to find the smaller businesses through more traditional media. This was confirmed by an article in Businessweek (2013) that says that only about three percent of advertising budget is spend online by small businesses. They believe small companies feel more for local advertising because the target group is scattered over the internet and not clustered locally anymore (Lacy 2009). Research also shows that small-business executives are not fully aware of the options they have online or are just not sure what to do with the options they have (BCG 2013).
Small and medium sized enterprises that do not have a large advertising budget will struggle to become well known. To be able to get national attention these companies can try and advertise using out of home advertising. Out of home advertising can be done at, for
7
al 2005). However, other small and medium sized companies will aim towards more national recognition to increase sales. As Wynarczyk et al (1993) say, small businesses face more uncertainty and are more vulnerable because they have a small customer base. Advertising on a national level may help deal with this vulnerability. Expansion of the customer base is one of the most important adjustments for survival of small businesses (Smallbone et al 1993).
One of the ways of advertising is in-store advertising, in which a company advertises inside a store using, for instance, video messages. This can be in the store where the company sells its products (point of purchase advertising), or be general advertising in a store or other place not directly selling the company’s products. Main advantages of using in-store advertising are that a company can reach its target audience relatively cheap and the company can select a store that fits their target audience best (Yim et al 2010). Also, with in-store advertising the retailer can select the best placement for the advertisement to prevent clutter. Point-of-purchase advertising can also lead to much wanted attention and it improves brand recall and purchase intent (Yim et al 2010).
1.1 Advertising using intermediary
Not every company, however, has stores in which it can advertise. This can be small companies that simply are only present in a small part of the country or companies selling online who do not have stores that can be visited. If they do wish to use in-store advertising they need to find another company that is willing to show advertisements in exchange for money or other rewards.
In the above situation a contract is signed stating that the first company will pay the second to place the screens, keep them turned on at the right times and update them regularly when the first company asks for it. The second company can include his own advertisements and determine what the best spots for placing the screens are. This creates an agency
relationship which is a contract under which the principal engages the agent to perform a service on behalf of the principal, this means that some decision making authority has to be delegated to the agent (Jensen & Meekling 1976). So one company acts as an agent for the other company which can be seen as the principal.
8
marketing programs. These agencies are companies that are contracted on commission basis. These contractual agreements involve an agency relationship. Facilitating agencies are “the groups of institutions that assist channel members in performing distribution tasks” (Rosenbloom 2012). These institutions provide services to the marketing channel members after the basic channel decisions have been made. So the important decisions that control the marketing of products remain with the principal. Bergen et al (1992) also state that a distribution channel consists of a set of agency relationships. In a distribution channel a manufacturer contracts resellers for local advertising and point of purchase promotion.
An agency relationship does bring problems, as described in the agency theory. An agent might act in its own interest and not that of the principal, this is called moral hazard. And the principal does not know if the agent is competent enough, this is called adverse selection. If one company relies on another company to perform parts of its marketing program, an agency relationship exists. Next to that it also pays a commission as specified in a contract. The agent has to place the advertisements, keep them up to date and in sight of customers. So, the agent takes over part of the marketing service and an outsourcing relationship is created.
Logan (2000) investigated how to design a successful outsourcing relationship. Until then, very little research had been done that links any type of outsourcing to agency theory. One exception is a research by Sharma (1997) that uses agency theory to frame business exchanges between a principal and a professional service organization. This is a different situation than that in the current research because in the current research the agent is not a professional service organization.
9
Tate et al (2010) used agency theory when examining the buying of marketing services. They also found that agency based literature on service outsourcing is limited. Tate et al (2010) mention two more researches that have used agency theory. However, they did not look at what contract would lead to the most preferable outcomes.A principal wants to deal with the problems existing in an agency relationship and make sure the agent is motivated to work towards the goal. One way of making sure the agent puts effort into his work is by creating commitment.
1.2 Commitment
Commitment exists when an exchange partner believes a relationship is important enough to warrant maximum effort towards maintaining that relationship (Morgan & Hunt 1994). A literature review by Fayezi et al (2012) investigated 86 articles of which just one partially analysed commitment in an agency theory setting. No research so far has looked at the choice of contract and its influence on commitment.
Fayezi et al (2012) point to the gap in research relating to agency theory and commitment. As they put it: “This becomes increasingly important when applied to supply chain
operations, where factors relating to knowledge, commitment and trust often outweigh contractual relationships”.
A factor that has regularly been found to positively influence commitment is pay satisfaction, which are the overall positive and negative feelings one has toward payments (Williams et al 2006). Pay satisfaction has, however, not been examined in agency theory (Fayezi et al 2012). The choice of contract, and the type of payments involved in these contracts, could have an influence on pay satisfaction.
1.3 Pay satisfaction
Only a few studies have examined the link between variable pay and pay satisfaction
10
The current research fills the gap in research by examining both pay satisfaction and commitment in an agency theory setting. It shows how a company can increase the pay satisfaction and commitment of its agents and what role fixed and variable payments play. It shows what contract can best be used to increase commitment and maintain a productive relationship. This can have a contribution to every principal-agent relationship, showing how to reward the agent so he is both retained and motivated to work hard. This leads to the following problem statement:
Does a higher degree of fixed pay lead to higher pay satisfaction in an agency theory setting, and does store size influence this relationship? And does this higher pay satisfaction lead to higher commitment?
In the current research one specific form of out of home advertising is used. Digital screens are placed in the stores of an agent. Larger companies with multiples stores or big stores will be paid to place multiple screens to make sure all customers see the advertisements. These screens will show various advertisements from several companies. Some of the advertised products can be sold by the agent but some advertisements show products the agent does not sale. The agent has to turn on the screens every morning and make sure they are in place for customers to see. The advertisements are renewed regularly so the advertising companies can show their latest products. The agent has to update the screens to replace the old advertisements with new ones.
This research starts with a theoretical framework in which existing research on the subjects of this research will be discussed. First the agency theory and the problems encountered in agency relationship will be described. Then the ultimate goal of commitment is discussed, followed by pay satisfaction and the link between the two. Next, the solutions for agency problems are discussed, focussing on compensation and the type of contract. Finally, a moderating variable, store size, is discussed. The theoretical framework is followed by the conceptual model. Next is the research design which explains how the hypotheses will be tested. This is followed by the results in chapter five and the conclusion and
11
2. Theoretical framework
Using another company to show out of home advertising can prove to be effective for many small and medium sized enterprises. These companies may not have the budget to advertise nationally but can use out of home advertising to reach a large audience. Grigorios & Alistair (2004) explained that an important marketing problem for small businesses is their lack of financial resources, both in absolute and in relative terms. These small businesses have relatively high fixed costs which prevent high investments in marketing. A 30 second
commercial on television on prime time can cost over 5000 Euro (Ster, 2013) and advertising in major newspapers can cost even more (TMG, 2012). The low budget of small businesses and high prices of national advertising force a small business to find other ways of
advertising nationally.
As explained in the introduction, when a company uses an intermediary for out of home advertising an agency relationship occurs. Therefore the agency theory is discussed in the first section. The problems that may arise in the relationship between and agent and a principal are discussed next.
Two aspects of agency relationships, namely adverse selection and moral hazard, are discussed separately. After which two other problems are discussed. A principal wants to deal with these problems and ultimately create commitment, so commitment is discussed next. Then pay satisfaction is discussed and the link between commitment and pay satisfaction. Then, compensation and the contract between principal and agent are
discussed. The form of compensation as described in the contract should ultimately lead to pay satisfaction and then commitment. Finally, a moderating variable, store size, is
discussed.
2.1 Agency theory
12
argued that corporations are structured to minimize costs involved in getting agents to act in the principal’s best interest.
Jensen and Meekling (1976) formally defined agency theory as a contract under which the principal engages the agent to perform a service on behalf of the principal, this means that some decision making authority has to be delegated to the agent. If both the principal and the agent are utility maximizers then the agent may not always act in the best interests of the principal. One important assumption of agency theory is that agents are said to be “selfish opportunists” who will exploit principals unless they are monitored and/or incentivized to not do so (Miller & Sardais 2011). Figure 1 shows the relationship between principal and agent.
Since the first formal use in literature the theory has been used in many areas like accounting, finance, political science, sociology, economics and marketing (Eisenhardt, 1989). Ross (1973) even says that “examples of agency are universal”, going on to explain that virtually all contractual arrangements contain at least some elements of agency theory, whether it be between employer and employee or owner and manager.
Figure 1: The Agency Dilemma: Information Asymmetry in the "Principal-Agent" Problem (Popovich 2012).
2.1.1 Moral hazard
13
A main source of moral hazard is information asymmetry (Holmstrom 1979). This is a situation in which one of both parties involved in a contract has information the other party cannot find out. The principal is not always able to fully observe the actions of the agent and has to rely on imperfect estimators.
When a company uses an intermediary to advertise, an information asymmetry exists too. The principal cannot at all times observe whether the agent actually uses the advertising in a place where the customers can see it and whether the agent uses the latest version of the advertising. The agent could also use the advertising screen to advertise its own products next to the principal’s advertisements. This means less screen time for the products of the principal. Or it could ask other companies to use the screen against a (small) fee.
In this case the principal also has to rely on imperfect estimators like taking a sample of the to check if all aspects are as agreed. However the principal may also want to prevent any problems from occurring and not just adjust when problems have occurred. Plus the principal cannot know whether an individual agent only shows advertising from the principal. As Popović et al (2012) say, control in this situation is never perfect and there is either a lack of information or a delay in information. When using a sample to check on the agent, the principal only has part of the necessary information because it cannot examine all agents. The sample also only gives information about a certain moment in time. An agent might feel less obliged to do as the principal wishes right after being checked because he will likely not be part of a sample again soon. Thus checking a sample of its agents will not provide the principal with enough information over time.
2.1.2 Adverse selection
14
When advertising using an intermediary, adverse selection can exist when the advertising uses modern technology and has to be updated regularly. The principal cannot know whether the agent is capable of doing this. The update process is done using special
software and a wireless connection and involves at least some technical knowledge from the agent. Not every agent can be assumed to have this knowledge.
2.1.3 Other Problems
There are two other problems that can occur in an agency relationship (Eisenhardt 1989). First the agency problem, which can exist when the goals of the principle and agent conflict and/or when it is difficult or expensive for the principle to monitor what the agent is doing. This is a problem because the principal wants to achieve a certain goal and wants to make sure the agent is working towards achieving that goal. A principal and agent will have a hard time working together when they do not set the same goals and act out of self interest. And even when they have the same goals an agent may shirk. To prevent this the principal wants to be able the monitor the agent. If this is difficult or expensive the principal will have to trust the agent and cannot know for sure if the goals are being met.
The second problem is the problem of risk sharing, which occurs when the agent and the principle have different attitudes towards risk. This can lead to the principal and the agent preferring different actions in a given situation based on their own risk preferences. Which again leads to the agent not acting in a way the principal would like him to.
An agency problem exists when a company advertises using an intermediary that will show their advertising. The company is a principal that pays the agent, the intermediary, to show the advertising in its stores. The principal pays the agent for use of its stores and for showing the advertising to the clients.
Main goal of the principal is to reach a larger audience than would be possible by other means of advertising with the same budget. An agent however has very different goals, it is concerned with helping its own customers in the best way possible and not with making sure the customers see the advertising. So a goal conflict arises between the advertising
15
There may also be a difference in risk preferences. The principal could be willing to take some risks to make sure the customers see the advertising and remember it. However the agent may not want to take a risk and prefer a modest advertisement. The principal wants to deal with the problems in the agency relationship and create commitment. So commitment will be discussed next after which the possible solutions for the agency problems are discussed.
2.2 Commitment
One way to deal with the problems that can exist between principal and agent is by creating commitment on the side of the agent. Commitment exists when an exchange partner
believes a relationship is important enough to warrant maximum effort towards maintaining that relationship (Morgan & Hunt 1994). Thus the agent believes the relationship to be important enough to put effort into it. Meyer & Allen (1993) argue that organizational
commitment consists of three components. First affective commitment, which is an emotional attachment to an organization. A person with affective commitment can identify with the goals of an organization. Next is continuance commitment, which exist when someone does not want to end a relation with an organization because this comes at a cost. Lastly there is normative commitment, which comes from a feeling of obligation towards the organization, because either the person or the organization has invested time and effort in the relationship. Affective, continuance and normative commitment should be seen as components of
organizational commitment and not as types of commitment (Meyer & Allen 1991).
Meyer et al (2002) conducted a meta-analysis to research the antecedents and
consequences of commitment and found affective commitment to have the strongest positive relationship with work relevant behaviours. These behaviours included performance and organizational citizenship behaviour. This is followed by normative commitment which has an influence but a less strong one. Continuance commitment did not have a positive relationship with these work relevant behaviours.
16
organization when they have the chance (Tufail et al 2012). So for a principal it is important to achieve and maintain high commitment from its agent. This will ensure that the agent will work toward the goals of the principal and not against them. One aspect that is said to lead to commitment is pay satisfaction, so this is discussed next.
2.3 Pay satisfaction
Pay satisfaction is the amount of overall positive or negative feelings one has towards their pay (Williams et al 2006). In this definition pay refers to all forms of compensation so not only direct payments but also non cash benefits and pay raises.
Pay satisfaction is usually said to be determined by four factors that were first described by Heneman & Schwab (1985). Later confirmatory factor analyses confirmed the existence of four factors in determining pay satisfaction (De Gieter et al 2006; Judge 1993; Judge & Welbourne 1994; Mulvey et al 1992).
The first factor determining overall pay satisfaction is pay level, which is the current direct compensation. Next is the satisfaction with raises, thus any changes in the pay level over time. Third is the satisfaction with any extra benefits like insurance. And the final factor is the pay structure and administration, which refers to the different pay rates and the
administrative procedures that have to be followed (Williams et al 2007).
Sturman and Short (2000) added on the four factors by stating that to keep up with current pay practises a fifth factor should be included. This fifth factor was satisfaction with bonuses. They found all five proposed factors to influence overall pay satisfaction significantly.
Williams et al (2008) in a later study used a fifth factor that was more broad than the bonus factor used by Sturman and Short (2000). This fifth factor represents the satisfaction with compensation forms like bonuses, incentives and commissions that are not part of the base payment. These forms of compensation are what Milkovich & Newman (2005) call variable pay. It is variable because the height of payments depends on the performance of the agent. The satisfaction with variable pay not only covers the height, but also the method of
17
Williams et al (2008) described the aspects that influence the fifth factor used in their
research, satisfaction with variable pay. The aspects are the way bonuses are used and the procedures in determining height and form of variable pay. Garcia et al (2009) also found the height of the last variable pay and the average height of previous variable payments to be of influence.
A lot of research has been done on pay satisfaction already, but Heneman and Judge (2000) said that more work should be done to explore the links between pay satisfaction and its outcomes. They also said that the pay satisfaction literature had to keep developing to keep pace with changes in pay practises. Modern pay packages are not the same as they were when the first four factors were developed. Current research may not capture satisfaction with important portions of modern pay packages. Williams et al (2008) also say that more insight into the factors that determine modern day pay satisfaction should be a focus of research. They go on to say that this has practical use as well because this “will allow managers to better understand their employees’ reactions to their compensation systems and to target interventions for problem areas.”
When a principal uses an agent to advertise, he will offer him a certain pay to entice the agent to work for him. So pay level satisfaction will be an important factor. As said in the introduction, part of the advertisements will show products that the agent already sells in its stores. The principal can set a goal for the amount of these products an agent has to sell to receive the variable payment. This goal is based on previous sales data to make sure it is achievable but still a challenge. The agent can share its monthly or quarterly sales data for these products and the principal can see if the goal has been met. Using variable payments means that satisfaction with variable pay and with the structure of the payment will also determine overall pay satisfaction. However, this research limits itself to a certain time period and no pay raises will be included in this period. So satisfaction with pay raises will not be included. Next to this, no benefits will be offered to the agents. Benefits respond to specific needs like security and personal welfare (Balkin & Griffeth 1993).
2.4 Pay satisfaction & commitment
18
feelings about the principal. The high payments that lead to satisfaction also lead to a feeling of obligation towards the principal thus enhancing normative commitment. Singh et al (2011) found pay satisfaction is related to continuance commitment too. This is because when the agent is more satisfied with its pay the cost of leaving increases. DeConinck & DeConinck (2010) also found that pay level satisfaction was positively related to organizational
commitment.
Not only pay satisfaction as a whole but also the aspects that determine pay satisfaction have been researched to find their effect on commitment. Dulebohn & Martocchio (1998) for instance looked at 6 different variables and the effect on commitment. They found that satisfaction with the base payment and the understanding of the structure of payments positively correlated with affective commitment. Williams et al (2002) also found that not only the level of pay but also the system and structure of pay correlate with commitment. This was confirmed by a meta-analysis (Williams et al 2006) adding that also the perceived
discrepancy between the amount of pay received and the amount that should be received. Shields et al (2012) later also found that pay differences had a negative effect on
commitment and confirmed that both pay fairness and pay understanding had a positive effect on overall commitment.
Concluding, an agent that feels satisfied with the payments received from the principal is more likely to show commitment, which leads to the following hypothesis.
H1: Higher pay satisfaction leads to higher commitment for an agent.
2.5 Solutions for the agency problem
19
information, especially no personal information, so it is assumed that this is not a concern to them. A principal cannot use information systems to check if the agent is in fact showing the advertisements. It also cannot tell whether the advertisements are visible and up to date. So using information systems is not the best solution for the agency problems in this case.
The second suggestion is to make the contract between the agent and principal outcome based, this is the focus of the current research. An outcome based contract would lead the agent to have the same goals as the principle because the reward for the agent and the principle depend on the same actions. The principal can set a goal for the amount of these products an agent has to sell to receive the variable payment. This goal is based on previous sales data to make sure it is achievable but still a challenge. The agent can share its monthly or quarterly sales data for these products and the principal can see if the goal has been met. This reduces the conflict of self-interest discussed earlier.
The types of contract that can be used in a principal agent relationship will be discussed further in the next paragraph. The main question remains if the agent has a preference for a certain type of contract and a certain type of compensation.
2.6 Compensation
20
In general the principal has two options for the type of compensation in this case namely fixed pay or variable pay. Fixed pay does not vary according to the achieved results or performance whereas variable pay changes directly with the results achieved (Madhani 2012). This also means that variable pay has to be re-earned every period and thus means more risk for the agent.
The literature on payment structure and satisfaction is far from unanimous. Mccausland et al (2005) for instance found that variable payments had a negative effect on satisfaction overall. Petrescu & Simmons (2008) however, found that paying workers for their performance lead to higher satisfaction. Artz (2008) says that variable pay can increase satisfaction because of increased earnings but it decreases satisfaction due to the increased risk. Overall, it
increases satisfaction only for certain groups, mainly working in larger firms. Parent (1999) agreed by stating that the increased pay that one can earn with variable payments increases satisfaction. One might also feel satisfied because they feel that their hard work is being recognized and rewarded (Brown & Sessions 2003). Some, however, feel that variable pay brings stress because of the increased monitoring necessary to determine if goals have been met (Fernie & Metcalf 1999).
Pouliakas &Theodossiou (2009) found that group performance incentives had a positive effect on satisfaction but individual incentives did not show a significant impact on
satisfaction. Berger & Schwab (1980) also looked at the link between fixed or variable pay and pay satisfaction. Earlier research had shown that fixed payment had lead to more
satisfaction. Berger & Schwab, however, found that variable payments could be as satisfying or even more satisfying than fixed payments when the agents do not have to deal with administrative difficulties.
So overall compensation has an impact on satisfaction, but the direction of this impact remains arbitrary. The way of compensating can be regulated in a contract with the agent which will be discussed next.
2.6.1 Contract
21
agent and rewards the agent when specific tasks have been completed. In an outcome contract the principal measures the agents’ output and holds him responsible for this. In a behaviour oriented contract the agent gets paid a fixed fee if he abides to the conditions of the contract. In an outcome oriented contract part of the reward is fixed but another part is variable and can change per period. Specifically, this means the agent will sell products the principal advertises. If the agent reaches a certain target, set by using sales data, it will receive the additional variable payment. So an outcome contract means part variable pay whereas a behaviour contract means fixed pay only, these wordings will be used
interchangeably.
The agent is rewarded when he reaches targets, how the agent does it is less important (Davies & Prince 2010). Outcome contracts shift risk from the principal to the agent whereas with a behaviour contract the risk is with the principal (Bloom & Milkovich 1998; Desmond et al 2011; Eisenhardt 1989; Jensen & Murphy 1990; Krafft 1999; Oliver & Anderson 1994). Because an agent is more risk averse than the principal, the agent will prefer a behaviour oriented contract (Eisenhardt, 1989). So even though the literature on compensation is arbitrary about the preference for either fixed or variable pay, the literature on contracts suggest that an agent prefers fixed pay only.
For an agent to be satisfied with overall compensation he will have to be satisfied with pay level, pay structure and the variable part of pay. When offering no variable pay this means that pay level will be the main determinant of satisfaction. If part of the payment is variable the structure of payment and the height of the variable pay will be important too.
When an agent receives more variable pay he also faces more risk (Madhani 2009; Mccausland et al 2005; Rendell & Simmons 1999). And, because the agent has no
guarantee he will receive the variable pay each period, offering a high degree of variable pay could decrease pay satisfaction (Cloutier et al 2013; Milkovich & Newman 2005).
This leads to the following hypotheses:
22
2.7 Store size
The agents in this research get paid to place screens and show advertisements on them. If a company has multiple stores or a large store it can place more than one screen. Every screen placed would be paid for. So a large company would be able to place more screens. A larger agent is expected to have a higher income because it has more customers and more space to sell products from.
Kissan & Thevaranjan (1998) found that agents who are more tolerant towards risk demand higher reservation wages. This is the lowest wage an agent would be willing to work for. However because these agents are more tolerant towards risk they will accept part of their pay in the form of incentives. Or as Desmond et al (2011) say it, more risk-averse agents demand a higher level of base compensation when facing uncertainty. So to meet these demands the principal will have to pay more fixed fees. Madhani (2009) also found that when the variable proportion of total compensation increases, agents will demand higher pay because they now face extra risks. He also found that agents with lower incomes are less willing to receive part of their pay variably. Caroli & Garcia-Peñalosa (2002) found the same saying that as income grows, agents become less risk-averse and agree to move to more variable pay. An empirical analysis by Shaw (1996) showed that indeed growth of income was positively correlated with risk taking. Dohmen & Falk (2011) found that risk-tolerant workers prefer variable pay which has a more uncertain payoff. Risk-averse workers on the other hand have a preference for fixed pay. So, agents that earn more money will be more inclined to take the risk associated with variable pay. This was supported by Mccausland et al (2005) who found that variable pay had a positive effect on satisfaction of high paid workers and a negative effect on low paid workers.
Next to this, Artz (2008) found that variable pay increases satisfaction for certain groups only. More specifically he found that workers in large firms are more satisfied with variable pay.
So an agent that has a higher income is willing to take more risks. And agents with lower income will be more risk averse and less likely to accept variable payments. Translating this to the situation with advertising using an intermediary; an agent that has room for a low amount of screens will want a higher fixed fee. This is because having room for a low amount of screens means having a lower income than an agent that has a lot of screens. And
23
not be as likely to accept variable pay as an agent with lots of room for screens. This leads to the next hypothesis
24
3. Problem statement
The hypotheses described in the theoretic review lead to the following problem statement and conceptual model. The conceptual model summarizes the different links between the described constructs and the direction of these links. First the problem statement::
Does a higher degree of fixed pay lead to higher pay satisfaction in an agency theory setting, and does store size influence this relationship? And does this higher pay satisfaction lead to higher commitment?
3.1 Conceptual model
- + + Degree of fixed payPay satisfaction Commitment
25
4. Methodology
The hypotheses described in the conceptual model were tested to see if they should be rejected or not. This chapter describes how they were tested. First the research design is described, followed by a description of the scenarios. Then, the questionnaire design is mentioned. Followed by the analysis method and an explanation of the tests used for the hypotheses. Finally, the sample size is discussed.
4.1 Research design
The hypotheses were tested using questionnaires accompanied by different scenarios. Malhotra (2006) says one of the advantages of using questionnaires is that they are easy to administer. Using fixed alternative questions makes the data obtained reliable and reduces variability of results. Finally, analysing and interpreting data is relatively simple when using a questionnaire. In these scenarios only the variables of interest change and the other aspects stayed the same. This was done to be able to compare the different groups of respondents based on the variables of interest.
To test the aforementioned hypotheses a questionnaire was designed and was answered by a number of respondents that worked in a barber shop. Barber shops were chosen because these shops lend themselves well for in-store advertising. A lot of people come in, who then sit in the same place for some time and have to look in pretty much the same direction the whole time while getting their hair cut. Because the customers look in the same direction, placing a screen showing different advertisements in front of the customers would lead to high visibility. Not a lot of advertising is shown in a barber shop yet so the advertising screens would not get lost in a swarm of advertisements. Barber shops usually also sell some hair products meaning they could get some variable reward for selling a target amount. A link to the questionnaire was sent to a number of barber shops. The accompanying e-mail explained what they had to do and that the person filling in the questionnaire had to have decision making authority in a situation like this and would be able to assess whether or not they would accept the advertising.
26
more luxurious, segment was not included. These choices were made to keep heterogeneity on variables that were not of interest low. This was done to make the comparison of the variables of interest more reliable.
4.2 Scenarios
The questionnaires were accompanied by a randomly assigned scenario to make
comparison between different scenarios possible. These questionnaires were filled in online and the respondents did not know about the other scenarios to make sure they were
unprejudiced and did not compare different scenarios. The respondents were explained a situation in which a principal would pay them to show advertisements using one or more digital screens. They would then have to decide whether or not they would show the
advertising in their store assessing the type and height of reward they would receive. The full scenarios can be found in appendix 1.
Three scenarios were used to divide the sample of respondents into different groups, this makes it possible to find out if the groups differ. The amount of money the agents would receive was the same in every scenario, only the division between fixed and variable changed. The payment changes in the three scenarios can be found in table 1 and will be explained after that.
Table 1
Payment changes in the different scenarios
Fixed
payment
1st screen
Variable
payment
1st screen
Fixed payment
every next
screen
Variable
payment every
next screen
Scenario 1
100
0
50
0
Scenario 2
70
30
30
20
Scenario 3
50
50
20
30
27
Then the groups that would receive variable pay. As said earlier, a positive factor of variable pay is the possibility to receive more when meeting goals. A negative factor, however, is the added risk of not receiving as much as someone with fixed payments when not meeting the goals. To account for both factors, the scenario described a situation in which a respondent would receive a base payment of 70 Euros. Next to this it could make 20 Euros extra in a bad month and 40 Euros extra in a good month. A good month would mean reaching the target amount of sales, while a bad month would mean not reaching the target. Because the chance of having a good month was given as being about 50 percent the average pay would again be 100 Euros. For every next screen they would get 30 Euros fixed payment and 20 Euros on average variable.
To test whether the base payment would influence the satisfaction with variable pay, two different groups were needed. One of these groups would receive a higher amount of fixed payment than the other. This means that one group would receive a larger part as variable and would thus face more risk. So next to the group that received 70 Euros fixed, another group existed that received only 50 Euros fixed. This group would receive either 40 or 60 Euros extra per month as variable pay. And thus receive about half of payments as variable pay. For every following screen this group would receive 20 Euros fixed and, on average, 30 Euros variable.
4.3 Questionnaire design
The questionnaire consists of three main parts, first asking about the pay satisfaction of the respondents, then the commitment and finally they were asked about the size of the store. The full questionnaire can be found in appendix 1.
4.3.1 Pay satisfaction
Whether the agent felt satisfied with the overall level of pay and the division between fixed and variable pay was investigated using question from the pay satisfaction questionnaire (PSQ). The questionnaire was originally developed by Heneman & Schwab (1985) and looks at 4 factors of pay satisfaction. The four factors are pay level, benefits, raises and
28
in that factor and had higher loadings than items not intended to form the factor. Judge & Welbourne (1994) also used confirmatory factor analysis on 2 samples and in both cases the four factor solution was supported. De Gieter et al (2006) tested the questionnaire in Belgium to see if the 4 factors of pay satisfaction would also exist outside the United States. Factor analysis showed that the 4 factor structure was supported. Williams et al (2008) later added a fifth factor to the PSQ. This fifth factor was satisfaction with variable pay.
For this research only selected questions concerning pay level, variable pay and pay structure & administration were used. This because the focus of the research is on these aspects and there were no extra benefits or raises of pay in the used scenarios.
Specifically the following questions were asked to the respondents:
Giving the described scenario, in how far would you say you would be satisfied with the following aspects?
- Base pay
- Height of variable pay - Total pay
- The division between fixed and variable payments
29
4.3.2 Organizational commitment
Meyer et al (1993) developed a 3 component model of organizational commitment. They also developed a questionnaire that can be used to measure the 3 components of commitment. Emphasising that the affective, continuance and normative commitment should be seen as components of organizational commitment and not as types of commitment (Meyer & Allen 1991). Measures of the three components were developed in an earlier study (Allen & Meyer 1990) and were then found to be psychometrically sound. Later studies have confirmed the existence of the 3 components of commitment using confirmatory factor analysis (Hacket et al 1994). These studies also confirmed that the three scales measure distinct constructs. So the three forms of commitment can be investigated separately with this questionnaire (Shore & Tetrick 1991). Analysis has also shown that the proposed antecedents correlate
differentially with the three constructs (Allen & Meyer 1990). Finally the three constructs measured in the questionnaire have been found to differentially correlate to job performance (Hackett et al 1992).
This questionnaire has been modified to Dutch by De Gilder et al (1997). In the
questionnaire, the forms of commitment are measured using a 5 point Likert scale ranging from totally disagree to totally agree. They also used confirmatory factor analysis to show that the Dutch version of this questionnaire consists of the same three constructs as the English version. Next to that the correlations with the proposed antecedents and
consequences were found to be very similar to those in the English version.
30
4.3.3 Store size
The respondents were also asked how big their store was, or more specifically how many chairs they had to seat the customers coming in for a haircut. A screen would be placed in front of each chair in use. So the amount of chairs determined the amount of screens that would be placed inside a shop.
This was done to determine whether respondents with bigger stores, and thus more room for screens, and higher income, would be more susceptible for a smaller base fee and a
relatively higher variable fee.
4.4 Analysis method
The data received from the questionnaires were then analysed to find out if the hypotheses could be confirmed or not. This part describes how this analysis was done.
Pay satisfaction is a formative scale. This means that an increase in any one of the items of pay satisfaction leads to an increase in overall pay satisfaction. An increase in pay
satisfaction does not mean that all items increase as well (Diamantopoulos & Winklhofer 2001). Pay satisfaction was made up out of the four items. These items were averaged to lead to a single pay satisfaction construct.
A problem that might exist in a formative scale is that off multicollinearity. This is because multicollinearity would make it hard to distinguish the influence the different items have on pay satisfaction. So the four items that make up the pay satisfaction were tested for collinearity. Using a variance inflation factor (VIF), which for all four items had to be lower than 10 (Diamantopoulos & Winklhofer 2001).
Commitment is used as a reflective scale (Gruen, Summers & Acito 2000). The questions were on the same 5 point Likert scale and all but one were used in the same direction. The one question, asking if respondents were likely to use an alternative was worded negatively. This question had to be reverse-scored first so all items were in the same direction. So, when someone was very likely to use an alternative, he scored low on commitment.
31
continuance commitment. The other questions were used for normative commitment. To see if the items did indeed have a high enough internal consistency to be used in these
constructs, a Cronbach’s alpha test was used.
Finally the scores for both types of commitment were averaged to give one composite commitment score as has been done in other studies (Lumley et al 2011 & Meyer et al 1993). This leads to one number for commitment for all respondents ranging from 1 (low commitment) to 5 (high commitment).
4.5 Hypothesis testing
After calculating the average commitment and pay satisfaction for the whole sample the hypothesis testing could begin. For every hypothesis the analysis method is described. In these tests the above described averages for commitment and pay satisfaction were used.
4.5.1 Hypothesis one
To test if higher pay satisfaction leads to higher commitment a regression analysis was conducted. Both pay satisfaction and commitment were Likert scales, containing multiple items that can be used as interval data in a test. Commitment was used as the dependent variable and pay satisfaction was used in the model to explain variance. For the regression to be significant the p-value had to be lower than 0.05. If this was the case the beta value of pay satisfaction was examined. The beta explains how much commitment increases or decreases with an increase of 1 for pay satisfaction (Malhotra 2006). This beta had to be positive to be in line with the hypothesis. If it was negative it would mean that higher pay satisfaction leads to lower commitment. Then the significance of this beta was examined to see if the relationship between pay satisfaction and commitment was indeed significant.
32
model (Colton & Bower 2002). So a regression with a low R square can still tell something about an independent variable.
4.5.2 Hypothesis two and three
Two multiple regression models were run, both containing pay satisfaction as the dependent variable. The independent variables in the first model were the amount of screens and the percentage of fixed payment. The number of screens an agent had room for was used as a way of determining store size. A bigger store would have room for more screens. This store size is hypothesized to moderate the relationship between the amount of fixed payments and pay satisfaction. To test this, an interaction effect between the two variables was added to the model.
To prevent multicollinearity the amount of screens and amount of fixed pay variables were centered first (Warner 2013). They were then multiplied to make a new interaction variable which was added in the second model. The second model had the same independent variables but an interaction effect between the two variables was included.
The percentage fixed pay depends on the number of screens an agent has. The more screens an agent has, the lower the percentage of fixed payment is. The amount of fixed payment also differs per scenario. The first group received fixed payments only. The amount of fixed and variable payments for the other two groups can be found in table 1 and 2. Table 2 shows the percentage of both type of payments for the group that received a high amount of fixed payment. Table 3 shows the percentage of both type of payments for the groups that received a low amount of fixed payment.
Table 2
Percentage of variable payment in high fixed payment group
33
Table 3
Percentage of variable payment in low fixed payment group
Again the significance of the regression was tested first using the F-value. This time the first and second model were compared too, using the change in R square. The change was tested with another F-test. If the change was significant it would mean that adding the interaction effect significantly improved the prediction of the model.
Then the different betas were examined. The beta for the amount of fixed payment is
hypothesized to be positive. If this was the case the significance of the beta was examined to see if the relationship was a significant one. The beta for the interaction effect would be negative, showing that larger companies do not prefer fixed payments.
4.6 Sample size
Having a high enough amount of respondents is important to achieve a high enough power of the tests. The power is the probability of rejecting a hypothesis when it is indeed false (Malhotra 2006).
For determining the sample size a number of factors are important. First, the alpha level or type I error. This error is the incorrect rejection of a null hypothesis. This alpha level is set at a tolerable level of risk of rejecting a hypothesis that is in fact true (Malhotra 2006). The alpha level is 0.05 in this case which is a common level. Using a lower alpha means needing
34
a larger sample. A larger sample means more data and less chance of wrongly rejecting a hypothesis.
Then the amount of independent variables used in the regression model. More variables means needing a bigger sample size, because more data is needed to detect all the different effects. Also the different levels of the independent variables are important. In this case, three scenarios were used with three different levels for the independent variable.
Next is power, which is the probability of achieving a significant R-square. The higher this power is, the larger the sample needs to be (Cohen & Cohen 1983). A common use of power is 0.80.
Cohen & Cohen (1983) said that with a power of 0.80, a significance level of 0.05 and 2 independent variables, a sample size of 50 would detect any R-square above 0.19.
Numerous researchers have looked at sample size in multiple regression, a number of which agree on a sample size around 50 for an analysis with two independent variables with three scenarios (Green 1991; Harris 1975; Tabachnick & Fidell 1989).
35
5. Results
A total of 41 respondents answered the questionnaire. They were fairly equally separated among the different scenarios, no scenario had significantly more or less than the others. 14 respondents received fixed pay only, 14 received a low amount of variable pay and 13 others received a high amount of variable pay. The results were first examined for any outliers, respondents that answered either 1 or 5 on every question. These outliers can distort the end results and would have to be identified beforehand. However, no outliers were determined so all respondents were used in the further analysis.
Most respondents can be classified as a small store, 63.4 percent had only 1 or 2 chairs occupied on average. 24.4 Percent had 3 chairs and another 12.2 percent had more than 3.
5.1 Commitment
As said earlier, for commitment the different questions were used to measure one construct, consisting of two components. The questions were on the same 5 point Likert scale. One question, asking if they would be likely to use an alternative, was scaled in a different direction than the others. So this question had to be reverse-scored first. To see if the items did indeed have a high enough internal consistency to be used in the two components, the Cronbach’s alpha test was used. The test showed a Cronbach’s alpha of 0.74 for
continuance commitment, which is higher than the required 0.7. The Cronbach’s alpha for normative commitment was 0.69, which was acceptable. The test results can be found in appendix 2. The high Cronbach’s alphas mean the internal consistency is high enough and the questions can be used to measure the two components of commitment that were used.
36
5.2 Pay satisfaction
For pay satisfaction multiple questions were also used to measure one construct. The answers to all four items of pay satisfaction were averaged to reach a single construct. Then the items were checked for multicollinearity. The results of this test can be found in appendix 3. The VIF scores were all lower than 10 meaning collinearity is not a problem.
The average pay satisfaction across all respondents was 3.48. The highest was 4.5, the lowest scored 2.
5.3 Hypotheses testing
Here, the aforementioned hypotheses resulting from the theoretical framework will be tested. The results of the test will determine whether a hypothesis should be rejected or not.
5.3.1 Hypothesis one
The data from all 41 respondents were used in a regression analysis with commitment as the dependent variable and pay satisfaction as the independent variable. The full results can be found in appendix 4. The regression was only significant using an alpha level of 0.1 because the test gave p=0.078. The beta of pay satisfaction was 0.279. This positive beta means that a higher pay satisfaction does mean higher commitment. However the R squared is low at 0.078. So only 7.8 percent of change in commitment is explained by the regression model. So even though pay satisfaction does significantly influence commitment it seems there are a number of other variables that have an influence as well.
5.3.2 Hypothesis two and three
37
two to model one gives a R square change of only 0.005, which is not significant (p=0.556). So adding an interaction between the store size and the amount of fixed payments does not increase the model significantly. However, model two is significant (p=0.00) so it can be used to assess the relationship between the variables in the regression. The analysis of VIF shows no problem with multicollonearity, all scores are well below 10.
The betas can be found in table 4. The beta for the amount of fixed payment is 0.639 and was found to be significant (p=0.00). This means hypothesis 2 is supported and higher fixed payment does lead to higher pay satisfaction.
That the interaction effect does not significantly improve the model is further confirmed by looking at the betas. The beta for the interaction is insignificant at p=0.56. Which means hypothesis three has to be rejected.
Table 4
Betas and their P-values
Beta P-value
Amount of fixed payment 0.639 0.00
Store size -0.041 0.76
Size * Fixed payment 0.073 0.56
38
6. Conclusion and discussion
Even small companies will want to advertise because it can lead to a increase in sales, consumers’ affect, cognition and word-of-mouth (Bruce et al 2012; Keller & Fay 2009; Vakratsas & Ambler 1999) . A way of doing this, that also takes into account the fact that small businesses do not have a lot of room to advertise, is advertising using intermediaries. This does however create an agency relationship.
Logan (2000) called for research investigating the choice between a behaviour or outcome oriented contract in an outsourcing relationship. Tate et al (2010) also called for a research looking at an agency relationship in service outsourcing. The analysis in the current research showed that agents did have a preference for a behaviour oriented contract in which all the payments were fixed. This is in line with earlier agency theory literature which suggests that an agent will prefer a behaviour oriented contract because the agent is more risk averse (Bloom & Milkovich 1998; Desmond et al 2011; Eisenhardt 1989; Jensen & Murphy 1990; Krafft 1999; Oliver & Anderson 1994). The less risky option of completely fixed pay is preferred above a outcome oriented contract.
This finding also adds to the body of literature concerning pay structure and satisfaction. Only a few studies have examined the link between variable pay and pay satisfaction so far (Cloutier et al 2013). The current research is in line with Mccausland et al (2005) who also found that variable pay leads to lower pay satisfaction. They also give a possible
explanation for this, stating that variable pay schemes might be seen as controlling. This could also be the case when advertising using an intermediary. The agent would have to present its administration, showing if and how much products he sold. The only way an agent could receive variable pay is if it lets the principal check him regularly. Agents may see this as controlling or will just not want to commit to the extra administrative difficulties. This explanation was also given by Berger & Schwab (1980). If the agent did not have to go through the extra effort, it may not have as clear a preference.
39
& Balkin 1989). That study however looked at group pay plans, so the current research adds to the literature by looking at individual pay plans. Analysis showed that pay satisfaction decreased when the percentage of fixed payments decreased. This finding was also in line with the aforementioned literature (Madhani 2009; Mccausland et al 2005; Rendell &
Simmons 1999). If the principal wants to shift risk to the agent, the agent wants this to be as low as possible. A higher base payment means a smaller part is variable and thus less risk for the agent.
This confirms what Artz (2008) found, variable pay increases satisfaction because an agent can earn more by his actions and feels rewarded. On the other hand it decreases satisfaction because the risk is higher for the agent. The group with low base payment were less satisfied overall because they had more risk. But satisfaction did increase when taking satisfaction with variable pay into account.
Another hypothesis that was tested was that of the positive link between pay satisfaction and commitment. Agents that showed higher pay satisfaction did indeed show higher
commitment. Feeling that one is rewarded apparently leads to an agent wanting to keep working for the principal. This higher commitment will lead to the agent putting in enough effort to keep the relationship with the principal at a high level. An agent that receives a high reward might also feel a sense of obligation towards the principal. Feeling he should put in the effort to earn a reward that high.
The significant effect of pay satisfaction on commitment was in line with earlier research (DeConinck & DeConinck 2010; Dwahan & Mulla 2011; Singh et al 2011). It is however the first to research commitment in an agency setting. Fayezi et al (2012) pointed to the gap in research relating to agency theory and commitment.
Artz (2008) also found that workers in large firms are more satisfied with variable pay. This same result was not found in an agency relationship. Large agents showed lower pay satisfaction, contrary to the expectation.
Larger agents did show higher satisfaction with variable pay but this difference was
40
6.1 Recommendations
As said in the introduction of this research, advertising can have a lot of positive effects for a company. Seeing sales and word-of-mouth increase can be the difference between surviving and going bankrupt for a small company. Companies tend to want a large audience and will look for new and not too expensive ways to advertise in a large scale. This research looks into one way of doing this, using an intermediary to advertise with digital screens. The research shows that the highest pay satisfaction can be achieved using fixed payments. So the first managerial implication is that if a company wants to use an agent to advertise it should use fixed payments.
41
7. Limitations and further research
This research limits itself to entrepreneurs working in a barber shop because the situation used in this research is best suited for these entrepreneurs. Because the customers look in the same direction when visiting a barber, placing a screen showing different advertisements in front of the customers would lead to high visibility. It is recommended to conduct this research in other sectors with principal-agent problems. Perhaps other sectors show different results or maybe these hypotheses hold true for other common business sectors as well.
The research also uses a questionnaire to examine pay satisfaction and commitment. These questionnaires have been developed after long research and tested vigorously but they can give problems. Some respondents may be generally more positive in their answers than others even though they do not feel different satisfaction. So there could be some
measurement error. However, no outliers or extreme answers were found and the sample size should be sufficient to deal with small amounts of measurement error.
This research focuses on commitment as a whole. It is the first research examining the type of contract in an agency relationship and commitment and could be seen as an exploratory research. It does not look into the effect of pay satisfaction on the different kinds of
commitment. Instead the research was conducted to first show if indeed a link between variable pay, pay satisfaction and commitment exists. Perhaps one form of commitment is more strongly affected by pay satisfaction than others. This has been researched before and continues to be an interesting subject. The found link between pay satisfaction and
commitment was not strong in this research. This suggests there are other aspects that influence commitment in an agency relationship. Further research should be done to determine these aspects to be able to keep commitment from the agent high.
It also for the most part limits itself to pay satisfaction as a whole and does not look into the different forms of satisfaction. Perhaps not every type of pay satisfaction leads to a higher commitment.
42
Literature
Allen, N. J. & Meyer J. P. (1990) "The Measurement and Antecedents of Affective, Continuance and Normative Commitment to the Organization". Journal Of Occupational Psychology. Vol. 63 (1) 1-18
Artz, B. (2008) “The Role of Firm Size and Performance Pay in Determining Employee Job Satisfaction Brief: Firm Size, Performance Pay, and Job Satisfaction”. LABOUR: Review Of Labour Economics & Industrial Relations. Vol. 22 (2) 315-343
Balkin, D. & Griffeth, R. (1993) “The Determinants of Employee Benefits Satisfaction”. Journal Of Business & Psychology. Vol. 7 (3) 323-339
BCG, Boston Consulting Group (2013) “Unlocking the Digital-Marketing Potential of Small Businesses” accessed March 2013 via
https://www.bcgperspectives.com/content/articles/digital_economy_marketing_sales_unlocki ng_digital_marketing_small_businesses/
Bergen, M., Dutta, S. & Walker Jr, O.C. (1992) “Agency Relationship in Marketing: A Review of the Implication and Applications of Agency and Related”. Journal of Marketing. Vol. 56 (3) 1-24
Berger, C.J. & Schwab, D.P. (1980) “Pay Incentives and Pay Satisfaction”. Industrial Relations. Vol. 19 (2) 206-211
Berle, A. & Means, G. (1932) “The modern corporation and private property”. New York: Harcourt Brace.
43
Bradley, M.T. & Brand, A. (2013) “Alpha Values as a Function of Sample Size, Effect size and Power: Accuracy over Inference”. Psychological Reports. Vol. 112 (3) 835-844
Brown S. & Sessions J.G. (2003) “Attitudes, Expectations and Sharing”. Labour.
Vol. 17 (4) 543-569
Bruce, N.I., Peter, K. & Naik, P.A. (2012) “Discovering how Advertising Grows Sales and Builds Brands” Journal of Marketing Research. Vol. 49 (6) 793-806
Businessweek (2013) “Small Businesses still aren’t Spending on Online Ads” accessed March 2013 via http://www.businessweek.com/articles/2013-03-19/small-businesses-still-arent-spending-on-online-ads
Carifio, J. & Perla, R.J. (2007) “Ten Common Misunderstandings, Misconceptions, Persistent Myths and Urban Legends about Liker Scales and Liker Response Formats and their
Antidotes”. Journal Of Social Sciences. Vol. 3 (3) 106-116
Caroli, E. & Garcia-Penalosa, C. (2002) ”Risk Aversion and Rising Wage Inequality”. Economics Letters. Vol. 77 (1) 21-26
Cloutier, J., Morin, D. & Renaud, S. (2013) “How does variable pay relate to pay satisfaction among Canadian workers?". International Journal Of Manpower. Vol. 34 (5) 465-485
Cohen, J (1988) “Statistical Power Analysis for the Behavioral Sciences (2nd ed.)”. New Jersey: Lawrence Erlbaum Associates
Cohen, J. & Cohen, P. (1983) “Applied Multiple Regression/Correlation Analysis for the Behavioural Sciences”. Hillsdale, NJ: Lawrence Erlbaum
44
Davies, M. & Prince, M. (2010) “Advertising Agency Compensation, Client Evaluation and Switching Costs: an Extension of Agency Theory”. Journal of Current Issues & Research in Advertising. Vol. 32 (1) 13-31
De Gilder, D., van den Heuvel, H. & Ellemers, N. (1997) “Het 3-componenten Model van Commitment. / A Three Component Model of Organizational Commitment”. Gedrag en Organisatie. Vol. 10 (2) 95-106
De Gieter, S., de Cooman, R., Pepermans, R., Caers, R., Du Bois, C. & Jegers, M. (2006) “Dimensionality of the Pay Satisfaction Questionnaire: a Validation Study in
Belgium”. Psychological Reports. Vol. 98 (3) 640-650
DeConinck, J.B. & DeConinck M.B. (2010) “An Examination of the Antecedents and
Outcomes of Pay Satisfaction among Retail Buyers” Marketing Management Journal. Vol. 20 (1) 40-48
Desmond, L. (Ho-Fu), Ghosh, M. & Lafontaine, F. (2011) "The Incentive and Selection Roles of Sales Force Compensation Contracts." Journal Of Marketing Research. Vol. 48 (4) 781-798
Dhawan, V. & Mulla, Z.R. (2011) “The Role of Pay and Leadership in Developing Organizational Commitment” South Asian Journal of Management. Vol. 18 (2) 60-75
Diamantopoulos, A. & Winklhofer, H.M. (2001) “Index Construction with Formative Indicators: An Alternative to Scale Development”.Journal Of Marketing Research. Vol. 38 (2) 269-277
Dohmen, T. & Falk, A. (2011) “Performance Pay and Multi-Dimensional Sorting:
