Name: Patrick de Graaf Track: Finance and Organization Student number: 10278311 Name supervisor: Eszter Czibor Programme: Economie en bedrijfskunde Date: 11-02-2015
The (non-)effect of non-monetary awards on
employee motivation in the Dutch food sector
ABSTRACT
This paper investigates the motivational effects of different non-monetary award factors. It is examined whether timeliness is a supplementary factor in determining employee motivation. This is done with the aid of a survey in the Dutch supermarket Albert Heijn. The survey includes vignettes, a method with several advantages over traditional surveys. With the help of this vignette survey it is found that non-monetary awards do not have a significant influence on motivation. What we did find was that motivational effects are highly situational.
Statement of Originality
This document is written by Student Patrick de Graaf who declares to
take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is
original and that no sources other than those mentioned in the text and
its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the
supervision of completion of the work, not for the contents.
Table of contents.
Section Page number(s)
1. Introduction 1-2
2. Motivation 2-3
3.0 Recognition in general 3
3.1 Monetary and non-monetary recognition 3-4
3.2 The motivational effects of non-monetary recognition 4
3.2.1 Reciprocity 4
3.2.2 Identity 4-5
3.2.3 Status 5
3.3 Awards 5
3.4 The motivational effect of awards 6
3.4.1 Publicity 6-7
3.4.2 Scarcity/maximum number of recipients 7
3.4.3 Timeliness 7
3.5 Most important papers in detail 8-9
4. Hypotheses about the different non-monetary factors 9
4.1 The degree of publicity 9
4.2 Scarcity/Maximum number of recipients 9-10
4.3 Timeliness 10
5.0 Method 10
5.1 Vignette method in general 10-11
5.2 Our survey with an example of a vignette 11
5.2.1 The vignettes 11-12
5.2.2 Indicating customer friendliness 12-13
5.3 Context and survey design 13-14
6.0 Results 14
6.1 Characteristics 14-15
6.2 Check for randomization 15
6.3 Quick overview of results 15-16
6.4 Regressions 17-21
7. Discussion and conclusion 21-22
Appendices A, B, C, D, E, F and G 22-28
1. Introduction
Employee motivation is more important than ever before. This is because motivated employees can differentiate a company from others, which is critical in achieving success (Hansen et al., 2002). One tool for motivating employees is to give them recognition. Employee recognition has three main dimensions. The recognition is formal or informal, monetary or non-monetary and individual or collective (Long and Shields, 2010). In this thesis we focus on formal, individual, non-monetary recognition in the form of awards. Awards are ‘extrinsic, predominantly non-material incentives allocated through an institutionalized recognition program’ (Neckermann and Frey, 2013: 67). Awards can be accompanied by material incentives, but that is not always the case (Neckermann and Frey, 2013).
One example that shows that awards can contribute to a company’s success is Tony’s (Van der Hulst, 2014). In addition to great fringe benefits like a joint ski vacation, Tony’s also has an ‘employee of the quarter award’ for recognizing great performance. All together, these benefits make Tony’s a great place to work and thereby work strongly motivational for the employees. However, awards do not always have the desired effect. Gubler et al. (2013) studied a laundry factory in which employees could win a monthly attendance award, including a gift card. Although the attendance of the employees who were not so punctual before improved, others who had perfect attendance before did worse because they lost intrinsic motivation. Another problem was that employees focused too much on the attendance goal and therefore lost focus on other important tasks. All together, the award led to lower productivity than before, which was of course not the intention of the company. These examples show that awards differ in the impact they have. This is mainly due to the influence of the award characteristics (Neckermann and Frey, 2013). For that reason these characteristics require further investigation.
Therefore, this thesis will study the effect of the different non-monetary award factors on employee motivation in the Dutch food sector. Thereby it will investigate whether timeliness is a supplementary factor in determining the effects of awards. Timeliness has been indicated as an important award characteristic (Saunderson, 2004). However, it has not received the attention yet like the non-monetary factors ‘degree of publicity’ and ‘scarcity’. The Dutch food sector is an interesting place to investigate these non-monetary factors, because companies here are forbidden by law to use monetary incentives on top of the base salary. So it is interesting to see how they can effectively motivate employees without the use of money. The thesis will investigate the Dutch food sector with the aid of a survey executed among employees in the Dutch supermarket ‘Albert Heijn’. This survey includes vignettes, which are short descriptions of hypothetical situation that combines the different characteristics of an award. More information about the methodology is provided in the method section (section 5).
The survey showed that none of the non-monetary characteristics have a significant effect on motivation in the Dutch food sector. So timeliness is not a supplementary factor in determining employee motivation. Anyway, there is some support that Dutch people prefer an award
accompanied by a lot of publicity and one that is not handed out frequently. What we did find to be significant was the effect on motivation of the situational context.
This thesis starts with a more extensive explanation of why the topic of non-monetary rewards is relevant. Then there follows an outline of the motivational effects of non-monetary recognition. On the basis of this the hypotheses will be given. Thereafter the method and its context will be explained. From there we proceed with the results of the study and we conclude with a
discussion of these results. This conclusion also describes the limitations of this study and ideas for future research.
2. Motivation
For a long time, the economic field assumed that people are homo-economicus. That means that people act totally out of self-interest and money was the main incentive device to solve motivational problems like the principal-agent problem (Frey and Jegen, 2001). However, despite the use of monetary rewards, motivational problems could often still not be solved. Sometimes monetary rewards even reduce motivation by crowding out intrinsic motivation (Frey and Jegen, 2001). It appeared that the economic field needed contributions from other fields like social psychology (Gibbons, 1998).
This field showed that people do not always act out of self-interest and that there are also other ways to motivate people (Palacios-Huerta, 1998). One tool, which has been studied only recently, is non-monetary recognition. This is used as a result of difficulties with cash recognition and because some firms are not able to give monetary incentives because of lack of funds or tax laws (Long and Shields, 2010). Non-monetary recognition can be used in the form of awards. Awards vary in their impact because the importance of the different award-factors varies per award (Frey and Neckermann, 2008). The different award factors, such as the non-monetary factors ‘scarcity’ and ‘publicity’, were earlier studied by Neckermann and Frey (2013). Further, Ashraf et al. (2014) investigated the mechanisms behind the non-monetary factors (More information about these articles can be read in literature section 3.5). Although Saunderson (2004) showed that timeliness is an important factor, these studies did not study the non-monetary factor timeliness. Therefore this thesis will investigate whether timeliness is a supplementary award-factor in determining employee motivation.
This thesis will investigate the importance of the different non-monetary factors in the Dutch supermarket ‘Albert Heijn’. Although many Western firms do have non-monetary recognition
systems (Long and Shields, 2010), they are hardly studied yet. Huberman et al. (2004) show that motivational effects are culture dependent. Further, the motivational influence of awards is also influenced by the company-context (Frey and Neckermann, 2008). Therefore it is interesting to see whether the findings in the Dutch supermarket will differ from the studies of Neckermann & Frey (2013) in the IBM Switzerland and the study of Ashraf et al. (2014) in Zambia.
Albert Heijn is a suitable company to study the use of non-monetary incentives, because the supermarket is prohibited by law to use monetary incentives on top of the base salary. Because of the relatively simple, homogeneous tasks performed by the Albert Heijn employees1, it can be assumed that they have little intrinsic motivation (Kosfeld and Neckermann, 2011). Therefore, there is little risk that extrinsic rewards crowd out intrinsic motivation and so Albert Heijn should use extrinsic rewards to motivate people (Gubler et al., 2013). Thus for this company (and for other companies in this sector too) it is important to understand the best way in motivating people with non-monetary incentives. What makes it even more interesting, is that this Albert Heijn just introduced a new non-monetary motivational tool (see method section 5.2). This thesis can help Albert Heijn to implement it effectively. This is probably necessary, since often managers do not have
1
The employees investigated in this thesis are the cashiers and the people who fill the shelves. They have indeed relatively simple and homogeneous tasks. Therefore these employees often do not enjoy their work, which is a requirement for intrinsic motivation (Amabile, Hill, Hennessey, and Tighe, 1994 in Gubler et al., 2013).
enough knowledge to effectively implement recognition programmes (Saunderson, 2004). A requirement for recognition being effectively implemented, is that one understands the concept of recognition. Therefore we will first give an overview of the literature so far on the topic of (non-monetary) recognition.
3.0 Recognition in general
Recognition is important for employees. Employees who receive too little recognition do not feel motivated to do the best they can. Having motivated employees is important, since it leads to more productivity and thereby the profit of a company (Wiley, 1997). Not recognizing enough can even lead employees to quit their job. Furthermore, recognition shows which behavior is desired in an organization (Nelson, 1996). So it is strikingly to see that many companies do not have a formal recognition plan. Furthermore, the companies that do have a recognition programme, often do not use them effectively. Usually, managers do not have enough knowledge to implement effective recognition programmes (Saunderson, 2004). Although many companies do not see the importance of using recognition, in the literature the topic of recognition in general has already received a lot of attention.
Recognizing employees has its theoretical basis in the reinforcement theory of Skinner (Luthans and Stajkovic, 1999). In short it means ‘that you get what you reinforce’ (Luthans and Stajkovic, 1999: 52). So when you recognize employees for what they do, it is likely that they will repeat their action. But when you do not recognize, so punish or ignore, the behaviour, employees will not be motivated to exert the same effort again. For a long time in economics, only monetary reinforcement tools were studied. Although it was proven that monetary recognition works in some cases, money could not solve all motivational problems. Therefore, the social psychology field also began to study recognition as a motivational tool. This field focused more on non-monetary, social recognition. We now turn to the comparison of monetary and non-monetary recognition.
3.1 Monetary and non-monetary recognition
Monetary recognition has its advantages and disadvantages in comparison with non-monetary incentives. Sometimes, monetary incentives can motivate people to do their best. For example, in the article of Lazear (2000) the productivity of employees increased significantly with the aid of monetary incentives. However, in other circumstances non-monetary incentives can be more
effective. Luthans and Stajkovic (1999) state that it is all about reinforcing the behaviour and that it is not necessary to reinforce by means of monetary tools. Hansen et al. (2002) also state that
companies must not rely on monetary rewards only. They should focus on other recognition programs too. One benefit of non-monetary recognition is that it is flexible and therefore can be executed as often as necessary. This thesis will now give some other reasons why companies should (also) rely on non-monetary recognition programs.
One of the potential difficulties with extrinsic recognition is the ´crowding out effect´ (Gibbons, 1998), which means that people lose intrinsic motivation because they get extrinsically rewarded for something that they understood as being normal (Gubler et al., 2013). The crowding out effect can also occur with non-monetary incentives (Gubler et al., 2013). However, the chance that intrinsic motivation is crowded out is smaller with non-monetary recognition than with monetary rewards (Frey, 2007).
it is important that people feel valued by their social environment. Even employees who are almost solely driven by money will not perform the best they can with monetary compensation, because the money they will get is never enough for them (Nelson, 1996).
Another disadvantage of monetary incentives is that it can let employees focus mainly on the goal that is rewarded with cash, whereas other goals are also important (Kerr, 1975). Frey (2007) states that this is less of a risk with for example non-monetary awards. This is because the criteria to win the award are more vaguely specified.
Of course, non-monetary recognition also has (general) disadvantages. An example is ‘manager bias’: when the manager has to choose the winner of a non-monetary award, he could earlier choose his favourite employees. Further, when every time the same people win the non-monetary awards, it may lead to irritations and demotivation for the losers (Long and Shields, 2010).
However, because of these aforementioned difficulties with monetary recognition and because some firms are not able to give monetary incentives because of lack of funds or tax laws, companies also look for non-monetary alternatives (Long and Shields, 2010). For that reason we will now outline the motivational effects of non-monetary recognition in depth.
3.2 The motivational effects of non-monetary recognition 3.2.1 Reciprocity
One way in which non-monetary recognition can motivate employees is through reciprocity (Grant and Gino, 2010). In economics they continuously assumed that people always act out of self-interest. But it appeared that when people receive something nice in exchange, they will act in a more
altruistic way (Fehr and Gächter, 2000). This phenomenon is called reciprocity. It results from the fact that people attach value to their social worth (to being appreciated by others). So when they (expect to) receive recognition, which means being appreciated by others, they are more motivated to act in a pro-social way (Grant and Gino, 2010). Tidd and Lockard (1978) were one of the first to empirically support this concept of reciprocity. They showed that when the smile of waiters is bigger, they receive more tip in return. So the more social recognition someone conveys, the more one gets in return. Non-monetary incentives are an example of giving recognition; therefore employees could reciprocate to it.
Gneezy and List (2006) study gift exchange to explain the concept of reciprocity in the workplace. They state that somebody will work harder than normal when they receive something in return for it. Ellingsen and Johannesson (2008) bring in the concept of intention-based reciprocity. This means that people behave like the perceived intention of others. So if employees believe that the manager will give recognition, they are more motivated to exert more effort themselves.
Reciprocity is not the only way in which non-monetary recognition motivates employees. It also motivates by means of an increased identity.
3.2.2 Identity
Recognition strengthens the identity of employees (Wiley, 1997; Gubler et al., 2013). Saunderson (2004) states that recognition increases the morale of the employees and thereby employees feel more identified with their company. Sorauron (2000) says that non-monetary recognition leads to an increase in employee engagement with the firm and thereby it enhances motivation. It enhances motivation to behave in line with the companies’ goals and to exert more effort (Akerlof and Kranton, 2005). Pfeffer (2007) showed that companies often do not use their recognition systems
effective enough to achieve this increase in identity of employees. A survey in the United States revealed that a majority (67%) of the employees do not feel identified with their company and are not motivated to work hard for the business goals. This again shows the importance of a sound recognition system. A third way in which a sound recognition system can motivate employees is through status.
3.2.3 Status
Status can be defined as ‘one's standing in a social hierarchy as determined by respect, deference and social influence’ (Ridgeway and Walker 1995:281 in Huberman et al., 2004). According to Huberman et al. (2004) people think it is important to be valued and recognized by their social environment. Therefore status works motivational. Status can be seen as a means in getting power, but often status on itself also is an end goal for people. Further, Huberman et al. (2004) found that the importance of status differs for different cultures (a more detailed description of the findings of Huberman et al. (2004) can be read in section 3.5). Kosfeld and Neckermann (2011) also studied the motivational effect of status. They show that an entirely symbolic (non-monetary) award can lead to performance improvement because it conveys social recognition and thereby status. These entirely symbolic awards are the topic of this thesis. These awards were not studied for a long time.
However, recently there is some literature about awards and we will give an overview of this.
3.3 Awards
Although not much studied yet, the use of awards is quite popular. They are handed out in a lot of countries, levels and in different sectors like the media.A famous example for the media sector are the Oscars/Academy awards, honouring filmmakers in different movie categories (Oscars.org | Academy of Motion Picture Arts and Sciences, 2015). Companies are another place where awards are handed out (Frey and Neckermann, 2008). Companies often use awards to indicate what kind of exceptional performance is expected on top of the base tasks (Frey, 2007). According to Kosfeld and Neckermann (2011) we can define awards by mentioning three common characteristics. The first one is that they are based on criteria that are vague, it is not clear that a specific effort will lead to
winning the award. The second one is that they create competitiveness, because there are losers and winners. The third characteristic is visibility: that others know about the award and that there is some publicity for the winners. Frey and Neckermann (2008) added a fourth characteristic. Namely, that awards bring some kind of value with them, for example money or the winner’s name on a poster. Neckermann and Frey (2013) stated that the latter three characteristics determine the biggest part of the motivational impact of awards on employees. So these three characteristics were the public recognition (visibility) an award involves, the scarcity of an award (which creates the competitiveness) and the value of an award, either cash or in the form of a gift (section 3.5 for more information about the findings of this paper). Because this thesis studies non-monetary, symbolic awards, we do not investigate the factor ‘value’. However, we do investigate the factors ‘publicity’ and ‘scarcity’ and we add to this the factor ‘timeliness’. Timeliness is added because Saunderson (2004) put forward that this is also an important award-factor in determining employee motivation and that it needs more investigation. We will give an outline about the impact of the three different, non-monetary characteristics of awards. But first we give some general motivational effects of awards.
3.4 The motivational effect of awards
The general motivational effects of non-monetary recognition (see section 3.2) also hold true for awards. Awards are an example of recognition and therefore they lead to reciprocity. Also, receiving an award leads to an enhanced identification with the company, which results in more motivation (Frey and Neckermann, 2008). Further, Neckermann and Frey (2013) state that awards provide recognition, status and feedback resulting in more motivated employees. Gubler et al. (2013) state that employees are motivated by awards because of the status, self-esteem and social comparison that it brings along.
This social comparison-mechanism is a concept that several articles mention in association with awards. Due to an award there is relative ranking information, because there are winners and losers. This leads to social comparison (Lazear, 1989). Frey (2007) states that people always want to distinguish themselves from others, therefore possible social comparison should increase motivation. Lazear (1989) confirms that employees could be motivated to exert more effort due to social
comparison, because people attach value to their relative ranking. On the other hand, he states that the downside risk is that it results in an over competitive environment, leading to demoralisation and demotivation among employees (Lazear, 1989).
Further, Neckermann, Cueni and Frey (2014) state that employees are already motivated to work harder due to the award because they know that their performance on a specific task is measured. When employees do more because they know they are measured, this is known as the Hawthorne effect (Adair, 1984). Whereas this is a general effect, Ashraf et al. (2014) state that the motivational effects of the other mechanisms (like social comparison) differ in magnitude for different situations (see section 3.5). The motivational effect of symbolic awards is influenced by, among other things, the organizational culture and the other goals employees must achieve (Frey and Neckermann, 2008). However, the magnitude of the motivational effect of awards is for the most part influenced by the aforementioned (section 3.3) award factors: ‘publicity’, ‘scarcity’ and ‘timeliness’. These factors will be discussed now.
3.4.1 Publicity
Different articles agree that publicity has a positive effect on motivation. According to Jeffrey and Shaffer (2007) people (thus employees) want other people to know it when their performance is great. Knowing that other people will recognize an employee’s performance, leads to more motivation. This process is called social reinforcement. Through publicity more people will know about the employee’s effort and therefore publicity leads to more motivated employees. This social visibility that is associated with the award is valued by the employees because it enhances their status and self-esteem (Huberman et al., 2004; Luthans, 2000). So another property of publicity is that it enhances status (Moldovanu et al., 2007). The importance of status for people and thereby its effect on motivation is already mentioned in section 3.2.3.
In fact, the psychological effect of an extrinsic reward might be more important in
determining motivation than the material value of it (Dawson and Dawson, 1990 in Wiley, 1997). The motivation theory of Maslow explains why people might value publicity more than the material value. Publicity could improve relationship and self-esteem needs, which are higher needs for people than the safety needs fulfilled by material value. Because it fulfills higher needs, people are more motivated by publicity than by material value (Markham et al., 2002).
There are different degrees of publicity for employees. Being mentioned in a newsletter is sometimes already enough recognition for employees to be motivated (Luthans, 2000).
Saunderson’s survey (2004) revealed that employees value it when awards are handed out on well-planned ceremonies. Neckermann and Frey (2013) showed that people rather receive public recognition than that they receive recognition anonymously. Further, they showed that employees prefer a lot of publicity over just some public recognition, but this result was insignificant (more information about this article in section 3.5). So it seems to be the case that people strongly value publicity. However, Huberman et al. (2004) state that the desired amount of publicity differs for different cultures. Public recognition is not considered desirable in all cultures (see section 3.5). Now we proceed with the second characteristic: scarcity.
3.4.2 Scarcity/maximum number of recipients
Scarcity has contradictory effects on motivation. We start with the positive effect on motivation. On the one hand scarcity increases the value of the award. This is because a scarce award, which is more difficult to gain, signals more quality and thereby status to others (Frey, 2006). People do make social comparisons and therefore it works motivating for them when they can achieve more status. Scarcity determines how much status an award conveys. The more scarce an award is, the greater relative performance it indicates and the more status it gives to the winners (Besley and Ghatak, 2008; Moldovanu et al., 2007).
On the other hand, Jeffrey and Shaffer (2007) emphasize that scarcity gives the award less value, because it reduces the chance of winning the award. This is especially true for people who have a lower chance to earn the award. They would be more motivated by the award, if it was not that scarce (Jeffrey and Shaffer, 2007). This can be explained with the aid of reciprocity (see section 3.2.1). The more award-winners, the higher the chance that an employee will receive recognition and therefore he will be more motivated to reciprocate by exerting more effort (Grant and Gino, 2010). The opposite occurs when an award is scarce and is called negative reciprocity. Negative reciprocity means that employees reduce their effort when they do not feel rewarded enough. So when the maximum number of recipients is low, there is more risk for negative reciprocity (Gneezy and List, 2006). Now follows the last factor, which also has to do with scarcity: timeliness.
3.4.3 Timeliness
Herzberg (1966 in Luthans, 2000) states that it is important that award programmes take place often and consistently. Nelson (1996) also states that recognition has a more significant impact when it is given often and consistently. This is confirmed by a survey of Saunderson (2004). The survey of Saunderson showed that timely distribution of awards is of prime importance for employees. Further, Pragya (2008) underlines that it is important that awards are handed out at the right time. According to Saunderson (2004), the right time is as soon as possible after the exerted effort.
However, it is important that the award is not handed out too often, because than it loses its scarcity value (Hansen and Weisbrod, 1972). So while there seems to be accordance about the statement that timeliness has a significant effect on motivation, there is debate about how frequent awards should be handed out to achieve the highest effect on motivation. Saunderson (2004) states that because of this, the concept of timeliness should be further investigated to see how often awards should be handed out. Before we turn to the method section in which this is investigated, we first have a detailed look on the three most important papers about the motivational effect of awards. Thereafter the hypotheses are given for the different factors.
3.5 Most important papers in detail
The paper that recently investigated the effects on motivation of the different award factors is the one of Neckermann and Frey (2013). In addition to the monetary factors, they studied the non-monetary factors public recognition and scarcity. They studied these by executing a survey among employees of an IBM research laboratory in Switzerland. This survey made use of vignettes, which are descriptions of hypothetical situations in which the different award factors vary in their dimensions (more information about vignettes in section 5). The vignettes included a cooperation award. To win this award the employees needed to make great efforts to improve cooperation between the labs, for example through sharing findings between labs. On the bases of the
randomized vignettes the surveyed employees had to indicate how motivated they were to share a finding. Thereafter the effects of the different award factors were isolated by means of an OLS regression. For the non-monetary factors it appeared that the factor scarcity (the maximum number of recipients) has no significant effect on motivation. On the one hand the scarcity increases the value of the award. This is because a scarce award, which is more difficult to gain, signals more quality and thereby status to others (Frey, 2006). On the other hand, the scarcity of an award has a negative effect because it reduces the chance of winning the award. The authors assume that these contradictory effects resulted in no effect at all. The other non-monetary factor, the degree of publicity, did have a strongly significant impact on motivation. Publicity is important, since people value that other people know of their quality. So the more publicity, the more motivated employees are to win an award. A limitation of this study is that there is no possibility to compute the total profitability of introducing an award, because the motivational effects of awards change over time.
Recently, with their research in Zambia, Ashraf et al. (2014) elaborated further on the mechanisms through which the non-monetary characteristics of awards influence behaviour. These mechanisms are: social visibility, social comparison and employer recognition. They investigated the influence of the different mechanisms with a field experiment among health worker trainees. The trainees took courses during the year and had to make a test on various topics in the beginning of the course and at the end of the course. In this experiment, there were one control group and four treatment groups. Each treatment group contained one of the mechanisms of non-monetary awards (with two groups for social comparison/relative ranking). For example, in the ‘social visibility’
treatment group, an interview with photo of the award winner (a top-performer on the test) was published in a newsletter. The results show that the negative social comparison effect
counterbalances the positive effects of the other mechanisms: employer recognition and social visibility. One likely reason for the negative motivational effect of social comparison is that people fear that they do the best they can, but then still not win the award. Then people are disappointed because their optimal ability is relatively not enough and lower ranked than they believed. Because people want to avoid to lose their belief about their ability, they may produce less effort than they could. In that way their ranking will indicate less information about their ability, because they did not do the best they can. This is called ‘information avoidance’ and is often done by low ability people. This negative effect of social comparison was apparently stronger than the positive competition effect social comparison leads too. For the other mechanisms, the positive effect of social visibility is the result of enhanced status and self-esteem and the positive effect of employer recognition is mainly due to the, earlier explained, concept of reciprocation. A limitation of this study is that the results are difficult to generalize, since the effects of the different mechanisms are context-dependent.
the value attached to status differs among different cultures. The method they use is an experiment in which the participants play a game in the first stage. In the experimental group, the winners of the first stage are recognized with applause of the losers. Besides winning itself, this public recognition also conveys some status to the winner. The results showed that people are willing to overinvest in stage one, indicating that people are even willing to exchange money for status. In other words, publicity is important and therefore a motivational factor for people. Furthermore, the status in this experiment had no direct benefits as means for the future, implying that people value status as an end goal on itself. Further, Huberman et al. (2004) studied with the experiment whether the value of status is dependent upon one´s culture. They studied five cultures (Hong Kong, Germany,
Sweden/Finland, Turkey and the United States) and regressed their investments in stage one with status as dependent variable. They found that the value of status is highly correlated with the power index of Hofstede’s study. 2 This power index indicates the acceptance of a hierarchy in a culture, so the acceptance of status. Countries that scored higher on the power index, also invested more in stage one to receive the recognition/status. This indicates that that the value of status is dependent upon one’s culture. Resulting, the need for public recognition also varies across cultures. The need for public recognition is low for countries with a modest, feminine culture: the Scandinavian countries. Employees in those countries also have need for status, but they do not want to be recognized with applause on a ceremony. Rather they are recognized with little publicity. Since The Netherlands are in between these countries on the masculinity/femininity rankings list of Hofstede (Clearlycultural.com, n.d.), presumably the same applies to the employees in the Dutch food sector. A limitation of this study was that the sample size of Germany was too small to draw conclusions from it.
4. Hypotheses about the different non-monetary factors
Based on the above literature review, this thesis will now give the hypotheses for the different factors.
4.1 The degree of publicity
Based on the literature so far we can assume that publicity is really important for people. Employees are motivated by means of the status and self-esteem that publicity brings along (see section 3.2.3). However, based on the paper of Huberman et al. (2004) I expect that employees in the Dutch food are motivated by publicity, but to a certain extent. Dutch people dislike too much publicity, just as the Scandinavian people. So my hypotheses for the factor ‘degree of publicity’ are:
H0: The degree of publicity has no significant effect on the motivation of employees in the Dutch food sector.
H1: The degree of publicity has a significant negative linear effect on the motivation of employees in the Dutch food sector.
4.2 The scarcity/maximum number of recipients
The literature so far (see section 3.4.2) describes counterbalancing effects of the award-factor ‘scarcity’ on motivation. It increases the status conferred by winning the award, but it decreases the chance of winning (amongst others Frey, 2006). This is confirmed by the results of the study of Neckermann and Frey (2013) (see section 3.5). So I expect that the same will apply to the Dutch food
sector and that overall, scarcity will have no significant effect on motivation. This leads to the following hypotheses for the factor scarcity:
H0: The maximum number of recipients has no significant effect on motivation in the Dutch food sector.
H1: The maximum number of recipients has a significant effecton motivation in the Dutch food sector. The expected direction of this effect is unclear.
4.3 Timeliness
The literature agrees on the statement that it is important that awards should be handed out often and consistently (see section 3.4.3). However, there is some debate about how often awards should be handed out. In line with the papers of Saunderson (2004) and Nelson (1996) I expect that awards have more impact when they are handed out as soon as possible. Therefore, my hypotheses for the factor timeliness are:
H0: Timeliness does not have a significant effect on motivation in the Dutch food sector. In other words, it does not matter how often awards are handed out.
H1: Timeliness has a significant positive effect on motivation in the Dutch food sector. Awards have more effect on motivation when they are handed out more often.
5.0 Method
To test these hypotheses this thesis makes use of a survey experiment subjected to the employees. Saunderson (2004) showed that employees want to have more voice in designing an award
programme. Knowing that the company takes account of them increases their morale. Therefore, using the results of a survey that they filled in, will motivate the employees. The survey includes vignettes, on which we now will elaborate further.
5.1 Vignette method in general
Neckermann and Frey (2013) describe the concept of vignettes. Vignettes are short narratives. They describe a hypothetical situation. Apart from some general text, the described situations contain various factors. These different factors are the elements that are investigated for their effect on behaviour. These factors have different values. A respondent only sees one value of each factor in a vignette and these values vary in between the different vignettes. After reading the vignettes, respondents have to choose what their behaviour would be for a specific, hypothetical task. With the aid of their answers the effects of changes in the various factors on behaviour can be estimated. In addition to the effects of changes in the individual factors, also the effects of changes in
combinations of those factors can be estimated. These estimations can be done with an OLS regression. To clarify the method we will have a look at an easy example.
A clear example of the vignette method is given in section V of the paper of Falk and Kosfeld (2006). They present the participants/employees with scenarios that only differ in one
condition/factor: they are either trusted or controlled by their supervisor (these are the values of the factors here). After reading the scenario, participants have to indicate their work motivation on a scale. In this way it can be found in which condition the work motivation of employees is higher: in the ‘trust’ or the ‘control’ condition. This is an easy example of the vignette method. The use of the vignette method is explained by mentioning some of its advantages over the traditional survey approaches.
comparison with traditional survey methods. One of them is that vignettes are more reliable, since they are often similar to real-life situations and therefore less vague/abstract than the traditional questions. Because traditional surveys have simpler and abstract questions, people are more inclined to answer biased. Another advantage is that the vignette is a narrative and therefore respondents are often not aware on which basis they ground their answers (Alexander and Becker, 1978).
Contrary, the traditional survey approach asks subjective questions about the influence of a factor on behaviour. People often have a wrong idea about their own attitude/behaviour, therefore vignettes work better than traditional survey questions (Bertrand and Mullainathan, 2001). Further, the vignettes allow respondents to answer to combinations of variables, whereas traditional surveys only ask for the effect of individual factors on behaviour (Neckermann and Frey, 2013).
Although more reliable, the vignette method is of course not always completely sound. Sometimes the behaviour people choose in the survey after reading the vignettes does not match the behaviour that is actually carried out (Neckermann and Frey, 2013). This could be due to the concept of hypothetical bias, which means that people overestimate their behaviour in hypothetical situations (Bohm, 1972). However, all the aforementioned benefits led to the use of vignettes in this thesis. We will now have a look at how the vignettes are used in this thesis.
5.2 Our survey with an example of a vignette
This thesis investigates the motivational effects of the different non-monetary award factors in the Albert Heijn. Recently, the Albert Heijn started with ‘The wall of Fans’, a board with notes recognizing employees that showed maximum customer-friendliness. This is a formal, non-monetary award handed out to individuals. The managers of the Albert Heijn do not know yet how to use this non-monetary tool effectively, making it a good place to investigate. To make sure all respondents knew what ‘The wall of Fans’ stood for, the survey started by explaining what ‘The wall of Fans’ actually is. Further, it was explained how employees could be nominated.3 After this introduction part, the survey proceeded with the vignettes and the subsequent questions.
5.2.1 The vignettes
In the survey each respondent reads three vignettes and after reading each vignette he/she has to indicate his/her effort in a specific scenario.4 The vignettes consist of the different factors ‘publicity’, ‘scarcity’ and ‘timeliness’. I incorporated these three factors as binary variables in order to make it easier to draw
conclusions for the different factors. This was necessary due to time constraints. The possible values these binary variables could take on were discussed with the Albert Heijn managers and can be read
3
The employees have to nominate themselves, because the managers do not want to bother the customer with this. After the nominations, the manager decides who will receive the award.
4
I chose for three vignettes per person because it leads to more results and thereby more power for the test outcomes. I did not use more than three vignettes per person, because that would probably be too much text
Table 1
Award factors and their levels.
Factor Factor levels
Degree of publicity: (categorical)
Wall of Fans (little publicity), Wall of Fans with Facebook and ceremony (a lot of publicity). Maximum number of
recipients: (categorical)
8 (scarce), 40 (not scarce).
Timeliness : (categorical) Each month (frequent), each half a year (non-frequent).
from table 1. These three binary variables make up eight different vignettes, which were randomly presented to respondents. An example of a vignette is shown below. The different award factors with their (randomly displayed) values are marked bold. The factor in place and its value are indicated in between brackets (this information is not shown to the respondents, to influence their answers as little as possible).
Example of a vignette:
Each month (factor 3: timeliness: value: frequent) the current wall will be emptied and 8 new employees (about 4% of the employees) (factor 2: scarcity; value: scarce) will be put on ‘The wall of Fans’ and furthermore they will be mentioned during the New Year reception and in addition to this their names will be stated on the company’s Facebook page (factor 3: degree of publicity; value: a lot of publicity).
This is just one example of a vignette that could be randomly presented to the participants. Now is explained why the randomization option is used.
Benefits of randomization
Unlike the study of Neckermann & Frey (2013), this survey did not include a question about baseline motivation. This was unnecessary since the vignettes are randomized to the different respondents, and therefore it can be assumed that on average each vignette is seen by groups with the same baseline motivation. This imposes no restrictions on investigating the effect of changes in the different factors.
Further, when using a survey, there could be a risk for order effects (Bertrand and
Mullainathan, 2001). This means that the order in which survey items are displayed influences the way people answer. The vignettes in this survey are randomly displayed and in that way control for these order effects.
5.2.2 Indicating customer friendliness
After reading a vignette the respondents had to indicate their level of customer friendliness for a specific scenario in a multiple choice question. In this multiple choice question option ‘A’ was the least customer friendly-action and option ‘D’ was the most customer friendly-action. So the idea was to investigate which factor-levels drive (motivate) people too option ‘D’, that is ‘maximum customer friendliness’. The survey contained six different situations, three for each group of employees (the cashiers and the people who fill the shelves). Afterwards, the respondents had to indicate how realistic they perceived the different situations. All situations appeared to be realistic enough (see graph A1 in appendix A), so that it does not bias the results. Below is given one example of a
situation for the cashiers. The rest of the situations and their level of realism are given in appendix A. Example of a situation for the cashiers:
It is a busy day. A customer has only one item of a product on the till, while the product is for sale, namely: ‘two for the price of one’. What do you do?
(A) You say nothing about the sale action.
(B) You tell the customer about the sale action, but let him/her wait in line again.
(C) You tell the customer about the sale action, so that he/she himself/herself can get the second item. Subsequently you wait until the customer gets the item and let him/her checkout immediately. (D) You call a colleague of the relevant section to get the second item for the customer and you wait until your colleague arrives at the till.
So respondents indicate their amount of customer-friendliness by answering a multiple choice question. The advantage of this, in comparison with using a scale as answer option (like Neckermann and Frey (2013) did), is that the answers become more imaginary for the respondents. Answering on a scale is more abstract and therefore the answers are less reliable.
A potential disadvantage of using multiple choice questions is that respondents are directed towards an answer and maybe their effort is not among the answers (Finch, 1987). However, this is not a big threat, since the answer options are thoroughly discussed and often there is little space for other answer possibilities.
Another potential disadvantage of using multiple choice questions, is that there is a
possibility that respondents answer socially desirable (leading to upward bias), because they do not want to look bad with their answers (Bertrand and Mullainathan, 2001). Of course, the survey
emphasizes that participation is anonymous to discourage this. Furthermore, the upward bias is not a threat here because this thesis investigates the differences in exerted effort for awards with different factor-levels and these differences stay the same (Neckermann and Frey, 2013). The survey ended with some questions about personal characteristics of the respondents. These characteristics allow to check for representativeness of the data and could be used as control variables. The results will be shown in section 6, but first we explain the context of the survey.
5.3 Context and survey design
Before executing the survey, first its content had to be developed. I thoroughly discussed the
content with the managers and other people with previous experience on the workplace. This helped to make the situations and the values of the different vignettes as realistic as possible. After
composing the survey, I did a pilot study as from the end of December, 2013. With the help of the pilot study I fixed some flaws and made the survey as sound and user friendly as possible.
Then the survey study was activated and subsequently executed during two weeks in
January, 2014. Respondents of the survey are employees of a supermarket in the Netherlands, called Albert Heijn. The Albert Heijn wants their employees to show exceptional customer friendliness, because this can differentiate the supermarket from others. This is important, since the competition of other supermarkets increases lately. Therefore they introduced ‘The Wall of Fans’ (more
information in section 5.2). This is unique for the Albert Heijn, because normally it does not make use of awards (more information about the choice for the Albert Heijn in section 2). The respondents of this Albert Heijn were either cashiers or people who fill the shelves. I choose these two groups, because they represent the biggest part of the workforce (respectively 19% and 49% of the 205 employees).5
I had to invite these people for the survey, since participation was voluntary. To invite people for the survey I made use of social media.6 The survey was placed by means of a link on the Facebook pages of the relevant employee groups (the cashiers and the people who fill the shelves). In this way, I could efficiently reach many participants at the same time. Thereby I asked people personally (in real life or via social media) to fill in the survey. I could invite people via social media, because it was an online survey. I used the software website of ‘Qualtrics’ to establish the survey. Qualtrics offers the possibility to make participation anonymous. Further, Qualtrics has a randomization option for
5
questions. Therefore I combined each question/situation with the eight different vignettes and then for each situation/question one of those combinations was randomly presented to a respondent. The situations were seen by the right employee group with the aid of the Branch option in Qualtrics. This Branch option makes it possible to lead different groups of respondents to different questions. One of the first questions was about the function of the respondent in the Albert Heijn and according to this answer the respondent would see the right situations (that is the cashiers’ situations for the cashiers). So due to the Branch option it was not necessary to make a survey for each group of employees. We will now turn to the results of the study, where we will see the influence of the different non-monetary award factors on motivation. But we start the results section with the characteristics of the respondents.
6.0 Results
6.1 Characteristics Table 2
Characteristics for the different functions Number of
women (men)
Median age.
Average number of years working for the Albert Heijn
Number of employees that work part-time (full-time)
People who fill the shelves
12 (27) 20 4.9 37 (2)
Cashiers 20 (0) 19.5 6.94 17 (3)
Total 32 (27) 20 5.69 54 (5)
In this survey there were 63 respondents who generated a total of 184 observations.7 Of these respondents, 22 were a cashier and 41 were people who fill the shelves (so a bigger fraction of cashiers than in the total workforce). Of these 63 respondents, 59 filled in the characteristics. The requested characteristics were ‘gender’, ‘age’, ‘part-time/full-time’ and ‘years worked for Albert Heijn so far’. These characteristics of the respondents can be read from table 2 for both of the employee groups. The characteristics make it possible to check for the representativeness of the employees.
The respondents are representative for the company and the two functions with respect to gender. The fact that there are no men among the cashier-respondents is not surprising, since there are only two of them among the total cashier group (so inclusive the non-respondents). For the characteristic age, the median is used instead of the mean in table 2, because the Albert Heijn puts everyone above 24 in the same group in its data records. So calculating a mean age is impossible, since we have no data about the exact age of people in the 24+ group. Therefore, using the median seems the best solution. Table B1 in appendix B gives a better overview of the number of employees per age group for each function. From this table it appears that with respect to age the respondents are representative at the function levels and also at the company level. Only age ‘17’ is
underrepresented, but that is not a real threat since the surrounding ages are well represented. Further, I cannot check for the representativeness of the characteristics ‘part-time/full-time’ and
7
So there are five observations less than the 189 observations expected with 3 situations per person, because a few respondents closed the survey earlier. Still we include them in the analysis, since they yield some observations.
‘years worked for the Albert Heijn so far’. Although they are important employee characteristics, Albert Heijn does not track the records of the characteristics. Anyway, these personal variables can be used as control variables for the regression. The personal characteristics were also used to check whether the randomization of the vignettes was successful.
6.2 Check for randomization
The successfulness of randomization is checked by performing t-tests on the means of the five different personal characteristics (inclusive function) for the binary-values of the different vignette factors.8 When the mean values were significantly different, it means that randomization was not completely successful. The results of the t-tests can be read from table C1 in appendix C. It appears that the factor scarcity is successfully randomized. Next, the factor timeliness is also quite
successfully randomized, since only the personal characteristic ‘gender’ has significant mean differences for the binary values of ‘monthly’. Contrary, for ‘ceremony’ (publicity), only the characteristic ‘function’ is successfully randomized. So there is unsuccessful randomization for the other four personal characteristics. The factors ‘Years working for the Albert Heijn’ and ‘part-time/full-time’ even have significant mean differences at the 1% level for ‘ceremony’.
Unsuccessful randomization is not necessarily the result of a bad randomization design, in this case it probably results from the small sample size. However, unsuccessful randomization has consequences for the estimated coefficients of the vignette factor in place. These coefficients change after adding the personal variable for which the vignette factors are unsuccessfully randomized, to the regression as control variable (see section 6.4). So the estimated coefficient of the factor ceremony will probably change substantially after adding the personal characteristics as control variables in the regression (see section 6.4). Before we turn to this regression, we will first give a quick overview of the results per situation and per vignette.
6.3 Quick overview of results
Before we state anything about the results of the answers, it is important that one understands how the answers are coded. In that way the meaning of the numbers mentioned hereafter can be understood. The numbers can be understood as following: 1= answer A, 2= answer B, 3= answer C and 4= answer D. Where answer A is the least customer friendly effort and answer D is the most customer friendly action. Now we know this, we can have a look at the results.
First something about the fraction each answer represents on the total of observations. This histogram can be seen in Graph D1 in appendix D. From this graph it can be seen that answer D (maximum customer friendliness) is the most chosen answer, representing 40,217% of the total of observations. When almost everyone chooses D, there is a high risk of social desirability. Further, there is no possibility to investigate the effects of the different award characteristics anymore, since there is no variance in the answers. However, it is not the case that almost everybody chose answer D, the other answers also represent a sizeable fraction. In fact, the more customer friendly is the answer, the more it is chosen. So answer A is the least chosen and answer D the most. This could suggest that the different factors have a positive effect on the motivation to be customer-friendly. However, this graph is not enough to draw any conclusions. Therefore we need regressions (see
8
section 6.4). But before, we will have a look at the average answer per vignette and per situation. Like section 5.2.2 stated, there are 6 different situations. Graph E1 in appendix E shows the mean observation for each situation. It appears that for the people who fill the shelves, the different situations have a similar mean. However, for the cashiers, situation 3 has a mean of almost 4
(maximum customer friendliness). This possibly indicates that the intrinsic motivation for this situation is really high.9 One could argue that this generates some bias for the estimations because people were already motivated for the task. However, the situation is still incorporated in the regression, because at least it does indicate that there is no ‘crowding out effect’ here. Furthermore, when we left out cashiers’ situation 3 of the regression, the estimated coefficients of the investigated award did not change significantly. Now we will have a look at the average observation/outcome for each of the eight vignettes, given in table 3 below.
When looking to the different vignettes there are no real outliers with respect to the mean of observations they have. However, there are three vignettes that score quite high. The first one is the one in which the award is handed out monthly (frequent), to 8 recipients (scarce), with even a ceremony (a lot of publicity). The second vignette which generates a high average is the same, except that there are 40 recipients (non-scarce) instead of 8. Thus according to this it seems important that the award is handed out frequently and that there is much publicity, but that the scarcity has no influence on motivation. However, in the vignette with the highest average, the award is handed out each half a year (non-frequent), to 40 recipients (non-scarce), with only a wall of Fans (little
publicity). So this is a bit contradictory to the other two vignettes that score high in terms of the factor-levels. Further, the standard errors are quite big, so it is likely that there is no one specific outstanding combination for the different vignette factors. Therefore we now turn to the OLS regressions to better understand which values of the vignette factors have the highest effect on motivation of employees. Here, we will get an answer on the question whether timeliness is a supplementary factor in determining employee motivation.
9
People probably see it as normal to help a disabled person (situation 3) and do not see it as exceptionally customer friendly.
Table 3
The mean of the observations per vignette
Different Vignettes M8W M40W M8C M40C H8C H40C H8W H40W Total Mean of observations 3.04 3.211 3.227 2.84 2.88 2.963 2.842 3.318 3.033 Standard deviation 0.735 1.032 0.869 0.943 1.092 1.224 1.015 0.945 0.991
Notes: The different vignettes are abbreviated with a letter/number for the value of each factor. First the factor ‘timeliness’: M= each Month (frequent), H= each Half a year (non-frequent). Second the factor ‘scarcity’: 8= 8 recipients (scarce), 40= 40 recipients (not scarce)
And third the factor ‘degree of publicity’: W= The wall of Fans only (little publicity), C= The wall of Fans with Facebook and ceremony (a lot of publicity).
So for example M8W stands for the following vignette:
‘’Each month (factor 3: timeliness: value: frequent) the current wall will be emptied and 8 new employees (about 4% of the employees) (factor 2: scarcity; value: scarce) will be put on the wall of Fans (factor 3: degree of publicity; value: little publicity).’’
-The means of observations can be understood as following: 1= answer A, 2= answer B, 3= answer C and 4= answer D.
6.4 Regressions
For the regressions we used OLS. Since the dependent variable (motivation to be customer friendly) is a categorical variable, an ordered probit regression seems to fit better. However, OLS makes it easier to interpret the size of the different variables. Therefore we discuss the OLS model in this result section (a comparison of the ordered probit regression with the OLS regression can be found in table G1 in appendix G). The estimations for the coefficients of the different factors can be seen from table 4 on the next page. We start by discussing model 1, which shows the regression with just the main award factors.
Model 1
Model 1 is a simple specification that measures the effect of award factors on the dependent
variable customer friendliness, without the inclusion of any control variables. For the factor publicity, the coefficient of the factor ‘ceremony’ implies that a higher degree of publicity has a negative, but insignificant effect on the motivation to be customer friendly. When there is a lot of publicity inclusive a ceremony, the motivation to be customer friendly is 0.137 points less than when the publicity consists of just the wall of Fans. So although not significant, the result is supportive of our hypothesis for this factor in section 4. Furthermore, the coefficient is of a reasonable size if we compare it to the standard deviation of the dependent variable ‘customer friendliness’, which is equal to 0.991 (see table 3). Thus this model let us conclude that employees in the Dutch food sector prefer little public recognition over a lot of public recognition. This is probably due to cultural
influences. Dutch people also like status, but do not like to receive too much attention (see section 3.5).
The amount of recipients/scarcity does not have a significant influence on the willingness to be customer friendly. This is according to the hypothesis and it is probably due to the countervailing effect discussed in section 3.4.2. A scarce award increases motivation because it demonstrates more quality and thereby status to people. On the other hand, a scarce award decreases motivation, because it reduces the chance to win the award. Here, this latter effect is somewhat stronger,
because the variable ‘8 recipients’ (scarce) decreases the motivation to be customer friendly by 0.073 point in comparison with a situation with 40 recipients.
Timeliness does not have (the expected) significant influence on the motivation to be
customer friendly. When the award is handed out more frequent (each month), the motivation to be customer friendly increases by 0.063 point in comparison with a situation in which the award is only handed out each half a year. So this is only slightly supportive of our hypothesis in section 4, but not significant at all. This could be due to the scarcity effect. When an award is handed out more often, it decreases the value of an award. So this counter-effect could account for the fact that the expected ‘the sooner, the better’ does not hold significantly here.
The constant shows the baseline motivation, so the award that is handed out each half a year, to 40 recipients and with only a wall of Fans. The combination of a high constant and the low coefficients of the factors imply that there is a high initial motivation to be customer friendly.
The adjusted R2 shows which part of the variance in the dependent variable (customer friendliness) is explained by the independent variables (the different factors) in the regression. The very low adjusted R2 does not necessarily imply that the regression is bad, however it does imply that there are other important factors that influence the motivation to be customer friendly.10 We will
now see whether the personal variables are among these important factors. First we will have a look at model 2 in which the different situations are added to the regression as control variables.
Table 4
Effect of award factors on customer friendliness
Model 1 Model 2 Model 3
Monthlya 0.063 (0.147) 0.013 (0.138) -0.083 (0.153) 8 recipientsb -0.073 (0.147) -0.024 (0.137) -0.012 (0.153) Ceremonyc -0.137 (0.148) 0.003 (0.139) 0.057 (0.158) Cashier situation 2d -0.974*** (0.283) -0.803*** (0.296) Cashier situation 3d 0.357 (0.283) 0.380 (0.307) Shelves situation 1d -0.764*** (0.246) -0.672** (0.306) Shelves situation 2d -0.521** (0.245) -0.490 (0.303) Shelves situation 3d -0.751*** (0.247) -0.794** (0.304) Womane -0.044 (0.221) Age 0.024 (0.042)
Years of service so far 0.027
(0.022) Parttimef 0.209 (0.433) Constant 3.111 (0.151) 3.551 (0.236) 2.734 (0.997) Number of observations 184 184 184 Adjusted R2 -0.010 0.135 0.134
Notes: Estimated coefficients from OLS regressions. The standard errors are between brackets. The dependent variable is the motivation to be customer friendly. This motivation was indicated by answering multiple choice questions and therefore could take the values 1 (answer A=least customer friendly), 2 (answer B), 3 (answer C) and 4 (answer D=most customer friendly). Because we did an OLS regression, it is incorporated as a continuous variable. We also performed a probit regression in which it is handled as a categorical variable (see appendix G).
a
Reference category: Half a year d Reference category: Cashier situation 1
b
Reference category: 40 recipients e Reference category: Man
c
Reference category: Wall of Fans only f Reference category: Fulltime * = p<0.1 **=p<0.05 ***=p<0.01
Model 2
It appears that adding the situations as control variables has a strong effect on the coefficients of the 3 main variables. For scarcity (variable: ‘8 recipients’) and timeliness (variable: ‘monthly’) we find the same effects as in model 1, only with an even smaller coefficient. The coefficient for ‘degree of publicity’ (variable: ‘ceremony’) does now have a negligible (insignificant) effect on motivation. This could imply that employees in the Dutch food sector do like public recognition, but do not care about what the magnitude of the public recognition is. It could also imply that they do not like public recognition at all, no matter how much it is.11 Next to the control function of the situation-variables, model 2 shows that these variables also have an effect on the stated willingness to be customer friendly. In comparison with the omitted category ‘cashier situation 1’, almost all situations have a significantly different effect on motivation. First, we discuss the cashier situations (for a description of all the situations, see appendix A).
What appears is that cashiers indicate a significantly (0.974 point) lower baseline motivation in situation 2 than in the omitted category (situation 1). This could be due to the fact that situation 2 is seen as less realistic than the other situations of the cashiers (see appendix A) and therefore the respondents could identify worse with being very customer friendly in that case. The estimated coefficient of ‘cashier situation 3’ is not significant, because ‘cashier situation 1’ (used as reference category) also has a high mean of observations for customer friendliness (see graph E1). Because situation 3 ensures a really high baseline/intrinsic motivation for the cashiers, it decreases the effect size of the different factors. So this model may give a somewhat more realistic view of the effects of the different factors on motivation to be customer friendly.
Further, model 2 shows that the people who fill the shelves have a significantly lower baseline motivation in all their situations than the cashiers in situation 1 and 3. This could either be because the situations invite less to demonstrate customer friendliness, or because the people who fill the shelves are less motivated to be customer friendly than the cashiers irrespective of the situation. Future research should address this question.
Because of these significant effects of the situations on motivation, the adjusted R2 increases substantially to 0.135. This indicates that the situations give the model more predictive power than model 1 and that the coefficients on the main factors are more reliable. Now we have looked at the employees’ function, we will turn to the other personal variables in model 3.
Model 3
Model 3 shows that there is still omitted variable bias in model 2, since the estimated coefficients of the main award factors change again. Because the personal variables are added as control variables, the coefficients of the award factors are more precisely estimated. This more precise estimations yield some unexpected, but still insignificant results. The factors degree of publicity and timeliness do have estimated coefficients that are contrary to the coefficients in model 1 and thereby to our hypotheses.
The coefficient on ‘ceremony’ implies that a lot of publicity has 0.057 points more effect on motivation than the wall of Fans only. This implies that Dutch people prefer a lot of publicity over just little publicity. So they are not like Scandinavians, but more as the other cultures in the article of
