READINESS FOR CHANGE IN THE HEALTHCARE SECTOR AND THE MEDIATING EFFECT OF PSYCHOLOGICAL SAFETY

THE INFLUENCE OF SOCIAL POWER ON

READINESS FOR CHANGE IN THE HEALTHCARE SECTOR AND THE MEDIATING EFFECT OF PSYCHOLOGICAL SAFETY

Master thesis, MscBA, specialization Change Management University of Groningen, Faculty of Management and Organization

14th May, 2011

AIMÉE HELEEN DEKKER Studentnumber: 1455044

Nieuweweg 20-3 9711 TE Groningen tel.: +31 (0)6 52004605

e-mail: a.h.dekker@student.rug.nl/aimeeheleendekker@gmail.com

Supervisor / university J. Rupert / C. Reezigt

Supervisor / field of study

R. Bakker / W. van der Kam

(Antonius Ziekenhuis, Sneek)

ABSTRACT

In this quantitative study, we focused on the relationship between social power and readiness for change, with psychological safety as a mediator in that model. The research was performed with 44 medical specialists, all working in a general hospital in the Netherlands. With the use of questionnaires, it was found that legitimate power is the only form of power that has a marginally positive relationship with readiness for change.

Psychological safety has not shown to be a mediator. Psychological safety appeared to have a very strong direct positive link with readiness for change, and this concept is therefore important to take into account when deciding upon a strategy to make medical specialists more ready for change.

Keywords: readiness for change, healthcare sector, social power, legitimate power,

psychological safety,

TABLE OF CONTENT

INTRODUCTION ...4

Management problem ...4

Research objectives ...4

THEORY AND HYPOTHESES DEVELOPMENT ...6

Organizational change...6

Readiness for change ...7

Social power ...9

Coercive power and readiness for change ...9

Reward power and readiness for change ... 11

Legitimate power and readiness for change ... 13

Referent power and readiness for change ... 15

Expert power and readiness for change ... 16

Psychological safety ... 18

Conceptual model ... 24

METHODOLOGY ... 24

Data collection and analysis ... 24

Measures ... 25

Readiness for change ...26

Social power ...26

Psychological safety ...26

Control variables ...26

Factor analysis ... 27

Readiness for change at the Antonius Hospital ... 29

Correlation and regression ... 30

DISCUSSION AND LIMITATIONS ... 33

Discussion ... 33

Limitations and directions for future research ... 36

CONCLUSION ... 37

REFERENCES ... 38

APPENDIX A – QUESTIONNAIRE ... 44

INTRODUCTION Management problem

This research is performed in the interest of the Antonius Hospital in Sneek, the Netherlands. The organization under study is a medium-sized general hospital; it is the second largest in the province, and therefore, has an important regional function. The target areas for this hospital are: a large part of south-western Friesland, Urk, the Noordoostpolder and the islands of Vlieland and Terschelling.

In today’s environment, the healthcare sector is under a lot of pressure to stay competitive, costs have to be cut while at the same time the quality of care has to stay equal and preferably even improve (Woodward et al., 1999). In order to stay competitive, the hospital decided to embark on a project called ‘Antonius aan de Wind’. The board of directors found that a new, more decentralized structure was needed because the old structure was seen as too bureaucratic and not focused on results. The starting points for the new structure were amongst others: be more result-oriented, involve the medical staff in decision-making, and create an organizational structure based on the process patients go through. In April 2009 the project officially started and now, approximately two years later, the question emerges if people participate in the project and how willing they are to embrace the changes.

Research objectives

This research has two objectives, a practical one and a theoretical one. The practical purpose of this research is to find out how willing the employees of the Antonius are to participate in the change. Consequently, with the results of this study, advice can be given to the managers of the hospital how to create and optimize readiness for change amongst their medical specialists.

The theoretical objective of this study is finding out how social power as a tool of

persuasion can be used in the healthcare sector. It would be interesting for management

professionals in this sector to see what the effect of power is on an individual’s readiness

for change, since high readiness for change can greatly improve the chance of a

successful change initiative (Armenakis, Harris & Mossholder, 1993).

Bass (1985) found that the use of power is related to the successful realization of a change event. In addition to that, Bradshaw and Boonstra (2004) say that the role of power within the context of organizational change is found to be important. However, within the existing literature on social power, inconsistencies have been found. Some authors claim a positive relationship between social power and readiness for change, while others found negative relationships between these concepts. As a reason for these inconsistencies, Rahim (1989) stated that these could be (at least partly) explained by some measurement failures, such as the use of single-item instruments to measure a form of power, the use of ordinal ranking instead of interval ranking, or the lack of strong validity. This makes this research even more interesting, as valid results from this research could increase the theoretical knowledge there is about social power and its relationship with readiness for change.

In addition to social power, psychological safety is measured. Edmondson and Woolley (2003) found a relationship between the psychological safety one feels and how one feels about a change. If one feels psychologically safe, a change is more often perceived as useful, and the employees will accept the change sooner. Even stronger, Schein (1993) found that psychological safety is needed for individuals in order to make them feel safe and skilled enough to change. This implies that this term is very relevant and interesting within the theory of organizational change: a positive relationship between psychological safety and readiness for change exists. In this research we will try to define this relationship even more by investigating if psychological safety functions as a mediating variable on the relationship between social power and readiness for change.

The objectives of this research lead to the following research questions:

1. To what extent are the employees of the Antonius willing to change?

2. How does social power influence readiness for change?

3. Does psychological safety have a mediating effect on the relationship between

social power and readiness for change?

THEORY AND HYPOTHESES DEVELOPMENT

In this section the theory supporting the research questions will be mentioned and described. First of all, we will look at organizational change followed by the readiness for change. Later on, social power will be defined. In that section hypotheses will be formed for the relationship between social power and readiness for change. After that, the concept of psychological safety is explained, followed by hypotheses for the mediating effect of psychological safety on the relationship between social power and readiness for change. This section will end with the conceptual model.

Organizational change

Organizational change is defined by Metselaar (1997:5) as “the planned modification of an organization’s structure or work and administrative processes, initiated by the organization’s top management, and which is aimed at improving the organization’s functioning”. Change is an ever-present feature in all types of organizations (Burnes, 2004b). Being able to change is of vital importance for the survival of companies today, since the pace of change is faster than ever before (Moran &

Brightman, 2001). However, attaining a successful change is not easy; high failure rates (sometimes even higher than 80%) are mentioned by different authors (Burnes, 2005).

Given this high failure rate, it is of great importance to learn more about the factors contributing to a successful change.

Within the literature of organizational change, two main approaches can be found:

planned change and emergent change (Bamford & Forrester, 2003). Planned change,

based on the work of Lewin (1951), has been the main theory for several decades, and it

views change as “a process that moves from one “fixed state” to another through a series

of pre-planned steps” (Bamford & Forrester, 2003:547). Emergent change on the other

hand is a relatively new concept; it came into being as a critique on the planned change

approach around the 1980s (Burnes, 2004a). Emergent change is seen as “ongoing

accommodations, adaptations, and alterations that produce fundamental change without a

priori intentions to do so” (Weick, 2007:237). In this case, the board of directors of the

Antonius Hospital decided to actively make a change, making the project a planned

change.

Planned organizational change is often described as a process of three stages; a first stage in which the old situation is reconsidered, a next stage in which the change is implemented and a final stage in which the change is institutionalized (Armenakis &

Harris, 2002). According to several authors (e.g. Armenakis, et al., 1993; Bernerth, 2004) the first stage of change is similar to the concept of readiness for change. Smith (2005) says that the risk of change failure is high when there is not enough individual and organizational readiness for change. Readiness for change therefore is of paramount importance for the survival of companies today (Moran & Brightman, 2001). Readiness for change is one of the most important factors contributing to a successful change.

Therefore, in this study we will focus on this concept.

Readiness for change

Despite the evidence that the first phase of change is important, a clear conceptual definition is not present in literature yet. Armenakis et al. (1993:681) mention readiness for change, which they define as the “[..] organizational members’ beliefs, attitudes, and intentions regarding the extent to which changes are needed and the organization’s capacity to successfully make those changes”. Metselaar (1997:34) mentions another concept: willingness to change. Willingness to change is “a positive behavioural intention towards the implementation of modifications in an organization's structure, or work and administrative processes, resulting in efforts from the organization member's side to support or enhance the change process”. Given the similarities between both definitions, we clearly see that the same concept is mentioned. In this research we will use the term

‘readiness for change’.

Readiness for change exists on two different levels; on the organizational level and the individual level (Smith, 2005). Organizational readiness for change cannot exist without individual readiness for change (Madsen, Miller & John 2005). Since individual readiness for change is needed in order to create organizational readiness for change, learning more about individual readiness for change is important for business professionals in order to create more successful change events (Cunningham et al., 2002).

Although readiness for change is an important construct in attaining successful

change, only a couple of authors have investigated factors that influence readiness for

change (e.g. Holt, Armenakis, Feild & Harris, 2007; Madsen, John & Miller, 2006;

Madsen et al., 2005). Knowing about and knowing how to influence readiness for change can help management professionals attain a successful organizational change. Therefore, this concept will be the dependent variable in this study.

Power, as was said in the introduction, is found to be important and related to successful organizational change (Bass, 1985; Bradshaw & Boonstra, 2004).

Nevertheless, this has not always been the case. Before the 1970s, in the early periods of organizational change, power was disregarded as being important. However, around 1970, a more critical view on change became apparent and the importance of change became visible.

In the healthcare sector, the view on power and control is also important.

Professionals, like medical specialists, function as a so-called clan (Ouchi, 1979). In a clan, strict rules for control are not present. Power and control are partly based on rules, but also on common agreements. Clearly, the quality of the work has to be high, but the medical specialist has a certain degree of freedom as to how he achieves this. However, the management functions different: they typically work based on formal rules, they have to account for all their actions (Witman, 2007). Immediately, we see the contradiction between the medical specialists and the managers: medical specialists prefer informal rules and agreements, while managers prefer standard, formal rules. Another related issue to this is the fact that control, by colleagues or a manager, is seen by medical professionals as an attack on their expertise. Medical specialists have the idea that because of their extensive training, they have to be flawless (Witman, 2007). One could think these two issues are unrelated to the organization of the hospital since it is the manager’s task to manage and the medical specialist’s task to ensure quality of care.

However, each of those two seemingly separate tasks may have an influence on the other.

The managers and the medical professionals have to cooperate with each other in order to reach the basic goal of the hospital: good, qualitative care within the financial boundaries.

Because of these two different tasks, that have to work together, struggles exist (Witman,

2007).

Because of the importance of power in relation to organizational change, and because of the intriguing role of power within the hospital we will investigate power more in-depth in this study; power is the dependent variable of this study. With this study we hope to find how power can be used to the benefit of the organization as a whole.

Social power

Power is defined by many authors, all with different variations. As Bierstedt (1950:730) puts it: “In the entire lexicon of sociological concepts none is more troublesome than the concept of power”. In its most basic sense, it could be described as the capability to pursue and achieve goals through getting a grip of one’s environment (Mann, 1986:6). Social power is, according to numerous authors (e.g. Busch & Wilson, 1976; Erchul & Raven, 1997), the ability of a person or group to influence a change in someone in an intended direction. An important issue that needs to be taken into account here is that, in the case of social power, the environment that needs to be changed is limited to persons.

According to Mintzberg (1983) the best way to study social power is by means of the framework designed by French and Raven (1959). This framework consists of five forms of social power: coercive power, expert power, legitimate power, referent power and reward power. In the next sections all five forms of power will be introduced and explained.

Coercive power and readiness for change

Coercive power is a form of power that is based on the perception of person B (the person on who power is exercised) that person A (the person who uses the power) can bring about negative consequences for him or her. Examples of negative consequences could be that A can fire B, or that B will not receive an extra bonus (Raven

& French, 1958b). Putting it in other words, coercive power is when one person forces another person to do something, with the possibility of using punishments for non- compliance.

An important characteristic of coercive power is the fact that person B must have

the idea that person A can observe B: surveillance is important (Shetty, 1978). When

coercive power is used to obtain an organizational change, person B complies with the change because he wants to avoid a negative consequence, not because he believes in the change: the behavior is not institutionalized in the values and beliefs of person B.

Because of that, without surveillance, the old behavior is taken back on again as soon as the control is gone. Because institutionalizing a change is seen as one of the crucial steps in achieving a successful change (Kotter, 1998), it is likely that using coercive power will not lead to more readiness for change.

The use of coercive power in the literature quite often has negative connotations.

Haffer (1986) for example, found that using coercive power to create a change is very likely to result in strong resistance. However, the fact that resistance (which is often referred to as the opposite of readiness, e.g. Armenakis & Harris, 2002) often occurs when using coercive power, does not mean that it is not often used. Shackleton (1995) says that despite the negative feelings towards the concept of coercive power, it is repeatedly used to ensure daily issues, such as meeting deadlines. This gives us the basic idea that coercive power might work for short term projects. This idea is supported by Erchul and Raven (1997). They found that the effects of coercive power are usually only gained on the short term because the use of this form of power calls for monitoring and follow-up. If the behavior of person B is not monitored, person B will not comply with the order of person A. If person B is monitored, but does not comply, person A has to follow up on what he said he would do. This results, for instance, in person B being fired or losing his chance on a promotion. As soon as person A does this, there is no change that person B will, in a later stage, comply with the wishes of person A.

Earlier research on the topic of coercive power and readiness for change shows

that coercive power is mostly seen as ineffective. Podsakoff and Schriesheim (1985) for

example, in their research on literature written on social power, found that the use of

coercive power was found to be the weakest reason (of all five forces of power) why

people would comply with supervisor’s requests. In line with that, Lidman-Adizes, Raven

and Fontaine (1978) found in their study that compliance was least likely to happen when

coercive power was used. This means that the hypothesis concerning coercive power and

readiness for change is that there will be a negative relationship between both concepts.

Within the context of the healthcare sector, another argument can be put forward for a negative relationship between coercive power and readiness for change. When managers decide to use coercive power to make medical professionals participate in a change event, this basically means that he forces the medical professional to participate in the change. When medical professionals are forced to make a change, it is very likely that this results in strong resistance against that change because research has shown that medical professionals do not feel that they need external steering and control (Witman, 2007). Control over their behavior, and external steering over their actions could make them unwilling to participate in the change. The tool used by the manager will be counterproductive: instead of participating; the medical professional might even be more unwilling to participate than before. Clark (2005:25) agrees to this, he says that ‘a coercive style often prompts a paradoxical response’.

Hypothesis one, the idea that coercive power style has a negative influence on the readiness for change, is based on what we have seen in this section.

H1a: Coercive power has a negative relationship with an individual’s readiness for change

Reward power and readiness for change

Reward power is the second form of power in the framework described by French and Raven (1958). The idea behind this form of power is simply that person A has the ability to reward person B for the change to be made, for example with a salary increase or with a promotion (Braynion, 2004; Frost & Stahelski, 1978).

Reward power is often seen as similar to coercive power, with one very important distinction. With reward power, person B should become more attracted to person A (because he receives something he wants to have), and therefore person B becomes more dependent on person A. In contrast, with coercive power, the attraction between person A and person B diminishes (because he loses something), resulting in a situation where person B tries to avoid future encounters with person A (Swasy, 1979).

A first similarity between reward power and coercive power is the fact that both

concepts require surveillance. In the case of reward power, the new behavior is shown in

order to acquire the reward that was promised by person A. in order for that to happen, person A must see that person B shows the new behavior. If surveillance is left out, person B might not behave correctly.

Another similarity between coercive power and reward power is found in Hersey et al. (2001). They mention that the effect of using reward power might not last long;

once the reward is given, person A will no longer have influence over person B. This gives us the impression that reward power might not be the best form of power to use when you want to implement a long-term change, because the change is not institutionalized within the norms and values of person B. However, for the short term that the reward was not given yet, reward power might be successful.

Reward power is found by different authors to be one of the least popular forms of power to reach a goal because people do not want others to think that their readiness can be bought (Handy, 1993; Podsakoff & Schriesheim, 1985; Yukl & Falbe, 1991). It is therefore often said to be unsuccessful in achieving a successful change (e.g Bachman, Smith & Slesinger, 1966; Brown, Lusch & Nicholson, 1995; Rahim, 1989). In addition to that, the use of reward power is restricted. A first limitation was already mentioned: the fact that it usually only works on the short term. Other limitations can be: rewards are not inexhaustive: the company usually only has a certain amount of reward to give. The question is if the amount of rewards is sufficient to make sure that everyone can receive a reward. Another question is if the rewards that are available are valued enough by the employees to make it a helpful tool (Shetty, 1978).

However, reward power has, on occasion, also been found to bring about positive

results, for example by Shetty (1978) and Kearney, Plax, Richmond and McCroskey

(1985). Shetty (1978) found that reward power leads to a positive performance. However,

performance is not the same as readiness for change. Performance is a part of readiness

for change, but we miss the institutionalization of the new values in the system of the

employees. Kearney et al. (1985) found that the use of reward power leads to compliance

with the will of the superior. What makes the study done by Kearney et al. (1985) less

strong than other studies is the fact that compliance by person B in this case was

indicated by person A. This means that person A commented on his use of power and on

the effectiveness of that behavior. As the information of Shetty (1978) is less relevant and

the information of Kearney et al. (1985) is less objective, we attach more value to the studies indicating a negative relationship between reward power and readiness for change.

Within the context of the healthcare sector the same argument can be found for a negative relationship between reward power and readiness for change as was already found by Handy (1993); the fact that people do not like the feeling of being bought: the use of reward power resembles a form of bribery. People working in the healthcare sector are professionals; this means that have high ethic and moral standard (Witman, 2007).

They will most probably not do something because they get something in return. Fiorelli (1988) supports this; he found that using reward power to reach a certain goal might be seen as unprofessional, and therefore it might not work.

Summarizing, the use of reward power by the manager will most likely result in a negative response by medical professionals. Therefore, we propose:

H2a: Reward power has a negative relationship with an individual’s readiness for change

Legitimate power and readiness for change

Legitimate power is based on the idea that person A has the legitimate right to exercise power over person B (French & Raven, 1958); for example person A has the contractual right to make B do something; he has a certain position within the organization that gives him the right to make others do something. A good example of someone with legitimate power is a judge; he has the right to sentence someone to imprisonment (Raven & French, 1958a). With regard to the organizational context, Raven and French (1958a) mention that legitimate power consist between offices rather than between persons. In this case, the offices of the organization under study are ‘the managers’ and ‘the medical specialists’.

What is important when someone wants to use legitimate power is that he makes sure that the employees accept his legitimate right to make decisions. When people believe in the formal right to make decisions, his power to influence the employees will be far greater than when people have not established the right to make decisions (Raven

& French, 1958a).

In the literature, a comparison is made between coercive power and legitimate power (Raven & French, 1958b). In both cases, even if person B does not see the rationale behind the change, he has to conform to that change. The difference is that with coercive power, he has to conform to avoid negative consequences. In the case of legitimate power, he will conform because he accepts the formal right of person A to make person B do something: person B will do something because the requests appeals to his system of values and beliefs, there is private acceptance of the change. Related to this is the amount of attraction between person A and person B. With the use of coercive power, we already found that person B will be less attracted to person A. When legitimate power is used, it was found by Raven and French (1958b) that person B will become more attracted to person A,

Another difference between coercive power and legitimate power was also presented by (Raven & French, 1958b). We have seen that, with coercive power, it is important that person A observes person B to see if the new behavior is shown. With legitimate power, this is not necessary: person B will display the new behavior because he feels that person A has the right to ask him to change, and because of that, he will behave accordingly.

In the literature it was found that legitimate power, together with expert power, is one of the two strongest reasons for employees to comply with the orders of their supervisor (e.g. Podsakoff & Schriesheim, 1985; Yukl & Falbe, 1991). Some even say (e.g. Shetty, 1978; Student, 1968) that legitimate power is the strongest of all five forces of power to comply with an order. The reason for this is that power based on legitimacy is the form of power we see in real life as well, for example: our government was given the (legitimate) right to make and enforce laws. We are taught from childhood on that legitimacy is important (Pfeffer, 1981). This is also the reason that legitimate power has been found to positively influence readiness for change by previous authors (e.g. Nesler, Aguinis, Quigley, Lee & Tedeschi, 1999; Rahim, 1989; Raven & French, 1958)

To summarize, the use of legitimate power will most likely influence the medical

professional’s readiness for change positively, in accordance with that we propose:

H3a: Legitimate power has a positive relationship with an individual’s readiness for change

Referent power and readiness for change

Referent power, the fourth form of power, is based on the perception of B that a certain attraction exists between A and B; this could be based on feelings of shared identity, or a feeling of sharing the same norms and values (Busch & Wilson, 1976). The basic idea is that person B wants to be like person A, because person A portrays himself as an attractive person (Braynion, 2004). One very important characteristic of referent power is that referent power can be used in lots of different settings, unlike for example reward power. With reward power, one reward is used for one situation. In a new situation, a new reward has to be given. With referent power, when person B likes person A, person A can influence person B in lots of different settings. The ways Hollywood celebrities can sometimes influence our political and social moral are a good example of this (Martin, 1978).

Referent power, unlike coercive power, reward power and legitimate power, is a form of power that cannot be established based on rules and regulations. Coercive power and reward power are established by having the rights to, for example, deciding upon giving a promotion or not. Legitimate power is something a manager has because of his job position. Coercive, reward and legitimate power are forms of power based on the position of the manager in the organization (Student, 1968). Referent power on the other hand has to be built in an informal manner, it cannot be formed by the organization: a good method to create referent power is participating in activities unrelated to the job, for example going to the annual Christmas party (Martin, 1978).

In previous research, referent power was found to be positively associated with

compliance and change because a request made by someone you respect is most probable

to internalize the request into your own values and beliefs, making you want to try harder

to achieve a successful realization of that request (e.g. Buchman, 1997; Busch & Wilson,

1976; Handy, 1993; Podsakoff & Schriesheim, 1985; Rahim, 1989). Additionally, Rodin

and Janis (1979) found that people who have referent power are able to bring about

genuine changes in attitude, values and decisions because they are seen as being

benevolent and caring. This gives us the impression that a positive relationship between referent power and readiness for change exists.

However, in spite of all these positive relationships that were found, research showed that although referent power has the desired effects, it is not the strongest reason for employees to obey their supervisor. Of the five different bases of power, referent power is said to be the 3

strongest reason to comply. This means that two forms of power are stronger reasons, and two forms of power are weaker reason to comply with a change (Podsakoff & Schriesheim, 1985). As was mentioned earlier, coercive power and reward power are the weakest reason to comply with a request. Legitimate power and expert power form stronger reasons to comply with the will of the supervisor. In both cases, this is because we live in a society that appreciates legitimacy and expertise.

To summarize, it is very likely that the use of referent power will have a positive influence on the readiness for change amongst medical specialist. Therefore, we propose:

H4a: Referent power has a positive relationship with an individual’s readiness for change

Expert power and readiness for change

Expert power is the last of the forms of power in the framework designed by French and Raven (1958). This form of power is based on the idea that person A has a certain expertise or specific knowledge or skills that allows him or her to exert power over person B (Busch & Wilson, 1976; French & Raven, 1958), for example; by giving certain information, an outcome can improve (Buchmann, 1997). A good example of someone with high expert power is a lawyer, he has studied the strict rules and regulations that exist in a country, and therefore he can give advice on issues related to legislation (Cho & Cameron, 2007).

In earlier studies, an interesting relationship between expert power and referent

power has been found. Aronson, Willerman and Floyd (1966) and Martin (1978) found

what is called by Martin (1978) the expert-referent power dilemma: when someone is

perceived to have expertise, the attraction, and thus the possibility to have referent power,

diminishes. The main reason for this is that when someone has a lot of expertise he has

superior knowledge, and therefore he is different to you. Referent power grows from a sense identification between two persons. When expert power is present, identification is that much harder, because one is clearly different than the other. This relationship also exists the other way around: when you like someone, or feel attracted to that person, it is unlikely that you will attribute that person with expert power. When you like someone you coexist on the same level, when someone has expert power he functions on another level because of his superior knowledge. While in some cases one can use multiple sources of power at the same time (French & Raven, 1959), for example emphasizing your legitimate right to make decisions (legitimate power) and at the same time being caring to your employees (referent power) could increase readiness for change, what was demonstrated by Aronson et al. (1966) and Martin (1978) gives us the clear indication that, when using power to make a change happen, it is wise to either choose expert power or referent power, and not both at the same time.

Shackleton (1995) says that of the five power bases, this is probably the least complicated one to implement, since we live in a society based on knowledge, and we are learned that experts know best, that is why they are the expert. However, what needs to be kept in mind is the fact that anyone can be an expert, as long as they know more than someone else (Handy, 1993). This means that person A needs to keep on acquiring knowledge and skills in order to maintain a certain level of expertise. In addition to this, French and Raven (1959) mention that expert power, opposite to referent power, is restricted in its use. Expertise is usually only present in a limited number or areas, only those areas in which expertise is proven. This means that for every new project, it is possible that the manager has to acquire new skills. What is important with respect to that is that the manager has to make clear to the employees that he has the needed expertise to deal with a change. When he does this, he can use his power to make people participate.

Knowledge and skills have been mentioned already as to increase someone’s expertise. Aside from that, other things seem to have an influence on someone’s perceived expertise. Martin (1978) mentions that a higher age, high levels of education and other indicators of social status can also increase someone’s perceived expertise.

In past research on this concept, it was found that, together with legitimate power,

expert power is one of the two strongest reasons to comply with orders (e.g. Podsakoff &

Schriesheim, 1985; Yukl & Falbe, 1991), because both legitimate power and expert power are installed in our society as important. Legitimate power because we live in a society where we give certain people the legitimate right to make decisions for us (e.g.

the police), and expert power because we live in an knowledge-based society, where expertise is appreciated (Shackleton, 1995). This argument is also put forward to explain the positive relationship between readiness for change and expert power that was found by many researchers: expertise is appreciated (e.g. Busch & Wilson, 1976; Dunne, Stahl

& Melhart, 1978; Nesler et al., 1999; Rahim, 1989, Yukl & Falbe, 1991).

When initiating a change the fifth option to persuade the medical specialists to participate in the change is to use expert power. In this case, the manager shows his expertise and knowledge about the managerial side of the change subject and convinces the medical professional that the change is needed. In order for this to work, the medical professional needs to believe in the expertise of the manager. Witman (2007) found that medical professionals appreciate the expertise of managers. They find that working together is needed to reach a decision. In the case of a managerial change, when the manager has the required expertise, the medical professional will comply with the ideas of the manager, in order to reach the goal. Especially in the case that the manager has relevant experience (e.g. he has coped with a similar change before), it is plausible that this will have a positive effect on the level of readiness for change. Knowing that you, and your organization, are capable of changing will increase readiness for change (Holt et al., 2007).

In this section we found that earlier research on expert power shows that a positive relationship exists between expert power and readiness for change. In the healthcare sector, the same relationship seems to exist. Therefore, we propose:

H5a: Expert power has a positive relationship with an individual’s readiness for change

Psychological safety

In the introduction of this study, it was already noted that psychological safety is a

very important and interesting term within the theory of organizational change. When

employees feel psychologically safe in the organization, a change is often perceived as more useful, and the change will be accepted sooner. Psychological safety is even said to be needed to make individuals feel that they are capable of change (Schein, 1993).

Thus, psychological safety feeds readiness for change. We now wonder if psychological safety mediates the relationship between social power and readiness for change. Therefore, in this upcoming section, we will define psychological safety, and we will hypothesize why it could have a mediating effect on the relationship between social power and readiness for change.

The concept of psychological safety is often used on the team-level (e.g.

Edmondson, 1999). In that context, the term is defined by Edmondson (1999:354) as “a shared belief held by members of a team that the team is safe for interpersonal risk taking”. Risk taking for example could be: asking a member of the team for help or not being afraid to talk about problems. Another definition of psychological safety, that is somewhat broader in its use, is found in Kahn (1990:708), he refers to psychological safety as “an employee’s sense of being able to show and employ one’s self without fear of negative consequences to self-image, status, or career”. Although on different levels, we immediately see the similarities between both definitions: you can show your inner self without being ridiculed for that. One might say that this is similar to the concept of trust. Psychological safety however, goes beyond trust: not only is trust an important issue, mutual respect and trust together form the basis of this concept.

In this study we focus on psychological safety at the organizational level. This was done earlier by Baer and Frese (2003). They define psychological safety within the organizational climate as the “formal and informal organizational practices and procedures guiding and supporting open en trustful interactions within the work environment” (Baer & Frese, 2003:50). Similar to the more general definition of psychological safety, the concept describes an environment where people can feel free to speak up, without any repercussions.

In a previous section, we assumed that coercive power leads to a lower readiness

for change. If one feels psychologically safe in the organization, a discussion can be

started about the newly proposed change event, since one is not afraid of being ridiculed or punished. This can help inform employees better which can improve their readiness for change. However, if coercive power is used in an organization, the climate in that organization will not be perceived as psychologically safe. Kahn (1990) and Nembhart and Edmondson (2006) suggested that supportive, clarifying managers increase the level of psychological safety experienced by others. This idea was empirically proven by Schepers, De Jong, Wetzels and De Ruyter (2008). They found that a supportive environment is an important antecedent of psychological safety. Turning that argument around: if the managers are not supportive (which they are not when they use coercive power), psychological safety will be low.

Additionally, Kahn (1990) also mentions that people feel more psychologically safe when they have some control over what happens. When people have control over their work, they feel that they are trusted with the responsibility to perform that task.

When the control is taken away, they feel that the trust is gone as well. When coercive power is used, all control is taken away by the manager. The employee has to do what the manager says, or negative consequences will be the cost. Hence, lowering the control of the employees will also lower the psychological safety one experiences because the employees will feel that they are not trusted with that task and consequently, they will experience lower levels of readiness for change.

As was said before, research showed that, if there is no psychological safety, no change will take place (Schein, 1996). Therefore, we propose the following mediating effect of psychological safety on the relationship between coercive power and readiness for change:

H1b: Psychological safety mediates the negative relationship between coercive power and readiness for change, such that coercive power leads to a decreased level of psychological safety and consequently, leads to lower levels of readiness for change.

We now wonder if psychological safety mediates the relationship between reward

power and readiness for change. As was said, a feeling of psychological safety creates

opportunities to ask questions, to discuss and to put yourself in a vulnerable position. In an organization where reward power is used as a tool of persuasion, psychological safety is probably low: the use of reward power means that a reward is given for compliance with the proposed change, no further explanation is given about the how and why of the change. Active participation (that is: discussions, feedback, etc.) is not asked for, and presumably not even wanted. In the section on coercive power we found two arguments for a mediating relationship between coercive power and readiness for change: non- supportiveness and having no control. These same arguments can be used for reward power as well. When managers use reward power to make the employees comply, there is no room for a supportive relationship between manager and employee: psychological safety will be low (Kahn, 1990; Nembhart & Edmondson, 2006; Schepers, et al. 2008).

When one is rewarded for behavior, having control is not possible, which also lowers psychological safety (Kahn, 1990). Based on this, we propose the following mediating hypothesis:

H2b: Psychological safety mediates the negative relationship between reward power and readiness for change, such that reward power leads to a decreased level of psychological safety and consequently, leads to lower levels of readiness for change.

In hypothesis 3a it was assumed that legitimate power increases readiness for change. We found earlier that psychological safety is needed in order to have readiness for change; we wonder if psychological safety functions as a mediator in the relationship between legitimate power and readiness for change.

In Kahn (1990) we find that working within your own role can positively

influence your psychological safety. Working within those roles is that each employee

fulfills the task they were hired for: the managers perform the managerial tasks and

medical specialists perform the medical tasks. The reason why this positively influences

psychological safety is that these roles are part of the internal norms and values each

employee has: it is acceptable to address your questions to someone with another role,

because he was given the right to function in his role.

If the managers use legitimate power to make someone participate in a change, the manager will emphasize his position in the organization; he will underline the fact that he has the right to make managerial decisions. In this case, it is likely that higher levels of psychological safety will be experienced, since it fits within his organizational role that the manager decides on managerial issues. Because of this, we assume a mediating effect of psychological safety on the relationship between legitimate power and readiness for change:

H3b: Psychological safety mediates the positive relationship between legitimate power and readiness for change, such that legitimate power leads to an increased level of psychological safety and consequently, leads to higher levels of readiness for change.

Additionally, we wonder if psychological safety could have a mediating effect on the relationship between referent power and readiness for change. In other words, could psychological safety be an explanation for that relationship? In a situation where a shared feeling of identity exists between manager and employee (i.e. referent power), psychological safety is high because the employee will feel that their manager will take the time to answer questions and explain issues relevant to the change: he will experience a more supportive relation. As said before, Kahn (1990), Nembhart and Edmondson (2006) and Schepers, et al. (2008) found that supportive managers create a more psychologically safe environment. Thus, referent power is likely to positively influence a person’s psychological safety, which in turn will lead to higher levels of psychological safety.

Another argument can be found by Walumbwa and Schaubroeck (2009). They found that when employees find their managers to be caring, they will experience more psychological safety. Earlier in this study we already found that when someone has referent power, this person is seen as being caring. This means that, when the manager is perceived to be caring to his employees, they will experience more psychological safety:

they will be not afraid to ask questions and start discussions. Therefore, we can conclude

that referent power is likely to positively influence readiness for change via an increased level of psychological safety.

Moreover, Carmeli, Brueller and Dutton (2009) found that quality relationships between individuals result in higher levels of psychological safety. A good example of a quality relationship is a relationship in which referent power is present, because quality relationships are present when the actors in the relationship feel valued and connected to each other. This means that, when the manager uses referent power, he will increase the level of psychological safety for the employees. All the argument put forward in this section lead us to the following hypothesis:

H4b: Psychological safety mediates the positive relationship between referent power and readiness for change, such that referent power leads to an increased level of psychological safety and consequently, leads to higher levels of readiness for change.

It is proposed in the previous section that the use of expert power will increase the readiness for change. We now wonder if psychological safety can explain this relation.

When the manager uses expert power, he tries to convince the employees to participate, based on his or her knowledge and experience. Kahn (1990) found that belief in the competence of the manager will positively influence the feeling of psychological safety experienced by an employee. If the manager can convey this knowledge to the employees, the employees will feel psychologically safe because they believe in the competence of the manager; they will feel free to ask questions and learn from the expertise of the manager This means that, if the manager shows his competence, for example, by using his expertise and relevant knowledge, psychological safety will improve, and with that, the readiness for change.

Similar results were found by Walumbwa and Schaubroeck (2009). They stated

that people feel more free to engage in interpersonal risk taking (one of the characteristics

of psychological safety) when they feel that their manager has sufficient ability to

successfully cope with that. Thus, we propose that psychological safety mediates the

relationship between expert power and readiness for change:

H5b: Psychological safety mediates the positive relationship between expert power and readiness for change, such that expert power leads to an increased level of psychological safety and consequently, leads to higher levels of readiness for change.

Conceptual model

A figurative summary of all hypotheses can be found in figure 1.

METHODOLOGY Data collection and analysis

The hospital under study employs 98 medical specialists in over 20 different medical specialties. We aimed at getting a response rate of at least 50%, because most professionals agree that a response rate of 50% or more is sufficient for analysis purpose (Church & Waclawski, 1998). To achieve this, we decided to use questionnaires. Reasons for this are that, first of all, Keller and Warrack (2003) mention that, when measuring attitudes, the use of a questionnaire is quite common. Another point that we took into account is the fact that using questionnaires is easy and not too time-consuming for the respondents. Since medical specialists are busy people, we feared that a qualitative way of questioning would results in a very low response rate. To increase the response rate some more, we focused on using short but validated scales.

Coercive power

Reward power

Legitimate power

Referent power

Expert power

Readiness for change Psychological

safety

| | + + +

Figure 1: Conceptual model

All 98 specialists who worked in the Antonius Hospital were approached to voluntarily fill in the questionnaire that was designed for this research. One week before the actual research took place, a note of interest (signed by one of the medical specialists and the author) was e-mailed to all specialists, to inform them about the research. One week after that, the questionnaire was distributed through the internal mail system. In order to improve the response rate complete anonymity was guaranteed. Respondents had two weeks to fill in the questionnaire.

The initial response was 50 out of 98 (response rate of 49%). However, six of the respondents were removed from the sample because they did not fill in 20 of the items, making their questionnaire not useful for the research. After the deletion of these six respondents the response was 44 (response rate of 43.12%).

The average age of the respondents was 49 (with a standard deviation of 8.32).

75% of the group was male, 25% was female. This matches the gender diversity of the entire group, making the sample representative for the group. All respondents had the Dutch nationality. Respondents were asked to mention their medical specialty. Because the answers to this question were too numerous to categorize, they were recategorized into three types of specialties: contemplative (e.g. cardiology), cutting, (e.g. gynecology) and supportive (e.g. anesthesiology). Rates for these types were: 47.7%, 22.7%, and 9.1%, respectively. Additionally, questions were asked regarding work experience and organizational tenure. Work experience had a mean of 19.51 years, with a standard deviation of 10.80. Organizational tenure had a mean of 11.86 years, with a standard deviation of 9.94. Missing values appeared in specialty (9 missing values), and in age, work experience, and organizational tenure (1 missing value).

Measures

For this research, existing scales were used to measure the variables. Likert scale

questions were used, with answer possibilities ranging from 1 (strongly disagree) to 7

(strongly agree). The questionnaire consisted of 37 items. In appendix A the

questionnaire can be found.

Readiness for change

Readiness for change was measured using a scale designed by Metselaar (1997).

In total, this scale contained 4 items. Examples of the questions were: ‘I intend to try to convince employees of the benefits the change will bring’, and ‘I intent to try to make time to implement the change’. It was mentioned in the introduction of these items that

‘the change’ represented the project ‘Antonius aan de Wind’. The scale had a Cronbach’s alpha reliability coefficient of .90.

Social power

The perceived social power was measured by 20 statements derived from Hinkin and Schriesheim (1989), four for each form of social power. Examples of the questions posed in the questionnaire were: ‘my supervisor can provide me with special benefits’

(reward power), ‘my supervisor can provide me with good technical suggestions’ (expert power), ‘my supervisor can make things unpleasant here’ (coercive power), ‘my supervisor can make me feel that I have commitments to meet’ (legitimate power) and

‘my supervisor can make me feel important’ (referent power). The Cronbach’s alpha for the forms of social power (reward, coercive, legitimate, expert, and referent, respectively) were: .90, .73, .83, .89, and .94.

Psychological safety

Psychological safety was measured with seven questions, designed by Edmondson (1999). Examples of the questions were: ‘it is difficult to ask other members of this organization for help’, and ‘if you make a mistake in this organization, it is often held against you’. Three of the seven items were reversed coded. After the recoding and the factor analysis, the Cronbach’s Alpha for this scale was .66. Although this is not perfect (Nunnally (1978) suggested a standard of .70), it is acceptable and sufficient for the continuation of this research because Kline (1999; in Stein, 2005) mentions that, when measuring a psychological construct (such as psychological safety) values below .70 can be expected, because of the diversity of the construct being measured.

Control variables

The control variables that were added were gender, age, nationality, work

specialty, relevant work experience and organizational tenure.

Factor analysis

In order to see if the variables were sufficiently distinct from each other we performed a factor analysis. A principal component analysis was performed, with a direct Oblimin with Kaiser Normalization rotation method. The analysis revealed 8 different components. In table 1, the results can be found.

In table 1, we see that the dependent variable, readiness for change, had high factors loadings on one single component. However, when looking at psychological safety, we encounter one problem: item 5 loads on the same components as readiness for change. This is a problem: one of the variables loads on the dependent variable. When doing any further analyses, these two variables are not perfectly distinct from one another, and this will lead to non-generalizable results. In dealing with this problem, we tried forcing the model into fewer components. However, this was not possible. Additionally, we tried removing the item from the scale. This was also problematic, since this resulted in a lower Cronbach’s Alpha for the concept of psychological safety, which can also give problems regarding the reliability of the results. In the end, it was decided to keep the scale. Although both options (removing the item, and keeping the item) are not perfect, removing any item from a scale needs to be done with caution. And given the fact that removing the item would also lead to problems, we did not see any reason for deleting it.

However, when reading and interpreting the results, we advice caution as in how applicable the results might be for any organization.

In table 1 it can be found that the factor analysis showed that reward power and

referent power were perfectly distinct from other variables; they all had high factor

loading scores and loaded on their own factor. A first glance at the factor loadings for

coercive power reveals a non-perfect factor analysis result. Two items score on one

component, the other two items score on another component. When looking at the

content of the items a clear reason can be found for this: item 1 and 2 were about coercive

power related to the job and item 3 and 4 were about coercive power related to the

atmosphere on the job. Because of this, nothing was done to change the factor scoring,

and no items were removed. When looking at legitimate power and expert power we find

something interesting. Legitimate power item 4 loads on both legitimate power and on

expert power, which is inappropriate. If we take a closer look on the content of item 4, we

see no reason for a double loading: it clearly focuses on the legitimacy the supervisor has, and it does not focus on his or her expertise. If a higher N would have been used in this research, it is possible that this problem would not have occurred. Therefore, in this case it was decided to keep the original scales.

Table 1: Factor Analysis for all Variables

Component

1 2 3 4 5 6 7 8

Readiness for change Item 1 -.80 .02 -.18 .00 .20 .15 -.07 .16 Item 2 -.80 -.18 -.13 -.11 .18 .09 -.00 -.07 Item 3 -.88 -.18 -.08 .05 -.01 -.07 .16 -.05 Item 4 -.93 .11 .19 .06 -.11 -.08 -.00 -.17

Reward power Item 1 -.03 .84 .01 -.08 .22 .04 .02 .08

Item 2 -.08 .91 .07 -.17 -.04 -.07 .05 -.12 Item 3 .08 .83 -.16 .11 .01 -.04 -.00 -.19 Item 4 -.01 .84 .09 .09 .05 .03 .16 .22

Coercive power Item 1 .14 .28 .42 .23 -.08 -.01 -.26 .28

Item 2 .02 -.02 .74 .25 .07 .05 -.05 .07 Item 3 -.00 .03 .01 .91 .00 .06 -.12 -.05 Item 4 -.03 -.07 .07 .94 .05 -.05 .02 .01 Legitimate power Item 1 -.07 .05 .02 .12 .78 -.01 -.09 -.17

Item 2 .04 .11 .05 .02 .92 -.06 -.02 -.08 Item 3 .04 .04 -.02 -.09 .84 -.05 .27 .12 Item 4 -.25 -.05 .10 -.08 .41 -.56 -.13 .29

Expert power Item 1 .18 .00 -.22 .29 -.05 -.74 .19 .08

Item 2 .06 .27 -.16 -.05 .01 -.84 -.04 -.14 Item 3 -.05 .15 -.14 -.19 .13 -.75 .05 -.09 Item 4 -.03 -.23 .16 -.00 .05 -.89 .15 .02

Referent power Item 1 .11 -.02 -.14 -.06 .08 -.09 .86 .08

Item 2 .02 .11 -.02 -.15 .08 .03 .87 .10 Item 3 -.02 .05 .07 .03 -.02 .01 .93 .21 Item 4 -.14 .08 .06 .02 -.03 -.21 .82 -.20 Psychological safety Item 1 -.09 .12 -.45 -.05 -.20 -.11 .20 -.05 Item 2 .04 -.10 -.66 .01 .09 -.33 -.12 .37 Item 3 -.01 -.05 -.04 -.05 -.14 .04 .25 .87 Item 4 -.34 .04 -.45 .14 -.02 -.10 .07 .32 Item 5 -.65 .25 -.00 -.04 -.09 .03 -.01 .30 Item 7 -.18 .05 -.48 .25 .14 .21 .35 -.25 Notes. Extraction Method: Principal Component Analysis.

Rotation Method: Direct Oblimin with Kaiser Normalization.

Psychological safety item 6 is deleted from the scale.

The factor analysis for psychological safety was fairly complex. As can be seen in table 1, the items now load on three different factors. On some of these factors other variables load as well. Looking at the Cronbach’s alpha, we found it wise to delete item 6 from the scale, this lead to an increase in Cronbach’s alpha from .62 to .66. Given the fact that Cronbach’s alpha for this scale is already quite low, this is a great improvement.

Based on this, it was decided to remove item 6 from the scale. As discussed before, deleting any other item from the scale would result in a lower Cronbach’s alpha and was therefore not considered.

RESULTS

The readiness for change at the Antonius Hospital is displayed in figure 2, by means of a pie chart. In table 2 the means, standard deviations and correlations of all variables under study are presented. To test the hypotheses, simple linear regression analyses were used.

Readiness for change at the Antonius Hospital

In figure 2, we can find the results for readiness for change at the Antonius hospital. The pie chart is divided into five categories, ranging from slightly negative to very positive. Slightly negative in this context refers to an overall score (item 1 + item 2 + item 3 + item 4 / 4) of less than four. Neutral refers to a score of exactly four. Slightly positive is a score between four and five. Positive is when respondents score anywhere between five and six, and very positive is anything higher than six. What we find is that only 2% of all respondents is slightly negative about the change. 2% is neutral, and 96%

is slightly positive to very positive about the change.

Correlation and regression

In table 2, the correlations between the variables under study can be found.

Table 2: Means, Standard Deviations, and Correlations for all Variables

M SD 1 2 3 4 5 6 7

1. Reward power 3.05 1.50 _

2. Coercive power 2.11 1.00 -.00 _

3. Legitimate power 4.60 1.15 .22 -.02 _

4. Expert power 4.27 1.40 .18 -.22 .40** _

5. Referent power 5.47 1.16 .29 -.41** .15 .40** _

6. Psychological safety 5.05 .74 .17 -.35* .15 .34** .54** _

7. Readiness for change 5.72 .86 .07 -.18 .27† .02 .21 .52** _ Notes. **p<.01. *p<.05. †p<.10.

N=44.

The Pearson correlation coefficients show an initial marginally positive correlation (r = .27, p < .08) between legitimate power and readiness for change. This means that higher levels of legitimate power result in higher levels of readiness for change. Other interesting correlations can be found between psychological safety and the following concepts: coercive power (r = -.35, p < .05), expert power (r = .34, p < .01), referent power (r = .54, p < .01), and readiness for change (r = .52, p < .01). This last

Figure 2: Readiness for change at the Antonius Hospital

relationship, between psychological safety and readiness for change is interesting. It states a direct positive relationship between both concepts, as was already mentioned in the theory. This correlation, together with the regression (ß = .52, p < .01, R

= .27) supports the findings of earlier research that psychological safety influences the readiness for change: when the level of psychological safety is high, levels of readiness for change is high as well. A final interesting correlation that can be found in table 1, is the positive correlation between expert power and referent power (r = .40, p < .01). This is interesting because earlier, we found that in recent literature it was written that expert power and referent power have a negative correlation with each other. In this research we found opposite results.

To test the hypotheses, regression analyses were performed. Hypothesis 1a stated that a higher use of coercive power leads to a lower level of an individual’s readiness for change. Regression analysis (ß = -.18, p = n.s., R

= .03) showed that this hypothesis cannot be confirmed and has to be rejected.

Hypothesis 2a stated that a higher use of reward power will lead to lower levels of readiness for change. Regression analysis (ß = .07, p = n.s., R

= .06) shows that this hypothesis also has to be rejected.

Hypothesis 3a stated a positive relationship between the use of legitimate power and readiness for change. The regression results (ß = .27, p < .08, R

= .07) showed that a marginally positive relationship exists between both variables. If legitimate power increases with one unit, readiness for change will increase with 0.27 unit. More practically, this means that about one quarter of the efforts put in increasing the perceived legitimate power is converted into readiness for change. Hypothesis 3a can be accepted because of this, although with some caution given the significance level of .08.

Hypothesis 4a stated that the use of referent power by the manager would lead to more readiness for change. Regression results (ß = .21, p = n.s., R

= .04) showed that this hypothesis also has to be rejected. Although a positive relationship exists, it is not significant.

The last hypothesis, hypothesis 5a, stated a positive relationship between the

manager’s use of expert power and the readiness for change of an individual. The