Financial scarcity: obstacle or opportunity for creativity?

Master thesis 22-7-2020

Elvira Lems s2252880

Master Public Administration: Public Management and Leadership Faculty of Governance and Global Affairs, Leiden University Institute of Public Administration

Dr. Joris van der Voet Dr. Kohei Suzuki Leiden University

Financial scarcity:

obstacle or opportunity

for creativity?

A study on the effect of financial scarcity in public

A

For several decades, scholars have been trying to find out if decision-makers react to organizational decline with innovative or rigid behavior. Since this puzzle has remained unclear so far, this study aims to unravel part of this management puzzle by examining the effect of financial scarcity in public organizations on creativity of individual decision-makers. Applied to the case of youth care in the Netherlands, a policy domain characterized by financial shortages, the following research question is formed: what is the effect of financial scarcity in public organizations on creativity of individual decision-makers? Through a survey experiment, 996 participating municipal councilors were asked to generate ideas to improve youth care in their municipality. Most participants worked for a municipality wherein 2018 a budget shortage was present for youth care. Before generating the ideas, municipal councilors in the treatment group were manipulated with negative financial performance feedback according to their budgetary status of 2018. The 3.272 generated ideas were analyzed on content and merged into 31 idea-categories. Following, these categories were scored by two groups of youth care-experts on creativity, effectiveness, and feasibility. The results show that, based on the scores of the aldermen, financial scarcity has a negative effect on creativity of decision-makers. Based on the scores of the managers, no significant result is found. There can be concluded that, according to this study, financial scarcity has a negative effect on creativity of individual decision-makers. This result supports propositions made by scholars that claim that necessity is the mother of rigidity. Since this result has important implications for the youth care-domain and some limitations of this study are identified, in the final chapter recommendations for future research will be done.

Table of contents

Abstract ... 1

Chapter 1. Introduction ... 4

1.1 Introduction ... 4

1.2 Research question and methods ... 5

1.3 Theoretical, practical and methodological relevance ... 5

1.4 Thesis outline ... 6

Chapter 2. Theoretical framework ... 7

2.1 Innovation in the public sector ... 7

2.1.1 Two logics to assess public sector innovation ... 7

2.2 The social psychology of creativity ... 8

2.3 Individual decision-makers in the public sector ... 9

2.3.1 Learning form performance feedback ... 9

2.4 Financial scarcity negatively affects creativity ... 10

2.5 Financial scarcity positively affects creativity ... 11

Chapter 3. Methodology ... 13

3.1 Design of the survey experiment ... 13

3.2 Case of the study ... 16

3.3 Unit of analysis ... 16

3.4 Variables ... 18

3.5 Methodology for measuring creativity ... 19

3.5.1 Techniques to measure creativity ... 19

3.5.2 The operationalization of creativity ... 20

3.5.3 Procedural requirements for creativity assessment by experts ... 20

3.6 Expert survey ... 21

3.6.1 Design of the expert survey ... 21

3.6.3 Data collection ... 22

3.7 Reliability and validity ... 25

3.7.1 Interjudge reliability ... 25

3.7.2 Discriminant validity ... 26

3.8 Data-analysis methods ... 26

Chapter 4. Results ... 28

4.1 Aggregated scores per idea ... 28

4.2 Frequency of the generated ideas per experimental group ... 31

4.3 Hypothesis tests ... 33

4.3.1 Expert group: Aldermen... 33

4.3.2 Expert group: Managers ... 34

4.4 Summary results ... 35

Chapter 5. Discussion ... 36

5.1 Discussion of the results and possible explanations ... 36

5.2 New insights to current literature ... 37

5.3 Limitations and recommendations for future research ... 38

5.4 Practical implications ... 40

References ... 41

Appendix 1. Inclusion and exclusion criteria and categorization process ... 45

Appendix 2. Expert survey questions ... 49

Appendix 3. Frequency of idea categories 1-31 per experimental group. ... 53

Appendix 4. Tables SPSS output: Independent t-tests. ... 54

C

1.

I

1.1 INTRODUCTION

A central puzzle that management scholars have been trying to solve for several decades, is if decision-makers react to organizational decline with innovative or rigid behavior (Mone, McKinley, & Barker, 1998). Organizational decline can be defined as “a substantial, absolute decrease in an organization's resource base occurring over a specified period of time” (Cameron, Kim, & Whetten, 1987, p. 224). Decline in organizations can exist in multiple forms: in performance, staff size, financial resources, and more. Similarly, it can be caused by various factors, like recession, austerity policies, or outdated strategies (Mone et al., 1998; Van der Voet, 2019). Over the last few decades, scholars have been trying to unravel the effect of organizational decline on individual, group, and organizational level. An important outcome variable in this field of research is the effect of organizational decline on innovation. Innovation is key to adaptation of organizations since it is crucial for the sustainable development of organizations (Brown & Eisenhardt, 1995). Although multiple studies have tried to solve this puzzle, they found different, even contrasting results. The inconsistent positions of scholars on this topic can be divided into two schools of thought. Some scholars claim that decline inhibits innovation based on a threat-rigidity theory. Other scholars base their point of view on the behavioral theory of the firm or prospect theory, by stating that decline stimulates innovation initiated by decision-makers (Mone et al., 1998).

Innovation can be defined as a “process through which new ideas, objects, and practices are created, developed or reinvented, and which are new for the unit of adoption” (Walker, 2007, p. 592). It involves therefore both improving and adopting innovations from other organizations that are new for the adopting organization, and generating ‘creative’ ideas (Anderson, Potočnik, & Zhou, 2014). The focus of this study is on innovation by using creativity. Creativity can be defined as “the production of novel and appropriate ideas by individuals” (Amabile, 1996, p. 230). In the public sector and especially at the governmental level, creativity in decision-making is important for the sustainable improvement of public services (Kruyen & van Genugten, 2017). At the same time, public sector organizations are often faced with a decline in financial resources, due to recession or austerity policies. Despite the frequency of financial scarcity in public organizations and the importance of creativity in public organizations, very few scholars have, however, studied creativity in the public sector and the effect of financial scarcity on creativity of public decision-makers (Amabile, 1982; Kruyen & van Genugten, 2017). Therefore, this study will focus on the puzzle if public

decision-makers react on financial scarcity with rigidity or creativity. In this study, financial scarcity is defined as “a substantial decrease in financial resources” (Van der Voet, 2019, p. 2). Since decision-makers in the public sector can individually have a significant impact on innovation and studying individual decision-makers has methodological benefits, this study will investigate the effect of financial scarcity on creativity of decision-makers on the individual level.

1.2 RESEARCH QUESTION AND METHODS

So far, it remained scientifically unclear if organizational decline inhibits or stimulates innovation by decision-makers. The goal of this study is to unravel part of this management puzzle by examining the effect of financial scarcity in public organizations on creativity of individual decision-makers. The research question of this study is: What is the effect of financial scarcity in public organizations on creativity of individual decision-makers? The study is based on a survey experiment and an expert survey. In the experiment, the 2066 municipal councilors that responded to the survey were randomly assigned to a treatment and control group. Participants in the treatment group received financial feedback based on the actual budgetary status for youth care of their municipality in 2018. Thereafter, the respondents were asked to generate ideas to improve youth care. In most municipalities, a budget deficit for youth care is present. Therefore, the youth care-domain in Dutch municipalities is a fitting case for studying the effect of financial scarcity on creativity. In the expert study, two groups of judges considered as experts on the youth care domain, 111 aldermen and 50 managers, are asked to rate the creativity, effectiveness, and feasibility of these ideas on a 1-10 scale. With statistical analyses via SPSS, the effect of financial scarcity is examined on creativity of ideas. The survey experiment allows to examine if the ideas of decision-makers in the treatment group, who receive negative financial feedback, are seen as more or less creative by the experts participating in the expert survey.

1.3 THEORETICAL, PRACTICAL AND METHODOLOGICAL RELEVANCE

Previous studies on the relationship between organizational decline and innovation examine this relationship with nonexperimental studies and on the organizational level. This study is theoretically relevant since it studies the effect of financial scarcity on creativity with an experimental design, which makes it possible to establish causality. Besides, this study

give a more objective perspective on the debate compared to previous studies that examine the effect of decline on e.g. the orientation of decision-makers against innovation measured through self-scoring surveys since self-scored data is subject to bias (Singla, Stritch, & Feeney, 2018).

This study is also of practical relevance. Research towards the effect of financial scarcity on public sector innovation is of continuing importance since budget cuts are common in the public sector. When it is clear how individual decision-makers react to financial scarcity in the public organization, this reaction can be taken into account. This study is also of practical relevance for the youth care-domain. Besides financial scarcity, several other problems exist with youth care in Dutch municipalities, like waiting lists, instability of youth care workers, and disagreement between municipalities on finances. Therefore, it is of practical importance to study the effect of present budget deficits on creativity of individual decision-makers, since creativity can help solve existing problems in youth care and other domains in municipalities.

Finally, this study has methodological relevance for studying concepts that are difficult to measure. Experimental studies that study “social and environmental influence on creativity” are very scare because of a criterion problem (Amabile, 1982). This problem is defined by Amabile as “the lack of a clear operational definition and an appropriate assessment methodology” (Amabile, 1982, p. 997). By inviting experts on the youth care-domain to score the creativity of the ideas according to their definition of creativity, this criterion problem can be bypassed. Therefore, this study contributes to studying concepts that are difficult to measure, like creativity of individual decision-makers, in an unveiled way.

1.4 THESIS OUTLINE

The remainder of this study consists of the following parts: after this introduction, a theoretical framework will follow in chapter two where concepts and definitions important to this study are discussed. Moreover, important theories and previous scientific findings relevant to the current study are outlined. In the methodology section in chapter three, the data collection process of the two surveys will be discussed further and there will be elaborated on the way the data is analyzed. Continuously, the results of the study will be presented in chapter four. In chapter five, the results as presented in chapter four will be critically discussed, a conclusion will be formed and recommendations for future research will be done.

C

2.

T

This chapter will present concepts, definitions, and theories that are relevant to the study. This chapter will be roughly divided into two parts. In paragraphs 2.1 to 2.3, the core concepts and definitions of this study will be presented: innovation in the public sector, creativity, and individual decision-makers in the public sector. In paragraphs 2.4 and 2.5, theories that inform the relationship between financial scarcity and creativity will be explained. Based on these theories, also the hypotheses will be formed in paragraph 2.5.

2.1 INNOVATION IN THE PUBLIC SECTOR

In the public sector, the main goal of innovation is to make the government more efficient and effective, by modernizing public services and coping with wicked challenges in society (Bekkers, Edelenbos, & Steijn, 2011). These wicked societal challenges make innovation harder in the public sector as compared to the private sector: for solving these issues, innovations are required that harmonize dissimilar, usually even conflicting public values in a legitimate way (Bekkers et al., 2011). This also has consequences for the way public sector innovations are assessed.

2.1.1TWO LOGICS TO ASSESS PUBLIC SECTOR INNOVATION

Public sector innovation differs from innovation in the private sector in two ways. The first difference is that innovation in the public sector aims to achieve the goal of legitimacy, instead of mainly economic growth in the private sector. The second difference is that for public sector innovation the “specific institutional context” in which innovations are implemented must be taken into account (Bekkers et al., 2011, p. 12). Therefore, two logics are important for public sector innovation. From these two logics, “values, norms and criteria” can be obtained as a basis to assess innovations in the public sector (Bekkers et al., 2011, p. 12).

First, innovations in the public sector can be assessed through what consequences they are expected to have, the “logic of consequences” (March & Olsen, 1989, p. 100). Innovations based on a logic of consequences are aimed at certain values, like efficiency and effectiveness. For example, the effectiveness of an innovation can be assessed by estimating if the innovation will succeed in achieving the goal (Bekkers et al., 2011).

Second, the “logic of appropriateness”. Public sector innovations can also be assessed based on their appropriateness to rules and routines part of the political and societal

Therefore, public sector innovations should next to effectiveness, efficiency and other alike economic values, also be judged on the extent to which the innovation is expected to be accepted “by specific groups or by the society as a whole” (Bekkers et al., 2011, p. 12).

Although ‘innovation’ and ‘creativity’ are related to each other, they are not indistinguishable. Whereas creativity mainly refers to generating completely novel ideas, innovation generally refers to the following phase of “implementing ideas toward better procedures, practices, or products” (Anderson et al., 2014, p. 1298). On the other hand, innovation does not necessarily have to emerge from developing a creative idea. Innovation can also emerge from adopting or reinventing ideas from other organizations that are new to the organization (Anderson et al., 2014).

2.2 THE SOCIAL PSYCHOLOGY OF CREATIVITY

For a long time, no agreement has been found in research about the exact way to define and assess creativity (Amabile, 1996). Researchers have been trying to solve the lack of a “clear operational definition and an appropriate assessment methodology” in many ways (Amabile, 1982, p. 997).

In early studies on defining creativity, researchers mainly focused on defining the process where creative outcomes should result from. Later, around 1950, creativity definitions became popular that defined creativity “in terms of the person” (Amabile, 1996, p. 21). Although most studies were implicitly guided by the “process approach” or the “person approach” to define creativity, most explicit definitions used in these studies refer to the creative product (Amabile, 1996, p. 21). Product definitions use the outcome of a creative process or person as the sign that distinguishes creative from noncreative work. Product definitions are believed to be most appropriate in studying creativity (Amabile, 1996). In this study, the following conceptual product definition will be used:

“A product or response will be judged as creative to the extent that it is both a novel (original) and appropriate response to the task at hand” (Amabile, 1996, p. 35).

Novelty and originality can be seen as necessary conditions of creativity: an idea will only be seen as creative when being novel and original (Nijstad, De Dreu, Rietzschel & Baas, 2010). However, according to the definition of Amabile (1996), these hallmarks of creativity are not sufficient conditions. For an idea to be creative, the idea must also be regarded as appropriate to the given task, otherwise an idea can be seen as crazy or irrelevant instead of creative (Nijstad et al., 2010). In this study, ideas are pre-assessed and filtered out for

assessment by the experts when seen as not appropriate to the given task. Therefore, creativity in the expert survey will be defined as the degree to which the idea is novel and original.

2.3 INDIVIDUAL DECISION-MAKERS IN THE PUBLIC SECTOR

Decision-makers in the public sector have an important impact. According to behavioral theory-scholars, policies in organizations arise from administrative behavior of individual decision-makers (Simon, 1997). Although public decision-decision-makers generally assemble in groups to make decisions collectively, major decisions often derive from decision-makers individually or in cooperation with some colleagues (Staw, Sandelands, & Dutton, 1981). Moreover, decision-makers in the public sector can generate and propose ideas to steer the main lines of public policy, because of their right of initiative (VNG, n.d.b.). In this way, decision-makers can individually have a significant impact on innovation in the public sector. Therefore, this study focuses on makers at the individual level, to examine how creativity of decision-makers – an important pre-stadium for innovation – is influenced by financial scarcity. 2.3.1LEARNING FORM PERFORMANCE FEEDBACK

The theory of learning from performance feedback by Greve (2003) theorizes about how individual makers adapt to organizational decline. This theory claims that decision-makers evaluate the performance of their organization based on aspiration levels. An aspiration level can be defined as “the smallest outcome that would be deemed satisfactory by the decision-maker” (Schneider, 1992, p. 1053).

Through aspiration levels, the performance of an organization can be ascribed to categories of success and failure. Aspiration levels can be generated from two sources of available information. The first source is historical performance information of the organization itself. This kind of aspiration level is called a “historical aspiration level” (Greve, 2003, p. 42). To form this aspiration level, the historical performance of the organization is used to evaluate the current performance of the organization (Greve, 1998). The second source is information about the current performance of comparable organizations, also referred to as a “social aspiration level” (Greve, 2003, p. 45). A social aspiration level can be formed by selecting an appropriate reference group, consisting of organizations that are similar to the organization (e.g. in size and proximity) and by collecting information about their performance (Greve, 2003).

2.4 FINANCIAL SCARCITY NEGATIVELY AFFECTS CREATIVITY

Multiple studies have previously examined the effect of organizational decline on innovation. Most of these studies fit in one of two contrasting schools of thought about the effect of organizational decline on innovation (Mone et al., 1998). The first school of thought, commonly labeled as “necessity is the mother of rigidity”, assumes that organizational decline inhibits innovation (Mone et al., 1998, p. 117). Rigidity is defined by Cowen (1952) as “the tendency to adhere to an induced method of problem-solving behavior when the induced solution no longer represents the most direct and economical path to the goal” (p. 518). Based on the threat-rigidity theory, scholars in this school claim that organizational decline can be seen as a threat, and threatening situations generally cause individuals to behave rigidly. According to this school of thought, two explanations for rigidity in behavior that people show in response to organizational threat are, one, a restriction in information processing, and two, adherence to dominant routines (Staw et al., 1981).

The first explanation for rigid behavior that decision-makers show in response to organizational threat according to this school of thought, is a “restriction in information processing” (Staw et al., 1981, p. 502). The study of Staw et al. (1981) proposes that threat generally enhances psychological stress at the individual level (Staw et al., 1981). Psychological stress causes that individuals are less able to distinguish between familiar and unfamiliar stimuli. Therefore, individuals in stressful circumstances tend to emphasize “prior expectations or internal hypothesis”. They also tend to focus more on dominant cues than on subordinate cues (Staw et al., 1981, p. 506). These reactions on psychological stress cause that decision-makers perceive and process less varied information, and thus creativity will be inhibited (Staw et al., 1981, p. 506).

Threat also enhances anxiety in decision-makers (Staw et al., 1981). According to a meta-analysis by Byron and Khazanchi (2010), anxiety negatively affects creative performance by disturbing the cognition of individuals. Threatening conditions cause individuals to focus their attention on the threat-source and therefore have a narrowing effect on the attention. Narrowed attention inhibits creativity since creativity requires a large scale attentional scope and broad cognitive search. Moreover, a narrow conceptual scope disturbs the ability to make connections between disparate ideas, which is essential for creative performance (Byron & Khazanchi, 2010). On top of that, this meta-analysis states that anxiety even has a larger detrimental impact on creative performance during verbal tasks. According to this

meta-analysis, ideas generated under stressful conditions are therefore more likely to be common than creative (Byron & Khazanchi, 2010).

An observation coming forward from a study of Cowen (1952) is that participants that were put under experimentally induced stressful conditions were less flexible to switch to other, more fitting solutions than participants under non-stressful conditions. Therefore, individuals experiencing stress tend to adhere to “dominant, well-learned or habitual” solutions, even when those solutions are no longer appropriate to the current problem (Staw et al., 1981, p. 506). Therefore, stressful conditions decrease the scope of alternatives considered and stimulates the amount of rigid behavior by decision-makers under stressful conditions (Staw et al., 1981, p. 502).

These reactions on threat, which can be provoked by financial scarcity, stimulate conservative behavior in decision-makers and reduce the capacity of decision-makers to propose creative solutions in response to financial scarcity (Mone et al., 1998). Therefore, based on these theories and observations, there can be hypothesized that financial scarcity has a negative effect on creativity of individual decision-makers.

Hypothesis 1: Financial scarcity has a negative effect on creativity of decision-makers.

2.5 FINANCIAL SCARCITY POSITIVELY AFFECTS CREATIVITY

In contrast, scholars part of the second school of thought, “necessity is the mother of invention”, claim that organizational decline stimulates innovation (Mone et al., 1998, p. 118). According to scholars of this school, organizational decline can be seen as an opportunity for innovation. This school of thought bases its reasoning on two theories: the behavioral theory of the firm of Cyert and March (1963), and the prospect theory of Kahneman and Tversky (1979).

The first theoretical ground for scholars of this school to assume that organizational decline stimulates innovation is the behavioral theory of the firm of Cyert and March (1963). This theory assumes that “the organization interacts with the environment through a performance feedback process” (Cyert & March, 1963, p. 14). One of the main concepts of this theory is “search”: a collection of activities to explore new procedures, strategies or technologies to increase organizational performance (Greve, 2003). Search does not imply direct organizational change. However, it can be seen as a pre-stadium of organizational change,

performance feedback according to an aspiration level (Greve, 2003, p. 54). According to the behavioral theory of the firm of Cyert & March (1963), problemistic search is stimulated after the performance of the organization turns out to be low compared to an aspiration level. The aim of problemistic search is to improve the performance. According to this theory, the extent to which problemistic search is stimulated moves up and down with the performance of the organization: problemistic search increases when the organization appears to “perform below the aspiration level” and decreases when the organization “performs above the aspiration level” (Greve, 2003, p. 55). Since problemistic search varies along with the organizational performance, problemistic search activities can be initiated by giving performance feedback on organizational goals. According to Greve (2003), performance feedback can be given based on aspiration levels. Therefore, based on these behavioral theories there can be hypothesized that financial scarcity stimulates decision-makers to search for creative solutions to improve the financial situation of the organization (Cyert & March, 1963).

Another theoretical ground is the prospect theory by Kahneman and Tversky (1979), which theorizes about risk-taking by decision-makers. According to this theory, the willingness to take risks changes in reaction to performance feedback (Greve, 2003, Kahneman & Tversky, 1979). This theory implicates a reflection effect, meaning that “risk aversion in the positive domain is accompanied by risk-seeking in the negative domain” (Kahneman & Tversky, 1979, p. 268). According to this theory, decision-makers therefore make choices based on the fact if the organization is in a current state of loss or gain. Decision-makers of an organization facing loss will tend to seek risk. In contrast, decision-makers of an organization achieving gains will be more averse towards risk (Mone et al., 1998). A study of Wiseman and Bromily has confirmed this theory and showed that especially a “decline in financial resources increases risk-taking” by decision-makers in declining organizations (1996, p. 538). In line with the prospect theory, it can therefore be expected that decision-makers of organizations facing financial scarcity are more open to trying creative solutions that generally entail more risk.

Hypothesis 2: Financial scarcity has a positive effect on creativity of decision-makers. In the next chapter, the design and data collection procedure of both the survey experiment and the expert survey will be explained. Moreover, the data analysis methods will be explained that will be used to test the collected data on the formed hypotheses.

C

3.

M

In this chapter, the methodology of the study will be explained. This study is based on two surveys: a survey experiment and an expert survey. In the survey experiment, historical or social performance feedback about financial scarcity is used as treatment and policy ideas are collected as qualitative data. The collected policy ideas are analyzed on content and appropriateness to the task and then categorized in 31 categories. In the expert survey, these idea categories are scored by two groups of experts on the dependent variable ‘creativity’ on a quantitative 1-10 scale. This chapter will be roughly divided into four parts. In paragraphs 3.1 to 3.3 will be elaborated on the design, case and unit of analysis of the survey experiment. Then, in paragraphs 3.4 will be elaborated on the variables of the study. Following, in paragraphs 3.5 to 3.7 the ways to measuring creativity, the design, and the reliability and validity of the expert survey will be discussed. Finally, in paragraph 3.8 the data analyses-methods used to test the hypotheses will be explained.

3.1 DESIGN OF THE SURVEY EXPERIMENT

The ideas used within this study are collected through a large-scale, online survey experiment distributed in November 2019. In the survey experiment, the environment for the treatment group is manipulated before collecting the data. In this way, it is possible to have “some control over the environment” (Toshkov, 2016, p. 167). In the survey experiment, the treatment group is manipulated with negative financial performance feedback to measure the effect of financial scarcity on creativity of the collected data.

A large-scale, online survey experiment design fits this study for several reasons. Through an online survey distributed on a large scale, it is possible to reach all Dutch municipal councilors and recruit them as respondents in an efficient way. An experiment-design is fitting since the random assignment of participants into the treatment and control group, “automatically controls for all possible alternative explanations and confounders”, even for confounders that are irrelevant for the study or that are unknown (Toshkov, 2016, p. 172). Through random assignment, groups are formed that only differ in the intervention (Toshkov, 2016). Second, experimental designs rule out reversed causation. It is possible to show that the experimental manipulation causes the perceived difference in the outcome and not vice versa since the intervention is shown to participants of the treatment group first before the outcome is collected (Toshkov, 2016).

Via e-mail, all Dutch municipal councilors were invited to participate in a survey about the view of municipal councilors on the implementation of youth care in their municipality. The participants were randomly divided into a treatment and control group. The treatment of the survey experiment communicated the actual budgetary status for youth care of the municipality to municipal council members based on three ranks (Table 1). As the youth care policy domain is characterized by severe resource scarcity, almost all respondents received negative financial performance feedback (92.7%). For municipalities that overspend the budget, but did not spend more than average based on municipalities of equal size, historical performance feedback was given that stated that the municipality overspent its budget in the past year (51.8%) (Table 2). For those municipalities that overspent the budget and also spent more than average compared to municipalities of equal size, social performance feedback was given, communicating that the municipality overspent its budget in 2018 and also spent more than municipalities of comparable size (40.9%). Participants randomly assigned to the control group received information unrelated to financial scarcity: they received information about the distinction between customized youth care services and escalated youth care.

Figure 1. Survey experiment-flow

Table 1.

Distribution of participants

Participants Treatment group Control group Total

Rank (1) No budget shortage 36 37 73

(2) Shortage but smaller than average 273 243 516

(3) Shortage and larger than average 210 197 407

Table 2.

Distribution of participants

Rank Description Percentage Feedback treatment group

1 No budget shortage 7.3% Positive

2 More spent than budgeted, but budget shortage was smaller than average

51.8% Negative historical performance feedback 3 More spent than budgeted and budget shortage was

larger than average 40.9%

Negative social performance feedback

After the intervention phase, all municipal councilors were asked to generate up to 5 ideas to improve youth care in their municipality. Finally, a manipulation check was done to test to what extent the participants are aware of the financial situation for youth care of their municipality in 2018. The manipulation check (Table 3-5) shows that within rank 2, more participants answer the question correctly in the treatment group (90.5%) than in the control group (81.9%). This difference is statistically significant; t(514) = -2.148, p = .032 (Table 3). Therefore, the manipulation of the treatment group in rank 2 can be seen as effective. As for the participants of rank 3, no difference was found between participants that answered the question correctly in the treatment group (91.0%) and the control group (89.8%); t(405) = -.317, p = .752 (Table 3). Table 3.

Mean difference answers on manipulation check between treatment and control group

Group Rank 2 p-value Rank 32 p-value

Control hist. (n=243) 1.79 (.485) .032 Control social (n=197) 1.87 (.420) .752 Treatment hist. (n=273) 1.87 (.422) Treatment social (n=210) 1.88 (.403) Table 4.

Results of manipulation check: treatment group

“What was the financial situation regarding youth care in your municipality in 2018?”

Total The realized expenses

remained within the budget

The realized expenses exceeded the budget

I don’t know

Rank No budget shortage 28 7 1 36

Shortage < average 17 247 9 273

Shortage > average 13 191 6 210

Table 5.

Results of manipulation check: control group

“What was the financial situation regarding youth care in your municipality in 2018?”

Total The realized expenses

remained within the budget

The realized expenses exceeded the budget

I don’t know

Rank No budget shortage 19 18 0 37

Shortage < average 36 199 8 243

Shortage > average 14 177 6 197

Total 69 394 14 477

3.2 CASE OF THE STUDY

The research question of this study will be applied to the case of youth care in the Netherlands. In 2015, the Dutch government decentralized the administrative and financial responsibility for youth care to the municipalities (Nijendaal, 2014). With this decentralization, more than 15% of the costs for youth care have been cut back by the government (VNG, n.d.a.). The reasoning behind this decentralization is that municipalities are closer to citizens and the local community than the central government. Therefore, municipalities could provide youth care more effectively and efficiently (Government of the Netherlands, n.d.a.). Moreover, the decentralization of youth care tasks should enable Dutch municipalities to develop integrated policies and offer well-coordinated care geared to local and individual situations and needs. The goal is to create more coherent, effective, transparent and less expensive services for children and families (Netherlands Youth Institute, n.d.).

Over the past few years, most municipalities appeared to substantially spend more on youth care than budgeted (Peeters, Batterink, Schumacher, & Tazelaar, 2019). Two-third of the municipalities even has a budget shortage of more than 20% and one in five municipalities has a deficit of more than 40% (Kieskamp, 2019). According to Dutch municipalities themselves, they suffer from deficits of more than 600 million euros collectively (Kieskamp, 2019). These deficits make youth care a fitting policy domain to study the effect of financial scarcity on creativity of decision-makers.

3.3 UNIT OF ANALYSIS

Decisions affecting the municipality are made by members of the municipal council. Municipal councilors are decision-makers that are representative for and chosen by residents through an election held every four years (Government of the Netherlands, n.d.b.). The unit of analysis in this study is the municipal councilor as an individual decision-maker at a Dutch municipality.

The relevant population consists of all municipal councilors working for a Dutch municipality (n = 8619). The 8619 municipal councilors together form the largest group of politicians in the Netherlands (Nederlandse Vereniging voor Raadsleden, n.d.). For this study, a voluntary response sample is used. Data of 2066 participating municipal councilors of Dutch municipalities that responded on the survey (total response rate = 24.0%) is used. After filtering out respondents with missing data, 996 respondents are left that collectively generated 3.272 ideas (effective response rate = 11.6%). The data shows that municipal councilors joined from 301 of the 355 municipalities. Therefore, 84.8% of the municipalities are represented in this study. The average amount of ideas generated by the different groups is presented in Table 6. The results of an independent t-test show that no significant difference is present in the average amount of ideas generated by the treatment group and control group, for the respondents of both rank 2 and 3 (Table 6). The distributions of the demographical characteristics of the respondents of rank 2 are presented in Table 7. The distributions of the demographical characteristics of the respondents of rank 3 are presented in Table 8. Independent t-tests show that no significant differences exist (all p-values > .05) on the demographical variables between the treatment and control group in both rank 2 and 3 (Table 7-8). Based on these balance checks, there can be concluded that the treatment and control group of rank 2 and 3 are equal based on the number of ideas generated and on the distribution of the demographical characteristics, also on demographical characteristics that were not observed with this survey. Concluding, this balance check ensures that differences that are found between treatment and control groups are caused by the independent variable.

Table 6.

Distribution of the number of ideas generated by the different groups

Respondent group Average ideas generated p-value

Control gr. historical (n = 243) 3.280 (1.433) .843 Treatment gr. historical (n = 273) 3.304 (1.342)

Control gr. social (n = 197) 3.239 (1.381) .235

Table 7.

Description of the sample – demographical characteristics rank 2

Characteristic Group Distribution p-value

Sex Man Woman Other/private 64.9% 32.9% 2.2% .958

Age (in years) Mean

Standard deviation Range 48.54 31.52 19 – 78 .931

Political orientation Strongly left Left

Not left, not right Right Strongly right Missing 6.2% 25.6% 45.7% 19.2% 0.6% 2.7% .773 Coalition or opposition Coalition Opposition Not applicable/other Missing 52.9% 43.6% 2.3% 1.2% .757 Years of working experience Mean Standard deviation Range 6.23 6.35 1 – 38 .268 Table 8.

Description of the sample – demographical characteristics rank 3

Characteristic Group Distribution p-value

Sex Man Woman Other/private 66.8% 31.0% 2.2% .242 Age (in years) Mean

Standard deviation Range 49.55 31.47 20 – 81 .936

Political orientation Strongly left Left

Not left, not right Right Strongly right Missing 7.6% 26.5% 40.07% 22.4% 1.0% 2.5% .871 Coalition or opposition Coalition Opposition Not applicable/other Missing 57.7% 37.3% 3.7% 1.2% .488 Years of working experience Mean Standard deviation Range 6.51 6.41 1 – 46 .602 3.4 VARIABLES

Independent variable. The independent variable in this study is financial scarcity. In the experiment is intervened with this variable, by giving municipal councilors in the treatment group negative financial performance feedback based on the budgetary status of 2018 for youth

care in their municipality. The independent variable is categorical (Toshkov, 2016). There are four different groups to differentiate between. The historical performance feedback-treatment group (rank 2) will be compared to the control group of rank 2 and the social performance feedback-treatment (rank 3) compared to the control group of rank 3.

Dependent variable. The main dependent variable in this study is creativity of policy ideas generated by individual decision-makers. Besides, also the effectiveness and feasibility of the ideas will be measured. In the survey, municipal councilors were asked to formulate up to five policy ideas to change or improve youth care in their municipality. The three dimensions are measured on a 1 to 10 scale by having the generated ideas rated by two groups of experts on the youth care-domain: alderman with youth care in their portfolio, and administrators and managers of youth care-organizations. The dependent variables in this study are continuous, interval variables (Toshkov, 2016). On the measurement of creativity and the expert survey will be further elaborated in paragraphs 3.6, 3.7, and 3.8.

To make the quality assessment of all ideas manageable, the qualitative output of the survey experiment is analyzed before included in the expert study. Usable ideas are categorized on content through an iterative process into 31 idea-categories. On how these categories were formed is elaborated in Appendix 1. An idea was categorized as unusable when the response was regarded as not appropriate to the given task, for example when the response was too vague to filter out the core message, when the response was a diagnosis of the problem instead of an idea to solve it, or when the response was too specific for the municipality of the municipal councilor to add it to a more general idea category. The inclusion and exclusion criteria and the categorization process are further explained in Appendix 1.

3.5 METHODOLOGY FOR MEASURING CREATIVITY 3.5.1TECHNIQUES TO MEASURE CREATIVITY

Previous studies that aimed to define and assess creativity through an experimental design, can be divided by two most used assessment techniques. The majority of experimental creativity studies used “creativity tests” (Amabile, 1982 p. 998). These assessment tests can be compared to standard intelligence tests. In these tests, the participant is given several, mostly verbal, assignments that are scored on scales of pre-determined creativity indicators (Amabile, 1982). The second, less common assessment technique for creativity is by using experts “to provide

problem of creativity, since experts are asked to assess creativity according to their definition of creativity (Amabile, 1996). Although being subjective, creativity is measured directly and without the intervention of creativity indicators (Amabile, 1982). In contrast, in creativity tests, creativity is measured through indicators that are claimed to be objective but have a subjective nature instead (Amabile, 1982). In this study, creativity will therefore be assessed through experts.

3.5.2THE OPERATIONALIZATION OF CREATIVITY

To measure creativity of the ideas directly and bypass the criterion problem, the consensual assessment technique of Amabile (1982) will be used. In this method, the generated ideas will be assessed by experts. Creativity is operationalized by experts themselves, by providing for subjective grading of the creativity of the ideas:

“A product or response is creative to the extent that appropriate observers independently agree it is creative. Appropriate observers are those familiar with the domain in which the product was created or the response articulated. Thus, creativity can be regarded as the quality of products or responses judged to be creative by appropriate observers […].” (Amabile, 1982, p. 1001).

The definition of creativity by Amabile (1982) is based on two important assumptions. First, for this definition there must be assumed that “it is possible to obtain reliable judgements of the creativity of ideas” (Amabile, 1982, p. 1001). This means that creativity in ideas, although hard to define, can be recognized by people familiar with the domain when they see it. Therefore, experts should be able to agree with another on the amount of creativity recognized (Amabile, 1982). Secondly, there is assumed in this assessment method that creativity exists on a spectrum. Therefore, ideas can be more or less creative than others (Amabile, 1982).

3.5.3PROCEDURAL REQUIREMENTS FOR CREATIVITY ASSESSMENT BY EXPERTS

Several procedural requirements exist when using this method. First, the experts participating in the expert survey should all have expertise in youth care. The amount of expertise does, however, not necessarily have to be identical (Amabile, 1982). Besides, the experts should assess independently and they may not be trained by the researcher to agree with other experts about the meaning of creativity. Also, the researcher may not give the experts specific criteria for judging creativity in ideas (Amabile, 1982). Randomization is also important: each judge should assess the ideas in a random order (Amabile, 1982). The expert survey of this study meets these requirements.

3.6 EXPERT SURVEY

3.6.1DESIGN OF THE EXPERT SURVEY

Creativity of the ideas is measured via an online expert survey, distributed in May 2020. Expert surveys offer opportunities when the aim is to collect data about “difficult-to-measure phenomena” (Meastas, 2018, p. 3). An expert survey-design is chosen for this study since creativity is difficult to measure through observational or literature study designs. Through an expert survey, experts with “specialized experience or knowledge” on the domain of interest can help to still be able to explore creativity (Maestas, 2018, p. 3). There is chosen to let experts score all 31 ideas. This design is beneficial since measures based on more raters per item are more valid and reliable. By obtaining an aggregated score from a larger group of individual scores, potential biases on the individual level can be compensated (Meastas, 2018).

However, asking experts to rate all items instead of randomly presenting a part, largely increases the estimated completion time of the survey to 20-25 minutes, which is quite long for surveys. Previous research based on experts as a tool of measurement shows that it is more important to increase the number of raters per unit when the concept that experts are asked to score is complex (Maestas, Buttice & Stone, 2014). Therefore, two means are used to stimulate responses of experts and to boost the overall response rate (Maestas et al., 2014). First, the benefit is outlined that completing the survey can be of interest to the experts and/or organizations themselves since it includes a summary of thousands of ideas to improve youth care. Second, the opportunity was offered to receive the results of the study afterwards.

Matrix questions are chosen for the expert survey to make scoring the three dimensions per idea more accessible. Every matrix question is introduced with the line: “Youth care in municipalities can be changed and improved by…” followed by the content of the

randomization can be advantageous for some reasons, this design is deliberately chosen to make scoring all 31 ideas more accessible. The order in which the ideas are presented is randomized. An example of a matrix question is presented in Appendix 2.

3.6.2DEFINITION AND RECRUITMENT OF THE “EXPERT”

In this study, “expert” is defined as a professional with specialized knowledge on the domain of youth care. Two target groups are selected for this study. First, aldermen that are portfolio holder of ‘youth care’, to provide for a political-administrative perspective. Second, directors and managers of youth care organizations holding a professional perspective. No further requirements were made to potential participants of these target groups. The term “expert” is therefore defined quite broadly. This is beneficial for two reasons. First, it makes it possible to form target groups consisting of experts with “diverse and independent perspectives” on the domain of interest (Maestas, 2018, p. 6). Second, to make sure sufficient potential raters could be invited to participate in the study.

The survey is distributed to 355 aldermen of youth care from all Dutch municipalities and to 128 youth care organizations located all over the country (via ‘Youth care Netherlands’ and ‘Association neighborhood teams’) (Associatie Wijkteams, n.d.; Jeugdzorg Nederland, n.d.). The youth care organizations were requested to forward the survey to directors and managers of the organization. The survey was distributed by sending information about the study and an anonymous hyperlink to the appropriate target-group survey via e-mail addresses found on the internet. One week after the first distribution, a reminder was sent to all potential respondents.

An exact amount of judge-response needed to reach reliable data is hard to decide in advance. When looking at previous studies using an expert methodology, the number of judges included varies considerably. The number of experts invited in previous studies of Amabile (1996) with offline assessment settings, is between 6 and 40. In expert surveys, the number of experts is in general much higher (Meastas, 2018).

3.6.3DATA COLLECTION

In the expert survey, experts are asked to assess the quality of the ideas presented, by assigning scores on a scale of 1 to 10 according to three dimensions: creativity, effectivity and feasibility. Before asking to score the ideas, the three dimensions were presented to the experts together with a “nonrestrictive definition” provided for each dimension (Amabile, 1996, p. 56).

Table 9.

Dimensions of assessment

Dimension Nonrestrictive definition given to experts

Creativity The degree to which the idea is novel and original

Effectiveness The degree to which you expect the idea to be efficacious

Feasibility The degree to which the idea is practically and politically achievable Asking the experts to assess the ideas on two dimensions besides creativity makes it possible to check if the assessment of creativity can be separated from judgements of other relevant aspects. In this way, there can be checked if the discriminant validity of the assessment is valid (Amabile, 1982). The dimensions effectiveness and feasibility can be linked to the two logics where public sector innovations are organized around, as described in paragraph 2.1.1: the logic of consequences and the logic of appropriateness (March & Olsen, 1989).

Two different expert surveys, both consisting of 40 questions, were developed for each target group. Each survey version consists of a consent question, six general demographical questions, two demographical questions specified for the target group and 31 matrix questions to score the ideas on creativity, effectiveness and feasibility. Before distributing, a general version of the survey was pretested and revised for context and clarity by two laypeople with no familiarity on the youth care-domain, three researchers of the Institute of Public Administration of Leiden University, one alderman for youth care and one manager of a youth care-organization. The content of the two final versions of the expert survey is attached in Appendix 2. The survey is designed and the data is collected with Qualtrics software, Version May 2020 of Qualtrics (Qualtrics, Provo, UT).

The expert survey was distributed in May 2020. After being online for 2 weeks and a reminder of the survey to the potential participants, the survey was closed and the data was collected. Table 10 summarizes the number of survey responses. Tables 11 and 12 present the distribution of general and specific demographical characteristics.

Table 10.

Summary of the expert-survey responses

Expert group Number of distributions

Bruto responses Netto responses Effective response rate

Aldermen 355 162 111 31.3%

Table 11.

Description of the sample – general demographical characteristics

Characteristic Group Aldermen Managers

Sex Man Woman Other/private 56.5% 38.7% 1.8% 48.0% 52.0% 0.0% Age (in years) Mean

Standard deviation Range 53.44 10.40 23 – 77 48.67 8.09 32 – 65 Education High school

mbo hbo/wo bachelor hbo/wo master Doctorate Missing 2.7% 6.3% 39.6% 47.7% 1.8% 1.8% 0.0% 2.0% 32.0% 62.0% 2.0% 2.0% Province (work) Drenthe Flevoland Friesland Gelderland Groningen Limburg Overijssel Noord-Brabant Noord-Holland Utrecht Zeeland Zuid-Holland Missing 5.4% 0.9% 2.7% 17.1% 1.8% 9.0% 5.4% 16.2% 9.9% 9.0% 3.6% 17.1% 1.8% 0.0% 4.0% 4.0% 8.0% 2.0% 10.0% 12.0% 16.0% 14.0% 10.0% 0.0% 18.0% 2.0% Years of working experience Mean Standard deviation Range 4.93 5.22 0 – 37 16.63 10.90 0 – 42 Table 12.

Description of the sample – specific demographical characteristics aldermen and managers

Characteristic Group Distribution

Aldermen

Political orientation Strongly left Left

Not left, not right Right Strongly right Missing 2.7% 31.5% 50.5% 11.7% 0.0% 3.6% Managers Function Director Manager Missing 40.0% 56.0% 4.0%

After the data collection, the individual expert-scores on the three dimensions are, per expert group, aggregated towards an average score on the dimension per idea (Toshkov, 2016). Although the measurements of creativity, effectiveness and feasibility are subjective, through accumulation of the many different expert opinions about these dimensions, the final score is

aimed to converge with the ‘true’ creativity, effectiveness and feasibility of the ideas (Toshkov, 2016).

3.7 RELIABILITY AND VALIDITY 3.7.1INTERJUDGE RELIABILITY

After the scores of the experts are collected, the scores on each dimension are “analyzed for interjudge reliability” (Amabile, 1996, p. 43). Although error to some degree is unavoidable when using experts as a tool of measurement, the scores on the dimensions must be sufficiently reliable (Meastas, 2018). The measurement of interjudge reliability is in this study used as equal to construct validity (Amabile, 1996). The importance of interjudge reliability for creativity is also implied by the operational definition used: “A product or response is creative to the extent that appropriate observers independently agree it is creative. […]” (Amabile, 1982, p. 1001). The interjudge reliability is analyzed by calculating the Cronbach’s coefficient alpha for the creativity dimension, through the reliability procedure in SPSS by including the creativity scores of the aldermen and managers apart for all ideas. This procedure is repeated for effectiveness and feasibility.

The results of the interjudge reliability are presented in Table 13. The creativity, effectiveness and feasibility scores in both expert groups have high interjudge reliability. The interjudge reliability of the creativity dimension is .927 for the aldermen and .942 for the managers. The interjudge reliability score of effectiveness by the managers is above .70. The rest of the interjudge reliability scores is above .80. Therefore, there can be concluded that the experts independently agree on the scores on all dimensions. Concluding, the subjective ratings by both expert groups on all dimensions can be assumed to be reliable.

Table 13.

Interjudge reliabilities per expert group for the three dimensions of assessment

Expert-group Dimension of Assessment Interjudge reliability

Alderman Creativity .927 Effectiveness .826 Feasibility .851 Managers Creativity .942 Effectiveness .742 Feasibility .821

3.7.2DISCRIMINANT VALIDITY

The interdependence between creativity, effectiveness and feasibility is measured through the bivariate Pearson Correlation of the three dimensions. By analyzing this interdependence, or the discriminant validity of the dimensions, there is examined if the three theoretically distinguishable dimensions are in the data also distinguishable (Hubley, 2014). This is done by checking the Pearson Correlation between the three dimension scores. Although it is likely that the judgement of creativity is strongly correlated with effectiveness and feasibility, it is still important that it is possible to separate these dimensions. This is important to make sure that experts score an idea high on creativity because they see it as creative, regardless of if the idea is perceived as effective and feasible or not. The interdependence between creativity, effectiveness and feasibility will be measured by the bivariate Pearson Correlation of the three dimension scores in SPSS. The correlations are presented in Table 14.

Table 14.

Correlations of the dimensions creativity, effectiveness and feasibility

Creativity Effectiveness Feasibility

Creativity _{Pearson Correlation 1 .728}** _.210

Sig. (2-tailed) .000 .102

N 62 62 62

Effectiveness Pearson Correlation .728** _{1 .433}**

Sig. (2-tailed) .000 .000

N 62 62 62

Feasibility Pearson Correlation .210 .433** ₁

Sig. (2-tailed) .102 .000

N 62 62 62

**. Correlation is significant at the 0.01 level (2-tailed).

As Table 14 shows, there exists a significant, strong and positive correlation between effectiveness and creativity (r = .728; p = 0.000; N = 62). Besides, also a significant correlation exists between effectiveness and feasibility. This correlation is positive and the strength is moderate (r = .433; p = 0.000; N = 62). Finally, the data shows that creativity and feasibility are positively and weakly correlated (r = .210; p = 0.102; N = 62).

3.8 DATA-ANALYSIS METHODS

After collecting the scores from the expert survey, average scores per idea for the three dimensions will be calculated for each expert group apart in SPSS. Thereafter, the average effectiveness, creativity and feasibility score per idea-category and expert group is added to all

the ideas in the SPSS-dataset of the survey experiment. Following, the frequency in which the ideas are generated in the four experimental groups is analyzed. Then, the hypotheses are tested through independent t-tests in SPSS. Recalling the hypotheses:

Hypothesis 1: Financial scarcity has a negative effect on creativity of decision-makers. Hypothesis 2: Financial scarcity has a positive effect on creativity of decision-makers. In the next chapter, the results of the independent t-tests will be presented.

C

4.

R

In this chapter, the results of the study will be presented. The average level of creativity, effectiveness and feasibility of the ideas generated in the four experimental groups that will be compared, depends on two aspects. First, the aggregated scores for creativity, effectiveness and feasibility per idea as scored by the alderman- and manager-expert group. These will be presented in paragraph 4.1. Second, the frequency in which the ideas are generated in the four groups. These will be presented in paragraph 4.2. In paragraph 4.3, the results of the hypothesis tests will be presented. All results will be presented apart for the alderman-expert group and the manager-expert group. The chapter will be closed with a summary of the results.

4.1 AGGREGATED SCORES PER IDEA

To be able to test the effect of financial scarcity presented by performance feedback on the creativity, effectiveness and feasibility of the ideas generated by decision-makers, the aggregated scores per idea for both expert groups are computed. The mean scores per idea are calculated with the descriptives function in SPSS. All scores per idea are presented in Appendix 5. The data shows that on average, the aldermen score the ideas higher on creativity (M = 5.803, SD = 0.784) than the managers (M = 5.435, SD = 0.786). This is also the case with the effectiveness and feasibility scores, although the mean differences between expert groups for the effectiveness scores (M_aldermen = 7.042, SD = 1.172; M_managers = 6.841, SD = 1.178) and for the feasibility scores (M_aldermen = 6.595, SD = 1.109; M_managers = 6.405, SD = 0.991) are smaller. The average scores on creativity, effectiveness and feasibility per expert group are presented in Table 15.

Table 15.

The mean scores of the ideas on creativity, effectiveness and feasibility, per expert-group

Expert group Dimension Aldermen Managers Mean SD Mean SD Creativity 5.803 0.784 5.435 0.786 Effectiveness 7.042 1.172 6.841 1.178 Feasibility 6.595 1.109 6.405 0.991

Most creative ideas. In Table 16, the three most and least creative, effective and feasible ideas according to the expert groups are presented. The table shows that category 24 is scored as the most creative idea by both expert groups. This idea calls for improving youth care by integrating or aligning the youth care-domain in municipalities with other municipal-level policy domains, for example education, housing, or poverty reduction. The second and third most creative ideas differ per expert group. Category 7 and 8 are scored high by the aldermen. These ideas argue

that youth care can be improved by making more use of prevention (category 7) and primary youth care by professional care actors (category 8). Category 14, stating that youth care can be improved by reducing overhead costs in municipalities and youth care organizations, and category 15, stating that youth care can be improved by changing the market forces that are currently present in youth care, are scored high on creativity by the managers (Appendix 5). Table 16.

Most and least creative, effective and feasible ideas according to each expert group

Expert/ Position

Aldermen Managers

Creativity Effectiveness Feasibility Creativity Effectiveness Feasibility

Highest Highest

1 Category 24 Category 25 Category 7 Category 24 Category 14 Category 7

2 Category 7 Category 7 Category 8 Category 14 Category 25 Category 30

3 Category 8 Category 5 Category 5 Category 15 Category 15 Category 14

Lowest Lowest

29 Category 18 Category 18 Category 20 Category 23 Category 12 Category 20

30 Category 9 Category 9 Category 19 Category 19 Category 20 Category 29

31 Category 29 Category 29 Category 29 Category 12 Category 19 Category 19

Least creative ideas. The least creative idea as scored by the aldermen is category 29. This idea advocates that youth care can be improved by cutting back on budgets of other policy domains within the municipality to save more budget for youth care. The second least creative idea according to the aldermen is category 9, which implies that youth care can be improved by focusing more on specialized youth care to prevent that a too low level of care is used and the youth care process therefore takes too long. The third least creative idea, category 18, advocates for decentralizing the organization and financing of youth care back to the provinces or the central government. Category 12, the least creative idea according to the managers, states that that youth care can be improved by monitoring the costs made by youth care-organizations. The second and third least creative ideas according to the managers are category 19 (“youth care can be improved by decentralizing all youth care rules to create policy freedom at the municipal level, rather than regional collaborations or joint arrangements) and 23 (“youth care can be improved by increasing the level of quality monitoring in youth care organizations, for example through result-oriented quality agreements”). The content of the ideas that are most effective and feasible according to both expert groups are presented in Appendix 5. The distribution of the creativity, effectiveness and feasibility scores of the ideas per expert group are presented in Figures 3, 4 and 5.

Figure 3. Distribution of the mean creativity scores, as scored by aldermen (1) and managers (2)

Figure 5. Distribution of the mean feasibility scores, as scored by aldermen (1) and managers (2)

4.2 FREQUENCY OF THE GENERATED IDEAS PER EXPERIMENTAL GROUP

Another aspect affecting the average level of creativity, effectiveness and feasibility per experimental group, is the frequency that ideas of an idea-category are generated in each experimental group. In Appendix 3, a table is presented with the frequency of the 31 ideas in each of the four experimental groups. In Figure 6, the distribution of the frequencies of the ideas present in the four experimental groups are presented. This figure shows that category 16 is by far most present within all experimental groups. This category proposes to improve the youth care situation by strengthening the access to youth care, for example by forming a stricter definition of what falls under youth care or by having more grip on given indications. Other ideas that are frequently generated in all experimental groups are categories 5, 8, 14 and 25 (Appendix 3, 5). The most striking result when analyzing the frequencies that the ideas are present in the treatment and control groups is that the distributions at first view seem to be quite similar. When analyzing the distribution per idea-category, the distribution is largely the same when comparing treatment and control groups. Only small differences can be identified between the treatment and control groups. These differences will be elaborated on in chapter 5.

4.3 HYPOTHESIS TESTS

In this paragraph, the results of several independent t-tests conducted via SPSS are presented. With these tests, the mean score for creativity, effectiveness and feasibility of the generated ideas by municipal councilors in the treatment group are compared with the mean scores of generated ideas by municipal councilors in the control group. The effect of financial scarcity on the creativity, effectiveness and feasibility of the ideas is tested with the aggregated scores of the aldermen and managers apart. First, the treatment and control groups are compared based on the scores of the aldermen. Then, the same is done based on the scores of the managers. 4.3.1EXPERT GROUP:ALDERMEN

In this paragraph, there is tested for differences between treatment and control groups based on the creativity, effectiveness and feasibility scores of the aldermen. To test the effect of financial scarcity on creativity, the average creativity scores of the treatment groups that received historical or social performance feedback are compared with the average creativity scores of corresponding the control groups. First, the treatment group that received negative historical performance feedback is compared to the corresponding control group. There is found based on the scores of the aldermen, that the average creativity score of the historical treatment group (M = 5.931, SD = .495) is lower than the average creativity score of the historical control group (M = 6.019, SD = .452) (Table 17). An independent samples t-test shows that this difference is significant on the 0.05 level; t(496) = 2.051, p = .041 (Table 17). Then, the treatment group that received negative social performance feedback is compared to the corresponding control group. Based on the scores of the aldermen is found that the average creativity score of the social treatment group (M = 5.950, SD = .436) is lower than the average creativity score of the social control group (M = 6.042, SD = .408) (Table 17). An independent samples t-test shows that this difference is significant on the 0.05 level; t(396) = 2.165, p = .031 (Table 17). Following, there is tested for differences in effectiveness and feasibility scores between the treatment and control groups. On all tests, the results showed that the average scores for effectiveness and feasibility are lower for the ideas generated by participants that received negative financial performance feedback (Table 17). The t-tests showed that these differences in all cases were not significant on the .05 significance level (Table 17). However, it is notable that the mean difference between treatment and control group, based on the scores of the aldermen, is larger for the historical treatment and control group (MD_effectiveness = .088, p = .151; MD_feasibility = .080, p = .138) than for the social treatment and control group (MD_effectiveness = .064, p = .310;