The influence of management concepts' language on change resistance: adopting agility in the public sector Open Access

The transfer of management concepts from the private sector into the public sector is a constant topic of debate (Pollitt and Bouckaert, 2017; Mathiasen, 1999; Radnor and Osborne, 2013). A recent example is the management concept of agility, which has attracted a great deal of attention from both public management practitioners and scholars (Soe and Drechsler, 2018; Nolte and Lindenmeier, 2023; Mergel et al., 2021). Past examples demonstrate that the adoption of management concepts from the private sector in public administrations can result in tensions, partial implementation or even complete failure (Pollitt and Bouckaert, 2017). In the current public management debate, these tensions are explained mainly by structural or normative differences at the macro level (Mergel et al., 2021; Baxter et al., 2023). At the same time, there is growing interest among researchers in studying language and its effects in an administrative context (Andersen and Jakobsen, 2017; Eckhard and Friedrich, 2024; Holzinger, 2020). This study argues that language, particularly its contextualization, plays a fundamental yet under-researched micro-level role in triggering resistance. By ignoring the impact of linguistic framing, public management research overlooks a critical factor that might influence the adoption of management concepts. The aim of this study is, therefore, to close this research gap and to empirically investigate the role of language in the adoption of management concepts. This study is based on an interdisciplinary approach, linking institutional translation theory with organizational behavior research, to empirically investigate the role of linguistic framing on resistance.

The main research question is: To what extent does the language of management concepts affect resistance to the adoption of such concepts among civil servants? This research question is investigated in a survey experiment that tests the impact of language on change resistance. In the survey experiment, civil servants were confronted with a change initiative involving the adoption of agile project management. In a 2 × 2 factorial design, they were randomly assigned to four groups. Each group was shown one of the four vignettes. The vignettes differed in terms of the language (agile vocabulary or public sector contextualized vocabulary) and the label (presence or absence of the label “agile”) used to describe the change initiative. 4,196 civil servants from German municipalities participated in the survey experiment. Multiple linear regression analyses were performed to investigate if there is a causal effect of agile language on resistance to change.

This study makes three key contributions. First, it contributes to the theory of the translation of ideas within institutional theory. Although the theory suggests that the contextual adaptation of management concepts is essential for their successful adoption (Røvik, 2016; Wedlin and Sahlin, 2018), thus far, research has primarily relied on qualitative case studies describing the translation process (Wæraas and Nielsen, 2016). Our research provides causal evidence for a central yet previously assumed mechanism of this process. By isolating language as a variable, we demonstrate that the way in which an idea is expressed linguistically is a powerful driver in overcoming employee resistance.

Second, this study offers a new explanation for implementation barriers in public sector reforms. The public management literature predominantly explains resistance to concepts imported from the private sector through structural and normative incompatibilities (Mathiasen, 1999; Mergel et al., 2021; Radnor and Osborne, 2013). Our study extends this established macro-perspective with a more granular level of explanation: linguistic incompatibility. We demonstrate that the foreign-sounding jargon of a concept can trigger resistance on its own, regardless of its substantive fit. While the intensity of this effect may vary across national administrative cultures, this finding is generalizable beyond agile management methods and offers a new, highly relevant practical perspective on the persistent challenges of implementing administrative reforms.

Third, our study bridges disciplinary boundaries by connecting macro- and micro-analytical perspectives on organizational change. We link a key concept from institutional theory (the translation of ideas) with a classic phenomenon in organizational behavior (resistance to change) within a public management context. This interdisciplinary approach allows for a richer understanding of how institutional macro-trends, such as the call for more agility, are mediated by linguistic micro-mechanisms that shape employee attitudes and behaviors at the individual level.

Debates on the reform of public administration are often closely linked to the adoption of management concepts. The drivers and outcomes of such reforms have been the subject of intense debate in public administration and public management research (Pollitt and Bouckaert, 2017; Nielsen et al., 2020). Agility is a current and prominent example of how management concepts are being taken up by public administration (Mergel et al., 2021; Neumann et al., 2024). The diffusion of agility is driven by digitalization and the call for a more adaptable administration (Soe and Drechsler, 2018). In this study, we understand agility as a management concept that provides a framework for iterative working, valuing people and evidence-based learning (Mergel, 2024).

While agility has changed various industries, the transfer to the public sector requires a nuanced understanding of its foundational elements and the specific challenges of cross-contextual application (Baxter et al., 2023; Hegele and Stoll, 2024). The concept of agility became particularly popular through initiatives in the information technology (IT) sector with the aim to make software development more flexible and efficient (Madsen, 2020). Agility is often associated with specific methods, such as Scrum. Originally invented for software product development, Scrum is now used as a framework for complex problems (Schwaber and Sutherland, 2020). It can be used in its pure form or be modified to specific conditions in the public sector (Baxter et al., 2023). Scrum is a process in which a part of a product can be planned, produced, and reviewed in a few weeks. The aim is to create value through these short iterations and gain experience by introducing adaptive solutions. These advantages of agility, such as speed, adaptability and interdisciplinary collaboration contrast with rigid, rule-based administration.

Introducing external management concepts into the public sector often encounters resistance due to the perceived structural and normative incompatibility with the inherent characteristics of these organizations. The public sector is characterized by features such as political influence, a legal framework, an asymmetric relationship with citizens, hierarchical structures, and risk avoidance (Mergel et al., 2021; Mathiasen, 1999). These macro-level tensions between private sector management concepts and public sector characteristics often manifest as resistance at organizational and individual levels (Nielsen et al., 2020). However, civil servants may respond in different ways to new management concepts. The range goes from those who adopt them enthusiastically to those who actively resist them (Powell et al., 2013). Due to the assumed tension between agility and bureaucracy (Mergel et al., 2021), we focus on resistance, which is defined as a negative attitude towards a change initiative, consisting of affective, cognitive and behavioral reactions (Piderit, 2000; Oreg, 2006).

This study argues that language plays a more fundamental yet under-researched role in mediating the adoption of management concepts by influencing how they are perceived and internalized, beyond the structural incompatibilities discussed in public management research (Baxter et al., 2023; Mergel et al., 2021). In line with institutional theory, and specifically the Scandinavian institutionalism approach, management concepts are not simply copied but must be translated to fit with the new organizational context (Sahlin and Wedlin, 2008; Czarniawska and Sevón, 1996). A key aspect of translation theory is the decontextualization of concepts from their origin and their contextualization within the adopting organization (Røvik, 2016). This process ensures that the concept fits the adopting organization better and reduces resistance (Wedlin and Sahlin, 2018). This study considers two interpretations of “translation”. The first is a broad contextual translation, viewing translation as the adaptation of the entire management concept to the organization's structures, processes, and culture. The second is a narrower, operational interpretation: linguistic translation, referring to the adaptation of specific terminology and labels to the host organization. The study's focus on the linguistic dimension (e.g. word choices and labels) is based on two important research gaps. While the public management debate acknowledges language as a point of conflict between civil servants and citizens (Eckhard and Friedrich, 2024; Döring, 2021), these linguistic tensions have not been sufficiently studied in the context of internal communication (e.g. change communication). Moreover, although contextual translation is believed to reduce resistance, the specific ways in which this occurs, particularly through linguistic adaptations, have not been explored enough at the empirical micro-level.

The linguistic incompatibility between the language of bureaucracy and the specialized vocabulary of modern management concepts creates a significant cognitive and social barrier during adoption. Both public sector organizations and the concept of agility have their own language. Redish (1983) concludes that the language of bureaucracy is an expression of a profession and has a specific style based on nouns, jargon, and laws. This language often causes conflicts when used with people who are not familiar with it. Agile language is also distinctive. Relevant publications such as the Agile Manifesto or the Scrum Guide are in English. These concepts are also commonly used in English, even in non-English speaking countries. However, the degree of openness towards such anglicisms differs significantly between national contexts. In addition, agile frameworks use a certain vocabulary (e.g. Scrum Master, Backlog, Sprints) and need their own glossaries to explain these words (e.g. Scrum.org). In contrast to the more complex bureaucratic language, the language of agile frameworks tends to be rather simple, with short sentences, and a more active writing style. However, the usage of agile vocabulary complicates this language. This is not just about English versus German, but about unfamiliar professional jargon versus established bureaucratic language, leading to comprehension difficulties for any professional language introduced from an outside context. The uniqueness and difference between agile and bureaucratic language could lead to conflict when adopting the concept of agility in public sector organizations. There is a cognitive burden, as unfamiliar technical language makes information absorption difficult, leading to confusion and potential rejection (Bullock et al., 2019). But also there is a social dimension, the use of specific language signals “in-group” versus “out-group” status (Bourdieu, 1991), potentially leading to feelings of exclusion or alienation for those unfamiliar with the agile jargon (Holzinger, 2020).

Drawing on framing effects from organizational behavior and psychology, we propose that the way a change initiative is presented linguistically can significantly influence its reception, independent of its substantive content. In the context of change, a frame influences how people perceive a change and react to it (Chreim, 2006). But conscious framing to achieve certain reactions (such as support for the change) also plays an important role (Andersen and Jakobsen, 2017). Frames are used to emphasize certain content elements of the change, such as the rationale for it or the connection to certain values of the organization (Andersen and Jakobsen, 2017; van der Voet, 2021). We argue that the selection of certain words and labels already evokes reactions and serves as a frame in itself. In a conceptual study, Vann (2004) theorized that differences in the language used by project management and public sector organizations can lead to resistance toward implementing project management in the public sector. The importance of language is also demonstrated by recent studies on agility in administration, which have observed “allergic” reactions to agile vocabulary (Kühler et al., 2024, p. 11) and reveal a fundamental linguistic conflict (Ewens, 2024).

Based on linguistic incompatibility, the psychological effects of unfamiliar jargon, and the benefits of contextualization, we hypothesize that explicit use of agile vocabulary will impact civil servants' resistance to change. Given the established cognitive burden (Bullock et al., 2019) and the “in-group” versus “out-group” dynamics (Bourdieu, 1991) stemming from unfamiliar agile vocabulary, we hypothesize that its use will activate negative affective and cognitive reactions. This is consistent with recent studies showing that language tailored to target groups is easier to understand (Eckhard and Friedrich, 2024), and with observations that there is resistance to agile language (Ewens, 2024; Kühler et al., 2024).

In line with the principles of translation theory (Røvik, 2016), which advocates for adapting concepts to the host context, we propose that using contextualized vocabulary (i.e. linguistic translation) will minimize cognitive load and foster a sense of consistency (Liguori et al., 2018). This linguistic fit is expected to reduce perceived threat and enhance acceptance. Thus, we formulate the following hypothesis:

Building on the analysis of framing effects and the symbolic power of management labels, we argue that explicitly labeling a change initiative with a management concept term, such as “agile”, influences resistance among civil servants, regardless of the specific vocabulary used. Here, the main mechanism is not cognitive load or issues of familiarity, as in H1, but the symbolic activation of institutional conflicts.

As they spread, management concepts can become institutionalized myths or fashions (Madsen, 2020; Sahlin and Wedlin, 2008). The label “agile” acts as a symbolic framework, providing a cognitive shortcut that instantly activates pre-existing mental schemas and emotional responses (Tversky and Kahneman, 1981).

In the context of the public sector, this label is strongly associated with the institutional logic of the private sector, particularly with values such as speed, disruption, and iteration. This logic is in direct opposition to the dominant institutional logic of public administration, characterized by principles such as legal certainty, stability, political influence, hierarchical structures and risk avoidance (Mathiasen, 1999; Mergel et al., 2021). Specifically, Baxter et al. (2023) highlight that the incongruence between agile practices and conventional governance methods in public sector IT projects leads to enduring tensions, supporting the notion of institutional conflict.

Thus, the mere use of the term “agile” signals not just a methodological change, but also a clash of fundamental institutions and the perceived imposition of an alien logic on the organization. Civil servants perceive the label as a threat to the institutional identity and traditional values of the bureaucracy. Consequently, resistance is primarily directed against this symbolic power, as well as the felt pressure to adapt to an external logic that is considered unsuitable (Neumann et al., 2024; Powell et al., 2013). This results in an emotional defensive reaction creating an “us versus them” dichotomy (Ewens, 2024).

Conversely, deliberately avoiding this loaded label neutralizes the symbolic tension. This approach shifts the focus from the provocative signal word to the substantive core of the change, enabling a more neutral perception of the initiative.

This leads to the following hypothesis:

We used a quantitative approach to test our hypotheses. The study was preregistered via AsPredicted (https://aspredicted.org/fh9k-gpst.pdf).^[1] A survey experiment was conducted to examine the relationship between the use of agile language and resistance to change. This approach is suitable for our research endeavor as it allows us to uncover causal relationships through experimental manipulation while leveraging the representativeness of a large-scale survey among civil servants to enhance external validity (Jilke and van Ryzin, 2017). The data were collected from civil servants in German municipalities. To improve the response rate, we collaborated with the German think tank Kommunale Gemeinschaftsstelle für Verwaltungsmanagement (KGSt). KGSt is an independent association that supports its members in the modernization of public administration through recommendations, indicator systems and the exchange of experience between municipalities. More than 2,600 German municipalities are members of the KGSt. The survey experiment was distributed to 28,261 civil servants through a mailing list of the KGSt. The survey experiment was conducted in November and December 2024 and one reminder was sent. 4,198 civil servants participated in the online survey, leading to a response rate of 14.9%. The participants represent 1,130 different local administrations, which vary in terms of state, size, and type (city, county, or municipality). Compared to the overall population of local civil servants, who are 45 years old (Statistisches Bundesamt, 2024) the participants are slightly older at 47 years. Furthermore, civil servants in management positions are overrepresented in the participant group (57% compared to approximately 5–15%). To address this overrepresentation, we included covariates and conducted subgroup analyses.

The case of German municipalities was selected for three reasons. Firstly, the German administration is law-oriented and tends to evolve incrementally (Pollitt and Bouckaert, 2017; Kuhlmann et al., 2021). Although the German administration has developed in the direction of a Neo-Weberian state, it has been observed in recent years that classical Weberian elements such as hierarchy and regulation are gaining in importance (Kuhlmann, 2024). These classical Weberian characteristics are at odds with agility, so we expect to see tensions making Germany a most likely case. Secondly, agility is a significant issue in Germany and especially among German municipalities (Kühler et al., 2024; Hegele and Stoll, 2024). We assume that there would be a wide range of responses, from civil servants who have not yet heard of this concept to those that rely heavily on it. Lastly, Germany has an official language that is not English. The German language is deeply embedded in the public sector (Ratzmann, 2018). Consequently, language tensions from the introduction of unfamiliar jargon, such as that found in agile practices, may be more readily observed. While this makes Germany a suitable case to investigate linguistic incompatibility, we are aware that the reception of anglicisms varies across different national contexts.

In the survey experiment, respondents were confronted with a hypothetical change initiative. They should imagine that a new way of working is being introduced for dealing with complex projects. We created four experimental groups, and each group was shown a different vignette. Respondents were randomly assigned to one of the four vignettes. We used a 2 × 2 factorial design that is detailed in Figure 1. Although the vignettes all depict the same change initiative –the introduction of agile project management– they differ systematically in their linguistic framing. This is reflected in the vocabulary used (whether agile-specific terms or public-sector-specific language is used) and the presence or absence of the “agile” label.

The agile label consists of a heading that describes that agility is to be introduced and two notes in the text that also refer to this. Vignettes without label did not have this heading and these notes. The other feature is agile vocabulary or contextualized vocabulary. Agile vocabulary was used from Scrum Guide which is one of the most popular agile project frameworks (Dikert et al., 2016). We have created contextualized vocabulary to replace the agile vocabulary. This was based on existing translations from municipalities and glossaries. With the help of experts on agility from different municipalities, four vignettes that are typical for the practical application were generated. For example, “Scrum Master” has been rendered as “Methodenverantwortlicher”, suggesting that this person is responsible for methods, while “Product Owner” has been conveyed as “Fachverantwortlicher”, suggesting that they are responsible for a specific subject. Other instances include the term “Sprint” that was interpreted as “Arbeitszyklus”, which denotes a cycle of work, and “Sprint Reviews” was expressed as “Ergebnispräsentation”, which indicates the presentation of results. The aim of the contextualization was to use German terms that were closer to the technical language of the civil servants, but also to try to keep the meaning identical. The vignettes can be found in Appendix A.

After seeing the vignette, respondents were encouraged to give their opinion on the change initiative described in the vignette. Respondents indicated their level of agreement with five statements using a five-point Likert scale, ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). The items included: (1) “I believe that change would harm the way things are done in the organization”, (2) “I think that it's a negative thing that we are going through this change”, (3), “I believe that the change would make my job harder”, (4), “I believe that the change would benefit the organization” and (5) “I believe that I could personally benefit from the change”. Items (4) and (5) were reverse-scored. The statements were drawn from Oreg's (2006) change attitude scale. Oreg differentiates three dimensions: affective, behavioral, and cognitive. We decided to use only the five items from the cognitive dimension, because we presented our respondents with a hypothetical change and did not ask about their behavior or feelings in situations of change that they had experienced. The scores on these five statements were summarized in a change resistance index. Cronbach's alpha was 0.88, which demonstrates acceptable scale reliability.

In addition, several questions were asked to obtain covariates. Age and tenure were measured in years. English proficiency was measured with the item “How would you rate your English skills?” Responses were recorded on a six-point scale ranging from “1: very poor” to “6: native speaker”. Management position was transformed into a dummy variable: being in a management position was coded as “1”, not being in such a position as “0”. In addition, respondents were asked about their experience with agile management methods. Respondents' self-reported experiences were captured on a five-point Likert scale, ranging from “1: mostly negative” to “5: mostly positive”. It was also possible to indicate whether there was no experience with agility and whether someone did not know what agile working methods are. The size of the municipality was measured by the population of the municipality in which the respondents work. The survey items are listed in Appendix B. The data on the size of the municipality was not collected in the survey as we were able to use a KGSt database linked to the respondents. A series of multiple linear regression models were used to analyze the data. This approach is suitable for testing the main effects of our two independent variables (vocabulary and label) and their potential interaction on resistance to change, while controlling for relevant covariates such as age, tenure, English proficiency, management position, and municipality size.

The total number of observations is 4,196. The data of two participants was deleted due to impossible answers regarding their age. Balance tests were conducted to check if the randomization of the experimental groups was successful. Table 1 illustrates relevant data for the randomization of the experimental groups. ANOVA was used to verify if the continuous covariates (age, tenure, agile experience, size of municipality) were balanced and a Chi-Square test was executed for the categorical covariates (English proficiency and management position). While most covariates were evenly distributed, the tests indicated potential imbalances regarding English proficiency and a lack of agile experience. These potential imbalances were further investigated (see Appendix C). A Tukey honestly significant difference (HSD) test was used for the continuous variable English proficiency. However, there was no further evidence of a statistically significant imbalance. Regarding imbalances concerning the absence of agile experience, the residuals of the Chi-Square test show that participants without agile experience are overrepresented in group 1 and underrepresented in group 4. As this might influence the results, additional models for the analyses of the data are used. The inclusion of the covariates is one solution (Gerber and Green, 2012), while another is to hold those covariates constant by using subgroups (Mutz, 2011). Therefore, our regression analyses include models with covariates and subgroup analyses to account for these potential imbalances.

Figure 2 shows mean scores for resistance to change across experimental groups with 95% confidence intervals. Those mean scores range from 2.24 to 2.40 on a five-point Likert scale and indicate that on average there is more support than resistance regarding the mentioned initiative to implement agile methods in the working environment of the participants. In addition, the confidence levels between groups with and without agile vocabulary (group 1 and group 3 compared to group 2 and group 4) are not overlapping, which suggests a statistically relevant difference between those groups. However, there is an overlap if groups are distinguished with or without agile labels (group 1 and 2 compared to group 2 and 4).

The distribution of agile experience is illustrated in Figure 3. Out of 4,196 respondents, 2,973 indicated they have experience with agility. 1,048 reported having no experience, 175 do not know what agility is. To summarize, among respondents who reported experience with agile working methods, 59% described their experiences as positive, 6% as negative, and 35% as mixed.

Table 2 shows the results of a regression analysis. Model 1 tests if the usage of agile vocabulary and agile label has causal effects on change resistance. A baseline score of 2.246 on a five-point scale further suggests that there is not a strong resistance to change around this agile initiative. Agile vocabulary (b = 0.150) has effects on change resistance and achieves statistical significance. However, agile labels do not have statistically relevant effects on resistance to change.

Models 2, 3, 4 and 5 differ in terms of agile experience and whether covariates are included. 175 participants who responded that they do not know what agility is were excluded from those models. In all models, agile vocabulary has a positive and statistically significant impact on change resistance. In general, models 2 to 5 show the effect of agile vocabulary is stronger for participants without prior agile experience than for those with such experience (b = 0.180 compared b = 0.109). Agile label has no statistically significant effect in any of those models. Model 3 and 5 show that the covariates age, tenure, and agile experience have a statistically significant relationship with resistance to change. While age and tenure have a positive relationship with resistance, agile experience has a negative relationship. Especially a positive agile experience demonstrates a strong negative correlation with change resistance (b = −0.416), indicating a significant relationship (see also Appendix F).

Hypothesis 1, which states that agile vocabulary increases resistance to change, is supported across all models, as indicated by the consistently positive and statistically significant effect for agile vocabulary compared to the usage of contextualized vocabulary. Hypothesis 2, expects that the agile label increases resistance to change, is not supported, as the agile label variable consistently shows no statistically significant effect across any model.

To verify the robustness of our results, further tests were conducted. First, the models were screened for outliers. These were subsequently removed, and the models were re-estimated and compared (see Appendix D). All models remained robust. Second, further subgroup tests were conducted (see Appendix E). The participants were divided into managers and non-managers. These results suggest that holding a management position significantly moderates the influence of agile vocabulary and agile label on resistance to change. However, this moderating effect is contingent upon agile experience. The significant interaction effects were only observed in the subgroup with agile experience. In a subgroup without agile experience, this moderation could not be statistically verified. Third, various interaction effects were examined (see Appendix F). The two dependent variables have no interaction effect. Furthermore, there was no interaction between the variable agile experience and agile vocabulary.

Our findings support the argument that the language of a management concept affects its implementation (Vann, 2004). The multi-case study from Kühler et al. (2024) and the case study from Ewens (2024) observed resistance against agile vocabulary in the context of public organizations. Our study reinforces these findings for a larger population. In addition, as we tested agile vocabulary against public sector contextualized vocabulary, we also provide evidence that linguistically contextualized management concepts create a better fit for the adopting organization (Røvik, 2016; Wedlin and Sahlin, 2018). This is also consistent with studies on change management and administrative reform, which assume that the way in which a change is communicated is relevant (Andersen and Jakobsen, 2017; Armenakis and Harris, 2002).

The second key finding is that the label agile does not affect resistance. We expected the agile label to have a positive effect on resistance to change. However, in all models the relationship was negative, although not statistically significant. This result contradicts findings from the qualitative study by Neumann et al. (2024), where respondents claimed that the label agility is negatively perceived in public organizations. The qualitative study of Kühler et al. (2024) also showed one case where an organization used agile methods but deliberately avoided the agile label. There are three explanations for our finding regarding the label. First, there is a group against agility that is vocal but not large enough to be statistically significant. This is also consistent with the literature, which generally assumes that, in addition to resistance, there is also support for change (Pollitt and Bouckaert, 2017). These critical voices may stand out in qualitative studies but may be lost in a quantitative setting. Second, it could be that because agility is already widespread and civil servants have had positive experiences, they may not feel any resistance to the management concept of agility and therefore do not resist the label. Third, another explanation is that the label has become so generic in its usage that it is no longer clear what it refers to (Madsen, 2020), which means that no clear effects can be identified.

Our study has implications for theory and practice dealing with public management. Our research contributes to the literature on how management concepts are translated and institutionalized in the public sector. We have shown that a linguistically contextualized management concept has a better fit towards civil servants and lowers change resistance. This is in line with Scandinavian institutionalism, which claims that management ideas get translated into the adopting organizations (Wæraas and Nielsen, 2016; Czarniawska and Joerges, 1996). Our findings provide empirical evidence for the micro-level mechanisms of translation, showing how linguistically contextualized concepts reduce cognitive burden and foster greater social inclusion, thereby leading to lower resistance. Following translation theory, it would be assumed that each host organization uses its own translation (Røvik, 2016). Our study already shows positive effects with a general translation and thus supports the study of Nielsen et al. (2020), which emphasizes the importance of so-called field translations for adopting organizations. Furthermore, Scandinavian institutionalists have a broad understanding of the term translation, which goes beyond translating words (Wedlin and Sahlin, 2018). The institutionalist argument that translation includes more than just linguistic translation should nevertheless incorporate the centrality of language in the adoption of management concepts, a factor which our study empirically proves to be causally relevant.

Furthermore, the literature assumes that management concepts and the public sector are fundamentally different, which gives rise to resistance (Mergel et al., 2021; Mathiasen, 1999; Radnor and Osborne, 2013). This also applies to the adoption of agile working methods, since agility and bureaucracy were presented as opposites. Studies generally highlight specific characteristics of bureaucracy, such as a strong hierarchy and a rigid organizational culture (Mergel et al., 2021; Baxter et al., 2023; Soe and Drechsler, 2018). However, we found empirical evidence that language serves as a crucial, yet often overlooked, additional feature at the micro-level. The specific linguistic culture of public administration presents a significant barrier to the introduction of specialized vocabulary from other fields, mediating the tension between agility and bureaucracy.

Our study also contributes to the body of literature dealing with change management in public administration. In their change model, Fernandez and Rainey (2006) recognize the importance of communication, but they focus on strategic communication and the use of specific content or arguments. Kuipers et al. (2014) pointed out in their meta-analysis that linguistic incompatibility leads to resistance, but possible solutions are not listed. Consequently, the current literature risks overlooking the impact contextualized language has on achieving readiness for change in public sector reform. This is also relevant because, due to the increasing pressure of digitalization, we can assume that further technical or management concepts will be introduced to the public sector that are not linguistically congruent.

Our findings also have implications for practice. Reforms in public administration, such as the introduction of agile working methods, are often associated with skepticism (Pollitt and Bouckaert, 2017). Our results illustrate that a significant number of civil servants in Germany have gained experience with agility and that they are more open than resistant to this concept. Practical advice for public managers could therefore be to create a framework in their organizations where people can experience agility. The study also has implications for change communication. Public managers should be aware that how they communicate is as important as what they communicate. Contextualizing vocabulary can help to reduce potential resistance. But it is also a dilemma. How much originality is needed to deliver the promised benefits, and how much translation is needed to keep resistance to change at an acceptable level? One possible solution to this dilemma could be to begin with contextualized language in order to overcome initial resistance. After initial experiences, managers should make these linguistic adjustments explicit and explain the original language usage.

While this study provides unique insights into the influence of agile language on resistance to change, it also has some limitations. As this survey experiment relies on vignettes of a constructed change initiative, different results may be found in the real world (van der Voet, 2021). Additional qualitative studies as well as field experiments in real-world settings could address these shortcomings. When researchers examine agile change in single or multiple case studies, they could explore the diverse use of language that goes beyond vignettes. Speeches by public leaders, the content of training courses and discussions in meetings can all be places to explore the impact of language in the real world.

Another limitation is that we cannot rule out the possibility that some of the factors causing resistance are due to civil servants perceiving agile vocabulary as a foreign language. German civil servants are encountering partly English agile vocabulary, so we attempted to account for this influence by using the control variable of English language proficiency. However, this control variable had no significant effect. Furthermore, the single-country focus of this study, which is based on German civil servants, limits the generalizability of the findings. German civil servants' openness towards the English language and jargon may differ from that of civil servants in other countries. To increase generalizability, future research should investigate the clash of such languages in a comparative setting. This should include an English-speaking environment to establish whether the effects persist when the vocabulary is not perceived as foreign. It should also include a non-English-speaking country with a different cultural affinity for English terminology, in order to better establish the role of national context and linguistic bias.

A final limitation is that we measure the effect of agile language on resistance at a specific point in time. Management concepts, especially those that are labeled as management fashions, undergo different phases (Madsen, 2020). We investigated this topic at a time when the concept had already become established in Germany. The impact of language may be dynamic. There may be more resistance at the beginning of implementing the concept, and less once it has already spread. Future studies should investigate the impact of language on change resistance in a longitudinal setting. As the concept of agility may already be diffused, a longitudinal setting may be particularly interesting for other management concepts that are relatively new and not yet widespread. This could entail a process-oriented longitudinal case study following the implementation, or the application of repeated survey experiments to analyze how resistance develops over time.

This study provides a novel perspective on the implementation of management concepts in the public sector by isolating the effect of language from the actual substance of a method and other external factors. Our findings reveal a significant nuance in administrative reform, as civil servants in our sample are generally open to the term “agility” but resist the use of non-contextualized agile vocabulary. This suggests that the perceived clash between agility and bureaucracy may be a linguistic incompatibility at the micro-level rather than a conflict of values or goals.

This research's distinctive value lies in identifying language as a barrier to reform implementation. By providing empirical evidence for the mechanisms of translation theory, we demonstrate that successful reform depends on linguistic fit. For scholars, this opens a new area of study by investigating why reforms fail even when their core ideas are accepted. For practitioners, our study offers a clear strategic pathway that encourages managers to prioritize the linguistic contextualization of new concepts before introducing specialized jargon to mitigate resistance to change. Ultimately, this research highlights that successful implementation of reform in the complex environment of public administration hinges not only on what the reform aims to achieve, but also on how it “speaks”.

The supplementary material for this article can be found online.