M.Sc. Business Administration
Organizational and Management Control
2018/19
EBM859A20
submission date: June, 24th 2019
University of Groningen - Faculty of Economics and Business
dr. Martijn van der Steen
master´s thesis:
“The intensity of sensemaking in crisis situations: a quantitative analysis of
the effects of crises on the sensemaking process”
words: 12.310
A
Content
Abstract
1 Introduction 1
2 Literature review 2
2.1 Federal Reserve System and Federal Open Market Committee 3
2.2 Crisis situations 3
2.3 The different dimensions and microprocesses of sensemaking 5
2.4 Noticing 6
2.5 Interpreting 7
2.5.1 Collective dimension 8
2.5.2 Retrospective dimension 9
2.6 Acting 11
2.7 Sensemaking and crisis situations 12
3 Methodology 12
3.1 Connection between language and sensemaking 12
3.2 Introduction into LIWC 13
3.3 Dataset 13
3.4 Selection of crisis situations 14
3.5 Measurement of variables 18
3.5.1 Intensity of the noticing microprocess of sensemaking 18
3.5.2 Intensity of the collective dimension of sensemaking 18
3.5.3 Intensity of the retrospective dimension of sensemaking 19
3.5.4 Intensity of the acting microprocess of sensemaking 19
3.5.5 Intensity of sensemaking 20
4 Findings 21
4.1 Structure of analysis 21
4.2 Analysis of mean comparison 22
5 Discussion 24
6 Conclusion 27
6.1 Limitations and further research 27
6.2 Overall Conclusion 28
Appendix A
B
Table of figures
Figure 1: S&P 500 in USD from March 2007 until November 2009 (daily data). ... 15 Figure 2: GDP of the USA in billions of USD measured quarterly between the years 1996 and 2012. ... 16 Figure 3: Intensity of Sensemaking over time. ... 21
Table of tables
Table 1: Overview of crisis selection and time periods. ... 17 Table 2: Overview of findings. ... 24
Abbreviations
FED Federal Reserve System
FOMC Federal Open Market Committee
GDP Gross Domestic Product
ISM Intensity of Sensemaking
ISMa Intensity of the Sensemaking microprocess of acting ISMc Intensity of the collective dimension of Sensemaking ISMn Intensity of the Sensemaking microprocess of noticing ISMr Intensity of the retrospective dimension of Sensemaking
LIWC Linguistic Inquiry Word Count
C
Abstract
1
1 Introduction
Word-work is sublime, … because it is generative; it makes meaning that secures our difference, our human difference – the way in which we are like no other life. We die. That may be the meaning of life. But we do language. That may be the measure of our lives (Morrison, 2019).
This extract from Toni Morrisons Nobel Prize acceptance speech in 1993 (Morrison, 2019) focusses on the creation of meaning and the importance of language to measure for this. As such, it suits as an installation into this research that investigates sensemaking by measuring it through the use of words. Sensemaking is about answering the question “what is going on?” (Colville, Pye, & Brown, 2016) and generally defined to be processual in nature (Gephart, Topal, & Zhang, 2010; Maitlis, 2005; Weick, 1995). This process works as a group activity in which actors exchange ideas and interpretations about what happened, and what is going on, in order to create meaning in uncertain environments (Maitlis, 2005; Maitlis & Christianson, 2014). However, much of the existing research investigates sensemaking only in the context of one specific situation. This results in treating it as it could be turned “on”. Theory suggests that it is constantly ongoing and never turned “off” (Louis, 1980; Weber & Glynn, 2006). In this research, sensemaking will be measured through the intensity of its process. This acknowledges that there are constantly sensemaking activities while emphasizing that the intensity of the process can be increased.
Theory suggests that one trigger that increases the intensity of sensemaking activities is the occurrence of crisis situations (Christianson, Farkas, Sutcliffe, & Weick, 2009). Every crisis situation is unique but similarities between different qualities of crisis situations exist. This study relies on two crisis situations to demonstrate differences between the qualities: The Financial Crisis that started off with a burst of the American housing market bubble in 2007 (Ramírez, Selsky, & Heijden, 2010) and the Iraq War in which an alliance of 46 countries lead by the United States of America invaded Iraq in 2003 (Newnham, 2008). This allows to distinguish between different qualities of crises and compare such differences with regards to their effect on the intensity of the sensemaking process.
2
As the subject of research, the Federal Open Market Committee (FOMC) is chosen. The FOMC is the monetary policy-making organ of the Federal Reserve System (FED) (Federal Reserve Board, 2017), and released transcripts of its meetings and conference calls over the past decades (Cannon, 2015). In this study, this unique data set is utilized and the sensemaking activities of the FOMC in crisis situations are analysed. More precise, the impact of qualities of the Iraq War and the recent Financial Crisis on the intensity of sensemaking is investigated using a computer-aided text analysis program (Linguistic Inquiry and Word Count). This program analyses every word of the transcripts and assigns words to pre-defined categories. These categories represent linguistic, psychological and time-related features (Mehl, 2006) that function as a linkage between the qualities of crisis situations and the intensity of the sensemaking process. This paper thereby explains how an important political institution varies in its intensity of sensemaking depending on the given situation. Implications of this research highlight which conditions intensify the sensemaking process. It is a step forward towards the improvement of sensemaking activities in difficult circumstances such as crisis situations. If beneficial conditions can be created and the effect of disadvantageous conditions can be minimized, the overall intensity and efficiency of sensemaking can be improved, increasing the probability of good decision-making.
The goal of the study is to implement the novel approach of measuring sensemaking through its intensity. Further goals are to strengthen our current understanding of the qualities of crisis situations as cues for sensemaking and how they affect the consecutive process. Additionally, a deepened insight on the cognitive and group processes of the FOMC is given. Therefore, the research question reads:
RQ: What are the effects of different qualities of crisis situations on the intensity of sensemaking?
2 Literature review
3
2.1 Federal Reserve System and Federal Open Market Committee
The FED is the central bank system of the United States of America (Ehrmann & Fratzscher, 2007). It helps to promote the effective operation of the country´s economy and serves the public interest. It conducts the monetary policy and stabilizes the financial system in general as well as individual financial institutions in particular. Additionally, it supports both payment and settlement system safety and encourages consumer protection and community development (Federal Reserve Board, 2017).
The FOMC is the committee of the FED that is concerned with monetary policy and consists of 19 members of which 12 have voting rights (Ehrmann & Fratzscher, 2007). Seven of them are appointed by the president of the United States of America (Abolafia & Hatmaker, 2013). These members have to be confirmed by the Senate and then keep their positions for a time span of 14 years. The remaining five members are presidents of one of the 12 regional banks. All presidents of the 12 regional banks attend the meetings but only five of them have voting rights at the same time (Abolafia, 2010). The president of the New York Fed has one of the five voting rights permanently whereas the presidents of Chicago and Cleveland alternate on an annual basis. The remaining nine presidents rotate for voting rights (Gerdesmeier, Mongelli, & Roffia, 2007). In their meetings that are held regularly every six weeks and take one- or two-days’ time, the committee discusses and decides upon monetary issues related to the economy (Abolafia & Hatmaker, 2013). Additionally, the FOMC has telephone conferences whenever it is deemed necessary, for example when the chairman senses that immediate action is required (Cannon, 2015; Federal Open Market Committee, 2016). The FOMC is responsible for open market operations. It helps the FED to influence the availability of money and credit and to achieve national economic goals (Federal Reserve System, 2019). Due to its focus on monetary policy, the FOMC has gained a lot of expertise in economic related topics, which is also the main field of knowledge of most of its members (Federal Open Market Committee, 2016).
2.2 Crisis situations
A crisis situation is defined as a disruptive event that only has a low probability of occurrence but which effect can be significant (Maitlis & Sonenshein, 2010; Pearson & Clair, 1998). The threat of such significant effects demands immediate attention. The ability to respond immediately, however, is reduced because of the ambiguity and complexity of the situation (Pearson & Clair, 1998). Crisis situations are accompanied by negative emotions such as sadness or guilt and also by more intense ones such as fear, panic and desperation (Heaphy, 2016; Weick, 1993).
4
the focus lies on the distinction between three main qualities: First, crises differ whether they happen suddenly or are more of an unfolding nature. Second, crises can be economical or political and therefore either be closely connected to the core competency of the FOMC (economical) or further detached from it (political). Third, the emotional quality of crises can differ so that one crisis has a more positive emotional state than the other one.
Unfolding crisis situations do not have one clear trigger event in the beginning. Instead, they consist of a series of cues that in retrospective can explain a path that led to the disruptive event (Pearson & Clair, 1998). Moreover, actors have more influence in shaping the situation through their enactment if a crisis is unfolding over time (Weick, 1988). Ramírez et al. (2010) provide an overview of the development of the recent Financial Crisis. They state that the crisis was preceded by a series of sequential events that started off when the “housing bubble” in the United States of America began to burst in 2006. As financial institutes lost trust in each other, the inter-bank lending of money and later on also lending activities to households were reduced. This phenomenon is described as a credit crunch. Due to the decreased availability of credit, the refinancing of leveraged activities became more insecure which led to liquidity problems of credit institutes that had a high debt to equity ratio. These liquidity problems occurring in mid of August 2008, are commonly described as the Financial Crisis. The bankruptcy of the major investment bank Lehman Brothers on the 15th of September 2008 can be seen as an accelerator, that spread panic and further increased the costs of borrowing (Ramírez et al., 2010).
Keeping in mind that the focus of the FOMC lies on monetary policy and the observation of the financial markets, the Financial Crisis unfolded over time for the members of the FOMC because they were closely connected to the preceding events. Contrary, the beginning of the Iraq War was sudden despite the existence of some indicators that already pointed at a military conflict (Paleari, Redondi, & Vismara, 2005). On the 27th of January 2003, the Blix report got published, reporting that the Hussein-regime had failed to disarm, which led to an increased possibility of a military operation. However, the first explosions in Baghdad on the 20th of March 2003 and the announcement of the American President George W. Bush, which stated the beginning of the armed conflict were unexpected (Paleari et al., 2005).
5
Moreover, the accessibility to resources such as oil and the political importance of such resources played an important role in the reasoning behind the outbreak of the war (Jhaveri, 2004).
The third quality that is used in this paper to qualify crisis situations is the emotional level. In general, crisis situations are accompanied by intense negative emotions (Christianson et al., 2009; Mirbabaie & Zapatka, 2017; Weick, 1993). Nevertheless, it is possible to differentiate on the emotional level. A military conflict that directly results in the loss of thousands of lives of soldiers and civilians is negatively emotional per se. The Financial Crisis had a negative influence on the economy but was missing the extreme negative trigger event. Moreover, because the core competencies of the FOMC are closely connected to the Financial Crisis, its members are better prepared for such a situation. This preparation leads to confidence which results in a more positive emotional state in the Financial Crisis.
2.3 The different dimensions and microprocesses of sensemaking
In its purest fashion, sensemaking is about finding answers to the question “what is going on?” by both making use of thinking and acting (Colville et al., 2016). The concept of sensemaking is often used as a lens to explain connections between phenomena in related fields (Sonenshein, 2009; Stensaker & Falkenberg, 2007). The literature defines sensemaking in varying ways and even uses different terminologies for the concept (Colville et al., 2016; Holt & Cornelissen, 2014; Mills & Weatherbee, 2006; Mills, Weatherbee, & Colwell, 2006; Schultz & Hernes, 2013; Stieglitz, Mirbabaie, & Milde, 2018; Weick, Sutcliffe, & Obstfeld, 2005). These differences in the perspective on sensemaking and wording about sensemaking lead to an extensive amount of definitions. Maitlis and Christianson (2014) point out, that the decision on “how to study sensemaking is itself an exercise in sensemaking” (p. 105). Despite widespread differences, many of the existing definitions share a common core which will be outlined in the following paragraphs.
6
However, most research on crisis situations only investigates sensemaking in one separated crisis (Christianson et al., 2009; Mirbabaie & Zapatka, 2017; Weick, 1993). Because sensemaking never stops, signs of sensemaking have to be visible in every moment, disregarding whether a crisis situation is present or not. The existent literature, however, often treats sensemaking as it would be an activity that switches its status from “off” to “on” upon the occurrence of a crisis situation. This implies that sensemaking activities would be turned “off” again, once the crisis situation is over. This theoretical inconsistency can be obviated by looking at the intensity of sensemaking. Thus, a constant level of sensemaking is ongoing, which can be intensified by various triggers (Maitlis, 2005; Weick et al., 2005). This research especially directs attention towards how the different qualities of crisis situations affect the intensity of sensemaking.
The following subsections aim to deepen the understanding of the dimensions and microprocesses of sensemaking1 and their interplay.
2.4 Noticing
A group of actors can constantly make sense related to one topic whereas the occurrence of another event focusses the sensemaking activities towards this event, increasing the intensity of the sensemaking process (Weick, 1995). To intensify the sensemaking process, a new sensemaking opportunity has to be noticed. This happens through cues or triggers. The cause of this noticing microprocess is rooted in the actors´ expectations (Weick, 1967). If expectations are not met, actors tend to look for possible explanations (Louis, 1980; Maitlis & Christianson, 2014). Arousal as a form of discharge in the autonomic nervous system is then provoked which redirects the actors´ attention (Maitlis & Sonenshein, 2010; Weick, 1995). This provides a stimulus to dissolve the gap between the actors´ expectation and their perceived reality to recreate predictability (Maitlis & Christianson, 2014).
In literature, the notion of identity threat as a sign of a clear violation of expectations is heavily discussed (Corley & Gioia, 2004; Pratt, Rockmann, & Kaufmann, 2006). It is argued that uncertainty about the identity of oneself or of the organization is closely related to the actor. Cognitive actions are undertaken to reshape the image of identity (Rao, Monin, & Durand, 2003) or adjust environmental conditions and interpretations to align with the actors perspective (Elsbach & Kramer, 1996).
Crisis situations can have such a disruptive character that they threaten identity and challenge existing institutions, routines and thinking patterns (Christianson et al., 2009), because “people suddenly
1 Differentiation between dimension and microprocess: In this paper, I use the term “microprocess” to refer to the
7
and deeply feel that the universe is no longer a rational, orderly system” (Weick, 1993, p. 633). As the quote shows, crises lead to a violation of expectations and confusion on how to react in the occurring situation. Two explanations for this confusion are the concepts of ambiguity and uncertainty. Ambiguity is defined as materializing in situations in which the actor cannot explain a deviance successfully because he is overwhelmed by too many possible explanations (Corley & Gioia, 2004; Weick, 1995). Situations are ambiguous if the actor has more than one possible option to proceed but is uncertain about which option would be the appropriate response (Weick, 2015). Ambiguity increases if multiple actors are involved because they can understand reality in different ways and therefore have a differing set of responses available (Merkus et al., 2017). Weick (1995) connects ambiguity and equivocality as both are about situations in which more than one interpretation is accessible. Following other scholars, this paper assimilates equivocality and ambiguity (Weick, 1995; Zong & Demil, 2015). The related concept of uncertainty occurs if the actor is facing a problem for which he is unaware of any possible interpretation (Weick, 1995).
Both ambiguous and uncertain situations have in common that they hinder effective decision making (Weick, 1995) and are present during crisis situations. Neither the cause of the situation, nor their effects, nor their resolutions are foreseeable at the moment of appearance (Pearson & Clair, 1998). They are a strong sign that expectations are not met and at the core of the noticing microprocess of sensemaking (Maitlis & Christianson, 2014). Instances of violated expectations, arousal, ambiguity and uncertainty in crisis situations, while keeping in mind that sensemaking activities are an ongoing process for which reason their intensity is the varying element, result in the following hypothesis:
H1: The sensemaking microprocess of noticing has a higher intensity in crisis situations than in non-crisis situations.
2.5 Interpreting
Interpreting is the microprocess that provides order to information extracted from cues (Maitlis & Christianson, 2014; Weick, 1995). It is about meaning creation for something that has occurred during the process but has not been defined yet. As such, interpreting is a bracketing activity leading to a simplification of the world (Weick et al., 2005). Complex and ambiguous characteristics of situations get identified, categorized, and labelled to enable the actors to get a grasp on one version of reality. Actors use this perception of reality to engage in “plausible acts of managing, coordinating and distributing” (Weick et al., 2005, p. 411).
8
Bartesaghi, 2016; Maitlis & Christianson, 2014; Mirbabaie & Zapatka, 2017; Nicholson & Anderson, 2005; Weick, 1995). The meaning of the situation is thus materialized and therefore made usable through the words in which it is expressed (Weick et al., 2005). Abolafia (2010) uses transcripts of the FOMC to identify important steps on how sense is made by the institution. He interprets the narratives to reconstruct how the group created meaning through discussion and debate (Abolafia, 2010). The word “reconstruction” and the emphasis on the FOMC as a group lead to two important dimensions of sensemaking that become visible in the interpreting microprocess: retrospective and collective (Maitlis & Christianson, 2014; Weick, 1995). The subsequent parts will explain both these dimensions that are established as two properties of sensemaking by Weick (1995) and are widely discussed and used in literature ever since (Heaphy, 2016; Holt & Cornelissen, 2014; McLeay, Merkl‐Davies, & Brennan, 2011; Merkus et al., 2017; Stieglitz et al., 2018). It is important to mention that these two dimensions are considered to be general characteristics of sensemaking and are not limited to be part of the interpreting microprocess. Nevertheless, as interpreting is at the core of sensemaking and most of the aspects of collectivism and retrospective align with this part of the sensemaking process, these concepts are explained here.
2.5.1 Collective dimension
Scholars have different perceptions on how to view sensemaking and how the process unfolds. There are two main perspectives of which one is grounded in the literature about social cognition (Maitlis & Christianson, 2014). This individualistic approach to sensemaking is based on the idea that an actor uses inherent cognitive frameworks to make sense of situations by relating the situation to known thinking patterns, adjusting these patterns and creating new patterns (Bingham & Kahl, 2012; Kaplan, 2008).
However, most scholars follow the perspective that sensemaking is a collective process in the way that meaning and reality are constructed through the interaction of multiple stakeholders (Heaphy, 2016; Holt & Cornelissen, 2014; Weick et al., 2005). Viewing sensemaking as a form of social construction was emphasized by Weick (1995) and aligns with the aforementioned importance of language and discourse: Only through collective action, words can be used to label, to bracket and to make sense (Castor & Bartesaghi, 2016; Maitlis & Christianson, 2014; Nicholson & Anderson, 2005).
9
interests and backgrounds which is a fruitful ground to start the discourse (Brown, Stacey, & Nandhakumar, 2008; Wachtendorf & Kendra, 2006; Weick et al., 2005). Because the occurrence of a crisis situation leads to an increase of the intensity of sensemaking activities, the collective dimension of sensemaking gets intensified. Thus, I hypothesize:
H2a: The collective dimension of sensemaking has a higher intensity in crisis situations than in non-crisis situations.
In the sensemaking literature, emotions have often been reduced to their role as a trigger in the form of arousal (Maitlis & Sonenshein, 2010; Weick, 1995). In recent times, attention was paid to both negative and positive emotions in the context of sensemaking. A high intensity of negative emotions impedes the actor. Scholars have investigated that strong negative emotions such as fear and panic preoccupy the actors cognitive capacity, leaving not sufficient resources to execute effortful sensemaking activities (Maitlis & Christianson, 2014; Maitlis & Sonenshein, 2010; Maitlis, Vogus, & Lawrence, 2013). A further reason why intense negative emotions hinder sensemaking activities is that the actor focusses on the emotion itself and is “distracted” from making sense of the situation, which reinforces his inability to attend to cues (Heaphy, 2016). Additionally, negative emotions can spread mistrust among group members, leading to actors falling back into a less collaborative defense state (Rafaeli, Ravid, & Cheshin, 2009).
With regard to positive emotions, Rafaeli et. al (2009) investigated in a study on virtual teams that “positive emotions attract and connect people” (p. 163), meaning that positive emotions enable processes towards intensified cooperation. The FOMC has more knowledge about the Financial Crisis because this crisis is closely connected to its core competencies. The confidence of being experts in the field results in a more positive emotional state of the members of the FOMC in the Financial Crisis than in the Iraq War, whereas the latter one is accompanied by more intense negative emotions. This leads to the following hypothesis:
H2b: The collective dimension of sensemaking has a higher intensity in the Financial Crisis than in the Iraq War.
2.5.2 Retrospective dimension
10
et al., 2010; Maitlis & Christianson, 2014). Despite their acknowledgment for this idea, many scholars argue, that sensemaking should be considered to have a retrospective dimension. Corley, Fabbri, & Gioia (2002) point out that actors can make a prediction about a future scenario based on the actions they imagine to take, and then react in retrospect on this imagination of the future. Put simply, it is about going two steps forward and then turning around and looking back one step: Sensemaking is about the future, but the way we look at it, is backward-oriented.
One of the primary examples that explains the retrospective dimension of sensemaking is how a nurse realizes that one child´s condition deteriorated over time while comparing her current perception of its health to a previous one (Weick et al., 2005). People are only completely aware of things that already happened and can then look back to try to find possible explanations for the occurrence of the current situation. Actors face a problem of equivocality when they are confronted with too many meanings which are only limited once the situation occurs (Weick, 1995). Crisis occurrences are situations in which expectations are violated (Christianson et al., 2009) and actors try to understand the reasons for the current situation and how, as well as why, it occurred. In order to do so, actors can try to interpret the current situation by referring it back to how it was before, and which reasons might have led to the variation. Therefore, in this paper the following hypothesis will be tested:
H3a: The retrospective dimension of sensemaking has a higher intensity in crisis situations than in non-crisis situations.
In part 2.2, three qualities of crisis situations are outlaid. One of them is the suddenness of the occurrence of the crisis. The Iraq War is categorized as being more sudden than the Financial Crisis whereas the latter one is more unfolding over time. This holds especially in the context of the FOMC, that is specialized in looking at data which is closely connected to the development of the Financial Crisis, such as data about the housing market. A second quality of crises can be used to make a distinction: The scope of the Iraq War is much broader because it is not only categorized as a political crisis but also a military conflict and influenced by many reasons, for instance, the accessibility to the resource oil (Jhaveri, 2004). This implies that a broader range of unanswered questions remains in the case of the Iraq War, resulting in a higher degree of equivocality. Therefore, FOMC members are looking back in time more frequently, comparing the current state not only with the most recent state but also analyzing prior wars and using them as a reference. Thus, I hypothesize:
11
2.6 Acting
The microprocess of acting is also an integral part of sensemaking (Rudolph et al., 2009). Only through action people have the possibility to affect their environment and shape it in their intended way (Weick, 1995). As examined in the retrospective part, action occurs before cognition. But only after the action, actors can look back and reflect upon the new situation. Acting is important for sensemaking because through the variation in the environment that it entails, new cues to make sense of are generated (Maitlis & Christianson, 2014). Additionally, only through action a feedback-loop can be established. In this loop, it is tested whether the premises that led to the execution of certain actions also hold in reality (Rudolph et al., 2009). This demonstrates another characteristic of enactment: Through actions, actors can shape the environment, but also does the environment sometimes vary in unintended ways. Neither is the environment completely constructed by actions, nor are our actions only the succession of how the environment is shaped (Weick, 1995).
It is rather difficult to take appropriate action during a crisis because of the high ambiguity and uncertainty. Nevertheless, the situation often requires a quick reaction which also supports the sensemaking process through the feedback-loop (Christianson et al., 2009; Maitlis & Christianson, 2014). Crisis situations create situations of equivocality in which immediate action is necessary to establish the feedback-loop and to start the process of the creation of new cues. Therefore, the next hypothesis reads:
H4a: The sensemaking microprocess of acting has a higher intensity in crisis situations than in non-crisis situations.
12
2.7 Sensemaking and crisis situations
The previous sections focus on the effects that crisis situations and their qualities have on the intensity of sensemaking. Several hypotheses test if the occurrence of crisis situations intensify the sensemaking microprocesses of noticing and acting as well as its collective and retrospective dimension. While every single developed hypothesis is interesting on its own and provides a more detailed view on the sensemaking process, the hypotheses H1, H2a, H3a and H4a can be aggregated to the following concluding hypothesis2: H5: The sensemaking process has a higher intensity in crisis situations than in non-crisis situations.
3 Methodology
This section guides through the analysis of the data. First, the connection between language and sensemaking is explained, followed by an introduction into the computer-aided textual analysis program “Linguistic Inquiry and Word Count” (LIWC). This program is used to transform the dataset, which consists of meeting transcripts of the FOMC, into a form that enables quantitative analyses. Second, the dataset is explained in detail. Third, the decision to focus mainly on the Iraq War and the Financial Crisis of 2007 as crisis situations is justified. Fourth, the measurements for the variables in the hypotheses are explained. Lastly, fifth, the steps followed to conduct an analysis using the statistic software IBM-SPSS are expounded.
3.1 Connection between language and sensemaking
The linkage between language and sensemaking is already mentioned in the literature review when the power of narratives is used to help explain the interpreting microprocess (Weick et al., 2005), as well as to introduce into the concept of collective sensemaking (Abolafia, 2010). The benefit of viewing sensemaking from a narrative perspective is that the actual process of meaning-construction can be the focus opposed to
2 For the collective and the retrospective dimension of sensemaking as well as for the microprocess of acting,
13
only identifying actors and their actions (Maitlis & Christianson, 2014). Some scholars focus on single aspects of language such as the importance of metaphors as a way of facilitating the understanding of complex linkages (Cornelissen, 2012). Various others more generally agree on the perspective that language can be used to demonstrate how sense can materialize (Maitlis & Christianson, 2014; Weick, 1995; Weick et al., 2005). This paper builds on that approach, using written transcripts of conversations to gain insights into the intensity in which different microprocesses and dimensions of sensemaking materialize.
3.2 Introduction into LIWC
In this research, the text analysis program LIWC is used to enable the usage of transcripts. In the last decades, computer-aided text analysis was facilitated through an improvement in computer technology and the increased availability of digitalized data (Mehl, 2006). LIWC is a computer program that goes through every word of the input files and assigns words to pre-defined categories, delivering percentages of the categories´ occurrences (McLeay et al., 2011). The categories are partly linguistic categories, but also psychological processes and time-related dimensions are formed (Mehl, 2006). The program has already been used in the field of sensemaking (McLeay et al., 2011; Milliken, Magee, Lam, & Menezes, 2012), although it is more recognized in other streams of research, especially in psychology and impression management (Aerts & Yan, 2017; Hooghiemstra, Kuang, & Qin, 2017; Kim, 2013). The comfort of automated text analysis provides certainty about the objectivity of the analysis as well as psychological validity (McLeay et al., 2011). The negative aspects of analysing single words without their context can be partly minimized by choosing the measurement scales wisely.
3.3 Dataset
14
between 1996 and 2012 is used. A total of 163 transcripts, of which 136 are transcripts of meetings and 27 are transcripts of conference calls, have been downloaded from the official website of the FED. The meetings are usually held over two consecutive days and the transcripts have an average wordcount of 48.894 whereas the conference calls are usually shorter, having an average wordcount of 11.461 (see appendix).
Every transcript is stored in a document that first lists the names and positions of the involved people. Thereafter, the conversation that took place is transcribed. No further changes have been made to the downloaded meeting and conference call files. The transcripts were used as an input for LIWC, whereas the output represents percentages of the different LIWC dimensions for every transcript.
3.4 Selection of crisis situations
In the observed timeframe between 1996 and 2012, many crisis situations took place. Next to the Iraq War and the Financial Crisis, there were several other crisis situations. One of the most shocking events was the terror attack in New York City on the 11th of September 2001. Next to this extraordinary crisis situation, which is commonly referred to as “9/11”, further examples of crises in that timeframe are major nature crises such as the hurricane Katrina in 2005 (Castor & Bartesaghi, 2016; Mills & Weatherbee, 2006) or the Sumatra Tsunami in 2004 (Chadha, Latha, Yeh, Peterson, & Katada, 2005). Disastrous events, such as the oil spill of the Deepwater Horizon drilling unit in 2010 (Joye, 2015), or the radiation release in Fukushima that followed after an earthquake and a tsunami in 2011 (Buesseler, 2012) and also terror attacks, such as the Madrid bombings in 2004 (Canel & Sanders, 2010) or the Boston Marathon bombing in 2013 (Gates et al., 2014), all occurred in the observed timeframe.
In order to select the most suitable crisis events for this analysis, two factors play a key role: Because conference calls are usually only held in case of special circumstances (Cannon, 2015; Federal Open Market Committee, 2016), and crisis situations are assumed to represent extraordinary circumstances, the occurrence of transcripts of conference calls can be used to define time periods of crises. Moreover, the quantitative approach of this investigation requires a minimum sample size. 13 of the 27 transcripts of conference calls are from the years 2003, 2007 and 2008, indicating that the Iraq War and the Financial Crisis are the most suitable events for further investigation.
15
crisis in 2007 are defined by substantial losses of the S&P 500, demonstrating how impactful the crisis was. The liquidation of two hedge funds of the major investment bank Bear Stearns on the 31st of July 2007 can be seen as an accelerator of the crisis and a sign that the market is under pressure (St. Louis Fed, 2009). Despite not having dramatic effects on the stock market at first, in retrospect, it can be seen as a major hint about the possible impact of the situation. The determination of the length of the crisis depends on the point of view. In this paper, the crisis period ends at the end of October 2008 when the stock market began to stabilize again despite not having reached its lowest point. This timeframe is supported by the argument that conference calls fall more often into crisis periods since 9 out of the 20 transcripts of this time period are transcripts of conference calls.
Figure 1: S&P 500 in USD from March 2007 until November 2009 (daily data). The red box visualizes the used time period of the Financial Crisis.
The Iraq War which is also referred to as the second gulf war started off with the bombing of the city of Baghdad on the 20th of March 2003 (O’Huallachain & Sharpe, 2005). One might argue that there have been cues of an increased probability of a military conflict. Examples are the publishing of the Blix report on the 27th of January 2003 in which was reported that the Hussein-regime had failed to disarm or when the President of the United States of America announced an ultimatum to the Iraq government just two days before the outbreak of the acts of war (Paleari et al., 2005). However, for this analysis, the first bombings are chosen as the beginning of the conflict because they signaled the inevitability of military action. The end of the crisis period is more difficult to assess: Already on the 7th of April 2003, the coalition forces around the American Army reached Baghdad and only a few weeks later, on the 1st of May 2003, President Bush declared that the major combat operations have ended (Paleari et al., 2005). However, the fighting
400 600 800 1.000 1.200 1.400 1.600 1.800 S& P 5 0 0 in USD
16
continued for several months and years, even after the former head of the Iraq government Saddam Hussein was captured on the 13th of December 2003. Therefore, this paper relies on the percentage of transcripts of conference calls once again to determine when the most heated phase of the Iraq War ended.
The development of both the American Gross Domestic Product (GDP) and the American stock market over time show a possible association to the two crisis situations. During the time period of the Financial Crisis, the stock market, as well as the GDP, showed a negative trend (figure 2; (Federal Reserve Economic Data, 2019)). Therefore, it might be possible that not the crisis situation itself but the development of the GDP leads to intensified sensemaking activities of the FOMC. However, the negative turn in the GDP is delayed for nearly one year after the beginning of the Financial Crisis as defined for this paper. The first major decrease in GDP happens between the third and the fourth quarter of 2008 when the GDP drops from 14,835 to 14,560 billion United States Dollar (USD). Fligstein et al. (2017) investigated the sensemaking efforts of the FOMC during the Financial Crisis and concluded that the members failed to connect the housing bubble with possible effects on the overall economy. The GDP as a measurement of the overall economic state is therefore not the main driver of sensemaking efforts of the FOMC, at least not in early stages of the crisis. During the time period of the Iraq War, no untypical variations in the GDP occurred. This reinforces the argument that the development of the GDP is not the driving force behind the sensemaking efforts instead of the occurrence of crises.
Figure 2: GDP of the USA in billions of USD measured quarterly between the years 1996 and 2012. The red boxes visualize the used crises periods of 9/11 and the Financial Crisis.
17
Besides the two selected crisis situations, further attention should also be paid to 9/11, which is widely recognized as the event that brought terrorism to America (Back, Küfner, & Egloff, 2010). On that day, two passenger airplanes got hijacked and hit both of the each 110 stories high towers of the World Trade Center, leading to the destruction of wide parts of the area and the loss of nearly three thousand lives (Shyam-Sunder et al., 2008; Sunder et al., 2005). The New York stock exchange was closed after the attack and remained closed for a series of days. The crisis negatively affected the stock market, but the market recovered relatively quickly. Only two days after the airplane attack, the FOMC held a telephone conference on the 13th of September, followed by another conference call on the 17th of September.
The three discussed crisis situations that would potentially fit for this study are further analyzed regarding their suitability. Table 1 shows the beginning as well as the end of each crisis, mentions the sample size per group and highlights the percentage of conference calls, also in comparison to the total sample of transcripts.
Table 1: Overview of crisis selection and time periods.
All three discussed crisis situations have a percentage of conference calls far above the average of 17%. However, 9/11 only consists of three transcripts because of its suddenness. This sample size is too small to deliver reliable results for which reason 9/11 is excluded and the three transcripts are categorized as occurring in the period of non-crisis. The two remaining crisis situations consist of 20 transcripts with a percentage of conference calls of 45% (Financial Crisis) and 9 transcripts with a percentage of conference calls of 44% (Iraq War) what promises a reliable basis for comparison. The total percentage of transcripts of conference calls in the overall sample is 17%, whereas the part of the sample that contains non-crisis transcripts only consists of 10% conference calls.
18
significant association between the two categorial variables (Denis, 2018). The crosstabs indicate a relationship between the occurrence of crisis situations and the presence of conference calls. This association is significant X2 (1) = 20,39 with a Phi-value of 0,354. Therefore, crisis situations are (statistically significant) characterized by a higher percentage of conference calls than non-crisis situations.
3.5 Measurement of variables
The independent variables are categorial variables, indicating whether a transcript falls into a crisis situation or not. For all the dependent variables the z-scores are taken and used to make all measurements comparable with a mean of zero and a standard deviation of one (Denis, 2018).
3.5.1 Intensity of the noticing microprocess of sensemaking
Theory suggests that the noticing microprocess of sensemaking intensifies when arousal occurs as a sign of violated expectations and uncertainty (Maitlis & Sonenshein, 2010). In uncertain situations and when expectations are not fulfilled, cues for sensemaking materialize (Maitlis & Christianson, 2014). Therefore, uncertainty is a sign of the presence of arousal and the intensification of the noticing microprocess of sensemaking. The following quote from a transcript during the period of the Financial Crisis is a suitable example on how uncertain some of the FOMC members are about what action to take next:
“I don't want to raise expectations, nor do I want to ignore questions. I don't know if there is a balance on that, or maybe we say just, "We'll wait and see." (Hoenig, 2007)”
LIWC offers the opportunity to measure the variable “certainty”. After the z-scores have been calculated, “uncertainty” can be measured as the negative value of the LIWC category “certainty” and is used as a proxy for the noticing microprocess of sensemaking.
3.5.2 Intensity of the collective dimension of sensemaking
19
“Following that, we would have a go-round and get comments and positions, and then we proceed to votes. So without further ado, let me turn to Bill in New York to start with a market update and then to talk a bit about the proposed facility (Bernanke, 2008).”
The psychological LIWC variable “social processes” measures the intensity of social concerns and social support which is supposed to be high in the case of fruitful group dynamics (Tausczik & Pennebaker, 2010). Therefore, “social processes” is chosen to measure the collective dimension of sensemaking.
3.5.3 Intensity of the retrospective dimension of sensemaking
Retrospective sensemaking is about comparing the current status of a situation with something that has been there before by looking back in time, as well as referring and recalling the past state of the situation (Weick, 1995; Weick et al., 2005). The following quote from the time period of the Iraq War represents an example of the usage of backward-looking behavior of the FOMC members:
“The markets over the last week were once again driven by headlines about the war. At times when market participants thought that the war would be short, they tended to be bullish for the dollar and for equities. (…) Interestingly, volatilities in the equity markets have actually fallen, which is counterintuitive. The VIX is down 3 percentage points; the pattern historically has been that, when equities have declined, the VIX has tended to rise (Kos, 2003).”
LIWC measures the use of past tense through the variable “past tense” which represents the percentage of words that were used in past tense and will be used as a measurement for the dependent variable “intensity of the retrospective dimension of sensemaking”.
3.5.4 Intensity of the acting microprocess of sensemaking
20
Stieglitz et al., 2018). The following two quotes from transcripts during crisis periods demonstrate how the members of the FOMC (1) update and (2) doubt each other:
(1) “Further questions for David? If not, as I indicated a couple of weeks ago, I thought it might be worthwhile to do a somewhat truncated roundtable today. So, for those of you who have remarks you'd like to make, what I plan to do is to go sequentially through the Banks, starting with the First District, and then turn to the Board members. Cathy, as President of the Boston Bank, do you have any comments to make now that you're on the line (Greenspan, 2003)?”
(2) “This gets to a broader concern that I have, which is the stick part of this. I can understand the carrot side of this thing, and we are doing it for the reasons that you stated, and I am very sympathetic to the argument. The question is, What [sic] do we get in return, and how do we make sure that, since we are not the regulator of these dealers, there is indeed discipline? I would like to know what your answer is. Do we have an arrangement with the SEC, or how do we work with the other regulators to make sure that discipline is applied here? (...) But I am just a little worried about being taken advantage of here. So [sic] the question is specifically about the "prudential regulation" of those people that we're dealing with. Then, I have other questions to follow (Fisher, 2008).”
All these aspects have in common that they consist of expressed questions. This dependent variable will be measured through the number of questions used, which is represented in the LIWC variable “questionmarks”.
3.5.5 Intensity of sensemaking
The complete sensemaking process consists of the three main sensemaking moves and is characterized by its different dimensions (Maitlis & Christianson, 2014; Weick, 1995). The intensity of sensemaking is measured through a variable that consists of measurements of the various aspects. Aligning with the aforementioned decisions regarding the measurement of the sensemaking moves, the variable “Intensity of Sensemaking” (ISM) is calculated as follows:
21
4 Findings
4.1 Structure of analysis
The first step is to transfer the dataset into the software IBM-SPSS to carry out statistical analyses. Afterwards, the hypotheses are tested conducting several mean-comparing analyses. LIWC delivers percentages as values for its categories, indicating how often words of one category occur in the transcript. The average of these percentages of all cases in the different qualities of crisis situations can be compared. The hypothesis H4 includes many theoretical aspects for which reason it is visualized in figure 3. This visualization of the dependent variable represented over time can give a first indication whether the intensity of sensemaking might be increased in times of a crisis.
Figure 3: Intensity of Sensemaking over time.
The figure indicates that the intensity of sensemaking seems to be higher around the time of crisis occurrences. Peaks during the Financial Crisis, the Iraq War and also around 9/11 are visible. Some transcripts are characterized by having negative values on the dependent variable “Intensity of Sensemaking”. This is not a sign of non-existence of sensemaking activities. The explanation for negative values lies in the methods used: The z-transformation makes the variables more comparable but results in an overall mean-value of zero. Therefore, a positive value has to be interpreted as representing an above average intensity of sensemaking activities. A negative value indicates a below average intensity of sensemaking activities, compared to the rest of the dataset.
22
The outliers remain part of the dataset and no cases are eliminated from the analysis. Outliers represent exceptional cases that differ from the expectation and are therefore often excluded because they might influence the outcome of analysis significantly. However, crisis situations are exceptional circumstances, so that the elimination of cases from such periods would minder the quality of the dataset.
The most common method to compare the means of two groups is the independent t test. To conduct an independent t test, normal distribution for both groups (crisis and non-crisis; Financial Crisis and Iraq War) would be required (Denis, 2018). The Shapiro-Wilk test for normality tests whether the underlying sample is normally distributed. In the test setup, normal distribution is hypothesized. For the hypotheses H1, H2b, H3a, H3b, H4b and H5 this assumption has to be rejected because at least one of the groups has a significance above 5% (Field, 2009). For the hypotheses H2a and H4a this assumption holds because both of the groups have a significance below 5% and normal distribution could be assumed. However, because of the existence of outliers as well as the varying sample size between the groups (crisis situations: 29; non-crisis situations: 134; Financial Crisis: 20, Iraq War: 9), for all the hypotheses the nonparametric Mann-Whitney U test is chosen as the most appropriate method of mean comparison. The Mann-Mann-Whitney U test works well with small sample sizes (Denis, 2018; ‘Nonparametric Tests’, 2018). Because the distributions of all the dependent variables for crisis and non-crisis situations have a different shape, the Mann-Whitney U test is used to compare mean ranks and not to compare medians (Field, 2009).
4.2 Analysis of mean comparison
The SPSS outputs of the following findings can be found in the appendix:
For hypothesis H1 the mean rank of the dependent variable “Intensity of sensemaking microprocess of noticing” is lower in crisis situations (M = 60.69) than in non-crisis situations (M = 86.61). It is important for this variable to keep in mind, that the variable measures uncertainty through the negative value of certainty. Because of that, the lower value in crisis situations aligns with the assumption, that the sensemaking microprocess of noticing is higher in crisis situations. This difference is significant (p = .007), so H1 is confirmed.
For hypothesis H2a, the mean rank of the dependent variable “Intensity of collective dimension of Sensemaking” is higher in crisis situations (M = 91.64) than in non-crisis situations (M = 79.91), aligning with the assumption. However, this difference is not significant (p = .225), for which reason H2a must be rejected.
23
A pretest using the LIWC variable “positive emotion” tests whether this assumption holds in reality. The mean rank of the dependent variable “positive emotion” is higher in the Financial Crisis (M = 15.58) than in the Iraq War (M = 13.72). This difference is statistically significant (p = .000), for which reason the pretest is confirmed. For hypothesis H2b itself, the mean rank of the dependent variable “Intensity of collective dimension of Sensemaking” is higher in the Financial Crisis (M = 17.10) than in the Iraq War (M = 10.33). This aligns with the assumption and the difference is significant (p = .049): H2b is confirmed. For H3a the mean rank of the dependent variable “Intensity of retrospective dimension of Sensemaking” is higher in crisis situations (M = 90.79) than in non-crisis situations (M = 80.10), aligning with the assumption. This difference is not significant (p = .268), so H3a must be rejected.
For hypothesis H3b the mean rank of the dependent variable “Intensity of retrospective dimension of Sensemaking” is higher in the Iraq War (M = 22.67) than in the Financial Crisis (M = 11.55), aligning with the assumption. This difference is significant (p = .001), so H3b is confirmed.
For hypothesis H4a the mean rank of the dependent variable “Intensity of sensemaking microprocess of acting” is higher in crisis situations (M = 96.74) than in non-crisis situations (M = 78.81), which aligns with the assumption. This difference is significant (p = .063), for which reason H4a is confirmed.
For hypothesis H4b the mean rank of the dependent variable “Intensity of sensemaking microprocess of acting” is higher in the Financial Crisis (M = 15.58) than in the Iraq War (M = 13.72). This aligns with the assumption, but the difference is not significant (p = .594): H4b must be rejected.
Concluding, for hypothesis H5 the mean rank of the dependent variable “Intensity of Sensemaking” is higher in crisis situations (M = 108.21) than in non-crisis situations (M = 76.33), aligning with the assumption. This difference is highly significant (p = .001), for which reason H5 is confirmed.
24
Table 2: Overview of findings.
5 Discussion
The research question of this paper is, what effects the different qualities of crisis situations have on the intensity of sensemaking. First and foremost, the confirmation of the concluding hypothesis H5 demonstrates that crisis situations have a significant positive effect on the intensity of sensemaking in groups. This finding confirms the merit of measuring the intensity of sensemaking. Measuring sensemaking activities through its intensity is an innovative approach that has not been discussed in the existing literature. It allows quantifying sensemaking as an ongoing process that can vary in its manifestations. Sensemaking should not be treated as being either in an active state or being in a passive state. The chosen approach aligns with consensus in the theory that highlight the ongoing processual nature of the sensemaking process (Louis, 1980; Maitlis & Christianson, 2014; Weick, 1995) but also adds to the theoretical research by introducing a novel approach to measure such a process.
25
This paper found how sensemaking is a process that has various levels of intensity and that crisis situations lead to an increase in this intensity.
In this study, it is highlighted that crisis situations have certain qualities that allow a differentiation between crisis situations. The study finds that such differences in the qualities are one reason for varying levels of the intensity of sensemaking. Three of these qualities are investigated in detail using two crisis situations as examples: The Iraq War and the Financial Crisis can be qualified (1) whether they are unfolding over time or are happening more sudden, (2) whether they are a political or an economic crisis, and (3) whether they have a more or less positive emotional level.
Discussing the latter quality, extant sensemaking literature is focused on negative emotions and how they hinder sensemaking activities (Maitlis & Sonenshein, 2010; Maitlis et al., 2013). Only a few scholars already investigated the enabling role positive emotions can play (Heaphy, 2016; Maitlis & Christianson, 2014; Rafaeli et al., 2009). This study adds to the literature by demonstrating, that even in the rather negative setting of a crisis situation, positive emotions increase the intensity of the collective dimension of sensemaking. Positive emotions improve group sensemaking despite emotional negative circumstances.
Elaborating on the other two qualities: The qualification as a political crisis and the sudden beginning of the Iraq War influence the intensity of the retrospective dimension. The manifestations of both these qualities seem to be associated with a high degree of equivocality. Extant literature emphasizes that sensemaking is retrospective per se (Corley, Fabbri, & Gioia, 2002; Weick et al., 2005). This study makes a distinction between qualities of crisis situations that affect the degree of equivocality. This finding adds that such qualities affect to what extent sensemaking activities are retrospective. While sensemaking in general is backward-orientated, different qualities of the crisis situation influence how strong this dimension is pronounced.
The findings related to the different qualities of crisis situations concede overall, that the sensemaking process is highly context-dependent and should be viewed with an open eye for details. This aligns with the perspective of Weick, who already emphasized that “the order in organizational life comes just as much from the subtle, the small, the relational, the oral, the particular, and the momentary as it does from the conspicuous, the large, the substantive, the written, the general, and the sustained” (Weick et al., 2005, p. 410). It adds to the literature which was focused on investigating one crisis situation separately without paying attention to general qualities of crises situations that can influence the intensity of sensemaking activities.
26
members of the FOMC are aware that they have to take immediate action in crisis situations. Even if the Iraq War is further detached from their core capabilities, they are capable of overcoming their resistance to engage in sensemaking activities regarding more loosely coupled topics. This finding may extend the literature by showing that the right group composition can lead to such a high intensity of a microprocess of sensemaking, that environmental factors cannot influence the intensity (statistically) significant. However, further research is required to expand these results. Another possible explanation for the insignificant results is that the Iraq War characterized by a high degree of uncertainty and ambiguity. This might lead to an extensive amount of proposed questions which coincides with the measurement of the dependent variable for the sensemaking microprocess of acting. A possibly significant confirmation of the hypothesis might have been offset by such an impact on the measurement.
Another important contribution of this research is, that the microprocesses of sensemaking can be affected separately. Some part of the literature makes a theoretical distinction between the three main sensemaking microprocesses noticing, interpreting and acting (Maitlis & Christianson, 2014; Weick, 1995). However, not every scholar differentiates to that extend and some view sensemaking more as a general process (Christianson et al., 2009; Mirbabaie & Zapatka, 2017). The confirmations of the hypotheses H1 and H4a reveal that the sensemaking microprocesses of noticing (H1) and acting (H4a) are intensified during times of crisis. So far, no assessment could be made regarding the importance of each of the microprocesses. This research cannot completely fill this gap. However, it demonstrates that it is possible to look at each microprocess individually. It even offers an approach to measure the importance of the microprocesses by looking at the intensity of each one.
27
intensity, because it is already on a high level during non-crisis periods. The same argument does not hold for the microprocesses, which are more specific and context-dependent.
6 Conclusion
6.1 Limitations and further research
The major contribution of this research is the approach of measuring the intensity of sensemaking. Further research can build on this approach. In close relation to this paper, different crisis situations and their impact on the intensity of sensemaking can be investigated. This helps to clarify qualities of crises that hinder sensemaking and allows the search for solutions to mitigate against such obstacles. Crisis situations are only one occasion where the intensity of sensemaking is of interest. Another one is time series analyses: These analyses open up interesting opportunities to observe how the intensity of sensemaking in one group varies over time and depending on the actors involved or on external circumstances.
Also, in this research, a detailed distinction between the three main sensemaking microprocesses noticing, interpreting and acting is made. Further analysis could aim at investigating in which settings and under which circumstances the noticing microprocess has a higher intensity. There may be factors, such as group size, group diversity or method of communication (personal meetings versus telephone conferences), that can improve our understanding of this aspect of sensemaking. Additionally, regression analyses could test which factors have the highest influence on the intensity of sensemaking. This would allow a prioritization of factors. The sensemaking process of groups can be improved if the most important factors are discovered and efforts are undertaken to influence them. This could have the practical implication that it enhances the design of crisis response groups. Therefore, in crisis situations when reaction time is an important factor, the right group composition can lead to quicker sensemaking which facilitates decision making.
28
Various scholars have called for more quantitative papers regarding sensemaking in the past (Maitlis & Christianson, 2014; Weick et al., 2005) to little avail. The quantitative approach has its limitations: Some of the analyses would have benefited from a bigger sample size. Especially the hypotheses about the comparison between the two different crisis situations can be criticized because of the low number of transcripts in the respective periods. Extending the sample would eliminate worries that the sample size of 29 for the group of crisis-situations or of 9 for the Iraq War affects the significance of results. Moreover, due to the low number of transcripts in crisis periods, the exact definition of the timeframe of such periods is of superior importance. The results of a robustness test, in which the 9/11 time period was included as part of the crisis sample as well, still support the results. Nevertheless, the aforementioned doubts cannot be eliminated completely and call for further research. Another limitation is the missing verification of some of the used measurements of variables. Due to the scare existing quantitative literature on sensemaking, some of the used measurements have never been confirmed in praxis. This research can be seen as a starting point to both, further theorize about different measurements for the intensity of sensemaking, and to test those measurements in different settings.
6.2 Overall Conclusion
This research makes three main contributions: First, the approach of measuring sensemaking activities through its intensity is introduced. Second, on a general level, crisis situations lead to an increased intensity of sensemaking. Moreover, this intensity is affected by the different qualities of the crisis situation. Third, every microprocess of sensemaking can be affected separately. Even more, this paper also offers an approach to measure such differences through the intensity of the sensemaking process.
A
Appendix
Wordcount comparison between meeting and conference call
B Chi-square test
Table 3:Crosstabulation of the variables "crisis occurrence" and "separation between meeting and conference call". Comparison between expected count and actual count of meetings and conference calls in crisis and non-crisis situations.
Table 4: Chi-Square test of the variables "crisis occurrence" and "separation between meeting and conference call".
C S&P500 during the Iraq War
Figure 5: S&P 500 in USD between January 2003 and December 2004. The red box visualizes the used time period of the Iraq War.
S&P 500 between 1996 and 2012
Figure 6: S&P 500 in USD between the years 1996 and 2012. This represents the development of the S&P 500 over the total time period of the dataset.
400 600 800 1.000 1.200 1.400 S & P 5 0 0 in US D
S&P 500 during Iraq War
400 600 800 1.000 1.200 1.400 1.600 1.800 S& P 5 0 0 in USD
D Boxplots for the hypotheses
H1:
Figure 7: Boxplot showing the median and outliers of the variable "Intensity of microprocess of noticing" in non-crisis and crisis situations. Non-crisis situations are characterized by the higher median. This variable has to be measured through its negative value because “uncertainty” is measured through the negative value of “certainty”. Therefore, the intensity of sensemaking microprocess of noticing is higher in crisis situations. Both groups contain significant outliers.
H2a:
E H2b pretest:
Figure 9: Boxplot showing the median and outliers of the variable "positive emotion" in the Iraq War and the Financial Crisis. The Financial Crisis is characterized by the higher median. The group of the Financial Crisis contains significant outliers.
H2b:
F H3a:
Figure 11: Boxplot showing the median and outliers of the variable "Intensity of retrospective dimension of sensemaking" in non-crisis and non-crisis situations. Crisis situations are characterized by the higher median. Both groups contain significant outliers.
H3b:
G H4a:
Figure 13: Boxplot showing the median and outliers of the variable "Intensity of sensemaking microprocess of acting" in non-crisis and crisis situations. Crisis situations are characterized by the higher median. Both groups contain significant outliers.
H4b:
H H5:
I Test of normality distribution
J H4b; Financial Crisis D(20) = 0,80, p = ,00
H5; Non-crisis: D(134) = 0,98, p = ,06 H5; Crisis: D(29) = 0,92, p = ,03
Mann-Whitney U test
K
Table 8: Results of the Mann Whitney U test. The significance levels (2-tailed) are displayed for all eight tested hypotheses. Wherever it was possible, the exact significance levels [2*(1-tailed)] are used.
Robustness test including 9/11 as a crisis situation
L
i
REFERENCES
Abolafia, M. Y. (2010). Narrative Construction as Sensemaking: How a Central Bank Thinks. Organization Studies, 31(3), 349–367. https://doi.org/10.1177/0170840609357380.
Abolafia, M. Y., & Hatmaker, D. M. (2013). Fine-Tuning the Signal: Image and Identity at the Federal Reserve. International Public Management Journal, 16(4), 532–556. https://doi.org/10.1080/10967494.2013.849167.
Aerts, W., & Yan, B. (2017). Rhetorical impression management in the letter to shareholders and institutional setting: A metadiscourse perspective. Accounting, Auditing & Accountability Journal, 30(2), 404–432. https://doi.org/10.1108/AAAJ-01-2015-1916.
Auerbach, R. D. (2009). Deception and Abuse at the Fed: Henry B. Gonzalez Battles Alan Greenspan’s Bank. University of Texas Press.
Back, M. D., Küfner, A. C. P., & Egloff, B. (2010). The Emotional Timeline of September 11, 2001. Psychological Science, 21(10), 1417–1419. https://doi.org/10.1177/0956797610382124.
Bernanke, B. (2008). Meeting of the Federal Open Market Committee on March 10, 2008.
Bingham, C. B., & Kahl, S. J. (2012). The Process of Schema Emergence: Assimilation, Deconstruction, Unitization and the Plurality of Analogies. Academy of Management Journal, 56(1), 14–34. https://doi.org/10.5465/amj.2010.0723.
Brown, A. D., Stacey, P., & Nandhakumar, J. (2008). Making sense of sensemaking narratives. Human Relations, 61(8), 1035–1062. https://doi.org/10.1177/0018726708094858.
Buesseler, K. O. (2012). Fishing for Answers off Fukushima. Science, 338(6106), 480–482. https://doi.org/10.1126/science.1228250.
Canel, M., & Sanders, K. (2010). Crisis Communication and Terrorist Attacks: Framing a Response to the 2004 Madrid Bombings and 2005 London Bombings. The Handbook of Crisis Communication, 449–466.
