THEORY AND APPLICATION OF QUANTITATIVE NARRATIVE ANALYSIS ...jpthomeczek.de/wp-content/uploads/2017/11/purpura... · QUANTITATIVE NARRATIVE ANALYSIS (QNA) Quantitative Narrative Analysis

THEORY AND APPLICATION OF QUANTITATIVE NARRATIVE ANALYSIS IN THE CONTEXT OF POPULISM Alberto Purpura - Università degli Studi di Padova

[email protected]

1

mailto:[email protected]

“[POPULISM IS] AN IDEOLOGY THAT CONSIDERS SOCIETY TO BE ULTIMATELY SEPARATED INTO TWO HOMOGENEOUS AND ANTAGONISTIC GROUPS, ‘‘THE PURE PEOPLE’’ VERSUS ‘‘THE CORRUPT ELITE’’, AND WHICH ARGUES THAT POLITICS SHOULD BE AN EXPRESSION OF THE VOLONTÉ GÉNÉRALE (GENERAL WILL) OF THE PEOPLE”

Cas Mudde (2004). ‘The Populist Zeitgeist’, Government and Opposition.

THEORY AND APPLICATION OF QUANTITATIVE NARRATIVE ANALYSIS IN THE CONTEXT OF POPULISM 2

THEORY AND APPLICATION OF QUANTITATIVE NARRATIVE ANALYSIS IN THE CONTEXT OF POPULISM

ANALYZING POLITICAL TEXTS- Classical content analysis

- Holistic grading

- Automated/Computer-assisted content analysis

3


CLASSICAL CONTENT ANALYSIS- Human coders (extensively trained)

- Often based on a codebook

- Different units of measurement according to topic

- Validity, are we measuring what we want to measure?

- Reliability, are the analyses consistent?

- In the case of populism the codebook should contain ALL the words that can refer to the entities we mentioned before, “the elite” and “the people”



HOLISTIC GRADING- Human coders (extensively trained)

- Grading scale (i.e. 0 = not populist, 1 = mixed; 2 = populist)

- Sampling units coincide with coding units (documents) or may use paragraphs

- Validity, are we measuring what we want to measure?


5


AUTOMATED CONTENT ANALYSIS- Fully or partially automated

- Usually dictionary based (some changes expected in the future)

- Words are usually the units of measurement

- Validity, are we really measuring what we want to measure?


- Matthijs Rooduijn & Teun Pauwels, Measuring Populism: Comparing Two Methods of Content Analysis, West European Politics, Nov 2011

- Example with populism:

- how to describe the aforementioned entities that may allow you to detect populism?

- synonyms

- the importance of context



QUANTITATIVE NARRATIVE ANALYSIS (QNA)

In order to understand collective dynamics we must study the collectivity as a whole, but we must not study it as a collective entity. Only by taking into account the individual entities, and most critically the relations between them and their activities, can we understand the macro structure we observe. Predicting and explaining macro-level dynamics is one of the most central goals of analytical sociology.

Hedström and Berman, 2009, The Oxford Handbook of Analytical Sociology

1. Key to social explanation are individuals' actions and the relations that link actors to one another.

2. The macro level can only be fully understood in terms of the micro level of individuals' actions and relations.



QUANTITATIVE NARRATIVE ANALYSIS (QNA)Quantitative Narrative Analysis or QNA is a methodological approach to texts that allows researchers to structure the information contained in narrative texts in ways that make possible a statistical analysis of the information.

- Identify and assemble a set of narrative documents for analysis (newspaper articles, police reports, personal narratives, interviews, speeches, …)

- Design a “story grammar” (a basic structure that will capture in a set of categories the information contained in narrative documents) namely information on actors, their actions and the attributes of actors and actions

- Choose a software to store the grammar and the coded information so as to make possible large-scale projects

<subject> → {<actor>} {<characteristics>} <actor> → crowd | mob | posse | Negro | Sheriff |…<characteristics> → <type> <number> <organization> <name> …

<verb>→ <verbal phrase> <circumstances> …<verbal phrase> → bring | burn | shoot | kill | hang |...< circumstances > → <time> <space> <reason> <instrument> <outcome>

…



QNA VS CLASSICAL CONTENT ANALYSIS

1. The coding categories are based on invariant structural properties of narrative, rather than the ad hoc, abstract, and theoretically defined categories of classical content analysis

2. The coding categories are tightly interlinked and interrelated through a set of rewrite rules (e.g., actors are linked to actions, and both actors and actions are linked to their attributes); in content analysis, coding categories are organized as lists, thus making it impossible to find out who did what, for instance or more in general the relations between different entities.

3. A grammar produces coded output that preserves much of the wording and narrative flavor of the original input text; coded output, then, will possess the same degree of semantic coherence of the original input.

4. It will give you the possibility to search for relations between the social actors in your data, using automated methods.

9


- Human coders or automated information extraction phase

- Information extraction phase based on a codebook

- Relations are the units of measurement

- Validity, are we really measuring what we want to measure?



QUANTITATIVE NARRATIVE ANALYSIS (QNA)

10


STORY GRAMMARS- Structured data VS unstructured data

- “[A story grammar] is nothing but the simple linguistic structure subject-action-object or actor-action-actor with their respective modifiers (for example, number of actors involved, type of actor, time and space of action, reason, outcome)” - Roberto Franzosi, From Words To Numbers; Narrative, data, and social science; Cambridge: Cambridge University Press, 2004.

- The approach of QNA exploits the invariant linguistic structural properties of narrative (namely, the chronological sequential order of narrative clauses and their simple linguistic structure SVO, or Subject-Verb-Object). In narrative, Subjects are typically social actors, Verbs are social actions, and Objects are either social actors or physical objects.

- Each SVO element can also have attributes, namely, the characteristics of both Subject and Object.

- The SVO and their attributes provide an invariant structure of narrative also known as “story grammar”



STORY GRAMMARS- Semantic triplets are reliable coding units compared to whole documents, paragraphs or single

words

- Strict rules of grammar guarantee reliability of the collected data and rigorous and systematic handling of textual materials, even in different languages (subject-verb-object structure)

- Compatible with texts of any length –> increased resolution than other approaches

- The structure of semantic triplets is very similar to the structure of populist discourse

- Preserve qualitative information of documents, facilitating mixed-method research

12


- Dealing with non-narrative documents

- Capture relations between words and concepts exploiting linguistic properties of statements

- In the social sciences, framing comprises a set of concepts and theoretical perspectives on how individuals, groups, and societies, organize, perceive, and communicate about reality.


SEMANTIC GRAMMARS

Taken together, the primary concepts developed within framing theory can be summarized asshown in Table 1.

The frame literature does not directly investigate the relations between framing tasks and framecomponents. And yet, I suggest we can draw a set of connections between the two. Let us see how.

The diagnostic task defines problems and faults; it attributes the causes of negative societalissues to specific actors. Hence, diagnostic elements define blames and responsibilities of unjustactions. Benford and Snow (2000: 615) suggest that diagnostic framing has to do with thearticulation of Gamson’s (1992a,b) ‘‘injustice frames’’ or ‘‘injustice components.’’ There is, infact, a direct link between diagnosis, as a framing function, and injustice, as a frame component.Elements playing a diagnostic task seem to generate the specific dimension of injustice. Benfordand Snow (2000: 616) also underline how ‘‘adversarial framing’’ (Gamson, 1992b) is oftendevised by social movement activists to define diagnostically what is just and what is unjust.

Prognosis, the second framing task, involves the elaboration of a ‘‘plan of action,’’ a set ofpossible solutions to societal problems. This requires the proposition of future ameliorativeactions able to change societal conditions. As Benford and Snow (2000: 617) put it: ‘‘prognosticframing activity typically includes refutation of the logic or efficacy of solutions advocated byopponents as well as a rationale for its own remedies.’’ Indeed, the prognostic task conveys asense of agency, the possibility for an action to occur and confute a normalized situation. Whiledescribing the agency component of collective action frames, Gamson (1992b: 59) underlinesexactly how this frame dimension tends to overcome the sense of ‘‘collective helplessness’’ thatmight be perceived by a general public. Thus, those prognostic elements framing possible futureactions are very likely to generate an agency dimension within a collective action frame.However, prognostic elements might express more than possible future actions. Prognosis doesnot only imply the possibility for an ameliorative action to occur, it also conveys the intention ofan actor to actually engage in that ameliorative action. In other words, prognostic elements do notonly express agency, they also convey identity elements.

While the identity component may be signaled by certain prognostic elements – thoseexpressing an actor’s intention to engage in action – it can also be delivered by motivationalelements. Motivational framing provides the rationale for engaging in prognostic action. Thisframing task must build salient elements to raise a movement’s participation levels (Snow andBenford, 2000: 617). Several factors might come into play in this process: a movement’s history,its credibility, and the resonance of its claims to the public. And these factors cannot but build partof a social movement’s identity. Hence, I suggest that identity, as a frame component, can beconsidered as a multilayered dimension generated by both prognostic and motivational elements.On the basis of these theoretical assumptions, framing tasks and frame components could then berelated as shown in Table 2.

S. Vicari / Poetics 38 (2010) 504–525 507

Table 1Central frame concepts.

Element Value Meaning Authors

Framing task Diagnostic Problem identification Snow and Benford (1988, 2000)

Prognostic Proposed solutionMotivational Call for action

Frame component Injustice Moral indignation Gamson (1992a,b)

Agency Potential of actionIdentity Self-recognition

Stefania Vicari, Measuring collective action frames: A linguistic approach to frame analysis

13


SEMANTIC GRAMMARS

- Frame analysis investigates processes of signification: it looks at the way meaning becomes functional to mobilize collective action.

- The scheme of signification that frames belong to can be understood as the ‘‘mental structure’’ (Johnston, 1995) that social movement participants adopt to get involved in collective action. Frames, as ‘‘schemata of interpretation’’ (Goffman, 1974), can be uncovered, reconstructed and made explicit through the analysis of their textual form.

14


FROM TEXT TO FRAMES, HOW LINGUISTS CAN HELP- If frames are represented by textual content, the search for a tool to conduct frame studies should

lead us towards textual analysis strategies.

- A semantic grammar is the essential semantic structure that a text can be translated into. It depends upon the linguistic properties of the text.

- As a consequence, every semantic element coded in the text is marked by a series of categories and subcategories it belongs to.

Abstract from the inauguration speech of Donald Trump (19-01-2017)

“For too long, a small group in our nation's capital has reaped the rewards of government while the people have borne the cost. “

<1> (Paragraph: (<1.1> Clause: (type of clause: adversarial) (Subject: (actor: a small group in our nation’s capital))(Process: (Type of process: (Verbal predicate: (non-modal: (Action: (verb: has reaped)))))) (Object: (non-actor: the rewards of government))) (clauses relation: while)(<1.2> Clause: type of clause: adversarial)(Subject: (actor: the people)) Process: (Type of process: (Verbal predicate: (non-modal: (Action: verb: have borne)))) (Object: (non-actor: the cost))))

15


If a verb has a passive voice, like:

“We will be protected by the great men and women of our military and law enforcement.”

It may be useful to invert the agent and the object coding as subject “the great men and women of our military …” and as object “We”.

<1> (Paragraph: (<1.1> Clause: (type of clause: self-referential) (Subject: (actor: the great men and women of our military and law enforcement)) (Process: (Type of process: (Verbal predicate: (modal: (Intention: ((auxiliary modal: will) (verb: protect))))))) (Object: (actor: We))))

FROM TEXT TO FRAMES, HOW LINGUISTS CAN HELP

16


MORE ON FRAMINGSocial movements frame discourse to provide movement members with a ‘‘collective us’’ in opposition to an ‘‘adversarial them’’ (Gamson, 1992).

In the analysis of social movements’ textual production, Stefania Vicari (Measuring collective action frames: A linguistic approach to frame analysis) suggests one should distinguish between ‘‘self-referential’’ and ‘‘adversarial’’ passages.

Self-referential: stories and descriptions whose subject is the social movement itself or one of its allies.Adversarial: all passages having an oppositional actor as the subject.

For the frame semantic analysis to be significant, subjects must be categorized as ‘‘collective us’’ or ‘‘adversarial them’’ and processes must be clustered by their modal or non-modal lexical semantics

Adversarial passages: i.e., with subject ‘’They’’, frame enemies and faults

i.e. From Donald Trump inauguration speech:

“And while they celebrated in our nation's capital, there was little to celebrate for struggling families all across our land.”

17


NON-MODAL SEMANTICS

All remaining self-referential processes delivered by non-modal predicates will also requirecareful investigation. State processes are central for a social movement to define and characterizeits being. ‘‘Narrative proper’’ (Franzosi, 2010: 29) processes provide the movement’s ownhistory. These processes, together with mottos and claims, all participate in the motivationalcause.

The analytical method so far introduced suggests a textual analysis able to evaluate thepresence of actors and processes and to distinguish among different types of processes. In order tomeet this goal we now need to elaborate a specific semantic grammar.

2.3. Frame semantic grammar

Drawing upon the previous sections, we can assume that the three core framing tasks can bereconstructed as shown in Table 4.

As noted, adversarial processes (i.e., with subject ‘‘They’’) frame enemies and faults. Bydepicting problems and attributing them to adverse actors they build the diagnostic dimension.Self-referential processes (i.e., with subject ‘‘We’’) of obligation complete the diagnosticdimension by depicting actions that must be done to address the problematic issues.

Self-referential processes of ability/possibility and of intention deliver future actions that can/will be done to address the adverse situation. They generate the prognostic dimension byintroducing future ameliorative actions.

To conclude, the movement’s history, definition and characterization build the motivationaldimension by providing salient elements to the public.

The modality-based semantic grammar we intend to adopt requires a set of linguistic codingcategories able to recognize and cluster different forms of modality. Our frame semanticgrammar will then need to focus specifically on actors, processes, and their relational patterns. Iborrow Franzosi’s (1989, 1994, 1998, 2010) use of rewrite rules to outline the skeleton of thissemantic grammar, and produce the coding scheme for the textual analysis. In the clause, whichis the lowest hierarchical level of analysis, a first actor, playing the role of subject and, possibly, asecond actor, in the role of object engage in a process. According to the researcher’s interests,actors and processes may be categorized through different modifiers such as gender, age, role,and tense, to name a few. But, most importantly for the frame semantic analysis to be significant,subjects must be categorized as ‘‘collective us’’ or ‘‘adversarial them’’ and processes must beclustered by their modal or non-modal lexical semantics. The following tree graph shows howprocesses can be categorized (for a whole skeleton of frame semantic grammar please refer toAppendix B in the online supplement) (Fig. 2).

S. Vicari / Poetics 38 (2010) 504–525514

Table 4

Extracting diagnosis, prognosis and motivation from a text.

Subject

We They

Process Modal Obligation Diagnosis Diagnosis

Ability/possibility Prognosis DiagnosisIntention Prognosis Diagnosis

Nonmodal Action Motivational Diagnosis

Character Motivational DiagnosisDefinition Motivational Diagnosis

- Non-modal semantics, linguistic tasks:

- Action (“he went”)

- Character (“he has power”)

- Definition (“this is a global movement”)


18


MODAL SEMANTICS- Modal semantics, framing tasks:

- Diagnosis: “We must take action”

- Prognosis: what can be done (ability), what will be done (possibility)

- Motivation: Intention

Table 3 shows a list of primary and quasi-modal verbs, delivering obligation, ability/possibility and intention in their specific uses.

The absence of one of the primary modal auxiliary verbs does not preclude the presence of amodal clause. We can extract modality by interpreting utterances of sentences with modaltendency (Klinge, 1993: 346). Roberts et al. (2007: 7) point out: ‘‘A modal clause is recognizablewhenever it conveys intentionality in a way that can be transformed (in a manner agreeable to anative speaker) to a form that includes a modal auxiliary verb.’’ Hence, for instance, the clause‘‘They have to stop’’ can be transformed into ‘‘They must stop’’ or ‘‘We want to fight’’ could beturned into ‘‘We are willing to fight.’’

Investigating the use of different modalities in a text can help the researcher analyse howintentionality is shaped through words. Roberts (2008) provides one of the rare applications ofmodality analysis for sociological purposes, defining modality as the ‘‘language-of-motivation.’’He presents an extensive evaluation of the potential of modality analysis to study intentionality indifferent cultures (Roberts, 2008; Roberts et al., 2007). By contrast, in this paper I addresswhether modality analysis could help build an effective frame analysis design. The next sectionwill show that the answer lies in the lexical semantics of modal passages.

2.2.2. Self-referential diagnosis, prognosis, and motivationWhile adversarial framing primarily conveys oppositional statements, self-referential framing

requires a more thoughtful investigation. I suggest that, by studying the linguistic tasks carried outby modal processes, we can try to define the framing tasks produced by self-referential content. Theresearcher should then distinguish among root-modal expressions dealing with the three coremeanings explained in the previous section: obligation, ability/possibility, and intention.

In specific terms, I assume that modal forms expressing obligation in first-person processesessentially deliver diagnostic framing since they imply a diagnostic reason for an action to occur.In the clause: ‘‘We must take action,’’ for instance, we assume that the necessity for the action tooccur is given by some external rationales signaled by the modal verb ‘‘must.’’ Linguists wouldcall it ‘‘deontic modality’’ while social scientists would name it ‘‘diagnostic framing.’’ In fact, themoral sense typical of obligation modality expresses the moral duty of diagnostic framing.

Ability/possibility and intention processes deliver future actions and states in two differenttypes of prognostic framing. While ability/possibility expresses what can be done, given internaland external conditions, intention draws the attention to what will be done to alter a problematiccondition.

S. Vicari / Poetics 38 (2010) 504–525 513

Table 3Linguistic and framing typologies.

Modal Example Linquistic task

Must, ought, should We must do something now Obligation

Have to, be to, need to, be They ought to stopbound to, . . . We need to do something

Can, could, may, might We can control our lives Ability/possibility

Be able to, be We could change the worldcapable to,. . . We are capable of making it work

Will, shall, We will do that for our country Intention

Want to, mean to, be going We want to start action

to,. . . We are going to fight


19


GLOBAL VIEW

Hence, this frame semantic grammar allows the researcher to show where, how and how oftenin a text a social movement portrays adversarial and own processes. The coding scheme alsoallows one to distinguish among passages dealing with self-attributed obligations, abilities/possibilities and intentions. Furthermore, it is able to cluster self-referential action and stateprocesses, with the first delivering the movement’s own history and the second providing eitherits characterization or definition (Halliday and Matthiessen, 2004).

The distance between discourses and collective action frames can then be shortened bytranslating linguistic tasks into framing purposes. This method allows the researcher to groundframe analysis into the investigation of ‘‘manifest (or objective) content,’’ away from qualitativecontent analysis techniques which require more ‘‘impressionistic judgments’’ (Roberts, 1989:148). Yet, in order to capture entirely a collective action frame, we still need to go from framingtasks to frame components.

3. The whole picture

Having found a way to read framing tasks in social movements’ texts, how can we approachframe components? The next level of analysis precisely focuses on collective action frames assystematically built on the three variable components: injustice, agency, and identity.

Given that the attributional function of diagnostic framing assigns causes to problematic issuesand therefore implies a moral definition of societal problems, it was argued that diagnostic elementstend to generate a sense of injustice. The agency component is rather built up by prognosticelements. Specifically, it defines what ‘‘collective us’’ has the ability and the possibility to do in thefuture. To conclude, the identity dimension delivers a movement’s salience, or characteristicfeatures. Gamson (1992a,b: 84) identifies three overlapping layers in the identity component:organisational, movement and solidary group. These different layers may or may not be embeddedin one another according to the history and the context of the collective social actor. I argue that theidentity component is generated by those accounts defining and characterizing ‘‘collective us,’’together with all self-referential narratives reporting its past and present history. As noted, theidentity component is also generated by all those prognostic actions delivering the movement’sintentions. Modality of intention, in fact, is ‘‘dynamic’’ (Klinge, 1993; Palmer, 2001) bydefinition—i.e., it reports the agent’s internal disposition towards an action.

S. Vicari / Poetics 38 (2010) 504–525 515

Fig. 2. Process tree graph.

This frame semantic grammar allows the researcher to show where, how and how often in a text a social movement portrays adversarial and own processes.


20


ANALYZE THE DATA

analyzing massive collections of text has been essentially impossible for all but the most well-fundedprojects.

We show how automated content methods can make possible the previously impossible in pol-itical science: the systematic analysis of large-scale text collections without massive fundingsupport. Across all subfields of political science, scholars have developed or imported methodsthat facilitate substantively important inferences about politics from large text collections. Weprovide a guide to this exciting area of research, identify common misconceptions and errors,and offer guidelines on how to use text methods for social scientific research.

We emphasize that the complexity of language implies that automated content analysis methodswill never replace careful and close reading of texts. Rather, the methods that we profile here arebest thought of as amplifying and augmenting careful reading and thoughtful analysis. Further,automated content methods are incorrect models of language. This means that the performance ofany one method on a new data set cannot be guaranteed, and therefore validation is essential whenapplying automated content methods. We describe best practice validations across diverse researchobjectives and models.

Before proceeding we provide a road map for our tour. Figure 1 provides a visual overview ofautomated content analysis methods and outlines the process of moving from collecting texts toapplying statistical methods. This process begins at the top left of Fig. 1, where the texts are initiallycollected. The burst of interest in automated content methods is partly due to the proliferation ofeasy-to-obtain electronic texts. In Section 3, we describe document collections which political sci-entists have successfully used for automated content analysis and identify methods for efficientlycollecting new texts.

With these texts, we overview methods that accomplish two broad tasks: classification andscaling. Classification organizes texts into a set of categories. Sometimes researchers know thecategories beforehand. In this case, automated methods can minimize the amount of laborneeded to classify documents. Dictionary methods, for example, use the frequency of key wordsto determine a document’s class (Section 5.1). But applying dictionaries outside the domain forwhich they were developed can lead to serious errors. One way to improve upon dictionaries are

Fig. 1 An overview of text as data methods.

Justin Grimmer and Brandon M. Stewart2

at Bibliothekssystem

Universitaet H

amburg on February 13, 2016

http://pan.oxfordjournals.org/D

ownloaded from

Grimmer & Stewart: Text as Data: The Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts, Political Analysis Advance Access, January 22, 2013

21


ANALYZE THE DATA

- Data preparation: involves collecting and checking your data for accuracy

- Descriptive statistics: are used to describe the basic features of the data in a study

- Inferential statistics: investigate questions, models and hypotheses (i.e. infer from the sample data what is the most significant reason of conflict between two entities)

22


DESCRIPTIVE STATISTICS- Helpful to simplify large amounts of data in a sensible way

- Each descriptive statistic reduces lots of data into a simpler summary

- Every time you try to describe a large set of observations with a single indicator you run the risk of distorting the original data or losing important detail

- Univariate analysis: (i.e. the frequency of adversarial clauses in multiple speeches)

- Distribution: the actual frequency in each speech

- Central tendency: mean, median, mode of the previous distribution

- Dispersion: range, standard deviation of the previous distribution

- Exploiting relations between elements in a story grammar helps to discover relations between coded entities such as who does what

- If your grammar contains location information you can compute general information about where did specific events took place and with what frequency and consequences

23


OTHER ANALYSES (1)Try to reach conclusions that extend beyond the immediate data alone. For instance, we use inferential statistics to try to infer from the sample data what the population might think. Or, we use inferential statistics to make judgments of the probability that an observed difference between groups is a dependable one or one that might have happened by chance in this study.

- We can exploit the relational properties of the story grammar and combine them with the characteristics of relational databases

- Usage of aggregate codes for specific elements such as actions and actors may help with computation of useful statistics



Focus on collective action frames as systematically built on the three variable components: injustice, agency, and identity.

1) Given that the attributional function of diagnostic framing assigns causes to problematic issues and therefore implies a moral definition of societal problems, it was argued that diagnostic elements tend to generate a sense of injustice.

2) The agency component is rather built up by prognostic elements. Specifically, it defines what ‘‘collective us’’ has the ability and the possibility to do in the future.

3) The identity dimension delivers a movement’s salience, or characteristic features.

Indeed, dealing with mutually exclusive framing categories involves a slight simplification ofcognitive processes: the exclusion of fuzzy categories might affect partially the results of a framestudy. However, by tracing the key theoretical links between the primary frame concepts, one canfirst look at textual elements as holding specific cognitive functions (i.e., framing tasks), and thenchart these cognitive functions as delivering the relevant cognitive dimensions (i.e., framecomponents).

1.2. Frame methodology

While we can refer to an extensive conceptual elaboration of frame dynamics, we still lack astrong methodological terrain. The long series of empirical studies that flourished from the 1990sonwards has mostly contributed to the explanation of frame concepts and processes (Benford,1997: 411). The central foci of investigation comprise frames’ tasks and components, theircharacteristic and variable features (Benford and Snow, 2000: 614–622; Snow and Benford,1992: 136–141), and master frame generative processes (Benford and Snow, 2000: 622–627;Snow et al., 1986: 467–476). The methods adopted in this first bulk of empirical analyses rangefrom historical case studies (D’Anjou, 1996; Mooney and Hunt, 1996; Noonan, 1995; Snow andByrd, 2007; Williams, 1995) to event-history analysis (McCammon, 2001), on to questionnairesurveys (Erwin, 1993), interviews (Benford, 1993a,b; Capek, 1993; Cress and Snow, 2000;Griffin, 1992; Jenness, 1995; Zuo and Benford, 1995), participant observations (Benford,1993a,b; Gamson, 1992b; Zuo and Benford, 1995), focus groups (Gamson, 1992b), textualanalyses of different sorts (Benford, 1993a,b; Berbrier, 1998; Cadena-Roa, 2002; Cress andSnow, 2000; Diani, 1996; Ellingson, 1995; Ferree, 2003; Gerhards and Rucht, 1992; McCaffreyand Keys, 2000; Rohlinger, 2002; Rothman and Oliver, 1999; Zuo and Benford, 1995), thematicanalyses of newspaper articles and official statements (Babb, 1996; Coy and Woehrle, 1996;Entman and Rojecki, 1993; Jenness, 1995; McCarthy, 1994), quantitative analyses of newspaperarticles or archival resources (Entman and Rojecki, 1993; McCammon et al., 2007; Snow et al.,2007), graphic representations of argumentative structures (Gerhards and Rucht, 1992), andrhetoric analyses (Williams, 1995). While addressing very similar research questions – focusedon frame dynamics – several authors have therefore devised the most different methodologicaltools (for an additional elaboration of the empirical studies here mentioned, please see theAppendix A in the online supplement).

Gerhards and Rucht (1992: 573) wrote: ‘‘Although the importance of framing to mobilizationprocesses has been emphasized time and again in recent years. . .there are no empirical analysesof frames and framing processes.’’ Drawing attention to the lack of empirical studies involvingframe analyses, Gerhards and Rucht actually proposed one of the first methodologicaldiscussions concerning frames and meaning construction in the sphere of collective action. Whywas frame analysis controversial on pragmatic grounds?

S. Vicari / Poetics 38 (2010) 504–525508

Table 2Relations between framing tasks and frame components.


OTHER ANALYSES (2)

25


DONALD TRUMP 19-01-17 INAUGURATION SPEECH“[…] For too long, a small group in our nation's capital has reaped the rewards of government while the people have borne the cost.

Washington flourished, but the people did not share in its wealth. Politicians prospered, but the jobs left and the factories closed.

The establishment protected itself, but not the citizens of our country. Their victories have not been your victories. Their triumphs have not been your triumphs. And while they celebrated in our nation's capital, there was little to celebrate for struggling families all across our land.

[…] We will bring back our jobs. We will bring back our borders. We will bring back our wealth. And we will bring back our dreams.”



DONALD TRUMP 19-01-17 INAUGURATION SPEECH“[…] For too long, a small group in our nation's capital has reaped the rewards of government while the people have borne the cost.

Washington flourished, but the people did not share in its wealth. Politicians prospered, but the jobs left and the factories closed.

The establishment protected itself, but not the citizens of our country. Their victories have not been your victories. Their triumphs have not been your triumphs. And while they celebrated in our nation's capital, there was little to celebrate for struggling families all across our land.”

Simplified Grammar:

<clause> -> {<semantic triplet>}

<semantic triplet> -> <subject>, <process>, <object>, type of clause

<subject> -> subject name, subject type

<object> -> object name, object type

<process> -> <verbal phrase>, process type, aggregate code

<verbal phrase> -> [modal], verb, negation



DONALD TRUMP 19-01-17 INAUGURATION SPEECH“Washington flourished, but the people did not share in its wealth.“

<1. semantic triplet> -> <subject>, <process>, [<object>], type of clausetype of clause -> adversarial

<subject> -> subject name, subject typesubject name -> Washingtonsubject type -> elite

<process> -> <verbal phrase>, process type, aggregate codeprocess type -> actionaggregate code -> prosper

<verbal phrase> -> [modal], verb, negation verb -> flourished negation -> false



“Washington flourished, but the people did not share in its wealth.“

<2. semantic triplet> -> <subject>, <process>, [<object>], type of clausetype of clause -> self-referential

<subject> -> subject name, subject typesubject name -> the peoplesubject type -> people

<process> -> <verbal phrase>, process type, aggregate codeprocess type -> actionaggregate code -> exclude

<verbal phrase> -> [modal], verb, negation verb-> sharenegation -> true

[<object>] -> object name, [object type] object name -> wealth


DONALD TRUMP 19-01-17 INAUGURATION SPEECH

29


“We will bring back our jobs.”

<1 semantic triplet> -> <subject>, <process>, [<object>], type of clausetype of clause -> self-referential

<subject> -> subject name, subject typesubject name -> Wesubject type -> people

<process> -> <verbal phrase>, aggregate code, process typeprocess type -> intention aggregate code -> accomplish

<verbal phrase> -> [modal], verb, negation [modal] -> will verb -> bring back negation -> false

[<object>] -> object name, [object type] object name -> our jobs


DONALD TRUMP 19-01-17 INAUGURATION SPEECH

30


DONALD TRUMP 19-01-17 INAUGURATION SPEECH“Washington flourished, but the people did not share in its wealth. Politicians prospered, but the jobs left and the factories closed.”

<Paragraph> -> {<Clause>}

<Clause> -> <Participant<S>>, <Process>, <Participant<O>>, Type of Clause

<+Participant<S>> -> <+Actor>, <+Object>

<+Actor> -> <+Individual Actor>, <+Collective Actor>, <+Organization>, Actor Aggregate Code

<+Organization> -> <+Institution>, Role in the Organization

<+Individual Actor> -> Individual Actor Aggregate Code, <+Personal Characteristics>

<+Collective Actor> -> Collective Actor Aggregate Code, <+Collective Characteristics>

<+Personal characteristics> -> Name, Sex, Nationality, Political party, …

<+Process> -> Type of process, Verbal Phrase, Aggregate Code

…



DATA MANAGEMENT- Large amounts of data with complex relational properties require appropriate software tools

- CAQDAS softwares: PC-ACE, MAXQDA, NVivo, …


www.pc-ace.com

32

http://www.pc-ace.com


SETUP YOUR GRAMMAR

33


DATA ENTRYOriginal sentence

SVO triplet Subject Subject-We Verb Verb-negation

object object-adjective1

object-adjective2

Circumstances-time

Circumstances-Space

Together, we will determine the course of America and the world for many, many years to come.

We The citizens of America

will determine the course of America and the world for many, many years to come

34


DATA ENTRY

35


DATA QUERYING

36


TERM FREQUENCY3-GRAM FREQUENCYwill make america 5we will make 5will bring back 4we will bring 4bring back our 4together we will 3of our country 3again we will 3you god bless 2women of our 2will be protected 2transferring power from 2

37


GEOGRAPHIC INFORMATION SYSTEMS (GIS)A geographic information system (GIS) is a system designed to capture, store, manipulate, analyze, manage, and present spatial or geographic data.

38


SENTIMENT ANALYSIS

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97

Sentiment

Spechsection

Sentiment values3:verypositive2:neutral1:negative0:verynegative

Sentiment analysis done with Stanford CoreNLP toolkit, smoothing: moving average, window size =10

39


SENTENCE ANALYSIS WITH STANFORD CORENLP


http://nlp.stanford.edu:8080/corenlp/

40


VERB ONTOLOGIES: FRAMENET

https://framenet.icsi.berkeley.edu

41

https://framenet.icsi.berkeley.edu


TOPIC MODELING

Lynchings in Georgia 1875-1935

42


REFERENCES- Matthijs Rooduijn & Teun Pauwels, Measuring Populism: Comparing Two Methods of

Content Analysis, West European Politics 34:6, 1272-1283, 01 Nov 2011

- Stefania Vicari, Measuring collective action frames: A linguistic approach to frame analysis, Poetics 38 (2010) 504–525

- Roberto Franzosi, From Words To Numbers, Narrative, data, and social science; Cambridge: Cambridge University Press, 2004.

- Sudhahar, De Fazio, Franzosi & Cristianini, Network Analysis of Narrative Content in Large Corpora, Natural Language Engineering, September 2013

- Grimmer & Stewart: Text as Data: The Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts, Political Analysis Advance Access, January 22, 2013

- Kirk A. Hawkins, Is Chávez Populist? Measuring Populist Discourse in Comparative Perspective, Comparative Political Studies, Volume 42, Number 8, August 2009

43

Documents

THEORY AND APPLICATION OF QUANTITATIVE NARRATIVE ANALYSIS ...jpthomeczek.de/wp-content/uploads/2017/11/purpura... · QUANTITATIVE NARRATIVE ANALYSIS (QNA) Quantitative Narrative Analysis