Operationalization of framing constructs in the NLP pipeline
| Framing dimension | Theoretical basis | NLP operationalization | Analytical technique | Interpretive role in this study |
|---|---|---|---|---|
| Problem definition | Entman (1993) | Identification of dominant ESG themes and issues | Topic modelling (LDA) + clustering | Captures how sustainability issues are defined and categorized in project discourse |
| Diagnostic framing | Benford and Snow (2000) | Detection of negative or neutrally framed issues (e.g., risk, compliance challenges) | Sentiment analysis (VADER compound, polarity scores) | Indicates how ESG issues are problematized or contested |
| Prognostic framing | Benford and Snow (2000) | Positive framing linked to solutions, innovation, or improvement | Sentiment analysis + co-occurrence with digital keywords | Reflects how ESG issues are positioned as solvable through actions or technologies |
| Motivational framing | Benford and Snow (2000) | Degree of evaluative intensity and rhetorical emphasis | Subjectivity scores (TextBlob) + sentiment strength | Captures how discourse encourages, justifies, or legitimizes action |
| Salience/Emphasis | Entman (1993) | Frequency and recurrence of keywords and topics within clusters | Keyword frequency analysis + cluster density | Identifies which ESG issues are prioritized or made more visible |
| Lexical framing | Framing theory + ESG discourse literature | Distinct vocabulary patterns across E, S, G (e.g., technical vs relational vs procedural language) | Keyword extraction + co-occurrence analysis | Reveals how ESG dimensions differ linguistically and conceptually |
| Digital framing (Technological mediation) | DSS and digital transformation literature (Marques and Ferreira, 2020; Power, 2022; Maibaum et al., 2024) | Co-occurrence of ESG topics with digital terms (e.g., AI, dashboard, automation) | Keyword dictionary + contextual filtering | Captures how digital technologies shape the presentation and perceived legitimacy of ESG issues |
| Framing intensity/Tone | Entman (1993) | Overall evaluative polarity and variation across ESG dimensions | Aggregated sentiment statistics + comparative tests (Kruskal–Wallis) | Enables cross-dimensional comparison of ESG discourse tone |
| Framing dimension | Theoretical basis | NLP operationalization | Analytical technique | Interpretive role in this study |
|---|---|---|---|---|
| Problem definition | Identification of dominant ESG themes and issues | Topic modelling (LDA) + clustering | Captures how sustainability issues are | |
| Diagnostic framing | Detection of negative or neutrally framed issues (e.g., risk, compliance challenges) | Sentiment analysis (VADER compound, polarity scores) | Indicates how ESG issues are | |
| Prognostic framing | Positive framing linked to solutions, innovation, or improvement | Sentiment analysis + co-occurrence with digital keywords | Reflects how ESG issues are | |
| Motivational framing | Degree of evaluative intensity and rhetorical emphasis | Subjectivity scores (TextBlob) + sentiment strength | Captures how discourse | |
| Salience/Emphasis | Frequency and recurrence of keywords and topics within clusters | Keyword frequency analysis + cluster density | Identifies which ESG issues are | |
| Lexical framing | Framing theory + ESG discourse literature | Distinct vocabulary patterns across E, S, G (e.g., technical vs relational vs procedural language) | Keyword extraction + co-occurrence analysis | Reveals |
| Digital framing (Technological mediation) | DSS and digital transformation literature ( | Co-occurrence of ESG topics with digital terms (e.g., AI, dashboard, automation) | Keyword dictionary + contextual filtering | Captures how digital technologies |
| Framing intensity/Tone | Overall evaluative polarity and variation across ESG dimensions | Aggregated sentiment statistics + comparative tests (Kruskal–Wallis) | Enables |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.