Elaboration of Underpinning Methods and Data Analysis Process of Directed Qualitative Content Analysis for Communication Studies

Introduction

This paper discusses the rationale and research design for a deductive qualitative study. It explains the rationale behind implementing a directed QCA method to analyze textual/video speech for trustworthiness. Moreover, the paper discusses recent methodological advancements/studies on directed QCA, employing it distinctively from previous studies proposing a redevised method. The study describes data collection and administration. The researchers analyze the text using a qualitative paradigm and an interpretative method (Strauss, 1987) or a study design to understand precise descriptions of specific phenomena (Miller et al., 2018; Maxwell, 2012) in political communication. A qualitative research technique explores these communicative intricacies, focusing on the perceptive meanings of a phenomenon (Creswell & Poth, 2018) and contextual explanations (Avgerou, 2019). Qualitative content analysis is often used to study communication (Mufwene et al., 2017). So, the present research’s design assisted in understanding the methodological approach that was employed to examine former Pakistani PM Imran Khan’s (hereafter, IK) spiritual leadership communication and its comparative analysis with King Abdullah II of Jordan from a Muslim political-cum-religious setting using content analysis from speeches employing directed qualitative content analysis.

The substance of a ‘message’ might refer to message or communication studies. It is called ‘elaboration of message’ through examining message components (see Petty & Cacioppo 1986) in communication analysis, and its description is debated (Dance, 1970; Craig, 1999). Because everyone communicates (Goldenberg et al., 2001), defining communication is challenging (Peters, 1986). One way to define ‘communication’ is to limit it to specific types of exchanges (Haythornthwaite, 2002), such as having a ‘conscious aim’ to persuade (Miller, 1966). Communication is how people, organizations, or both use language, signs, symbols, and semiotic conventions understood by all parties (Gretzel & de Mendonça, 2019; Olurotimi, 2022; Peluso, 2021). The study used this method focused on verbal (human) communication, in which leaders send their message (spiritual leadership) to followers (the receiver) through a channel (speeches). The study aims to evaluate the ‘message’ using qualitative content analysis. "Human (symbolic) communication that alters the attitudes and behaviors of others to accomplish shared group goals and needs" (Johnson & Hackman, 2018, p. 12). The previously mentioned researchers/scholars elaborated the definition based on leadership communication studies. The researchers found earlier academics researching leadership from communication perspectives considered transformational leadership elements equal/parallel to communication (see Aslam, 2024; Aslam et al., 2023). Suppiah et al. (2022) examined political leadership transformational leadership communication. They considered transformational leadership communication equal/parallel communication without using any communication theory. They stated that leaders communicate their vision, trust, honesty, and transparency to their followers, emphasizing Tun Mahathir’s vision and beliefs. Aslam (2024) studied political communication in a transformational and spiritual leadership context yet employed the transformational leadership communication of Johnson & Hackman (2018). So, the present researchers examine how IK and King Abdullah II of Jordan modified and transmitted spiritual leadership (see Table 3.1) through their communication on an international platform. The present research would be innovative and novel in the context of the following aspects;

  1. The study aims to present a novel method and process of analysis within the framework of a directed qualitative content analysis approach. For instance, previous research devised to employ the DCQA 16, 8, and 7 steps (Assarroudi et al., 2018; Aslam et al., 2023; Kibiswa, 2019; Mayring, 2014; Zhang and Wildemuth, 2009 for more details) methods. However, the present article offers a four-step conceptual method (see section 3 for more details) that would be understandable to a layperson/researcher.
  2. While the third co-author of this research had previously adapted and utilized DQCA in communication, this paper presents the first-ever four-step conceptual model in communication.
  3. Prior studies on methods have deemed interviews necessary for data collection in the QCA process (Assarroudi et al., 2018; Mayring, 2014; Zhang & Wildemuth, 2009). The current study posits that including transcribed or direct textual content is a necessary component that can be derived from various sources, such as videos, audio recordings, social media posts, and tweets.

Literature Review

Various researchers have expanded methodological advancements in qualitative content analysis across different fields. However, some scholars have introduced a method known as directed QCA, a subset of QCA, which employs a distinct approach. For instance, it can be employed inductively, deductively, or summatively (Elo & Kyngäs, 2008; Hsieh & Shannon, 2005; Mayring, 2000; Mayring, 2014). QCA evaluates textual material using a systematic coding procedure (Assarraudi et al., 2018), encompassing words, phrases, statements, documents, images, or even entire communication without quantification (Kibiswa, 2019; Mayring, 2000). Data analysis yields patterns, themes, or categories (Assarroudi et al., 2018; Zhang & Wildemuth, 2009; Elo & Kyngäs, 2008). Experts and researchers apply QCA inductively, deductively, or summatively (Hsieh & Shannon, 2005; Mayring, 2000; Mayring, 2014). The inductive approach to QCA aids in developing schematic models, theories/themes, conceptual/theoretical frameworks, or theories as the most commonly used data analysis approach (Vaismoradi et al., 2013, 2016); it could be tested and refined by employing it with QCA (Aslam, Alsharairi, Barzani et al., 2024; Elo & Kyngäs, 2008). DQCA is employed in various domains, but its application in political communication remains relatively less explored (Elo & Kyngäs, 2008).

Elo and Kyngäs (2008) determined that content analysis aims to construct a conceptual model that accurately describes the phenomenon. Both inductive and deductive analytical methods consist of three primary phases: preparation, organization, and reporting. The preparatory phase exhibits similarities in both techniques. Concepts are derived from the data through inductive content analysis. Deductive content analysis is employed when the analytical framework is defined based on pre-existing information. Inductive content analysis is employed when there is a lack of prior research addressing the issue or when existing research is incomplete or disjointed. A deductive approach is beneficial when the primary objective is to examine a previous theory in a distinct context or to compare categories across different periods.

Assarroudi et al. (2018) state that qualitative content analysis encompasses traditional, guided/directed, and summative methodologies for analyzing data. These methodologies provide comprehensive knowledge and insights into the investigated topic. They aimed to describe and integrate the data analysis process in directed qualitative content analysis. They utilized multiple databases to access publications about directed qualitative content analysis following a systematic approach for analyzing data in directed qualitative content analysis. They presented a comprehensive 16-step method for conducting directed qualitative content analysis. This method might fill a gap in the existing international literature by explaining the practical qualitative data analysis process. Their proposed method of directed qualitative content analysis appeared to be a dependable, transparent, and thorough approach for qualitative researchers. Implementing this approach would enhance the precision of qualitative data analysis, enable the comparison of findings from various studies, and produce tangible outcomes.

Kibiswa (2019) suggested the Qualitative Content Analysis (QCA) method as a research approach that combines deductive and inductive reasoning. In the inductive technique, the researcher derived categories/themes from the acquired data to initiate the study. In contrast, the deductive approach, also known as the directed approach, derived these categories/themes from existing theories to provide the framework that guides the research, verifying or supporting the relevance of the theory/theories that guide the study or expanding the application of the theory/theories to different contexts or cultures from those in which they were initially produced. The research sought to address the limitations in the qualitative research tradition by introducing an 8-step DQCA (Directed Qualitative Content Analysis) framework. This framework is divided into three phases: Study Preparation, Data Analysis, and Results Reporting. The purpose of this framework was to enable researchers to achieve the same objectives using data that the researcher does not generate.

Deductively, the present paper described and integrated the DCQA approach. The qualitative field rarely used the deductive technique in this strategy (Kibiswa, 2019). The researchers aimed to assess if political communication (debates or speeches) in the international venue (UNGA) incorporates the spiritual leadership model from a Muslim political-cum-religious perspective adapted from Aslam et al. (2023). As indicated, the researchers deduced and redevised the DQCA model from Aslam et al. (2023), Assarraudi et al. (2018) (communication in nursing), Kibiswa, 2019 (conflict analysis), Rasool et al. (2022) (scholarly communication), and Rasool et al. (2023a, 2023b) (applied linguistics) to employ it in political communication. The DQCA has been used exclusively in communication (Berelson, 1952) but not in political communication. Thus, the current study used DQCA to identify unique characteristics of political messages (Holsti, 1968) and designate patterns or categories (Assarroudi et al., 2018).

In previous research, researchers devised to employ the DCQA 16, 8, and 7 steps (see Aslam et al., 2023; Assarroudi et al., 2018; Kibiswa, 2019; Mayring, 2014; Zhang and Wildemuth, 2009 for more details) method. However, in this paper, the researchers proposed a four-step model (a. Sample design, b. Data collection process, c. Specifying unit of analysis, and d. data analysis process) yet deduced from previous research. The present researchers reviewed all possible articles while devising a DQCA strategy for political communication using the previous studies in a newly modified DQCA model.

Implementation of DCQA

Sample Design

After the preparation phase of directed QCA (from Assarraudi et al., 2018) and the acquisition of general skills, the researchers devised a sampling strategy that included ‘key informants’ (Elo et al., 2014) and YouTube debates/speeches using purposive sampling (Coyne, 1997), taking into account unique socio-demographic characteristics (Sandelowski, 1995). Speech data were sampled until saturation or redundancy (Cleary et al., 2014). Following the previous reasoning, textual data from participants’ speeches that met inclusion/exclusion criteria were selected following the sub-steps:

Figure 1.Sample Delimitation Process (adapted and modified from Aslam, 2024).

  1. The study used directed qualitative content analysis on 8 English-language speeches of IK and King Abdullah II of Jordan. Notta Premium transcribed YouTube speeches. The research sample design operationalized the following steps:
  2. The researchers preserved the delimitation of sampling to IK and King Abdullah II of Jordan.
  3. For this delimitation, the researchers created inclusion/exclusion criteria (see Table 2) and provided evidence-based reasoning. ‘The Royal Islamic Strategic Studies Center’ released “The 500 Muslims: The World’s 500 Most Prominent Muslims” from 2018 through 2023 (see Schleifer, 2018; 2019; 2020; 2021; 2022; 2023), categorizing 500 influential Muslims by profession after extensive research.
  4. The selection sample criteria focus on political leadership communication. The book also emphasizes political leadership in the second and first parts. The context favoring the political spectrum ranks them among the top 50 prominent political personalities (see Figure 1 for details).
  5. The research sample comprises the top fifty political figures. The researchers confine it to current (Western) democratic political leadership that is internationally accepted and the Middle Eastern kingship system. Emir(s) and Supreme Leader(s) were excluded from this study.
  6. According to the book ‘The 500 Muslims’, King Abdullah II of Jordan, Tun Mahathir Mohammed, IK, and Muhammadu Bukhari spoke in English at UNGA sessions. The study overlooked Malaysia and Nigeria due to a lack of political-cum-religious knowledge and instead focused on King Abdullah II of Jordan and IK.
  7. Finally, the content analysis includes English speeches/debates/talks. King Abdullah II of Jordan and IK’s speeches/debates/talks were not analyzed in Arabic and Urdu.

Data collection process

Directed QCA requires researchers to choose a text for analysis. Manifest data may comprise a researcher’s readings of the selected text or textual interview data of participants (Elo & Kyngäs, 2008; Assarroudi et al., 2018) or even video data (Aslam et al., 2023).

Researchers did not include pertinent interviews with the speakers as secondary data. Only their speech data were included, as explained in step one. The data were collected from their video speeches, which are available on YouTube. Notta software transcribed the video data into textual data. For these political leaders/figures, interviewing them directly or contacting specialists/relevant personalities (in Pakistan and Jordan) for primary data would be complex. Moreover, interview practice was also impossible in the present case, and there was not colossal funding support. However, any extension research may involve more Jordanian and Pakistani co-authors for interviews or focus group interview tools while gathering the public’s perspective.

Specifying the Unit of Analysis

A diary, interview, conversation, abstracted and coded text, or transcript might be the unit of analysis (Graneheim & Lundman, 2004; Assarroudi et al., 2018). Hence, as guided by the DQCA of Assarroudi et al. (2018) and Aslam et al., 2023), the transcribed text (from video speeches) was the analysis unit for the current study. The present research analyses speech transcriptions as text, and for the discussion portion, King Abdullah II of Jordan and IK’s English interviews with worldwide media are incorporated for more trustworthiness.

Units of Segmentation Means of Qualitative Data for Qualitative Content Analysis (QCA)
Content(s) Analysis
Units of analysis Speeches
Units of coding (Elements in) Speech Text(s)
Context Unit (Other Elements in) Speech Text(s)
Table 1.Units of Analysis and CodingSource: Aslam et al., (2023)

Data Analysis Process

The researchers studied and analyzed the data as often as needed to answer questions about the speaker, place, time, or reason for communication (Assarroudi et al., 2018). Imagine a researcher answered the preliminary questions and retrieved linked meanings from immersed data (Aslam et al., 2023; Kyngäs, 2020; Elo & Kyngäs, 2008; Elo, 2014).

The ‘data analysis phase’ was deduced and redevised from Assarroudi et al. (2018), who represented the ‘organization phase’ and the ‘preparation phase’. So, the present research adapted possible key categories from Aslam et al. (2023) utilizing directed QCA. Mayring (2014) suggested deducing primary associated and classed categories matrix deductively. This creative matrix could also inductively develop new major categories from existing theories or research (Elo & Kyngäs, 2008);

Based on current theory or linked research, categories/subcategories should be objective, exact, and correct (Mayring, 2000, 2014). If the researchers created a category called ‘Vision,’ it should be honestly and accurately defined along with ‘Inspired Vision’ or ‘Vision.’

Major Categorization Operational categorization matrix Definitions/Meanings for Operational Categories
Elements of Spiritual Leadership for Political Communication - “inspired vision” (Hatcher, 1991), Vision (Fry, 2003) “Speech/text or writing suggests a tendency of communication with followers to reach out and embrace something more expensive than themselves.- by clarifying the general direction of change, simplifying more detailed decisions” (Aslam et al., 2023)
Centering-Focus (Hatcher, 1991) “Speech/text or writing suggests a tendency to Center and Focus on the vision” (Aslam et al., 2023).
Integrity (Fry, 2003) “Speech/text or writing suggests a tendency of communication to remain on the path, regardless of where it leads, and communication as the vehicle of change” (Aslam et al., 2023).
Desire-Courage (Hatcher, 1991) “Speech/text or writing suggests a tendency of courage and drive to manifest desires and vision in the world” (Aslam et al., 2023).
“Empowerment of others” (Hatcher, 1991) “Speech/text or writing suggests a tendency to ‘give away power,’ making others feel powerful” (Aslam et al., 2023).
Hope-Faith (Fry, 2003) “Speech/text or writing suggests a tendency to hope and faith in the vision” (Aslam et al., 2023).
Table 2.Coding protocol Source: Aslam et al. (2023)

The researchers created coding criteria based on theoretical definitions to describe the primary/main categorization (Mayring, 2014). Coding rules showed the matrix’s primary categories and subcategories, which may increase the study’s credibility. Hence, theoretical definitions provide theoretical coding rules.

Published research validated the classification matrix (Aslam et al., 2023). In a study with several coders, it may be crucial. Discussing autonomous job challenges is crucial. So, the present research improves the unit-of-analysis interpretations according to its own needs. Elo et al. (2014) suggest improving the classification matrix, which may help improve the inter-coding schema's reliability (Vaismoradi et al., 2013). In this type of research, pre-testing was not possible, and the categorization matrix was also trusted/validated/improved by consulting the second and third authors or peer researchers. Prepublication reviewers of a significant project by Aslam et al. (2023) also suggested improvements for the categorization matrix.

Using meaning units, each primary category was given clear examples (Mayring, 2014). After pre-testing and validating the categorization matrix for this study, samples were selected for each primary category to be anchored. The classification matrix and study objectives recommended selecting meaning units and summing and assigning preliminary codes (Graneheim & Lundman, 2004; Mayring, 2000, 2014) through analysis. Semantic content, similarity, and dissimilarity arrange the preliminary codes. Generic categories result from categorizing (Elo & Kyngäs, 2008). Throughout time, generic and major categories become conceptually and logically linked. This process may also be used to generate categories repeatedly. The study continuously compared data and validated the categorization matrix at various stages (Zhang & Wildemuth, 2009).

Finally, in implementing the reporting phase, the final section details the data analysis and listed findings (Elo & Kyngäs, 2008). To show the raw data classification matrix relationship, the research displayed the findings systematically detailing sampling, data collection, analysis, and participant characteristics specifying trustworthiness standards and how they were met. Elo et al. (2014) created a comprehensive QCA reporting checklist also utilized in the present study.

Discussion

Previous research has identified interviews as an essential element (key informants) for data gathering in the QCA process (Assarroudi et al., 2018; Mayring, 2014; Zhang & Wildemuth, 2009). According to the present study, it is proposed to include transcribed or direct textual content (as key informants) from several sources, such as videos, audio recordings, social media posts, and tweets. Moreover, in this sub-step, the present research offers a data delimitation process not represented in previous research. Although this process was used for political speeches, future researchers can amend it according to their needs.

In previous research, the researchers suggested various steps for the analysis process, i.e., categorization, subcategorization, pre-testing of categorization, and sample anchoring for directed QCA (see Assarroudi et al., 2018). However, the present research conceptually argues that these steps can be done parallelly while deductively analyzing data. The present/mentioned conceptual method is devised explicitly for a deductive approach deducing the categories/themes from the existing theory(ies). The process is explained according to traditional layouts of the research presentation (see Assarroudi et al., 2018; Kibiswa, 2019). However, this may be different in an inductive approach. For instance, for the inductive approach, the research steps could be interchangeable, such as deriving/deducing categories/themes from the acquired data to initiate the study (Kibiswa, 2019).

As argued in step one, the critical data for directed QCA would be textual, so the present research argues that all the data video/audio could be converted into text. Therefore, the sample design of the current investigation included a component in which the coded text was transcribed through third-party software. The research ensured the accuracy of the transcription of the video, and the researchers ran the source material via an online program called Notta, which also enhanced the researchers’ trustworthiness. It was first assumed that the research would take a deductive approach to the aim at hand. Because of this, the researchers utilized a deductive categorization matrix in the study (see Assarroudi et al., 2018; Kibiswa, 2019).

In the present case, when the researchers reached the phase of reporting findings, they realized that if the analysis of spiritual leadership were presented independently, it would not be fully comprehended by the readers. Researching spiritual leadership was a novel approach to political communication. So, researchers agreed to use elements from another leadership style, transformational leadership, for the categorization matrix, which would help readers better understand the matrix. For this purpose, the researchers utilized an inductive methodology to construct an extended categorization matrix. For example, experts and researchers may use qualitative content analysis in an inductive, deductive, or summative manner (Mayring, 2000; Mayring, 2014; Hsieh & Shannon, 2005). Consequently, the primary category for spiritual leadership and the subcategories were formed. Deducing from the previous research, Aslam (2023) and Aslam et al. (2023) employed this model in 16 and 8 steps, respectively. Hence, the present research offers a simplified approach, delimiting it to a 4-step model.

Trustworthiness

Qualitative study validation has many names in the literature (Creswell, 2013). Qualitative studies include dependability, transferability, and trustworthiness, according to Polit and Beck (2020). Moreover, confirmability was also added as the fourth term by Lincoln and Guba (1985). Furthermore, Polit and Beck (2020) have recently listed the term 'authenticity.' Qualitative research on trustworthiness has used these phrases (Polit & Beck, 2020).

So, based on the previous practices by the researchers, i.e., presenting the DQCA in phases—preparation, organization, and reporting (Elo et al., 2014), or steps (Assarroudi et al., 2018; Kibiswa, 2019)—the present research also re/developed four steps maintaining the reliability/trustworthiness, dependability, transferability, or authenticity of the directed QCA for future research. The phases/steps were essential to describe each step taken by the researchers for the formation of the dependable and transformable matrix (Graneheim & Lundman, 2004; Vaismoradi et al., 2016; Elo & Kyngäs, 2008). Tables were also shown to highlight the reduction process for better and more vivid understanding (Elo et al., 2014; Elo & Kyngäs, 2008). The researchers also maintained truthfulness during data analysis (Polit & Beck, 2020). Hence, Assarroudi et al. (2018) suggest using steps in directed QCA as a model to increase its credibility. Therefore, the present research also designed steps to implement the method to maintain credibility.

Conclusion

The present research offers the elaboration and underpinnings of DQCA, or Directed Qualitative Content Analysis, which is a method of content analysis that has been recently elucidated and put into practical use by a limited number of academics on an international scale. They utilized logical reasoning in certain instances, primarily after qualitative research protocols. Unlike the inductive method, which begins with a broad hypothesis and gradually accumulates data to support it, the deductive method, also referred to as the directed approach, involves researchers creating categories/themes and subcategories/subthemes for their research based on pre-existing theories. The current research additionally elucidates this conceptual method within the framework of the deductive approach. Deductive or directed QCA aims to examine, confirm, or expand the range of the study’s underlying theory(s) by utilizing data obtained from sources that are not directly related to its original creation. Researchers used deductive Qualitative Content Analysis (QCA) in prior investigations with 16, 8, or 7 steps. Therefore, this article suggests a four-step conceptual model derived from the researchers’ past works and other studies to fill the gaps in the qualitative research tradition and accomplish the same objectives with the study data. Furthermore, the current research presents a data delimitation technique in this sub-step that has not been seen before. This approach was employed for political speeches, but future researchers can adapt it.

Limitations / Weaknesses

The present paper offers illustrative sample data in tabular format as exemplars for researchers specializing in political communication who may employ DQCA in their forthcoming research. However, this data was transcribed in 'text' from video data. The present method may or may not be implacable on other data types. In previous studies, the researchers proposed using the DCQA 16 or 8-step technique (see Assarroudi et al., 2018; Kibiswa, 2019; Aslam et al., 2023 for further information) that would also be implacable replaceable. Moreover, the conceptual model primarily focuses on and lays out the steps for the deductive approach that would not be implacable for inductive yet with some amendments.

Implications

Nevertheless, the researchers in this work put out a four-step approach (a. Sample design, b. Data collection method, c. Specifying unit of analysis, and d. data analysis procedure) derived from earlier research. The researchers thoroughly examined all available publications to develop a DQCA (Directed Qualitative Content Analysis) method for political communication. This strategy was based on earlier studies and implemented in a newly updated DQCA model. Future researchers may utilize this method for deductive approaches in other fields of study; however, it may be limited to textual data. Likewise, the researchers may utilize this short method for an inductive approach; however, the four steps would be interchangeable.

Acknowledgment Statement: The authors thank the reviewers for their comments, which helped complete this manuscript.

Conflicts of interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Authors' contribution statements: Author 1 contributed to methodology, writing—original draft, project administration, and validation. Author 2 contributed to the investigation, validation, and resources. Author 3 contributed to conceptualization, methodology, investigation, writing—original draft, validation, writing-reviewing, and editing.

Funding statements: This research received no external funding, so it was conducted without financial support from any funding agency or organization.

Ethical consideration statement: Since this study did not involve human or animal subjects, ethical considerations related to participant welfare, informed consent, and privacy were not applicable. However, ethical standards regarding academic integrity, transparency, and proper citation were upheld throughout the research process.

Data availability statement: The paper is methodological in nature, so the data analysis section is not present. However, the raw data will be available at the editor's or reviewers' request. Moreover, coded data through Atlas will also be available if required.

Disclaimer: The views and opinions expressed in this article are those of the author(s) and contributor(s) and do not necessarily reflect JICC's or editors' official policy or position. All liability for harm done to individuals or property due to any ideas, methods, instructions, or products mentioned in the content is expressly disclaimed.