Analysis of research on disability and diversity indexed in Web of Science in communication

1. Introduction

Since the turn of the century, the discipline of Communication is experiencing unstoppable development as an academic field, largely justified by the hegemonic role of digitalisation and information technologies in current society. More proof of the interest aroused by this speciality is its growing academic output periodically published in scientific journals, monographs and symposiums. This prodigious activity has in turn led to the need for researchers to sort and catalogue this vast bibliography as the only way to take a snapshot of the fundamental attributes that determine research in a field of knowledge and also to pinpoint any strengths, weaknesses, opportunities and threats (Caffarel-Serra et al., 2018) for subsequent advancement.

This has led to ‘meta-research’ (Goyanes, 2020), an introspective, self-referential current that, according to Saperas & Carrasco-Campos (2019), ‘enables us to quantitatively describe how researchers carry out their research practices’ (p. 227). The purpose is therefore to understand essential aspects such as dominant theories (Bryant & Miron, 2004; Carrasco-Campos & Saperas, 2022), methods used (Demeter & Goyanes, 2020; Lovejoy et al., 2016), topics studied (Günther & Domahidi, 2017; Walter et al., 2018) or funding declared in papers (Martín-Algarra et al., 2018; Gaitán-Moya et al., 2021).

Based on a wide range of documents and using automated techniques, bibliometric analysis also makes it possible to map different dimensions of a discipline, such as authorship patterns (Trillo-Domínguez & De-Moya-Anegón, 2022), descriptors (Aguado-Guadalupe et al., 2022) or citation structure and dynamics (Segado-Boj et al., 2023); these aspects are no less significant when attempting to understand the mechanisms that govern a certain discipline.

Given the relevance of the systematic observation of academia, the purpose of this study is to analyse outputs addressing disability and diversity matters based on communication by means of bibliometrics and content analysis. The section below reviews the state of the art and then details the methodology used.

2. Disability and diversity in research in communication

As with disability and diversity in real life, relegated and barely visible, these topics have not been a priority in studies in the field of communication.

Two models of the social paradigm of disability were taken into consideration when preparing this paper (Díaz-Jiménez, 2003; Díaz Velázquez, 2009; Jiménez & Huete, 2010). Terminology has varied over the years from the social model of disability (Barton, 1998), which identifies it with political and social issues (Díaz-Jiménez, 2003; Brogna, 2006; Ferreira, 2008; Jiménez & Serrato, 2014), until the adoption of a scheme where society determines what is normative and anything not fitting in this will be equivalent to functional diversity (Moscoso, 2015; Palacios & Romañach, 2006; Romañach & Lobato, 2009; Palacios & Romañach, 2008). Although the two terms are not synonymous, they will be used as such for operability, and to delve deeper into existing research and how it has varied.

A leading paper in this field is by Díaz-Jiménez et al. (2021); a systematic review of scientific output in the field of mediation, social work, disability and functional diversity. After searching in nine databases, a sample of 17 articles were analysed, leading to the conclusion that output on disability and mediation first emerged in the 21st century. Authorship is female (65%), Spanish (41%), and primarily from academic fields (80%) compared to institutions (20%). Regarding topics, few papers can be found on physical or organic disabilities compared to cognitive functional diversity. Types of paper include: 47% research articles, 23% theoretical essays, 18% case studies and the rest are bibliography reviews.

Educational research into functional diversity and digital competence featured in the bibliometric study by Delgado-Vázquez et al. (2019). These researchers observed a sample of 263 papers taken from Web of Science (WoS) and Scopus published between 1982 and 2019. Their analysis pinpointed that scientific output increased as of 2016 and the main source of dissemination is articles published in journals (90.7%). Countries addressing these issues include: United States, Spain and United Kingdom; English is the predominant language of publication. Likewise, Delgado-Vázquez et al. (2019) indicate that there are three thematic trends: how people with functional diversity interact with technology; the relationship between technology and the communication of these same individuals; and, finally, observing links between e-inclusion and digital competence.

Two studies were located that are more related to cultural and leisure industries: video games (López-Gómez et al., 2022) and museums (Llamazares de Prado & Arias, 2020). The first is a literature review of 769 documents, including 384 conference lectures, 279 articles and 40 book chapters. English is the predominant language (734 documents) and output usually comes from the US (179 documents). López-Gómez et al. (2022) conclude by emphasising ‘the need to favour the acquisition of digital skills, the emotional development of players, and the priority of delving deeper into the creation of video games that promote inclusive education’ (p. 2). Llamazares de Prado & Arias (2020) selected 490 articles published between 2002 and 2017, noting that ‘the role of museums as a place for reflection and bringing together different intergenerational groups is evident’ (p. 137).

Based on the scarce evidence found, there is a clear need to map output limited to the field of Communication addressing matters of disability and diversity. The study objectives are specifically:

SO1: detect keywords and bibliographic references most used in articles on disability and diversity from communication.

SO2: analyse output context in terms of authorship, internationalisation and funding, as well as the theoretical-methodological instrumentalisation of this research.

The next section details the methodological procedures used to address these goals.

3. Methodology

Two closely related studies were carried out to meet the fundamental goal and resulting specific objectives. They are described below:

3.1. Study 1: bibliometric analysis

WoS was used to compile all articles indexed between 2012 and 2022 in the ‘communication’ category from both the SocialSciences Citation Index (SSCI) and Emerging Sources Citation Index (ESCI) databases. Despite other databases such as Scopus traditionally being viewed as more representative as they include a greater number of non-English documents and from social science disciplines (Mongeon & Paul-Hus, 2016), since the inclusion of the ESCI, WoS has cushioned this bias and increased its representation in terms of discipline and language (Huang et al., 2017; Repiso & Moreno-Delgado, 2022). Nevertheless, WoS is still the more restrictive source (Singh et al., 2021), so we understand that it reduces noise and offers an adequate representation of the most relevant scientific results.

The search term formula, grouped by Boolean operator, was: ‘disability’ OR ‘functional diversity’ OR ‘disabled person’ OR ‘disabled people’ OR ‘disabled persons’ OR ‘physical disease’ OR ‘mental disease’. Results were downloaded on 2 November 2022 and a database was generated with N = 316 articles, representing the population universe of manuscripts addressing these topics from the perspective of communication.

VOSviewer (van Eck & Waltman, 2010) was then used to create networks based on bibliographic information, and Pajek (Batagelj & Mrvar, 1998) to map co-words (figure 1) and co-citation (figure 2). These graphs show relationships (arcs) between elements (nodes). Therefore, when two words are simultaneously mentioned in the same article or two words are cited in the same list of references, they are connected with a line. The co-word map can identify the thematic structure of a discipline while the co-citation network reveals its foundation and theoretical organisation (McGowan et al., 2022; Moreno-Delgado et al., 2020; Newman, 2001; Page & Capizzo, 2021). Keywords are also harmonised to identify variants of a single concept (e.g., ‘social media’ and ‘social networking sites’).

The principles of representativeness were followed for a clearer visualisation so that only co-appearances (arcs) with a minimum frequency detailed at the bottom of each figure are included. Line thickness is proportional to relationship frequencies and node size to the number of appearances (times each word is used or each reference cited), a figure listed in square brackets. Node colour corresponds to communities identified (clusters). In other words, sets of vertices that tend to be connected. The Louvain algorithm was used to detect them (multi-level thickening, simple refinement, parameter resolution = 1, number of random restarts = 1, maximum number of levels in each iteration = 20, maximum number of repetitions in each level = 50). In the case of the co-citation network, figure 2 only shows the lead author and year of publication; full references are provided in an appendix available on the Open Science Framework. ¹

3.2. Study 2: content analysis

The second study consisted in applying content analysis as a meta-research tool. A probabilistic sample was randomly selected (95% confidence interval, 5% margin of error) from the population universe of study 1 (N = 316). The content analysis corpus thus amounted to n = 174 papers, coded by two authors of this paper by applying a codebook based on similar studies (Bryant & Miron, 2004; Carrasco-Campos & Saperas, 2022; Demeter & Goyanes, 2020; Fernández-Quijada & Masip, 2013; Leydesdorff et al., 2013; Martínez-Nicolás et al., 2019; Walter et al., 2018). The codebook comprises 22 variables: ²

Data on the impact of each article, the personal identity of the corresponding author, their university affiliation (and country) or manuscript language were taken directly from WoS, a procedure that enabled triangulation (Denzin, 2012) with manually inferred variables. The 174 articles were coded during the month of February 2023. To calculate intercoder reliability, a subsample of 34 manuscripts (∼20%) was selected and analysed simultaneously by both judges. We were thus able to calculate Krippendorff’s alpha (Krippendorff, 2011) for each variable by using the Kalpha macro for SPSS (Hayes & Krippendorff, 2007). Mean reliability for the 11 variables requiring manual inference (5 output context and 6 instrumental) amounted to a more than acceptable value³: M (αk) = 0.74.

The following block shows data grouped according to the two studies conducted.

4. Results

4.1. Study 1: bibliometric findings

4.1.1. Keywords

Firstly, table 1 includes the list of most frequent keywords in literature on disability and diversity in communication:

The most frequent term is logically ‘disability’, followed by ‘accessibility’, ‘disability studies’, ‘autism’ or ‘intellectual disability’; all these terms undoubtedly refer to the topics studied. Other terms more specific to the discipline can also be identified: ‘digital divide’, ‘technology’, ‘television’, ‘information and communication technologies’, ‘broadcasting’, ‘Internet’ or ‘Twitter’. Likewise, ‘content analysis’ and ‘critical discourse analysis’ refer to common methods in the field; ‘frame’ or ‘narrative’ to concepts related to framing theory and narrative paradigms respectively (film theory, transmedia storytelling, etc.). The co-word network below (figure 1) graphically recreates the five different clusters:

The first and most numerous is a broad set of largely unrelated terms such as ‘sports’, ‘virtual worlds’, ‘ethics’, ‘social exclusion’ or ‘inclusion’. The heterogeneity and scarce internal connection between the nodes of this community can be interpreted as the cluster contains the topics studied in an isolated manner, with no connections beyond the general link with disability. On the contrary, four smaller communities with greater internal cohesion can be observed. A green cluster comprised of concepts related to how television represents disability such as the Paralympic Games. The remaining three refer to the link between disability and social media, especially Twitter (pink), racial and gender aspects (blue), and the digital technology divide (red).

As for co-appearance frequency, the term ‘disability’ is usually preferably linked with ‘accessibility’, and in descending frequency, with ‘children’, ‘representation’, ‘Paralympics’, ‘television’, ‘inclusion’ and ‘social networks’.

4.1.2. Citation structure and dynamics

As clearly illustrated in figure 2, four major communities of references can be observed regarding co-citation:

The largest community has in turn three different clusters. This is based on the work of Goggin & Newell (2003) on disability and new media, with others orbiting it dedicated to aspects such as accessibility (Adam & Kreps, 2009; Black, 2014), the digital divide (Dobransky & Hargittai, 2006) or other means in which technologies hinder the integration of people with disability (Alper, 2017; Ellis & Kent, 2017). Another (pink) cluster links gender studies and queer theory with disability issues (McRuer, 2006; Alison, 2013). The blue conglomerate includes references on media representation of disability (Haller, 2010), indicating Goffman's Theory of Stigma (1963).

Another major community (green) combines references to the media construction of paralympic athletes, regarding both news coverage and media projection (e.g., Schantz & Gilbert, 2011; Buyesse & Borcherding, 2010), as well as their production (Howe, 2008) and reception (Berger, 2008; Hardin & Hardin, 2003). It also includes inherent and specific concepts and stereotypes of this field such as ‘supercrip’ (e.g., Silva & Howe, 2012). Meanwhile, the white cluster includes a community focused on applying principles of accessibility to teaching (Wilson 2000; Walters, 2010), especially from the technical communication perspective (Palmeri, 2006). Finally, the blue group is dedicated to the use of Internet and social networks by people with functional diversity (Caton & Chapman, 2016; Chadwick et al., 2013).

4.2. Study 2: content analysis results

4.2.1. Output context

Firstly, see figure 3 for a timeline with the evolution of both average annual number of articles published and their average impact in number of citations: ⁴

The most prolific year was 2017 with 33 manuscripts published while publications in 2014 achieved the greatest impact (M = 2.14 citations). General data for the sample set: 15.81 manuscripts per year (SD = 11.54) and with an average impact of 1.33 citations (SD = 1.83). There are no statistically significant differences in mean annual citations during the period analysed [F (10, 163) = 0.43, p = 0.93].

A total of 80 journals were identified as publishing these outputs. The most prolific are shown in table 3.

In view of the data shown in table 3, output on disability and diversity from communication is usually published in the US with another two countries following at a considerable distance: United Kingdom and Spain. This Anglo-Saxon dominance also translates into manuscript publication language: 85.6% English compared to 14.4% Spanish.

Regarding authorship, on average papers are signed by 2.05 authors (SD = 1.36), but a single author is most common (41.4%). Furthermore, 153 different corresponding authors were identified (60.9% female, 29.3% male and 9.8% unknown gender), these include Meryl Alper (n = 5), Gerard Goggin (n = 4), Nicole Matthews and Damian Mellifont (n = 3). The most prolific author affiliated with a Spanish university is Antonio Iáñez-Domínguez (n = 2). From a comparative perspective, differences in average impact of articles led by men (M = 1.55, SD = 2.09) compared to women (M = 1.1; SD = 1.69) are not statistically significant although they do indicate a trend [t (155) = 1.42, p = 0.079]. As would be expected, 97.1% are affiliated with universities; the most significant of the 140 institutions detected are the University of Sydney (Australia, n = 9), Northeastern University (USA, n = 5) and Macquarie University (Australia, n = 3). Spanish institutions with most participation are CEU San Pablo University, Complutense of Madrid, Pablo de Olavide, Basque Country, Salamanca, Santiago de Compostela and Seville (n = 2). Extending the approach to the entire sample, affiliation bodies are primarily located in: the USA (n = 65), Spain (n = 25), Australia (n = 24) and United Kingdom (n = 18).

Peer collaboration is absent in 39.7% of the cases (remember: Mo = 1 author), in 29.3% of the cases it is intra-institutional (same university), 27% inter-institutional (universities in the same country) and only 4% international. To conclude the output context, only 23% of studies receive additional funding. In this sense, there are no significant differences in terms of paper impact regardless of whether they have support (M = 1.6, SD = 1.94) or not (M = 1.24, SD = 1.8) [t (172) = 1.09, p = 0.139]; nor are there gender gaps in terms of receipt or lack of funding [χ2 (2, n = 174) = 4.44, p < 0.109].

4.2.2. Theoretical-methodological instrumentalisation

Type of work is distributed as follows: empirical (75.3%), theoretical-essay (23%) and methodological (1.7%). Observation of messages (38.5%) is the predominant object of study, followed by simultaneous analysis of various objects (21.3%), communication policies and structure (15.5%), source of information (14.4%) and, finally, audiences (10.3%). The following methodological techniques were identified (figure 4):

According to figure 4, and considering table 2, the most common methods are content analysis (13.8%) and discourse analysis (11.5%). On the other hand, the ‘technique’ variable can be recoded into a new variable with the following aggregate categories⁵: quantitative (44.4%), qualitative (42.1%) and methodological triangulation (13.5%). When comparing impact in number of citations, there are no statistically significant differences [F (2, 123) = 0.177, p = 0.838] between articles using quantitative (M = 1.35, SD = 1,91), qualitative (M = 1.26, SD = 1.84) or combined methods (M = 1.58, SD = 1.99). There are also no mismatches when comparing aggregate types of technique and funding [χ2 (2, n = 126) = 2.75, p = 0.253]. Continuing with matters related to methodology, most papers resort to probabilistic samples (42.5%) and, in the case of content analysis —except for one case— they have no intercoder reliability report.

Figure 5 shows all the coded information in this regard from a theoretical perspective:

Of the wide variety of concepts, the most common conceptual scaffolding concerns health communication and framing theory (9.2%) and media literacy (8.6%). Theory triangulation also stands out, present in 12.6% of cases.

To complete this results report, we must compare certain critical factors according to the ranking where articles are indexed: SSCI (65.5%) or ESCI (34.5%). As might be expected, SSCI manuscripts (M = 1.78 citations, SD = 2.07) yield a greater impact than ESCI (M = 0.46 citations, SD = 0.65), revealing statistically relevant differences [t (172) = 4.82, p < 0.001, d = 0.86]. Table 4 recovers other highly relevant analysis parameters that serve to compare articles based on indexing ranking:

In view of the data in table 4, there is no gender gap in terms of publication of articles in both rankings. Significant differences are observed in paper funding, more frequent in SSCI than ESCI, and in language use as English predominates in SSCI and Spanish in ESCI. Finally, quantitative research is more recurrent in ESCI while qualitative is in SSCI.

5. Discussion and conclusions

As pointed out in the introduction, the purpose of this paper was to map academic output limited to the field of Communication addressing matters of disability, diversity and their multiple derivatives. Specifically, the first objective was to detect keywords and bibliographic references most used in the field. A bibliometric analysis was conducted on a total of 316 articles indexed in WoS between 2012 and 2022. Apart from the concept of ‘disability’, the most common descriptors referring to the preferred topics are: accessibility, autism, intellectual disability, Paralympic Games, digital divide, childhood, diversity, inclusion and mental health. The most important media or supports identified are social networks, television, ICTs and Internet. They thus represent a series of terms that clearly dialogue with the topics identified in the study by Delgado-Vázquez et al. (2019): functional diversity, ICT, digital divides and competence; these in fact appear to be the most representative currents of study in the field.

Meanwhile, these descriptors are closely related to the bibliographic references in the field and that refer, among other aspects, to disability and new media (accessibility, digital divide and use of ICTs) or media representation of diversity. The frequency of keywords reflects scientific output on the topic analysed. Although this is not external to the political and social spheres (Smelser, 1989), the scientific agenda cannot automatically be inferred to coincide with the social agenda. Merely studying output cannot verify how political dynamics influence or, in this specific case, act as evidence of social perceptions of disability. Future studies could use techniques such as surveys to detect the reasons why scientists analysing disability choose the topics of their studies, and to what extent they are guided by what they perceive as social needs.

On the other hand, the second objective was based on analysing output context and the theoretical-methodological instrumentalisation of this research by analysing the content of a representative sample of 174 articles. Data collected point to a growing annual output in line with the findings of the study by López-Gómez et al. (2022), in US journals, indexed in SSCI and written in English. However, according to Delgado-Vázquez et al. (2019), the United Kingdom and Spain play a prominent role in the dissemination of papers in the journals of both countries, and especially in origin of the affiliated institutions of the most prolific researchers. Authorship is clearly female, as noted by Díaz-Jiménez et al. (2021), and the vast majority from universities. International collaboration networks are few and far between as the most common form of co-authorship is for manuscript authors to belong to the same university. Likewise, additional funding for this type of research is also uncommon, a fact that should be corrected in the future not just by academia, but especially by institutions responsible for granting this kind of aid. Agenda 2030 is expected to have a favourable impact on the roll-out of financial aid to bring visibility, also from research in communication, to disability and diversity.

Regarding theoretical-methodological instrumentalisation, the prototype paper in this field is empirical (as indicated by Díaz-Jiménez et al., 2021) and focuses on observation by means of analysing content (at quantitative level) or discourse (in qualitative terms), information and messages broadcast on social networks, Internet and television. Given the meagre reports available, it is imperative to stress the need to calculate intercoder reliability data when dealing with content analysis as this is a critical step for its correct application (Krippendorff, 2011; Lovejoy et al., 2016). Conceptual paradigms that would provide theoretical support for this research are related to health communication, framing and media literacy, consistent with the core bibliographic references in the field.

To complete a substantive review of the study findings, we must once again stress that there is no gender imbalance in terms of proportion of male and female authors disseminating their articles in the two WoS rankings, nor is there a gap regarding impact or additional financial support. Significant differences can however be observed between manuscripts indicated in SSCI and ESCI at funding and language level, although the most striking differences lie in methodology as quantitative techniques are more frequent in ESCI while qualitative are in SSCI.

The limitations of this study should be resolved with a view to future research. Specifically, it would be highly convenient that variables alluding to theories and methodologies reduce the number of cases in the ‘others’ category, although an attempt was made to control this aspect by detecting keywords. Likewise, content analysis could have been applied to the population universe of articles and extended to manuscripts indexed in other databases such as Scopus. SSCI and ESCI could also be differentiated in the bibliometric approach in order to map descriptors and core bibliographic references; this may have led to differences between one ranking and the other, as in previous studies (Segado-Boj et al., 2023). In any case, and given the scarce bibliometric evidence and, above all, content analysis in the field, this paper is expected to contribute to sorting and gaining knowledge in a field of study that should be made more visible, just like the phenomena it addresses.

Author’s contribution

Valeriano Piñeiro-Naval: conceptualisation, methodology, statistical analysis, visualisation, original draft-writing, writing-revision and edition. María Marcos-Ramos: conceptualisation, research, data collection, original draft-writing and supervision. Teresa Martín-García: conceptualisation, research, data collection and original draft-writing. Francisco Segado-Boj: methodology, visualisation, original draft-writing and writing-revision. All authors have read and accepted the published version of the manuscript. Conflict of interest: the authors declare there are no conflicts of interest.

References

Notes

1 Complete references included in the co-citation network (figure 2) are available for consultation as an appendix at: https://doi.org/10.17605/OSF.IO/V9XHM

2 Full codebook and main data matrix available at: https://doi.org/10.17605/OSF.IO/V9XHM

3 Variable reliability and matrix created ad hoc to calculate this are available at: https://doi.org/10.17605/OSF.IO/V9XHM

4 WoS only offers the aggregate total of manuscript citations (see original database at: https://doi.org/10.17605/OSF.IO/V9XHM). Average annual citations, which is a more reliable indicator, was calculated by dividing the aggregate total by the numbers of years of the manuscript.

5 The following values were used to recode the ‘technique’ variable: 1 = quantitative (manual and automated content analysis, survey, experiment, meta-analysis and bibliometric analysis); 2 = qualitative (discourse analysis, interviews, focus groups and case studies); 3 = triangulation (various techniques at once) and ‘missing’ when not applicable or in the case of other methods not pre-coded.

Author notes

* Assistant Professor, Department of Sociology and Communication

** Associate Professor, Department of Sociology and Communication

*** Assistant Professor, Department of Sociology and Communication

**** Assistant Professor, Department of Journalism and Global Communication

Additional information

To cite this article : Piñeiro-Naval, Valeriano; Marcos-Ramos, María; Martín-García, Teresa; & Segado-Boj, Francisco (2023). Análisis de la investigación sobre discapacidad y diversidad indexada en Web of Science en comunicación. ICONO 14. Revista Científica de Comunicación y Tecnologías Emergentes, 21(2). https://doi.org/10.7195/ri14.v21i2.2016

ICONO 14, Revista de comunicación y tecnologías emergentes

ISSN: 1697-8293

Vol. 21

Num. 2

Año. 2023

Analysis of research on disability and diversity indexed in Web of Science in communication

Valeriano Piñeiro-Naval ¹, María Marcos-Ramos ¹, Teresa Martín-García ¹, Francisco Segado-Boj ²