Skip to main content
NCBI home page
International Journal of Developmental Disabilities logoLink to International Journal of Developmental Disabilities
. 2021 Aug 9;69(2):291–303. doi: 10.1080/20473869.2021.1961204

A representation of students with intellectual disabilities in South Korean online newspaper articles using keyword network analysis

Yeji An 1, Dongil Kim 2,✉
PMCID: PMC10071973  PMID: 37025345

Abstract

The purpose of this study was to analyse online newspaper articles on students with intellectual disabilities (IDs) in order to identify related social phenomena to derive implications for inclusive education. Such study has traditionally practised through content analysis and/or discourse analysis manually, which is prone to subjective interpretation. Thus, this study implemented automated analysis to objectively select and interpret a big data. A total of 8,890 online newspaper articles that were published from 1990 to April 2019 were collected through automated parsing. The entire period and decade-phase based keyword and keyword network analysis were practised in order to determine how the social perceptions and related issues had changed over time. The results indicated that there was a rapid growth in scope of articles on students with IDs over the past 30 years. The attention of media gradually expanded from special education to improving quality of lives of students with IDs and their families. Moreover, online newspaper articles seemed to focus on social controversies and incidents such as sexual assaults that are related to students with IDs. Based on the results, ways to support inclusive education as well as social inclusion of students with IDs were discussed.

Keywords: intellectual disabilities, online newspaper articles, media, inclusive education, analysis, network analysis, big data, computational social science

Introduction

Considerable efforts have been made to understand and define the general construct of intellectual disability (ID) over the years (Schalock et al. 2007). The construct of disability has been discussed in the context of social, cultural, educational and political views, which has led to the change in terminologies consistent with the discussions. In particular, the terminology of ‘mental retardation (MR)’ has been used for the longest period of time since 1961, replacing the earlier devalued terms such as idiocy and mental subnormality (Park 2010, Finlay and Lyons 2005, Harris 2013). MR includes the meaning of disability which exists inside the individual as an invariant trait. However, as the paradigm of support occurred, the importance of an early identification for supporting individuals potential have gradually expanded since 1992 and the new terminology, ID, began to be used more often in the name of related organizations and laws (Kang and Kim 2012, Min 2013). The new terminology emphasizes the importance of the interaction between an individual’s environment and social-ecological construct rather than focusing on person-centred neurobiological deficits as a trait (Harris and Greenspan 2016). Understanding such change is notable because it reflects the social views of disability and the related issues around students with IDs (Kang and Kim 2012, Park 2010, Lee 2007).

ID is a disability defining as having significant limitations in both intellectual functioning and adaptive behaviours (AAIDD 2010). Deficits of adaptive behaviour can affect the various skills that are required for independent living as members of society. These difficulties tend to be discovered in various forms, depending on the individual, in the early stages of life development and persist throughout their lives (Chadwick et al. 2005, Matson et al. 2009). As previously mentioned, emphasizing interactions between individuals and their environment has merged inclusive education as an effective educational environment for students with IDs for learning various skills necessary for improving their quality of lives (Stoiber et al. 2016). Inclusive education refers to including students with IDs in general education classrooms with individualized support for their special needs. The concept of inclusive education was supported and adopted by many countries across the world, including South Korea (United Nations 2006, Vislie 2003). According to the 2020 Special Education Annual Report of South Korea, 72.1% of all students with disabilities are receiving inclusive education. Considering that students with IDs account for 53.1% of all students with disabilities and 71.9% of them receive inclusive education, inclusive education is an important issue for students with IDs. The school age is a period where students with IDs can naturally learn adaptive behaviour skills that are essential for independent living in the small society of school; however, it is also a critical period where students with IDs can experience rejection and/or stigmatization for their maladjustment. Thus, experiencing inclusive education at this time is important for both students with and without IDs for future inclusion. In order to provide positive experiences of inclusive education for all students, it is important to acknowledge and create positive attitudes and environments for students with IDs. The 2019 national report of school violence in South Korea reported that school violence against students with disabilities has increased by about five times in the last four years (Lee et al. 2020). Previous studies targeting teachers perceptions on inclusive education also reported that teachers have a low level of awareness of inclusive education, and thus educational support is provided in a passive form (Kim and Lee 2020, Park et al. 2019). These conditions of inclusive education in Korea are often understood as a lack of awareness of the disability and negative perception of significant others such as teachers, parents, and students towards students with disabilities. The necessity of education for understanding disabilities is constantly being emphasized, but more specific measures for improving awareness should be sought (An and Kim 2020).

Media is one of the important ways of forming and maintaining people’s social attitudes toward a particular group in society. Media has such an influence because it serves as an interpretative schema that applies to classifying and interpreting information to the public according to how the media emphasizes and reports on a specific topic (Goffman 1974, Ryu 2018). This is called news framing which can be described as three models: framing effect, agenda-setting effect, and priming effect (Scheufele and Tewksbury 2007). First, the framing effect assumes that the way the audience understands can be affected by how the media has characterized the issue in the report. Second, agenda setting assumes that repetitive reports on specific issues will lead the public to perceive these issues as important. Last, priming effect assumes that the information exposed to the media can be used as a basis for evaluation in the decision-making process of related issues. These effects can be formed by various factors such as politics, economy, society, culture that can lead to various changes. Public’s social perception of people with IDs can also be influenced by the media’s representation of them, which can lead to changes of related policies and services (Jo and Berkowitz 1994, Wilkinson and McGill 2009). Therefore, a review of the media on people with IDs can provide useful information to understand the public’s social perspectives as well as how the media has potential to promote inclusiveness.

Despite the fact that it can function as a main means for grasping the perspectives of both individuals with and without disabilities, studies analysing media on them are relatively incomplete. Research on the topic reported that the use of medical terms has decreased compared to the past, and the use of people-first languages, the number and size of articles, and positive narratives of ID have increased over the years (Wilkinson and McGill 2009, Yoshida et al. 1990). Yet, individuals with IDs are often described as objects of welfare and/or support, and the media only covers simple events and accidents rather than reporting on a variety of topics (Kim and Park 2012, An and Kim 2020). However, the issue of previous research is that the news coverage the findings were based on were limited in number. People are more likely to support inclusion if they consider the individuals with disabilities as more capable (Siperstein et al. 2003). This study analyses the media that affect public awareness while simultaneously showing public perception of people with IDs, which can be used to create a society where everyone lives together.

Today's media, which includes much data, has the characteristics of big data which Laney (2001) conceptualizes including a large volume, a variety of data structure, and the speed of rapidly generated velocity. Media coverages are traditionally analysed manually through content analysis or discourse analysis, but these methods have limitations of analysing large amounts of data with objectivity. With the recent development of technology, techniques and programs that can analyse and interpret large amounts of data such as data mining, text mining, opinion mining, and predictive analytic have appeared. In the meantime, it became possible to expand the research by analysing the vast amount of media coverage. Text mining, also called text analysis, is increasingly being used as a method of gathering unstructured text data such as social network feeds, comments, blogs, and news that were difficult to target for analysis, and deriving meaningful information or knowledge (Gandomi and Haider 2015). Especially, keyword network analysis is a method that can derive and visualize a new semantic graph by extracting standardized data from unstructured data. The graph allows not only descriptive but also inferential analysis by analysing the structure and pattern of the network as well as the roles and associations between keywords in the network (Borgatti et al. 2009, Kim et al. 2020). Thus, this study intends to uncover societal phenomena and media representation of students with IDs by analysing online newspaper articles published in the last 30 years in South Korea using keyword network analysis. The study assumed online newspaper articles as proper type of media due to its influence as more 85% of Korean today get information through internet rather than other types of media (An and Kim, 2020). The study enables a perceived reality to emerge, leading to practical discussion of social phenomena which leads to a more salient understanding. The questions that inform this study are:

  1. What are the keywords of online newspaper articles of students with intellectual disabilities?

  2. What are the relationships between keywords of online newspaper articles of students with intellectual disabilities?

Methods

Data collection

In South Korea, the use of portal-type internet news had increased significantly around the year 2000 as people began to use the Internet to quickly access the news related to significant incidents around the world such as the 9/11 terror attacks in the United States and the 2002 Korea-Japan World Cup (Ban and Kwon 2007). In particular, large portal sites in South Korea have rapidly changed the patterns of news users, and their influence has expanded so that various performances of individual news media can be determined depending on whether news is distributed through the portal site (Kim and Kim 2012, Chae 2014). According to the News Concentration Survey, which analysed news outlets from 2016 to 2018, the occupancy rate of portal-type internet news increased from 85.6% to 89.3% while the share of usage of daily online, news agencies, news channels, and others decreased (Public Opinion Concentration Survey Committee 2018). Although various portal news services have repeatedly appeared and disappeared, Naver News service is consistently evaluated as the portal news service that has the greatest influence in Korea (Song and Yang 2017). According to 2019 Korean Click Survey, Naver is the number one site in Korea with more than 2,700 unique visitors and an 89% reach rate. Unique visitor is a visitor measurement index value that removes duplicate visits from visitors who access the portal during a specific period (Korean Click 2019). In this regard, this study’s data set included the affiliated articles provided through Naver News Service (https://news.naver.com) for analysis.

The dataset used for this study is obtained by using Python web-scraping technique, BeautifulSoup4 package (Richardson 2017). In order to collect online newspaper articles related to students with IDs, prior literatures were reviewed and identified both current and previous terms referring to students with IDs. The initial search using selected terms yielded 9,066 articles. By implementing web-scrapping techniques, the study was able to collect selective information that are needed for analysis. Selected information included the title of the article, the date of the report, the name of the press, and the news contents. The selected data were screened to confirm the suitability for the study. Data that did not meet the purpose of this study were excluded by the following criteria: (a) Articles that did not include search terms in both the title and the body were excluded from the analysis; (b) Identical articles with the same report time and contents were excluded from the analysis; (c) Articles consisting of ten or fewer words were excluded due to lack of information for analysis. A total of 8,890 articles from 1999 to April 2019 were finally identified to meet the eligibility of the study.

Data analysis

Keyword analysis

Keyword Analysis is a type of text mining that analyses the knowledge and new information inherent in unstructured text by using the number of word occurrences (Feldman and Dagan 1995). Keyword analysis may be analysed based on TF (Term Frequency) and TF-IDF (Term Frequency-Inverse Document Frequency). TF represents the number of occurrences of a word in a document, and TF-IDF is a word extraction method that reduces the importance of hackneyed expressions that appear in almost all documents that do not provide meaningful information for interpretation. The collected data were pre-processed in the following steps: 1) normalization was performed to unify words that need to be analysed in the same sense, by altering spaces, synonyms, and abbreviations; 2) words and phrases unnecessary for analysis were treated as stopwords; and 3) morphological analysis was performed to convert text into the smallest unit of words to extract the nouns needed for analysis. The Python program was used throughout the data pre-processing steps. The R program was then used to generate a Document-Term Matrix (DTM) suitable for analysis to calculate the TF and TF-IDF.

Keyword network analysis

Keyword Network Analysis is another analysis method of text mining that extracts relationships based on the frequency of co-occurrence between keywords contained in unstructured data (He 1999). Each keyword is formed of one node, and the relationship between them is represented by an edge. The structure of the generated network can be interpreted through a degree indicating the number of specific keywords connected to other keywords and centrality indicating a relative position in the entire relationship. Nodes located at the centre of the network can facilitate or interfere with the relationships between nodes, so the centrality of nodes provides important information to understand the network structure (Johnson et al. 2013).

In this study, considering the importance of the entire data set, a node was set for 30 high frequency words of TF-IDF that can represent the unique characteristics of each document (Shin et al. 2017). In addition, the co-occurrence matrix was calculated through the Python program to represent the relationship between keywords as a complete graph network. The co-occurrence matrices assume that keywords have semantic relevance when they appear together in the same document (Leydesdorff and Welbers 2011). The network was analysed using degree centrality, eigenvector centrality, and beta centrality (Bonacich 1972, 1987), which weights the centrality around the nodes. Degree centrality indicates the number of connections for each node, and eigenvector centrality is an indicator that can explain the influence of a node on the network because it is affected by the degree of centrality of the connected node. Beta centrality is an index that considers the structure of the entire network and provides information on nodes located on the network, considering the total amount of potential impacts on the nodes through direct and indirect paths. Formulas to obtain the mentioned centrality index are shown in the table below (Table 1).

Table 1.

Centrality index formulas.

Centrality index Formula
Degree centrality di=∑jxij
di: degree centrality of node i, xij: Items (i, j) in adjacency matrix
Eigenvector centrality ei=λ∑jxijej
e: Eigenvector centrality index value; λ: Eigenvalue (proportional constant)
Beta centrality c=(I−βA)−1A1
c: Sum of rows of matrix
Open in a new tab

Finally, the network is mostly composed of clusters of nodes with strong connections, so that the structure of the entire network can be understood by analysing these clusters (Cho and Ahn 2016). Therefore, this study attempted to understand the network structure pattern through CONCOR (CONvergence of iterated CORrelations) analysis. CONCOR analysis is a hierarchical clustering of keywords based on the Pearson’s correlation coefficient of the co-occurrence matrix data, and is the most commonly used method in structural equivalence analysis (Kim et al. 2020). UCINET6 (Borgatti et al. 2002) was used for keyword network analysis, and the network was visualized using the NetDraw program (Borgatti 2002). Data analysis was practised in Korean, and the results were translated and presented in English.

Results

Keyword analysis over the entire period

The results of keyword analysis based on TF of 8,890 online newspaper articles for students with IDs from January 1990 to April 2019 are shown in Table 2. The most frequent words were 'school', followed by ‘society’, ‘parents’, ‘education’, and ‘Seoul’. The majority of information in online newspapers for students with IDs is related to education and activities centred on schools. Words related to social support facilities, programs, and crime related words such as ‘police’, ‘investigation’, ‘incident’, and ‘charges’ were also identified.

Table 2.

Keywords based on TF over the entire period (1990 ∼ 2019).

No. Keywords TF No. Keywords TF No. Keywords TF
1 School 11,480 11 Living 4,174 21 Provision 3,588
2 Society 7,420 12 Special school 4,112 22 Incident 3,570
3 Parents 7,102 13 Police 4,016 23 Center 3,558
4 Education 6,694 14 Event 3,998 24 Program 3,480
5 Seoul 6,170 15 Investigation 3,854 25 Problem 3,450
6 Teacher 5,798 16 Offer 3,812 26 Family 3,386
7 Region 5,518 17 Facility 3,768 27 Plan 3,362
8 Support 5,404 18 Nationwide 3,758 28 Outcome 3,216
9 Activity 4,698 19 Participation 3,742 29 Culture 3,200
10 Proceeding 4,254 20 Operation 3,670 30 Charges 3,192
Open in a new tab

Note. TF = term frequency.

The results of the analysis based on TF-IDF are as shown in Table 3, and words representing government support such as ‘offer’, ‘government’, ‘improving’, ‘organization’, and ‘growth’ showed a high frequency. The result also showed words related to education, experiences, activities of students, as well as the word ‘sexual assault’, which appear in a context similar to ‘investigation’, ‘incident, and ‘charges’ in the result of TF analysis.

Table 3.

Keywords based on TF-IDF over the entire period (1990 ∼ 2019).

No. Keywords TF-IDF No. Keywords TF-IDF No. Keywords TF-IDF
1 Offer 1.111 11 Counseling 0.507 21 Employee 0.426
2 Government 0.662 12 Experience 0.504 22 Service activity 0.423
3 Improving 0.652 13 Outcome 0.493 23 Special education 0.406
4 Organization 0.580 14 Rehabilitation 0.489 24 Family 0.399
5 Growth 0.553 15 Institution 0.462 25 Declaration 0.393
6 Specialty 0.542 16 Busan 0.462 26 Business 0.387
7 Teacher 0.539 17 Hospital 0.456 27 Center 0.387
8 Special school 0.528 18 Inhabitant 0.445 28 Living 0.380
9 Participation 0.516 19 Occupation 0.441 29 Contest 0.369
10 Condition 0.513 20 Sexual assault 0.431 30 Therapy 0.368
Open in a new tab

Note. TF-IDF = term frequency.

Keyword analysis by phases

To examine the changes of the collected data over 30 years, the period was divided and analysed by each decade, as shown in Table 4. The newspaper articles published from 1990 to 1999 were divided into phase 1, the articles published from 2000 to 2009 were divided into phase 2, and those published from 2010 to April 2019 were divided into phase 3. As a result of analysing phase 1 including 62 articles based on TF-IDF, the most frequently used words were ‘occupation’ followed by ‘Busan’, ‘resource room’, ‘program’, and ‘problem’. It is inferred that the media interests related to the careers of students with intellectually disabilities were relatively high through words such as ‘occupation’, ‘rehabilitation’, ‘facilities’, ‘ability’, and ‘acceptance’ analysed during the phase. Words such as ‘performance’, ‘dance’, ‘contest’, ‘choral’, and ‘impression’ were analysed as well, interpreting that the media perspective of students with IDs in the 1990s was somewhat sympathetic. It is interesting that specific Korean regions such as ‘Busan’ and ‘Gyeongnam’ were keywords. In phase 2, which included 1,708 articles, the keywords ‘rehabilitation’, ‘experience’, ‘service activity’, ‘study’, and others emerged with similar frequency. The similar occurrence frequency of keywords implies that articles on various topics have been reported rather than specific topics. On the other hand, in phase 3, which includes 7,120 articles, the frequency of ‘offer’ was the highest and the words such as ‘establishment’, ‘government’, ‘improvement’, and ‘group’ were analysed as keywords. It means that there were many articles related to improving government supports for students with IDs. It was also found that many articles related to crime were reported, which can be seen through analysed keywords such as ‘accident’, ‘police station’, ‘assault’, and ‘sexual assault’. Lastly, the keywords analysed in all phases are ‘occupation’, ‘Busan’, and ‘special education’.

Table 4.

Keywords based on TF-IDF for each phase.

Phase
no. 		Phase 1 (1990 ∼ 1999)
 Phase 2 (2000 ∼ 2009)
 Phase 3 (2010 ∼ 2019)
Keyword TF-IDF Keyword TF-IDF Keyword TF-IDF
1 Occupation 0.302 Rehabilitation 0.401 Offer 1.064
2 Busan 0.274 Experience 0.394 Establishment 0.817
3 Resource room 0.269 Service activity 0.382 Government 0.658
4 Program 0.249 Study 0.366 Improving 0.639
5 Problem 0.245 Hospital 0.352 Organization 0.585
6 Learning 0.199 Daejeon 0.345 Participation 0.550
7 Office of Education 0.191 Contest 0.343 Growth 0.546
8 Impression 0.186 Special education 0.331 Experience 0.539
9 Installation 0.180 Therapy 0.331 Teacher 0.539
10 Performance 0.167 Employee 0.331 Specialty 0.536
11 Principal 0.166 Office of Education 0.329 Special school 0.524
12 Rehabilitation 0.164 Busan 0.324 Condition 0.498
13 Class 0.159 Female 0.316 Accident 0.498
14 Instruction 0.149 Living 0.315 Outcome 0.473
15 Emotional disorder 0.148 Volunteer work 0.312 Busan 0.472
16 Facility 0.148 Deliver 0.312 Police station 0.462
17 Commemoration 0.147 Counseling 0.300 Hospital 0.454
18 Special Education 0.142 Training 0.293 Institution 0.449
19 Creation 0.142 Happiness 0.293 Occupation 0.434
20 Dance 0.141 Game 0.293 Inhabitant 0.433
21 Physical disability 0.140 Exercise 0.292 Employee 0.424
22 Ability 0.138 Seoul 0.282 Special education 0.416
23 Acceptance 0.138 Family 0.281 Family 0.405
24 Contest 0.135 Principal 0.277 Assault 0.395
25 Choral 0.132 Proceeding 0.275 Center 0.382
26 Teacher 0.132 Field 0.275 Sexual assault 0.381
27 Society 0.130 Implementation 0.268 Business 0.381
28 Culture 0.125 Preparations 0.266 Therapy 0.370
29 Operation 0.122 Occupation 0.264 Declaration 0.366
30 Gyeongnam 0.118 Composition 0.263 Culture 0.354
Open in a new tab

Keyword network analysis by phases

Phase 1 (1990 ∼ 1999)

As a result of keyword network analysis of phase 1, the network was composed of 30 nodes and 684 edges. A simplified network with edges of more than 3 connections is shown in Figure 1. The network was organized around ‘teacher’ and showed a high co-occurrence rate with keywords such as ‘special education’ and ‘Office of Education’.

Figure 1.

Figure 1.

Open in a new tab

A simplified keyword network of phase 1 based on TF-IDF.

Table 5 shows the results of analysing the effect of each keyword on the entire network through centrality analysis. In all the results of the centrality analysis, ‘teacher’ showed the largest centrality index, followed by ‘class’, ‘Office of Education’, ‘operation’, and ‘special education’. The network of phase 1 is mainly described by ‘teacher’ and ‘Office of Education’ with administrative responsibilities in relation to education for students with IDs.

Table 5.

Centralities of the keyword network of phase 1 (1990 ∼ 1999).

No. Degree centrality Eigenvector centrality Beta centrality
1 Teacher 0.507 Teacher 45.812 Teacher 1.775
2 Class 0.407 Cclass 42.850 Class 1.659
3 Office of Education 0.393 Office of Education 40.382 Office of Education 1.564
4 Operation 0.390 Operation 39.826 Operation 1.542
5 Special education 0.383 Special education 39.546 Special education 1.531
6 Society 0.372 Facility 35.482 Facility 1.374
7 Facility 0.338 Society 33.385 Society 1.294
8 Rehabilitation 0.307 Ability 31.034 Ability 1.202
9 Ability 0.307 Learning 30.581 Learning 1.184
10 Emotional disorder 0.286 Emotional disorder 30.057 Emotional disorder 1.164
11 Llearning 0.283 Resource room 27.478 Resource room 1.064
12 Contest 0.262 Rehabilitation 26.420 Rehabilitation 1.024
13 Resource room 0.252 Installation 26.407 Installation 1.022
14 Installation 0.245 Creation 21.893 Creation 0.847
15 Creation 0.200 Contest 20.467 Program 0.800
Open in a new tab

CONCOR analysis was conducted to understand the structural equivalence of the phase 1 network, as shown in Figure 2, and resulted in three groups. The first group is composed of keywords with high centralities, such as ‘special education’, ‘operation’, ‘class’, and ‘Office of Education’. It indicates that contents related to future direction of special education for students with IDs were mainly dealt with in this phase. The second group covers the topics related to educational programs provided for students with IDs which reveals keywords including ‘resource room’, ‘program’, ‘learning’, ‘problem’, ‘teacher’, and others. This group demonstrates that discussions have been made on supporting students with IDs in both general and inclusive education environments in accordance with special education policies aimed at placing all students in the field of education. Finally, the third group consisted of keywords related to culture and art events centred around students with IDs. Various events involving students with IDs were reported through the media in phase 1, and it is assumed with the keyword of ‘impression’ that the media had a form of reporting that promotes feelings of compassion toward students with IDs. In summary, the articles on students with IDs that were reported in phase 1 were on cultural and artistic events as well as administrative and practical discussions on special education.

Figure 2.

Figure 2.

Open in a new tab

A result of CONCOR analysis on keyword network of phase 1.

Phase 2 (2000 ∼ 2009)

As a result of analysing the 30 keywords from newspapers on students with IDs, a network was formed with 870 edges. As can be seen in the simplified network shown in Figure 3 edges of less than 60 connections were deleted, ‘living’, ‘Seoul’, and ‘family’ are located at the centre of the network, and are analysed to understand if they play an important role in explaining the network while other centrality of other keywords is relatively small.

Figure 3.

Figure 3.

Open in a new tab

A simplified keyword network of phase 2 based on TF-IDF.

A centrality analysis was conducted to grasp the degree of influence of each keyword on the network in more detail. As shown in Table 6, the ranking of the centrality index of the top 15 keywords were analysed very similarly in all the analyses. The keyword with the highest centrality index in the network of phase 2 was ‘living’, followed by ‘Seoul’, ‘family’, ‘experience’, ‘proceeding’, and ‘therapy’. Unlike the results of phase 1, where education-related keywords were dominant, in phase 2, it was found that the main topics were related to therapy and rehabilitation, as well as aspects of quality of life enjoyed by students with IDs. The result of ‘Seoul’ as the top keyword also supports the idea that the numerous therapies and rehabilitation services were carried out in the capital city of Korea, Seoul, where the population density is very high. Through these results, changes in the media’s interest in certain topics over time were observed.

Table 6.

Centralities of the keyword network of phase 2 (2000 ∼ 2009).

No. Degree centrality Eigenvector centrality Beta centrality
1 Living 0.543 Living 47.234 Living 1.830
2 Seoul 0.504 Seoul 44.413 Seoul 1.720
3 Family 0.415 Family 37.758 Family 1.462
4 Experience 0.360 Experience 32.891 Experience 1.274
5 Proceeding 0.352 Proceeding 32.655 Proceeding 1.265
6 Therapy 0.350 Therapy 31.817 Therapy 1.232
7 Implementation 0.326 Implementation 29.871 Implementation 1.157
8 Rehabilitation 0.315 Rehabilitation 29.255 Rehabilitation 1.133
9 Exercise 0.271 Exercise 25.605 Exercise 0.992
10 Field 0.270 Field 25.175 Field 0.975
11 Contest 0.259 Contest 23.892 Contest 0.925
12 Preparations 0.255 Preparations 23.726 Preparations 0.919
13 Service activity 0.242 Employee 23.110 Employee 0.895
14 Employee 0.242 Service activity 22.517 Service activity 0.872
15 Counseling 0.236 Counseling 22.210 Counseling 0.860
Open in a new tab

As a result of CONCOR analysis, a total of 4 groups were formed as shown in Figure 4. The first group is created around ‘living’ and ‘family’, and other keywords with high degree centrality, including ‘rehabilitation’, ‘field’, ‘female, ‘study’, and ‘happiness’. Of these, ‘happiness’ has a relatively small degree centrality, but it seems to play an important role in explaining other keywords by being located in the middle of the network. Therefore, one of the main contents of the phase 2 network is related to improving the overall quality of life of students with IDs. The second group is composed of keywords such as ‘proceeding’, ‘implementation’ and ‘contest’ centred on ‘Seoul’, which means that various events related to students with IDs were held in Seoul and promoted through the media. This finding can be supported by the third group which consists of ‘service activity’, ‘volunteer work’, and ‘exercise’ that are centred around ‘experience’, ‘service activity’, and ‘volunteer work’, which show that various social support was requested along with the mentioned events. The final fourth group consisted of keywords related to therapeutic support services such as ‘therapy’, ‘training’, and ‘counselling’. The keywords of ‘special education’ and ‘Office of Education’, which played an important role in the network of phase 1, were not included in any of the groups in phase 2, but ‘therapy’ seemed to have a more significant role. These results represent the changes in the Special Education Act for Persons with Disabilities, in which the therapeutic education system was abolished and the therapeutic support system was newly established, and the Third Special Education Development Five-Year Plan of 2008, which includes related services.

Figure 4.

Figure 4.

Open in a new tab

A result of CONCOR analysis on keyword network of phase 2.

Phase 3 (2010 ∼ 2019)

Thirty keywords extracted from online newspapers in the third phase reporting on students with IDs were connected to 786 edges to form a network. To understand the structure of the network more clearly, a simplified network is presented as shown in Figure 5 by deleting edges with less than 180 connections. It can be seen that the network is centred on ‘teacher’, and compared to the relationship between other keywords, ‘sexual assault’ and ‘special school’ showed a higher co-occurrence rate with higher centrality. The keywords that are relatively central and analysed to be important in describing the network are ‘center’, ‘offer’, ‘family’, ‘culture’, ‘outcome’, and others.

Figure 5.

Figure 5.

Open in a new tab

A simplified keyword network of phase 3 based on TF-IDF.

As a result of the centrality analysis presented in Table 7, ‘teacher’ and ‘provision’ showed the first and second rank in all centrality analysis. In addition, the keywords’ ranking of centrality differed according to the analysis method. For instance, ‘outcome’ and ‘sexual assault’, which were seventh to thirteenth, rose to fifth and eleventh from the result of eigenvector centrality and beta centrality.

Table 7.

Centralities of the keyword network of phase 3 (2010 ∼ 2019).

No. Degree centrality Eigenvector centrality Beta centrality
1 Teacher 0.414 Teacher 45.022 Teacher 1.744
2 Offer 0.340 Offer 37.744 Offer 1.462
3 Center 0.334 Special school 35.994 Center 1.450
4 Special school 0.315 Center 37.447 Special school 1.394
5 Family 0.287 Outcome 32.445 Outcome 1.257
6 Culture 0.271 Family 31.843 Family 1.233
7 Outcome 0.292 Culture 30.651 Culture 1.187
8 Institution 0.269 Institution 30.444 Institution 1.179
9 Business 0.260 Business 29.604 Business 1.147
10 Occupation 0.218 Occupation 25.520 Occupation 0.988
11 Organization 0.217 Sexual assault 25.151 Sexual assault 0.974
12 Therapy 0.217 Organization 24.635 Organization 0.954
13 Sexual assault 0.216 Therapy 24.262 Therapy 0.940
14 Inhabitant 0.209 Inhabitant 23.820 Inhabitant 0.922
15 Condition 0.195 specialty 22.151 Specialty 0.858
Open in a new tab

The structural equivalence of the network was analysed with CONCOR analysis as shown in Figure 6. The results confirmed that the phase 3 network is composed of three groups. The first group included the most keywords such as ‘culture’, ‘business’, ‘experience’, centred around ‘offer’, ‘center’, and ‘family’. These keywords are related to business and programs for students with IDs and their families. The second group consists of ‘teacher’ and ‘outcome’, suggesting that incidents and accidents related to ‘sexual assault’ and ‘assault’ on students with IDs appeared to be one of the main topics in the network of phase 3. The keywords of this group are deeply related to the incident that caused great social controversy in 2005, which was a habitual sexual assault targeting students with both hearing and intellectual disabilities at Inhwa School, a special school in Gwangju, South Korea. The last group consists of ‘special school’, ‘institution’, ‘establishment’, ‘special education’, and ‘inhabitant’, which can confirm social controversies related to the establishment of special schools. This group represents a conflict in the community regarding the establishment of special schools. For instance, in September 2017, at the residents' debate for the establishment of a special school in the Gangseo area of Seoul, parents of students with disabilities kneeled down and appealed to residents who opposed the establishment of a special school. It was extensively reported through the media and received wide public attention, as evidenced in this study.

Figure 6.

Figure 6.

Open in a new tab

A result of CONCOR analysis on keyword network of phase 3.

Discussion

Considering that media on students with IDs can have an impact on the public’s perceptions and attitudes to interfere or facilitate inclusion, the research aiming to understand societal phenomena and media representation in the field of education can provide essential information to promote inclusion of students with IDs. This study implemented keyword network analysis to analyse online newspaper articles to uncover real-world issues involving students with IDs that were particularly prominent over the past 30 years. Furthermore, 30 years of data were extracted and analysed by decade to show salient change over time.

The results revealed that the number of articles reporting on students with IDs in three phases increased to 62 in phase 1, 1,708 in phase 2, and 7,120 in phase 3 which reflects the rapid growth in the scope of the articles on students with IDs. There is a possibility that such a change is not a change in news related to students with IDs, but rather an increase in the number of online newspaper articles reported in general. However, the number of press articles has also increased across phases confirmed that media attention increased over time in reference (1 press in phase 1; 64 press in phase 2; 93 press in phase 3). The rate of increase in the number of press is not high compared to the rate of increase in the number of online newspaper articles related to students with IDs, which can be interpreted as increasing media interest. As inclusive education was emphasized as a major task in the 3rd Special Education Development Plan in 2008, inclusive education was greatly expanded at this point. Various discussions took place while including students with IDs, and this seems to have been also reported through media. In fact, these social discussions were related to special education operations and events in phase 1 which extended to positive change of discussion about improving the quality of life of students with IDs and their families in phase 2. In addition to the contents of the programs for students with IDs in phase 3, they were found to consist of various social controversies and conflicts that hinder their inclusion. Especially, the keyword analysis over the entire period based on both TF and TF-IDF showed the importance of words related to public support and sexual assaults and/or incidents in special education.

The results have further led to discussions as follows. First, there is an increasing need to provide differentiated services for students with IDs in various fields including education and living. The results of this study represented discussions on services and supports for students with IDs and their families, as well as the establishment of special schools for students with disabilities. Effective forms of education can vary depending on the level of disability and individual characteristics. Inclusive education has also been recognized for its effectiveness for a long time, but there is still a high demand for the establishment of special schools in terms of guaranteeing the right to education for students with severe and multiple disabilities and the right to choose the education (Lim and Kim 2019). Despite the demand for the establishment of special schools, special schools are still recognized as ‘hate facilities’, and are facing considerable difficulties from the basic stage of establishment in Korea (Lim and Kim 2019, Park et al. 2019), and this seems to have been revealed through the relationship between the keywords of this study. Moreover, people with IDs, classified only as severe disabilities in Korea, have limitations in independent living, and the nation’s accountability for support along with their life cycle is constantly being discussed. As a result, 44 provisions were included under the Developmental Disabilities Act announced in 2014 to provide welfare and family supports for each life cycle reflecting security and welfare needs (Kwon 2016). In spite of these changes, the fact that there is an increasing need for education and living stands out represents the issues of Korean society that requires continuous interests and change of actions. To this end, communication and cooperation between important stakeholders surrounding these issues need to be facilitated, and related research should also be conducted.

Second, the vulnerability of students with IDs suffering sexual assault has been observed through the findings. Previous studies also report that students with IDs are obedient, are dependent on instinctive needs, and have difficulty distinguishing between causes and effects due to their limited intellectual abilities and adaptive behavioural skills. Moreover, they argue that the limited opportunities for independent external contact that cause the lack of knowledge of the right to choose or not choose a sexual partner increases the likelihood of presenting sexually inappropriate behaviour or being subject to sexual crimes (Carmody 1991, Reiter et al. 2007, Ryu et al. 2016). The results of this study also support previous studies through empirical analysis. In the case of Inhwa School's sexual assault case analysed in this study, the permission to establish a school corporation was revoked in 2011, seven years after the incident was known. This was evaluated as a hesitant decision to avoid social controversy and was called the 'late administration' with significant criticism. Afterwards, the aforementioned situation has raised public awareness of crime of sexual violence against people with disabilities. This led to a national response involving the release of comprehensive security measures to eradicate sexual violent crimes against people with disabilities, and the deletion of the provisions for parental offences and inability to protest, which received great attention from the media, as shown in this study. Despite the increasing awareness of students' vulnerability to sexual assault and recurring incidents, it is difficult to find meaningful educational changes to prevent such crimes. The sexual education program conducted in the educational field for students with IDs is criticised as it does not provide adequate information compared to the program for students without disabilities, and thus is only conducted in a bureaucratic manner in South Korea (Park 2015, Ryu et al. 2016). Considering the findings of previous studies that reported on the effectiveness of sex education and/or self-advocacy skills training to increase the sex crime prevention rate (Schaafsma et al. 2015, Whitehouse and McCabe 1997), active and systematic preventive efforts of professionals are required to lower the rate of sexual assault targeting students with IDs.

Third, the media's attention was very focused on stimulating issues and controversies, and it was often seen as a tendency to present students with disabilities as objects of social protection and welfare rather than in the perspectives of their rights. Due to the aforementioned news framing, issues that are repeatedly dealt with in the media can bring significant changes by making the public perceive the issue as important. The external stimulus related to students with disabilities can also make people recognize and evaluate them as issues that have been strongly dealt with in the media due to ignition effect. In other words, the way media emphasizes and reports on specific issues can make meaningful changes for students with IDs, but it affects not only the way the public sees and understands social phenomena, but also the way people perceive students with IDs in general. Therefore, the media should provide objective and sufficient information in a careful manner to enhance the proper understanding of the legitimate rights of students with IDs (An and Kim 2020, Keller et al. 1990). This study provided information on salient issues related to students with IDs, as well as how they were described across 30 years of media. This result is expected to be used as an important basic data for improving public perception and attitude towards students with IDs and promoting the formation of an inclusive culture in South Korea.

Limitations and future directions

There are several limitations in this study with regard to future research. First, this study highlights social phenomena related to students with IDs from a macro perspective due to the nature of big data. For this reason, it is limited in grasping the details surrounding students with IDs. In order to understand social phenomena related to students with IDs, it is necessary to take into account the various aspects such as public policy development in future research. Second, although this study collected data from Naver News Service which is the most influential portal in Korea, there are other portal news services that receive different media affiliated services. Media also have various types, such as video and newspapers, so future studies can expand the scope of analysis targets and explore additional information that were not explored through this study. Last, this study represented social discourse that influences the formation of the perceptions and attitudes of students without disabilities toward students with IDs, yet it cannot provide positive or negative insights on this matter. Thus, sentiment analysis of media coverage can be conducted in the future to grasp not only social phenomena but the public’s sentiment toward students with disabilities.

Disclosure statement

No potential conflict of interest was reported by the authors.

References

  1. AAIDD . 2010. AAIDD. [online] Available at: <https://www.aaidd.org/intellectual-disability/definition> [Accessed 11 August 2019].
  2. An, Y. and Kim, D.. 2020. Keyword and topic analysis of online news coverage on students with developmental disabilities. Korean Journal of Special Education, 54, 27–50. [In Korean] [Google Scholar]
  3. Ban, H. and Kwon, Y. S.. 2007. The study of the usage correlation between portal and traditional news media. Korean Journal of Journalism & Communication Studies, 51, 399–426. [In Korean] [Google Scholar]
  4. Bonacich, P. 1972. Factoring and weighting approaches to status scores and clique identification. The Journal of Mathematical Sociology, 2, 113–120. [Google Scholar]
  5. Bonacich, P. 1987. Power and centrality: A family of measures. American Journal of Sociology, 92, 1170–1182. [Google Scholar]
  6. Borgatti, S. P. 2002. NetDraw: Graph visualization software. Harvard, MA: Analytic Technologies. [Google Scholar]
  7. Borgatti, S. P., Everett, M. G. and Freeman, L. C.. 2002. UCINET 6 for Windows: Software for social network analysis. Harvard, MA: Analytic Technologies. [Google Scholar]
  8. Borgatti, S. P., Mehra, A., Brass, D. J. and Labianca, G.. 2009. Network analysis in the social sciences. Science (New York, N.Y.), 323, 892–895. [DOI] [PubMed] [Google Scholar]
  9. Carmody, M. 1991. Invisible victims: Sexual assault of people with an intellectual disability. Australia and New Zealand Journal of Developmental Disabilities, 17, 229–236. [Google Scholar]
  10. Chadwick, O., Cuddy, M., Kusel, Y. and Taylor, E.. 2005. Handicaps and the development of skills between childhood and early adolescence in young people with severe intellectual disabilities. Journal of Intellectual Disability Research : JIDR, 49, 877–888. [DOI] [PubMed] [Google Scholar]
  11. Chae, J. 2014. Study on the viewpoint diversity of the portal sites news by a type of remediation: Focused on the 18th election news. Korean Journal of Broadcasting and Telecommunication Studies, 28, 237–284. [In Korean] [Google Scholar]
  12. Cho, Y. and Ahn, D.. 2016. Examining news report research trends using keyword network analyses. The Journal of the Korea Contents Association, 16, 278–291. [In Korean] [Google Scholar]
  13. Feldman, R. and Dagan, I.. 1995. Knowledge discovery in textual databases (KDT). In: 1rst International Conference on Knowledge Discovery and Data Mining (KDD-95), Montreal, Canada, 20–21 Aug. [Google Scholar]
  14. Finlay, W. M. and Lyons, E.. 2005. Rejecting the label: A social constructionist analysis. Mental Retardation, 43, 120–134. [DOI] [PubMed] [Google Scholar]
  15. Gandomi, A. and Haider, M.. 2015. Beyond the hype: Big data concepts, methods, and analytics. International Journal of Information Management, 35, 137–144. [Google Scholar]
  16. Goffman, E. 1974. Frame analysis: An essay on the organization of experience. Cambridge, Mass: Harvard University Press. [Google Scholar]
  17. Harris, J. C. 2013. New terminology for mental retardation in DSM-5 and ICD-11. Current Opinion in Psychiatry, 26, 260–262. [DOI] [PubMed] [Google Scholar]
  18. Harris, J. C. and Greenspan, S.. 2016. Definition and nature of intellectual disability. In: N. N. Singh, ed. Handbook of evidence-based practices in intellectual and developmental disabilities. Basel, Switzerland: Springer International Publishing, pp.11–39. [Google Scholar]
  19. He, Q. 1999. Knowledge discovery through co-word analysis. Library Trends, 48, 133–159. [Google Scholar]
  20. Jo, E. and Berkowitz, L.. 1994. A priming effect analysis of media influences: An update. In: Bryant J. and Zillmann D., eds., Media effects: Advances in theory and research. Hillsdale, NJ: Erlbaum, pp.43–60. [Google Scholar]
  21. Johnson, J. C., Borgatti, S. P. and Everett, M. E.. 2013. Analyzing social networks. London: Sage Publications. [Google Scholar]
  22. Kang, J. and Kim, R.. 2012. Perspectives: Social construction approach to intellectual disabilities. Journal of Intellectual Disabilities, 14, 1–23. [In Korean] [Google Scholar]
  23. Keller, C. E., Hallahan, D. P., McShane, E. A., Crowley, E. P. and Blandford, B. J.. 1990. The coverage of persons with disabilities in American newspapers. The Journal of Special Education, 24, 271–282. [Google Scholar]
  24. Kim, D., A., Ahn, Y., Lim, J. and H, 2020. A study on development of in-service training for primary and secondary school principals: Focusing on educational administration leader’s program experience. The SNU Journal of Education Research, 29, 109–130. [In Korean] [Google Scholar]
  25. Kim, D. and Kim, S.. 2012. Legitimate expectation and unjustifiable actualities - Evaluation of news contents in Korean newspapers. Journal of Political Communication, 27, 35–73. [In Korean] [Google Scholar]
  26. Kim, S. and Lee, E.. 2020. The administration of full-time inclusive elementary classroom and support needs of teachers. The Journal of Inclusive Education, 15, 1–27. [In Korean] [Google Scholar]
  27. Kim, J. and Park, J.. 2012. Analysis of newspaper coverage on intellectual disabilities. Journal of Intellectual Disabilities, 14, 125–141. [In Korean] [Google Scholar]
  28. Korean Click. 2019. Website ranking by PC domain. [online] (updated Aug 2019). Available at: <http://www.koreanclick.com/top10/pc_domain_ranking.html> [Accessed 20 August 2019].
  29. Kwon, J. 2016. A survey and analysis on the acceptance attitudes and perception of the parents of students with disability on the “The Act Support and Guarantee of Right for Persons with Developmental Disability”. The Journal of the Korean Association on Developmental Disabilities, 20, 23–42. [In Korean] [Google Scholar]
  30. Laney, D. 2001. 3-D data management: Controlling data volume, velocity and variety. Meta Group Research Note, 6, 1–4. [Google Scholar]
  31. Lee, Y. C. 2007. Discourse on deconstruction of “mental retardation” labeling. Journal of Mental Retardation, 9, 23–42. [In Korean] [Google Scholar]
  32. Lee, M., An, Y., and Kim, E.. 2020. Analysis of school violence research trends related to students with disabilities. Journal of Education & Culture, 26, 795–819. [In Korean] [Google Scholar]
  33. Leydesdorff, L. and Welbers, K.. 2011. The semantic mapping of words and co-words in contexts. Journal of Informetrics, 5, 469–475. [Google Scholar]
  34. Lim, J. and Kim, J.. 2019. A study on the perception and development direction of special education related persons about special education school establishment. Journal of Education & Culture, 25, 155–175. [In Korean] [Google Scholar]
  35. Matson, J. L., Dempsey, T. and Fodstad, J. C.. 2009. The effect of autism spectrum disorders on adaptive independent living skills in adults with severe intellectual disability. Research in Developmental Disabilities, 30, 1203–1211. [DOI] [PubMed] [Google Scholar]
  36. Min, C. 2013. The current status and challenges of inclusive education in Korea. The Journal of Special Education: Theory and Practice, 14, 49–69. [In Korean] [Google Scholar]
  37. Park, J. 2010. Implication on definition and classification of intellectual disability in AAIDD 11th. Journal of Intellectual Disabilities, 12, 1–19. [In Korean] [Google Scholar]
  38. Park, S. 2015. Current states of sex maladjustment behaviour and sex education in middle school students with disabilities who are attending special class. Journal of Intellectual Disabilities, 17, 185–207. [In Korean] [Google Scholar]
  39. Park, J., Koh, H. and Park, Y.. 2019. Perception comparison of special education teachers, parents of children with disabilities, general education teachers, and residents in local communities of special education facilities. Korean Journal of Special Education, 54, 165–184. [In Korean] [Google Scholar]
  40. Park, S., Moon, J., Lee, S. and Han, K.. 2019. Current status of supports for students with special education needs and teachers perceptions on the inclusive education in Private school in Seoul. Special Education Research, 18, 111–141. [In Korean] [Google Scholar]
  41. Public Opinion Concentration Survey Committee. 2018. News usage concentration survey report 2016 ∼ 2018. [pdf] Seoul: Public Opinion Concentration Survey Committee. Available at: <https://www.mcst.go.kr/kor/s_notice/press/pressView.jsp?pSeq=17050> [Accessed 11 August 2019]. [Google Scholar]
  42. Reiter, S., Bryen, D. N. and Shachar, I.. 2007. Adolescent and intervention. Seminars in Speech and Languages, 4, 63–77. [Google Scholar]
  43. Richardson, L. 2017. Beautiful Soup. Available at: https://www.crummy.com/software. [Accessed 11 August 2019].
  44. Ryu, J. 2018. An experimental survey research on framing effects. Journal of Future Politics, 8, 35–64. [In Korean] [Google Scholar]
  45. Ryu, J. M., Bang, M. Y. and Zhang, Y. F.. 2016. The effects of a situation-based sexual abuse prevention program on the knowledge of sexual abuse and sexual attitude of middle school students with intellectual disabilities in special class. Journal of Intellectual Disabilities, 18, 99–117. [In Korean] [Google Scholar]
  46. Schaafsma, D., Kok, G., Stoffelen, J. M. T. and Curfs, L. M. G.. 2015. Identifying effective methods for teaching sex education to individuals with intellectual disabilities: A systemic review. The Journal of Sex Research, 52, 412–432. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Schalock, R. L., Luckasson, R. A., Shogren, K. A., Borthwick-Duffy, S., Bradley, V., Buntinx, W. H. E., Coulter, D. L., Craig, E. M., Gomez, S. C., Lachapelle, Y., Reeve, A., Snell, M. E., Spreat, S., Tassé, M. J., Thompson, J. R., Verdugo, M. A., Wehmeyer, M. L. and Yeager, M. H.. 2007. The renaming of mental retardation: Understanding the change to the term intellectual disability. Intellectual and Developmental Disabilities, 45, 116–124.2.0.CO;2] [17428134] [DOI] [PubMed] [Google Scholar]
  48. Scheufele, D. A. and Tewksbury, D.. 2007. Framing, agenda setting, and priming: The evolution of three media effects models. Journal of Communication, 57, 9–20. [Google Scholar]
  49. Shin, A., Baek, S., Yu, Y. and Kim, Y.. 2017. Keyword network analysis on ‘Free Semester Policy’ with Korean newspapers articles. The SNU Journal of Education Research, 26, 17–42. [Google Scholar]
  50. Siperstein, G. N., Norins, J., Corbin, S. and Shriver, T.. 2003. Multinational study of attitudes towards individuals with intellectual disabilities: General findings and calls to action. Washington, DC: Special Olympics. [Google Scholar]
  51. Song, H. and Yang, J.. 2017. Online news portal service and changes in news distribution: Big data analysis of Naver news in 2000–2017. Korean Journal of Journalism & Communication Studies, 61, 74–107. [In Korean] [Google Scholar]
  52. Stoiber, K. C., Purdy, S. and Klingbeil, D. A.. 2016. Evidence-based practices. In: N. N. Singh, ed., Handbook of evidence-based practices in intellectual and developmental disabilities. Cham: Springer, pp.41–68. [Google Scholar]
  53. United Nations. 2006. United Nations Convention on the Rights of Persons with Disabilities. Available at: <http://www.un.org/disabilities/convention/conventionfull.shtml>. [DOI] [PubMed]
  54. Vislie, L. 2003. From integration to inclusion: Focusing global trends and changes in the western European societies. European Journal of Special Needs Education, 18, 17–35. [Google Scholar]
  55. Whitehouse, M. A. and McCabe, M. P.. 1997. Sex education programs for people with intellectual disabilities: How effective are they? Education and Training in Mental Retardation and Developmental Disability, 32, 229–240. [Google Scholar]
  56. Wilkinson, P. and McGill, P.. 2009. Representation of people with intellectual disabilities in a British newspaper in 1983 and 2001. Journal of Applied Research in Intellectual Disabilities, 22, 65–76. [Google Scholar]
  57. Yoshida R. K., Wasilewski, L. and Friedman, D. L. 1990. Recent newspaper coverage about person with disabilities. Exceptional Children, 56, 418–423. [DOI] [PubMed]

Articles from International Journal of Developmental Disabilities are provided here courtesy of The British Society of Developmental Disabilities

RESOURCES