On January 22, 2001, an elderly couple in a wheelchair fell from a lift for the disabled at Oido Station on Seoul Subway Line 4 and died. The last 22 days were the 22nd cycle of death. 22 years have passed, but the issue of 'guaranteeing the right to move for the disabled' is still ongoing. The National Solidarity for the Elimination of Discrimination Against Persons with Disabilities continues to protest by riding the subway, and the Seoul Metropolitan Government declares its will to take a hard line with a public opinion poll showing that 56% of citizens oppose the protests of the Jeonyeon Yeon. What are the citizens thinking? Is the 'public opinion' known so far correct? SBS Data Journalism Team Mabujachim and Knowledge Content Startup Underscore have diagnosed public opinion in our society looking at the protests of Jeon Jeon-yeon.
In Part 1, through an analysis of opinion polls, we looked at how support for the Jeon Yeon-yeon protests was affected by factors such as residence, subway usage, political beliefs, and demographic factors.
In this episode, I'm going to look into public opinion hidden in online news comments.
What was the difference between online comment public sentiment compared to public opinion polls?
I analyzed like this.
Among the 41 major media articles written from December 1, 2021 to January 12, 2023 using the Korea Press Promotion Foundation's 'Big Kinds', the titles include <Jeon Jang-yeon>, <Jang Cha-yeon>, <Disabled Demonstration>, and <Subway Demonstration>. >, <Demonstration for the Disabled>, and <National Coalition for the Elimination of Discrimination against the Disabled> were included in 1,165 articles.
We collected 233,000 top 200 comments for these articles.
Among the users who wrote the comment, 2,500 people were randomly selected for each portal, and 1,971 people who were confirmed to have commented on articles related to Jeon Ji-yeon for at least two days based on the most recent 3,500 comments were finally used for analysis.
We classified them into conservative and progressive types through a political orientation classification model.
Jeon Yeon-yeon's article comments... When was the most
First of all, we looked at how many comments were posted on the article Jeon Jeon-yeon, and the number of comments on articles related to the subway getting on and off protests by organizations with disabilities.
Between March and May of last year (period 1), the public opinion of comments fluctuated greatly, and from November to December (period 2) of last year, public sentiment of comments fluctuated again.
It exactly coincides with the time when public opinion for and against the protests on getting on and off the subway was heating up.
What is the background that public opinion was hot only at two points?
Analyzing media articles at the time, the Jeonjeon-yeon protests before March 2022 were largely out of people's attention.
At the time, there were protests for getting on and off the subway, but the reaction to the related article was not great.
The average number of comments on an article was less than 5.
However, since March, the number of comments on articles written by Jeon Yeon-yeon skyrocketed, as shown in the graph above.
"...If we tolerate the method of struggle that the greatest number of unspecified inconveniences are special interest in us', the order of our society will collapse..."
- March 25, 2022 Excerpt from Lee Jun-seok, former representative of People's Power on Facebook -
The reason cannot be explained with data, but it seems true that public opinion began to boil right after Lee Jun-seok, former representative of People's Power, criticized the method of protests for the mobility of people with disabilities on his Facebook page for three days from March 25 to 27 last year.
The controversy over Jeon Yeon-yeon, launched by former CEO Lee, quickly began to fall into the logic of the camp, and the battle against the Jeon Yeon-yeon protests began to intensify in earnest.
The problem is that malicious comments have also increased.
As shown in the table above, if expressions such as various abusive language, hatred, and criticism are included, the malicious comments classification algorithm classifies them as malicious comments.
Although the frequency of comments has soared, the number of malicious comments has also increased, so it is inevitable that the comment forum has failed to create a healthy discourse.
In particular, the percentage of malicious comments from conservative users is noticeably high.
Looking at the comments on the article at the time, it was easy to see comments that defended former CEO Lee's remarks, including hatred of the disabled.
Comments that have been quiet since May fluctuate once again at the end of the year.
This is because the Jeon Yeon-yeon resumed the subway protests demanding a budget to guarantee the rights of the disabled.
However, when comparing the frequency of comments between the two periods, there is one notable difference.
Right after Lee's remarks about Jeon Yeon-yeon, there are a lot of comments from both progressive and conservative users.
Of course, the number of comments from conservative users is much higher, but the number of comments from progressive users was also significantly higher.
Regardless of ideology, it can be seen that all of them actively responded to the issue of Jeon Jeon-yeon at this time.
However, when looking at the trend of comments on the articles of Jeon Yeon-yeon, which resumed at the end of the year, changes in the pattern of progressive users are observed.
You can see that the number of comments from progressive users has decreased noticeably.
Rather, the number of comments from conservative users has increased, but is it because the number of comments from progressive users has decreased because they have lost interest?
Progressive users who chose to comment instead of comment… why?
That's not it.
The reason can be seen in the graph above.
Looking at the graph, the proportion of large comments out of all comments was higher for progressive users.
In other words, progressive users wrote more large comments instead of comments.
The large comment rate of progressive users was 15.8%, which was 3.6%p higher than the 12.2% large comment rate of conservative users (p<.001).
But what about comments?
For those who are unfamiliar, let's start with the nature of the big comment before interpreting it.
Large comments are slightly different in nature from the general comments mentioned above.
While general comments are mainly used to give direct opinions about the body of an article, large comments are a way to express opinions while relatively reducing exposure.
Applying this to the controversy over Jeon Yeon-yeon's protests, if conservative users criticized or attacked the protests and organizations getting on and off the subway through general comments, progressive users refuted it with large comments or defended Jeon Yeon-yeon.
If you write a general comment, you can see that conservative users' attacks and controversy are growing, so it can be seen as a response by posting a large comment under the comment of a conservative user.
In the comments above, when one user left a negative comment about Jeon Yeon-yeon, another user refuted with a large comment.
Shall we also look at the time series?
It can be seen that both the <Graph of Comments vs. Large Comments by User Political Inclination> and <Ratio of Bad Comments vs. Comments by User Political Inclination> increased significantly at the end of the year.
As a rough guess, the surge at the end of last year appears to be partly due to a change in the political environment.
The government and the person in charge of the city of Seoul, who are responsible for resolving the controversy of Jeon Yeon-yeon, have changed.
In fact, if you look at the large comments from progressive users at the end of last year, there are a lot of criticism and abusive language against the current government and the mayor of Seoul.
**If the 'Go to View' button is not pressed, you can view the address by moving it to the address bar.
► This project was supported by the Korean Journalism Society-SNU Fact Check Center.