papers

2021

2021

1. Master Thesis

   Vocational Education Under The Structure Of Credentialism: A Case Study On Taipei Municipal Daan Vocational High School

   Shih Peng Wen

   Tunghai University , 2021

   Abs

   In Taiwan, vocational education is not preferred. The students of Taipei Municipal Daan Vocational High School (abbr. TAIVS) are exceptions, because they choose to go to vocational high schools instead of academic high schools. The reason why those students apply for TAIVS is worthy of investigation. We use the sequential analysis as our research method which originates from Objective Hermeneutics.

   According to our analyses, we find that the student’s motivation is basically based on Parents’ demand. Furthermore, parents’ demand is conducted by the pursuit of “Good School”, and we named that the “Structure of Credentialism.” Under this structure, TAIVS is an alternative to “Good School”, so people choose vocational education.

   Keywords: Tracking Education, Vocational Education, Credentialism, Structure of Credentialism, Objective Hermeneutics, Sequential Analysis

2022

2022

1. Conference Paper

   探索網路輿情與真實民意之間的關係：以關注議題為例

   謝佑明 ,  王晧軒 ,  and  溫士朋

   2022

   Abs

   近年來，網路上的討論已經變成一股不可忽視的聲量。眾人藉由網路論壇，成功發起多次社會運動；政治人物紛紛到臉書、YouTube等平台建立自己的帳號，以便宣傳自己的理念與想法；新聞記者也從網路找尋新聞故事或事件題材，並做進一步的報導；這些現象皆說明在網路上的輿論資料是值得研究的對象。本研究希望藉由調查問卷及網路輿論資料，來探討民眾對議題的重視程度，與新聞媒體及社群媒體的議題討論熱度是否相同。在調查問卷方面，將進一步分析民眾關注議題之間的關連性，以及是否因為接收訊息來源的不同而有一定的程度差異；在網路輿情資料方面，分別對新聞媒體與社群媒體進行資料探勘與主題模型分析，以取得關注的討論主題以及討論熱度，進一步分析這兩大平台所重視的議題為何。最後與調查問卷結果進行關注議題的比較，以瞭解各項議題的重視程度及異相之處。

   關鍵字：調查問卷；資料探勘；主題模型；網路輿情。

2023

2023

1. Conference Paper

   Unraveling the Correlation between Perceived Issue Importance and Issue Salience On the Internet among Users with Different Media Repertoires

   You Jian Wu ,  Hao Hsuan Wang ,  Shih Peng Wen , and 3 more authors

   2023

   Abs HTML

   With the advent of online media and social platforms, people now have access to a plethora of channels for receiving information. This has led to a diversity of media diets, which in turn may be associated with distinct perspectives on various issues. Consequently, a growing body of research has adopted a media repertoires approach to identify patterns of media use and explore how perceptions or attitudes vary among different types of media users. This study aims to provide evidence of the relationship between media diets, the perceived importance of various issues, and online public opinion on issues, using a representative sample from Taiwan. By conducting an in-depth analysis, we seek to shed light on the complex interplay between these variables. We employed latent class analysis (LCA) to identify four distinct classes of media diets among survey respondents (n = 3,982). A binary observation variable was derived from the four-point scale frequency of news information consumption by using the median split, with a frequency above the median considered high usage. Based on the three criteria: 1) the lowest AIC and BIC values, 2) entropy values above 0.6, and 3) a moderate share of each class, we identified a four-class model. Omnivores, consisting of 30.85% of the respondents, have the highest probability of high usage of all platforms. Old School, 27.37% of the respondents, have a lower probability of high usage of forums (PTT, Dcard) and Instagram. Netizens, 27.19% of the respondents, mainly use forums and have lower usage of newspapers, TV, and online news. Traditionalists, 14.59% of the respondents, have a higher probability of high usage of traditional media but a lower probability of usage of online news and social media platforms. We also asked survey respondents to rank thirty-two issues according to their perceived level of importance across three major topics: eighteen long-term social issues, eight communication issues, and six hot issues. To gauge online public opinion on the importance of issues, we detected the number of articles retrieved through keyword searches related to each issue. For each class of media diets, we used Spearman’s rank correlation to assess the correlation between the perceived level of issue importance and online public opinion as observed in online news and forums. Among Netizens, who browse forums to a greater degree, the correlations between their perceptions of issue importance and public opinion observed in online news and forums are .62 and .58 respectively, which are the highest among all four classes. Among Traditionalists, the two correlations were the lowest (.40 and .37). Extending prior literature, this paper suggests that whether peopleís perception of the importance of issues reflects online opinions depends on their media diets.

2. Conference Paper

   Online Mediaís Agenda-Setting Effect: A Method Triangulation Approach with Survey and Text Mining

   Hao Hsuan Wang ,  Shih Peng Wen ,  You Jian Wu , and 3 more authors

   2023

   Abs HTML

   Social media have become important platforms for people to receive news and express their opinions, which could influence others’ views on the importance of various social issues. Agenda-setting theory suggests that audiences perceive the issues which are most frequently discussed as the most important. This implies that the salience of an issue can transfer from social media users’ agenda to the public agenda. To investigate the agenda-setting effect, this research explores the relationship between the salience of issues in social media and their importance as perceived by the public. We also examine the role of active users on different social media platforms. Against this background, this work addresses the following three research questions: first, whether the importance of issues perceived by social media users correlates with the overall salience on the same platform; second, if the most active users exhibit a similar pattern compared to other users in general; and third, if the importance of issues perceived by the most active commenters is in line with the overall salience on the same platform. This work employs a mix of survey and text-mining approaches to address the above research questions. To measure the perceived importance of the issues to the general public, we adopted the Comparative Agendas Projects (CAP) framework as the main agenda structure. In a survey, we asked respondents from a representative sample (N = 5,102) of internet users in Taiwan to rank the importance of the issues in the framework. To aid in the measurement of issue salience in the next step, we also asked the respondents to write down relevant words for each issue. We also measure the respondents’ social media usage and their degree of participation in comments on social media. To measure the issues’ salience in social media, we harvested text data from Dcard, PTT, and LineToday, Taiwan’s three most popular news discussion forums. Next, we use a combination of custom dictionaries and topic modeling to analyze the salience of different issues in social media. First, following established methods in dictionary construction, we build word vector models from each platform and use the words from respondents to filter out the top ten words correlated with each CAP issue. Next, we recruit annotators to verify the relevance of the words. In addition, the topic model method transforms text into vectors using the BERT embedding model, which was then clustered using the K-means clustering algorithm. We hired annotators to certify every cluster’s agenda with the most representative words identified using the term frequency-inverse document frequency (TF-IDF). The words were then combined into an issue dictionary. Finally, we use the dictionary to measure the salience of each issue in the CAP framework. We then address the three research questions by testing the association between the issue salience on social media platforms and the importance of the issue as perceived by all survey respondents, the most active users, and the most active commenters.

3. Conference Paper

   家長的教養焦慮：以文字探勘技術分析「家有中學生」臉書社團

   溫士朋

   2023

   Abs HTML

   教育改革自1990年代開始，迄今已有 30 餘年，其中升學主義、升學壓力是主要的討論對象。然而，鮮少以「家長」的角度來探討「面對升學的反應」的相關研究。本研究試圖以「家長」作為出發點，針對組成成員為家長的Facebook 社團「家有中學生」的貼文進行分析，並試圖探索家長的在網路上，會希望與其他家長進行何種討論。為了希望得知家長們的討論有哪些「主題」，本研究採用自然語言處理（Natural Language Processing）的「主題模型」（Topic Model）將貼文進行分類。本研究發現，「家有中學生」社團的主要討論內容，分別為學業、親子關係、醫療，與學業相關的貼文，在六月時會變多，且多為討論會考成績；並且，與會考相關的討論內容，皆高於其他類型的貼文。最後，根據主題模型的結果，可以得知，雖然教育改革不斷的進行，但是，教育改革並未考量到家長的「行動邏輯」為何。換句話說，即便制度一直改變，但家長們仍然在在意的是「考試成績」的高低。