Analyzing College Experience Through Social Media and Python: A Mixed-Method Comparison of Chinese and Non-Chinese Students

• In Event: International Students and Social Media and Digital Learning

Proposal

With over 3.6 billion users in the digital society nowadays and presumably 4.41 billion by 2025, social media has become an indispensable part of modern life, particularly among the younger demographic, such as college students (Kanchan & Gaidhane, 2023). Studies show that students have a high level of intention to share information in the social media and that social media is among the most preferred media of information. (Kim, 2020) Therefore, social media is an ideal portal to observe students’ real experience and thoughts of college life, which is a key predicting factor of their retention rate and academic performance.
Studies show that students’ satisfaction with college life and learning experiences positively correlates with graduation rates and indirectly influences academic performance (Strahan & Credé, 2015). Despite increased higher education enrolment rates, college students around the world are facing various challenges: Chinese college students’ adaptability in universities is disappointing (Shao, 2019), while 64% of the undergraduate students who began seeking a 4-year bachelor’s degree in the USA in fall 2014 managed to graduate within 6 years by 2020 (NCES, 2021). In addition, prevailing depression and high stress rates are found in countless studies(Ramón-Arbués et al., 2020). Ssocial media platform today gave those population a place to share.
Python is a powerful tool for text analysis and has been widely used in research on social media, but this approach has rarely been combined with educational research. This explanatory mixed-method study incorporates python-based text processing and sentiment analysis of university experiences on English and Chinese social media along with in-depth interviews with domestic and international students in China, aiming to gain insights into similarities and differences in higher education between cultures, as well as the impact of cultural factors on students' subjective university experiences.
For this study, we adopt Hettler's(1980) Model of Wellness. Hettler defined wellness as an overall state of physical, mental, and social well-being and developed a holistic wellness model integrating six dimensions of a student's life (physical, intellectual, social/emotional, spiritual, environmental, and occupational). In order to achieve a balanced level of such wellness and fully enjoy positive, healthy, and complex learning and developing experiences, a student should devote time and personal effort in each dimension. This theoretical framework enables us to intentify assessment dimensions while conducting sentiment analysis and interview questions, and systematically analyze potential factors contributing to the differences between students’ university experience across different cultural contexts.
Considering that students may have reservations about school-based surveys and thus are more willing to speak up to peers on social media, we chose Zhihu, China's largest online Q & A platform, as our source for Chinese comments, and Reddit and Quora, both popular English online Q & A communities with over 300 million active users, as our source for English comments. We retrieved a number of questions from popular Chinese and English social media using keywords “University Experience (life)”, “Satisfaction”, “How would you rate” etc. (in both languages). While collecting texts for analysis, we followed the following principles: We only fully exported answers roughly below 200 words (characters); for high-quality answers that are too long or descriptive, we extracted leading and/or concluding paragraphs and leave out obviously off-topic sentences after carefully reading. Eventually, 321 English and 306 Chinese comments were collected, totaling over 50,000 words and characters.
After data and text cleaning processes, sentiment analysis was conducted using Python. Sentiment analysis technology can quantify and evaluate emotions in natural language and social interactions (Wang & Dong, 2020). Currently, this technology plays a significant role in market research, business strategy, public opinion management, and educational research (Guan, 2021). In education, sentiment analysis can quickly categorize students’ feedback on courses and school services and predict their emotional trends. Therefore, this technology is effective in improving teaching approaches, campus environment, and overall learning experiences. It brings new perspectives and methods to educational research in the digital society, contributing to the enhancement of education quality and student satisfaction. The technology is hitherto more commonly used in education for evaluation of teaching, educational products’ marketing management (Zhou & Ye, 2023) and teaching material analysis (Chen, 2022), and is not yet widely used in analyzing education-related discussions on social media. In this study, we created training and testing sets in order to select the most accurate model for the collected comments before visualizing major findings and conducting a more targeted round of sentiment analysis based on high-frequency words and key phrases.
To address the limitations of quantitative results and further examine the impact of culture in students’ subjective college experience. We used convenience sampling and criteria sampling to recruit 14 domestic and international undergraduate students (7 each) with diverse cultural backgrounds (such as social economic status, birth place, campus location etc.) seeking a bachelor’s degree in a comprehensive research institution in Shanghai.
We found through the sentiment analysis, college students on English social media tend to have a more positive attitude towards their university experience than those on Chinese social media. This phenomenon may stem from the different focuses in students’ college lives, as well as varied influence of national situations on aspects of higher education. Major differences especially occur in intellectual, social/emotional and occupational dimensions. Moreover, comments in both languages demonstrate distinct logic and emotional patterns, reflecting the values of different cultural circles. This offers insights into varying definitions of college satisfaction. The conclusion is further verified in interviews with domestic and international students in China, during which other influencing culturally-related factors are discovered. The different ways students express themselves offers insight into how higher education institutions should take cultures into consideration when communicating and working with international students. By comparing and analyzing these patterns, we can better recognize the strengths and limitations of different cultures and learn to better understand and enhance students' university experiences.

Authors

• Yirong Yu, Shanghai Jiao Tong University

• Xiaoqiao Zhang, Shanghai Jiao Tong University