On social media platforms, people sharing similar stances or attitudes often cluster into “opinion groups” that dominate the formulation and dissemination of perspectives in public opinion. Within these opinion groups, members tend to accept information that resonates with their own views in interactions, potentially leading to the echo chamber effect. The presence of echo chambers within opinion groups may affect members’ exposure to and engagement with heterogeneous opinions, which may affect group influence and information dissemination. However, it remains unclear whether there are significant differences between opinion groups within and outside the echo chamber effect. Therefore, this study aims to comparatively analyze the influencing and disseminating power of opinion groups within and outside echo chambers in the social network structure.This study conducts a multi-level analysis to investigate the problem. First, a clear definition of “opinion group” is proposed on the basis of an “opinion-attitude” structure, and opinion groups in online public opinion events are precisely identified using large language models. Second, a novel method based on homogeneity in retweeting behavior for identifying echo chamber effect is introduced, which is used to identify opinion groups within echo chambers. Finally, the HITS algorithm is improved to measure the influencing and disseminating power of opinion groups, and a comparative analysis of the influencing and disseminating power between groups within and outside echo chambers is conducted.The result reveals no significant difference in influencing power between groups within and outside echo chambers. However, a significantly lower disseminating power is shown in opinion groups within echo chambers. This suggests that echo chambers primarily affect the performance of opinion groups by constraining information dissemination, rather than directly enhancing or weakening the overall influence of the groups. The disseminating power of opinion groups is closely tied to the social media environment, whereas influencing power is more dependent on the inherent characteristics of groups themselves. Further sentiment analysis reveals that opinion groups within echo chambers display more neutral expressions, with fewer emotionalized expressions. Since emotional content is more likely to spark discussion and resonate with audiences, the observed differences in disseminating power between the two types of groups may stem from the limited outward permeability of information in groups within echo chambers, where information circulates frequently within the group but struggles to extend beyond its boundaries.This study proposes a novel definition of opinion groups centered on the “viewpoint-attitude” structure, introduces the use of large language models for hierarchical identification of opinion groups, and presents a modeling and identification method for echo chamber effects based on retweet behavior. Furthermore, the HITS algorithm is improved to measure the influencing and disseminating power of opinion groups. Through this approach, the study reveals, from the perspective of echo chamber effects, the differentiated disseminating power exhibited by opinion groups due to their respective communication environments, aiming to contribute to a better understanding of the dissemination mechanisms and governance strategies of online public opinion.
LIUS, GUANX, XIAH. Exploring group opinion polarization based on an opinion dynamics model considering positive and negative social influence[M]//CHEN J, HUYNH V N, TANG X, et al. Knowledge and systems sciences. Singapore: Springer Nature Singapore, 2023: 260-273.
[2]
DONGJ, ZHANGY C, KONGY. The evolution dynamics of collective and individual opinions in social networks[J]. Expert Systems with Applications, 2024, 255: 124813.
[3]
GUX, CHENF, YANGX, et al. Evolutionary trend and network structure characteristics of publicity information dissemination about waste separation by different opinion leaders[J]. Resources, Conservation and Recycling, 2023, 194: 106991.
[4]
POGSONM, NICHOLLST. Agent-based modeling of diversity, new information and minority groups in opinion formation[J]. Computational Communication Research, 2024, 6(1): 1-40.
[5]
DEL VICARIOM, BESSIA, ZOLLOF, et al. The spreading of misinformation online[J]. Proceedings of the National Academy of Sciences, 2016, 113(3): 554-559.
[6]
KRATZKEN. How to find orchestrated trolls? A case study on identifying polarized Twitter echo chambers[J]. Computers, 2023, 12(3): 57-78.
[7]
FEBRIA LINAL, HARTINIS, MARDHIYAHD, et al. Understanding opinion leadership in social media: the role of perceived fit with personal interests in purchase behavior and social media engagement[J]. International Journal of Economics and Management, 2024, 18(3): 343-357.
[8]
MALINENS, KOIVULAA. Influencers and targets on social media: investigating the impact of network homogeneity and group identification on online influence[J].First Monday,2020,25(4):10453.
[9]
FIRMANSYAHY, RAMADHONYF, ROHANIAHY. Opportunities and challenges towards the 2024 election: the “echo chamber” phenomenon and the influence of social media on the levels of political literacy of beginner and millennial voter[J]. ICCD, 2023, 5(1): 243-249.
[10]
CONOVERM, RATKIEWICZJ, FRANCISCOM, et al. Political polarization on Twitter[J]. Proceedings of the International AAAI Conference on Web and Social Media, 2021, 5(1): 89-96.
[11]
GRUZDA, HAYTHORNTHWAITEC. Enabling community through social media[J]. Journal of Medical Internet Research, 2013, 15(10): e248.
[12]
MCPHERSONM, SMITH-LOVINL, COOKJ M. Birds of a feather: homophily in social networks[J]. Annual Review of Sociology, 2001, 27(1): 415-444.
[13]
SONGK, WANGD, FENGS, et al. Detecting positive opinion leader group from forum[M]//GAO H, LIM L, WANG W, et al. Web-age information management. Berlin, Heidelberg: Springer Berlin Heidelberg, 2012: 95-101.
[14]
YANGJ, LESKOVECJ. Defining and evaluating network communities based on ground-truth[J]. Knowledge and Information Systems, 2015, 42(1): 181-213.
DUBOISE, BLANKG. The echo chamber is overstated: the moderating effect of political interest and diverse media[J].Information, Communication & Society, 2018, 21(5): 729-745.
[17]
YES, TUL L, WANGX J,et al. Opinion consensus incorporating higher-order interactions in individual-collective networks[J]. Chinese Physics B, 2024, 33(7): 070201.
[18]
CHOID, CHUNS, OH H, et al. Rumor propagation is amplified by echo chambers in social media[J]. Scientific Reports, 2020, 10(1): 310-319.
[19]
JASNYL, WAGGLEJ, FISHERD R. An empirical examination of echo chambers in US climate policy networks[J]. Nature Climate Change, 2015, 5(8): 782-786.
[20]
AGUILARE, FUJISAKIY. Reaching consensus via coordinated groups[J]. SICE Journal of Control, Measurement, and System Integration, 2021, 14(1): 20-26.
[21]
LIW, GUOC, DENGZ, et al. Coevolution modeling of group behavior and opinion based on public opinion perception[J]. Knowledge-Based Systems, 2023, 270: 110547.
[22]
VENDEVILLEA, DIAZ-DIAZF. Echo chamber effects in signed networks[J]. Physical Review E, 2025, 111(2): 024302.
[23]
CHENX, LIANGH, ZHANGY, et al. Consensus manipulation in social network group decision making with value-based opinion evolution[J]. Information Sciences, 2023, 647: 119441.
[24]
BOUTYLINEA, WILLERR. The social structure of political echo chambers: variation in ideological homophily in online networks[J]. Political Psychology, 2016, 38(3): 551-569.