在數位時代，網路聲量已成衡量選舉結果的重要指標之一，也是政治傳 播研究的新趨勢，可惜本土的選戰實證付之闕如。有鑒於此，本文分析183 萬筆跨平台數據，探討2020 年臺灣區域立委選舉中，410 位候選人的網路 聲量與其得票率及選舉結果的關聯性。結果顯示，候選人間的網路聲量落差 甚大；且在性別、選區、年齡及選舉結果上有顯著差異。尤其是選戰後期的 正面新聞聲量與得票率呈現高度正相關，而且在73 位勝選者中有70 人的後 期正面新聞聲量在選區中名列前二。本研究凸顯網路聲量在選戰實務的重要 性，並從選舉進程、聲量來源、與情緒等多維度提供實證支持，可作為後續 得票率預測研究之基礎。

In the digital age, online buzz has become a crucial indicator for predicting election outcomes. However, empirical studies on local election campaigns remain scarce. This study addresses this gap by analyzing 1.83 million cross-platform data points to investigate the relationship between online buzz, vote shares, and election outcomes for 410 candidates in Taiwan’s 2020 legislative elections. The results reveal significant disparities in online buzz among candidates, with notable differences observed in gender, electoral district, age, and election outcomes. Particularly, a strong positive correlation was found between positive news websites buzz in the later stages of the campaign and vote share. Among the 73 winning candidates, 70 ranked within the top two in terms of positive news web site buzz in their respective districts during the final phase of the election. This study underscores the practical importance of online buzz in election campaigns and provides empirical evidence from multiple dimensions, including election progression, online buzz sources, and sentiment. These findings offer a foundation for future research on vote share prediction.

大數據;立委選舉;網路社群;網路選戰;網路聲量

big data; legislator election; social media; online campaign; online buzz