This research explores how portrayals of Chinese companies in American and British economic news tweets on X (formerly Twitter) change in response to shifts in China's macroeconomic conditions between 2007 and 2023. China's economy experienced rapid growth with fluctuations during this period, making it crucial to understand how these fluctuations are reflected in international media coverage, especially given the significant role of Chinese companies in global trade and the increasing reliance on social media for news consumption. The study utilizes van Dijk's framework on ingroup and outgroup ideologies, examining how sentiments, emotions, and themes shape the ideological depiction of Chinese firms. Existing literature shows a link between economic conditions and media sentiment, with a tendency toward negative bias, even when positive economic trends occur. This study investigates how China's macroeconomic indicators, such as GDP and PMI, affect the sentiment expressed in tweets about Chinese companies. The performance of Chinese companies is intrinsically linked to China's macroeconomic performance, influencing investor decisions and therefore impacting how international media frame them. Social media's engagement-driven nature also influences news production, with journalists often employing emotive language to attract audiences. Therefore, analyzing sentiments and emotions in tweets provides insights into journalistic strategies and their impact on the economic value of news. The research questions address how sentiments correlate with macroeconomic indicators and ideological representation; how strengths, weaknesses, opportunities, and threats (SWOT) are evaluated through sentiments; and how emotions influence sentiment shifts and ideological transitions. The findings offer valuable insights for business journalists and researchers, improving the quality of economic news reporting on Chinese companies.

Existing research reveals varied portrayals of China's economy in American and British media. American media often present a negative view, especially during crises and trade tensions, sometimes employing war metaphors to depict China as a threat to American interests. British media portrayals are more nuanced, showing a mixture of ethical concerns, risks, and opportunities. Studies have observed a dichotomous representation of China as both a market disruptor and a potential economic catalyst, with a shift toward risk and human rights concerns after 2005. However, recent research also notes favorable depictions of Chinese economic achievements, particularly concerning initiatives like the Belt and Road Initiative, alongside concerns about global trade inequity. Previous studies primarily focus on traditional media and China's macroeconomic landscape, or on specific companies like Huawei, using corpus-assisted discourse analysis. This study bridges the gap by examining social media representation of Chinese companies, employing sentiment analysis and evaluative target annotation.

The study compiled a corpus of 55,394 tweets (934,155 words) from 35 verified X accounts of 19 prominent US and UK economic news outlets between August 2007 and December 2023. Data collection used the Python-based tool Twint and 22 seed words related to Chinese companies. The corpus was divided into American and British sub-corpora. China's quarterly GDP and monthly PMI data from CEIC Data were also collected to contextualize the tweets. The tweets were categorized into six macroeconomic scenarios reflecting China's economic performance (e.g., GDP UP/PMI > 50, GDP DOWN/PMI < 50). The analysis involved a four-step procedure. Step 1 used two RoBERTa-based models (twitter-roberta-base-sentiment and twitter-roberta-base-emotion-multilabel) to analyze sentiments (positive, negative, neutral) and emotions (anger, anticipation, disgust, fear, joy, love, optimism, pessimism, sadness, surprise, trust) in the tweets. Step 2 developed an annotation scheme to identify evaluative targets (strengths, weaknesses, opportunities, threats) in positive and negative tweets, consistent with frameworks emphasizing internal and external company factors in economic news. Step 3 utilized ChatGPT-4 (with validation from Claude 3 and Microsoft Copilot) for LLM-assisted annotation of evaluative targets in the 17,074 positive and negative tweets, using a two-stage process of prompt optimization and annotation. Step 4 employed Spearman correlation and multiple linear regression analyses (SPSS 29) to examine correlations between sentiments and macroeconomic indicators, and the influence of emotions on sentiment changes.

The Spearman correlation analysis revealed significant (p < 0.01) correlations between sentiments in tweets and China's macroeconomic indicators. Positive sentiment decreased as the economy shifted from expansion to contraction, while negative sentiment increased. American news tweets showed the highest positive sentiment during economic balance and the highest negative sentiment during prolonged downturns. British news tweets showed increased positive sentiment during stability or early recovery and increased negative sentiment when downturns were predicted. Analysis of evaluative targets (strengths, weaknesses, opportunities, threats) in positive and negative tweets showed that positive tweets focused on strengths and opportunities, while negative tweets emphasized weaknesses and threats. In American media, emphasis on strengths decreased during economic downturns, while emphasis on opportunities increased. Positive tweets often highlighted technological advancements and opportunities like IPOs. Negative tweets focused on weaknesses (legal, financial difficulties) and threats (trade tensions, security concerns). In British media, emphasis on strengths decreased during downturns, while emphasis on opportunities increased during recovery. Positive tweets highlighted corporate successes and market appeal, while negative tweets focused on financial shortcomings, safety issues, and external threats. Multiple linear regression analyses showed that nine emotions significantly (p < 0.0001) impacted positive and negative sentiment variance. In American media, joy and trust positively influenced positive sentiment, while surprise had a negative influence. Fear positively influenced negative sentiment, while anticipation had a negative influence. In British media, love and joy positively influenced positive sentiment, while fear positively influenced negative sentiment and anticipation had a negative influence.

The findings demonstrate a significant correlation between the sentiment expressed in economic news tweets about Chinese companies and China's macroeconomic performance. The portrayal of Chinese firms shifts from ingroup to outgroup representations as economic conditions worsen. The study's results contrast with some previous findings, showing increased positivity during economic expansion, potentially due to social media's engagement-focused nature. The amplification of positive sentiment during favorable conditions may attract investors, while increased negativity during downturns serves as a cautionary signal. The study also contributes to the literature by providing a nuanced understanding of the portrayal of Chinese companies on social media, showing that negative representations intensify during economic downturns, potentially reflecting media efforts to caution audiences about risks. The emphasis on strengths and opportunities during specific macroeconomic scenarios suggests strategic framing by news outlets to balance investor awareness and market interests. The interplay between economic dynamics and ideological shifts highlights the influence of economic pressures on media representations. The use of emotions like joy, trust, and fear in shaping the portrayal of Chinese companies underscores the importance of considering the emotional impact on audience engagement and investment decisions.

This study reveals a dynamic relationship between China's macroeconomic performance and the ideological portrayal of Chinese companies in American and British economic news tweets. The findings highlight the significant influence of economic indicators, journalistic practices, and emotional language on media representations. Future research could explore additional macroeconomic indices, expand the scope to include media from other regions, and investigate the long-term effects of these portrayals on investor behavior and cross-national trade relations.

The study's limitations include the focus on only two macroeconomic indicators (GDP and PMI) and the restriction to US and UK media. Future research should incorporate additional macroeconomic indices and expand the geographical scope to capture a broader perspective on the portrayal of Chinese companies in international media.