Daily released news and analyst reports can affect economic activities such as product marketing and stock trading. At the same time, we often interpret news by using our domain knowledge to see what influence an event might have. Therefore, it is an important analytical task to evaluate hypotheses or opinions by considering simultaneously the public opinion formed by these news sources and one’s own causality knowledge. In this paper, we present HOPE-Graph: a hypothesis evaluation service that considers news and the knowledge of causal relationships between hypotheses. This service is implemented by integrating news analysis obtaining pro/con rates for hypotheses into a hypothesis evaluation framework that propagates pro/con rates based on positive/negative causal relationships. The framework is conceived as an abstract argumentation framework that evaluates arguments on the basis of the support/attack relationships between them. By using the causality knowledge, it is possible to obtain a score that preemptively incorporates the score of the causal hypothesis. To investigate the effectiveness of our service, we checked the Granger causality from the score of the hypothesis of a company’s performance to the stock price and found that proper consideration of causality knowledge can lead to scores that are more relevant to stock prices in more cases.