Poverty alleviation crowdfunding is an "E-commerce + Agriculture + Crowdfunding" innovative mode, which promotes the targeted poverty alleviation and rural economic development in China. Previous researches mainly focus on non-content elements (e.g., the number of comments and praise) while publications are rarely related to content elements. We suggest that content elements (e.g., text of products) also have an important role in helping users to make different decisions (i.e., purchasing or donating). We select antipoverty programs in JD as a sample and adopt the linear regression model to analyze the influence of text type on the behavior of crowdfunding users. Specifically, the experimental results indicate that the emotion text (ET) make users tend to donate while the product text (PT) make users tend to purchase. This study will help project sponsors to achieve better effects on poverty alleviation by adjusting the ratio of ET to PT.