With advancements in natural language processing, intelligent virtual agents (IVAs) are increasingly integrated into various sectors like education, customer service, personal assistance, and healthcare. Medical counseling and digital therapy, fields that require trusting relationships between patient and practitioner, profit immensely from the use of IVAs. A key component of these social relationships is empathy, which helps build trust and understanding. This paper investigates whether simulated empathy through emotional expressions in IVAs can improve interactions and influence users' emotional contagion. Additionally, it explores the correlation between self-reported empathy and users' expressiveness. Participants alternate reading an emotional story with a virtual agent (VA) which mirrors the users' emotional expressions in one condition, while remaining neutral in the baseline condition. The results show that simulated emotions can animate participants to elicit more facial expressions in response to the VA's, while correlating with the users' self reported empathy.