Background and aimsAlthough social networking sites brought giant convenience, many negative effects on users’ psychological well-being need more investigation. This study used a survey to examine Facebook addiction and state anxiety without Facebook. As research shows gender is related to trait anxiety and may interact with trait anxiety to influence state anxiety, we also assess the interaction effect between gender and trait anxiety.MethodsA total of 526 college students in the US participated in the survey. A systematic sampling method was used and an e-mail invitation with the link of the survey was sent to every third student on the students’ e-mail list. Study measures included demographics, trait anxiety, Facebook intensity, different Facebook activities, Facebook addiction, and state anxiety without Facebook. Hierarchical multiple regression was run to test how trait anxiety, gender, Facebook intensity, and different types of Facebook activities predict Facebook addiction and state anxiety.ResultsFacebook use intensity predicts Facebook addiction (β = 0.573,p < .001) and state anxiety (β = 0.567,p < .001). Facebook use for broadcasting positively predicts Facebook addiction (β = 0.200,p < .01) and state anxiety (β = 0.171,p < .01). Trait anxiety positively predicts Facebook addiction (β = 0.121,p < .05) and state anxiety (β = 0.119,p < .05). Gender interacts with trait anxiety and jointly predicts Facebook addiction (β = 0.201,p < .01).Discussion and conclusionsTrait anxiety, Facebook intensity, and broadcasting behavior on Facebook positively predict Facebook addiction and state anxiety. Moreover, gender interacts with trait anxiety, so that the gender difference in Facebook addiction is significant only when trait anxiety is low.