The cognitive–emotional theory of esteem support messages predicts that message style will affect the outcomes of esteem support interactions. However, little research has focused on the effects of message style; that is, how esteem support messages are delivered. The present experiment addresses this lacuna by manipulating message style in a laboratory study examining face-to-face esteem support interactions. Confederates were trained to provide emotion-focused esteem support to naïve participants (N = 173) in four styles along the assertive–inductive dimension, in addition to a listening-only control condition. We then assessed the effect of the interaction on participants’ state self-esteem. Results indicated that emotion-focused esteem support improved state self-esteem more than listening support; however, there was no significant effect of message style. Post-interaction state self-esteem improvement was positively associated with the quantity of emotion-focused esteem support content provided during the interaction.