Individuals with mental illness endure pervasive stigma from society because of their condition, which puts them at elevated risk for distress, diminished self-esteem, social maladjustment, and worsened treatment outcomes. Therefore, treating only the symptoms of mental illness is not sufficient to improve the quality of lives of affected individuals. Rather, societal change in reducing stigma of mental illness must occur. This study tested the efficacy of a commonly used strategy (education-contact intervention; EC), relative to a novel intervention (loving-kindness meditation; LKM), and a control condition on reducing cognitive, affective, and behavioural stigma toward the mental illness of bipolar disorder. This addressed a number of limitations in the existing literature regarding measurement of stigma components and assessment of intervention efficacy. Three hundred seventy-six participants watched a 15-minute video corresponding to one of the three conditions: EC, LKM, or control. Then, they completed a self-report measure of affective, cognitive, and behavioural stigma toward a hypothetical person with bipolar disorder, and an implicit stigma measure. Participants also interacted with a confederate who presented as a fellow participant and disclosed being diagnosed with bipolar disorder. After the interaction, participants completed measures of affective stigma toward the confederate. Behavioural stigma toward the confederate was measured by participants’ behavioral intentions to interact with the confederate in the future, the confederate’s impression rating of the participant, and naïve coders’ thin-slice coding of the interaction. Results indicated that the EC condition led to improvements in some aspects of stigma toward a hypothetical person with bipolar disorder, and greater intentions to interact with the confederate. Importantly, knowledge of bipolar disorder, positivity toward others, and negativity toward others were significant mediators through which EC impacted improvements in stigma toward a hypothetical person with bipolar disorder. Although LKM did not reduce affective and implicit stigma as expected, it was associated with greater intentions to interact with the confederate. Moreover, higher positivity toward others due to LKM mediated positive outcome in some aspects of stigma. Implications of findings for stigma interventions, as well as clinical implications for individuals with mental illness and mental health care providers, are discussed.