There are many barriers to the delivery of evidence-based treatment, including geographical location, cost, and stigma. Self-help may address some of these factors but there is a paucity of research on the efficacy of self-help for many problems, including social anxiety. The present research evaluated the efficacy of a mindfulness and acceptance-based self-help approach for the treatment of social anxiety. Individuals seeking help for social anxiety or shyness were recruited from the community. Participants (N = 117) were randomly assigned to a book (n = 58) or wait-list control condition (n = 59) on a 1:1 ratio. Hierarchical linear modelling results supported the efficacy of the self-help condition with between-group effect sizes on social anxiety outcomes ranging from .74 to .79. Significant change was also observed on self-compassion, mindfulness, acceptance, and depression. Some variables, including social anxiety and acceptance, were assessed weekly for those in the book condition. Additional participants (n = 35) were recruited for the book condition increasing the sample size to 93 for the latent change score modelling analyses. A unidirectional model was supported: increases in acceptance were associated with subsequent decreases in social anxiety. Overall these results support the use of a mindfulness and acceptance-based self-help approach for social anxiety.