Paradoxical effects of thought suppression: a meta-analysis of controlled studies