Managers confronted with evidence of employee misconduct were biased in their recommendations for disciplinary action depending on whether they liked or disliked the accused employee. As expected, when evidence was strong, managers more frequently recommended that the subordinate receive assistance-focused and punishment-focused disciplinary action than when evidence was weak. When managers indicated they liked the employee, they were not inclined to suggest use of disciplinary actions but endorsed use of discipline when they disliked the accused employee. Joint effects of evidence and liking were found for assistance-focused and punishment-focused actions. Implications and the relevance of the biasing effects of managers' recommended actions are discussed.