An optimization model is employed to determine the outcome of within-group conflict in cooperative societies. The model assumes that selection acts simultaneously on the stronger, dominant members of the group to secure more benefits for themselves at the expense of subordinates, and on subordinates to leave the group in response to this usurpation if they can do better elsewhere. When it is in the interest of the dominant to maintain the group, the dominant will be limited in its opportunity to bias the benefits in its favor by the options available to subordinates outside the group. The resulting differences in fitness or net pay-off among group members due to usurpation by dominants define the degree of skew. Ecological factors affect both the net benefit of grouping and the options for leaving, and thus establish the limits of skewing. The assumptions and rules governing this interaction are first outlined using a game theory approach. A continuous model for predicting the exact degree of skew is then developed. This analysis shows that the degree of skew increases when the benefit of group breeding relative to solitary breeding increases, and when the degree of relatedness among group members increases. Finally, the trade-offs between breeding with relatives versus nonrelatives are analyzed from a dominant relative's point of view. The results of two empirical studies on a social bird and a social wasp are compared to the predictions of these models. The exceptionally good fit to the models is interpreted as evidence in favor of the “social manipulation” approach to “helping” and “altruism” in cooperatively breeding species.