Pro-social and ethical behavior primarily understood as the ability to sacrifice one’s own interest to the benefit of others leads to welfare improvements and social progress, whereas anti-social and unethical actions result in irrevocable welfare losses and destruction. This dissertation includes three independent laboratory experiments that investigate specific factors behind people’s ethicality and prosociality. The research framework of the thesis is based on the moral psychology theories of self-image maintenance. The key idea of this approach is that people strive to maintain a positive self-image of themselves and use situational excuses to pursue a self-interest action they would otherwise consider inappropriate. In this respect, pro-social preferences are context dependent. The main emphasis of the experiments included in the thesis is on the dynamic nature of pro-social and ethical behavior, where one choice is studied in the context of another morally relevant action. Thus, the preferences to behave pro-socially are shown to be not only context but also path dependent. The thesis discusses the circumstances under which in the sequence of morally relevant choices consistent vs. balancing behavior emerges and seeks to provide empirical support for this discussion. The specific research questions addressed in the dissertation are as follows: is people´s willingness to donate affected by the ex-post information about the true need of the recipient? Are people that have been engaged in lying trying to compensate their wrongdoing afterwards by sharing more in the dictator game? To what extent the compensatory behavior is affected by the indirectness (delegation) of lying? Do participants with different mind-sets (outcome-based vs. rule based) account their previous ethical actions differently? The main findings of the dissertation corroborate the self-image maintenance paradigm. Participants who choose not to disclose the true need of the recipient shared significantly less, i.e. the ex-post blindness serves as an excuse for selfishness. Delegation of choice increases the incidence of lying, however, participants who delegate share more in the dictator game, i.e., they strive to compensate more than direct liars. Participants with rule-based mind-set (those who consider it inappropriate to sacrifice one person life to save five people in the trolley-dilemma task) and not the outcome-based participants react to examples of the unethical behavior of others by reducing their donations. The conclusion section of the dissertation outlines the areas for future research: the interplay of responsibility behind the action and the tendency to balance or compensate for it; the role of information acquisition in emotion-driven reciprocal behavior in social interactions; the effect of the moral account (previous un-ethical actions) on the social evaluation of the actors’ behavior.