A social dilemma is a situation of a conflict between self-interest and collective interest, in which a person is often forced to choose between personal and collective gain. Research shows that contributions to the collective good can be enhanced by a change in the payoff structure and especially by punishing free-riders. Free-riders tend to maximize their personal gain, while profiting off of others’ cooperation. In contrast to them, cooperators are willing to contribute to the collective good. In the current research, we examined whether there are individual differences in the reaction to the punishment threat in a social dilemma simulation game. Our hypothesis was that some players (i.e., those low in Honesty, high in Machiavellianism, Narcissism, Psychopathy, and Extraversion) will cooperate in a calculating manner, strategically shifting to cooperation when threatened with punishment. Also, we hypothesized that some players will be cooperators (i.e., those high in Honesty, Agreeableness, and Introversion, and low in Emotionality, Machiavellianism, Narcissism, and Psychopathy) or free-riders (i.e., those high in Sadism), regardless of the threat of punishment. The research sample included 105 participants (average age: 21 years, 86 women). Half of them played the Public Goods game without the punishment option and then with it (*n*=52); for the other half the order of the games was reversed (*n*=53). In the sample as a whole, there were no correlations between cooperation shift or contribution to the public good and personality traits. Although significant correlations were found in the subsamples in which the punishment threat was introduced before or after the no-punishment condition, such correlations were inconsistent and only of moderate strength; therefore, stronger studies are needed in order to draw more reliable conclusions. In the sample as a whole, the punishment threat enhanced cooperativeness (*t*(104)=2.48, *p*=.015); but its effectiveness, surprisingly, depended on the order in which it was introduced. Players were more cooperative under the threat of punishment if they played the punishment game before the no-punishment game (*t*(52)=7.15, *p*=.000); this was not the case when the order of the games was reversed. We discuss the results in the light of the findings on conditional cooperation and we consider their implications for interventions in the social environment. In addition, we suggest that future research include macroeconomic and macrosocial factors, which could more strongly affect public good contributions in a social dilemma situation, on top of or even despite the personality of the players.