This paper presents a Coalitional Game Theory based approach to investigate the privacy levels of data to be published. Privacy means, unlinking the individual’s identity from his sensitive data. Anonymization is one of the technologies often used for providing privacy. In this paper, to accomplish the anonymity we have considered each tuple as a player, allocated a payoff for each of them using concept hierarchy tree (CHT) of quasi identifiers (QID) of the table and established a coalition among them based on their payoffs. We present our results after analyzing the constraints for k-anonymization and in revealing the boundaries of k.