According to some authors, party identification (PID) may not only increase with time, but can also decrease as a consequence of evaluation of parties’ performance. Theories related to ‘active’ learning state that people could learn from both positive and negative inputs. Campbell and his colleagues acknowledged that PID could consist of positive and negative feelings, although most research has focused on the positive side. It is not only that positive performance will lead individuals to prefer one party to another, but it is also the case that poor performance will lead them to choose a second best party, or in some cases no party at all. In these cases, the ‘hostility’ toward parties may be the main guide in defining PID. It has been argued that there are more chances of disliking one or more parties without liking any other in multiparty systems relative to two-party systems. Evaluations of party performance over time may give rise not just to positive feelings that enhance PID, but also may produce negative sentiments that decrease PID, or in some cases, increase hate toward a party, what has been defined as ‘negative’ PID. Two-party systems, as Britain or the United States, offer a partial setting for studying parties’ negative evaluations, mainly because competition is polarized, and even perhaps this is the reason why very few studies have been found on the subject. Conversely, a multiparty system facilitates the analysis of negative evaluations, expecting to find both people who have developed an attachment to a party and people that may hate other political parties. An ideal context to study such as the impact of negative evaluations is either the one-party systems, as post-Communist countries where it is expected to find certain degree of aversion to the former ruling party, or countries where long-ruling parties had governed but recently lost, like Taiwan or Mexico. Previous studies have distinguished positive from negative PID, by demonstrating the existence of two different kinds of attachment. Those respondents who mention that they would never vote for some party are said to have developed a negative PID. In contrast, I hypothesize that there is one sort of party attachment (that is PID) that may be guided by both positive and negative party evaluations. I hypothesize that positive and negative feelings are determinants of PID. For both Panistas and Perredistas the PRI should be the party they would never vote for. For the case of independents, I also hypothesize that even though these individuals do not hold a clear partisan attachment, they should hold a strong anti-PRI feeling, even stronger than towards any other party. In sum, looking at countries that ended up a one-party period or an authoritarian period gives an interesting but overlooked tweak that may advance research on PID by developing the concept of negative party identification. [ABSTRACT FROM AUTHOR]