Despite increasing attention to online campaigning and social media use, television debates are still the center stage for political debate in western democracies. The literature on the effects of such debates has noted that watching TV debates enhances political learning about candidates and their issue positions (Abramowithz, 1978; Pfau & Everlad, 1994). A less travelled road addresses relational effects, this is, nonverbal cues that convey ideas about candidates and how people relate to them. Such relational cues might overpower verbal content (such as in the Bush/Dukakis debate, see Pfau & Kang, 1991), priming people to rely more on candidate’s personality perceptions (Druckman, 2009).

While researchers concerned about candidates’ relational communication abilities have addressed short and long-term effects (Pfau, Driedich, Larson & Van Winkle, 1993; Pfrau & Everland, 1996), research on the persistence of television debates is still scant. In the words of Pfrau (2017): “Debate scholars need to assess relational influence across time, much as Bernett (1981) had recommended, albeit about content effects of debates, more than two decades ago.”

This work attempts to assess the influence of two TV electoral debates across time. In doing so, it posits a hard test on the lasting effects hypothesis by overcoming two limitations of past works: the fact that they are mostly confined to the US (and therefore, seldom have addressed TV debates’ effects in multi-party systems), and that very few attention has been paid to second-order elections.

More specifically, this work assesses whether participating in two electoral debates hold prior to the 2017 Catalan regional election ( 17 and 18 december 2017) affected the perceptions of the two main candidates in the opposition, Inés Arrimadas (Ciudadanos) and Miquel Iceta (Partit dels Socialistes de Catalunya), regarding competence, honesty, responsiveness and expertise. To do so, the paper relies in a student-based sample (N=370) surveyed before and after the debates. The research design relies in a novel technique named “Unexpected Event during Surveys Design” (Muñoz, Falcó-Gimeno, Hernández , 2018), used to estimate causal effects in similar scenarios.