Partisan preferences usually stand out as the major driving force behind voters’ expectations about election outcomes. Although other variables, such as political knowledge or education, are found to exert a positive influence on the ability of citizens to anticipate the results of political races, the effect of these variables is quite modest in most cases. Hence, it seems that purely individual-level factors (apart from party identification and political leanings) hold limited explanatory power. Some studies argue that we need to move from the strictly personal sphere to the interpersonal one if we are to better understand the underpinnings of forecasting abilities. This paper leverages data from 48 elections at the district, regional, and national levels in nine different countries (i.e., Austria, Canada, Germany, Israel, Italy, New Zealand, Spain, the United Kingdom, and the United States) to assess the impact of ‘contact networks’ and network characteristics (such as network size or the frequency of political discussion and disagreement) on citizens’ electoral expectations. The analyses also rely on data obtained from two question modules that were included, specifically for this study, in the 2019 Canadian Election Study and the 2020 Saskatchewan Election Study. The results cast some doubts on the capacity of social interactions to influence citizens’ forecasting skills.