What do international election observation missions think (and feel) about the elections they observe? An election observation report represents a written judgment by one group regarding a given election. While elections are fundamentally a technical process, the normative salience of democratic processes and the partisan nature of politics means that the content of election observation reports may reflect the normative or political priorities of the observers drafting the report. This project examines how the content and normative/emotional affect contained in election observation reports varies as a function of regional, observer-specific, and election-specific characteristics. Using quantitative text analysis, we find significant variation in the topic content, normative stance, and affective stance election observation reports from over 1000 elections over the past 30 years. More specifically, we find that the language used to describe elections in Africa is more negative than language used to describe elections elsewhere: we expect that observers more frequently emphasize the risk of electoral fraud, even when fundamental electoral characteristics (e.g., pre-election violence) are similar. Our results shed new light on the role of international electoral observation missions in ensuring the quality of elections in developing democracies.