This study explores how researchers’ analytical choices affect the reliability of scientific findings. Most discussions of reliability problems in science focus on systematic biases. We broaden the lens to emphasize the idiosyncrasy of conscious and unconscious decisions that researchers make during data analysis. We coordinated 161 researchers in 73 research teams and observed their research decisions as they used the same data to independently test the same prominent social science hypothesis: that greater immigration reduces support for social policies among the public. In this typical case of social science research, research teams reported both widely diverging numerical findings and substantive conclusions despite identical start conditions. Researchers’ expertise, prior beliefs, and expectations barely predict the wide variation in research outcomes. More than 95% of the total variance in numerical results remains unexplained even after qualitative coding of all identifiable decisions in each team’s workflow. This reveals a universe of uncertainty that remains hidden when considering a single study in isolation. The idiosyncratic nature of how researchers’ results and conclusions varied is a previously underappreciated explanation for why many scientific hypotheses remain contested. These results call for greater epistemic humility and clarity in reporting scientific findings.

Protest against authoritarian rule is a well-studied phenomenon in the social sciences, but mass rallies in favor of authoritarian regimes have received only limited scholarly attention. While previous work has portrayed authoritarian regimes as characterized by mass apathy and political demobilization, we show that this is only partially true today. We argue that autocrats mobilize their supporters selectively as a strategic response to political threats. Rallies increase collective action costs for rivaling elites, opposition movements, and bystanders because they signal regime strength (deterrence) and curb mobilization efforts against the regime (repression). Nevertheless, the mobilization of supporters is costly, as autocrats have only imperfect information about current levels of support, rallies require organizational capacity and clashes between supporters and opponents can get out of control. Drawing on the first global data set with information about pro-government rally events in all authoritarian regimes from 2003 to 2015, our quantitative analysis reveals systematic patterns in the occurrence of rallies in line with our theoretical framework. We find systematic increases in pro-government mobilization during episodes of large domestic and regional opposition mobilization, high coup risk, and prior to elections.

This article analyses the state of democracy in the world in 2019. We demonstrate that the "third wave of autocratization" is accelerating and deepening. The dramatic loss of eight democracies in the last year sets a new record in the rate of breakdowns. Exemplifying this crisis is Hungary, now the EU's first ever authoritarian member state. Governmental assaults on civil society, freedom of expression, and the media are proliferating and becoming more severe. A new and disturbing trend is that the quality of elections is now also deteriorating in many countries. Nevertheless, there are also positive signs: pro-democracy protests reached an all-time high in 2019. People are taking to the streets to protest the erosion of democracies and challenge dictators. Popular protests have contributed to substantial democratization in 22 countries over the last ten yearsincluding Armenia, Tunisia, and Ecuador. This was before the Covid-19 pandemic. Responses to the crisis, including many states of emergencies, risk further accelerating autocratization.

Following its global diffusion during the last decade, the Internet was expected to become a liberation technology and a threat for autocratic regimes by facilitating collective action. Recently, however, autocratic regimes took control of the Internet and filter online content. Building on the literature concerning the political economy of repression, this article argues that regime characteristics, economic conditions, and conflict in bordering states account for variation in Internet filtering levels among autocratic regimes. Using OLSregression, the article analyzes the determinants of Internet filtering as measured by the Open Net Initiative in 34 autocratic regimes. The results show that monarchies, regimes with higher levels of social unrest, regime changes in neighboring countries, and less oppositional competition in the political arena are more likely to filter the Internet. The article calls for a systematic data collection to analyze the causal mechanisms and the temporal dynamics of Internet filtering.

This article analyses the state of democracy in 2020. The world is still more democratic than it was in the 1970s and 1980s, but a trend of autocratization is ongoing and affecting 25 countries in 2020, home to 34% of the world's population. At the same time, the number of democratizing countries has dwindled by nearly half, reducing to 16 countries, home to a mere 4% of the global population. Freedom of expression, deliberation, rule of law and elections show the most substantial net declines in the last decade. A major change is that India, formerly the world's largest democracy, turned into an electoral autocracy. The V-Dem data suggests that direct effects of the Covid-19 pandemic on levels of liberal democracy were limited in 2020. Still, the longer-term consequences may be worse and must be monitored closely. Due to the pandemic and state restrictions on the freedom of assembly, mass mobilization declined to its lowest level in over a decade, yet the decline in pro-democracy protests in 2020 may well prove to be short-lived once the pandemic subdues.