Figure 1.1. Four Stages of Influence Production

Figure 1.2. Entry Control and Lobbying Access among Selected Nondemocracies, 1990–2018 Data source: Coppedge et al. (2021), N = 2,528 autocratic country-years. Note: Turkey is classified as democratic until 2012, Belarus from 1991–1996, and Montenegro in 1990–1997, 2003–2005, 2009, and 2015.

Figure 2.1. Information Demands in Political and Expertise-Based Exchanges

Figure 2.2. Overview of Respondents in Sample by Country’s Regime Type Note: N = 292. Regime type is based on Varieties of Democracy indicator. In the statistical analysis we merge the two autocratic regimes and the two democratic regimes. In total there are 102 unique countries.

Figure 4.1. Cosine Similarity Measures Between Think Tanks Note: Cosine similarity scores calculated from the aggregate document-frequency matrix for each think tank. Higher scores (and darker shading) indicate that the think tank pair produces more similar texts. Table 4.1 lists the think tank labels.

Figure 4.2. Pearson Correlations between Topics. Note: Pearson correlations between topics. A more negative correlation (and darker shading) indicates a lack of topic co-occurence in the same document.

Figure 4.3. Differences in Topic Proportion by Think Tank Type Note: Estimates further to the right indicate that the think tank type discusses a given topic with greater frequency.

Figure 4.4. Changes in Topic Proportions over Time Note: Changes in expected topic proportions over time. Higher values indicate that a greater proportion of the corpus is devoted to a given topic.

Figure 4.5. Topic-Term Similarity over Time Note: Similarities in words used to describe different policy topics over time, with a fitted Loess curve to ease trend visualization. An increase on the y-axis indicates an increase in pairwise similarity in texts.

Figure 5.1. Ecology of Advocacy Groups, and the Target Institutions Source: Author.

Figure 5.2. Shared Images of Areng Cambodia on Google and Facebook Source: Author.

Figure 7.1. Conditional Effects of Alignment on Outside Lobbying Note: Based on M3 (Table 7.1), other variables held constant at mean, 95% confidence interval. N=267 groups

Figure 7.2. Predicting Outside Lobbying in Autocracies Note: Based on M4 (Table 7.1), other variables held constant at mean, 95% prediction intervals. N = 133 groups.

Figure 8.1. ENGOs’ Engagement in Different Advocacy Tactics, Percentages Note: ENGOs were asked how frequently they had contacted members of the People’s Congress and Political Consultative Conference at both local and national levels. Those that were reported as barely were coded 0, and those that had contacted officials at the legislative organs at least once a year over the past three years were coded 1. For the rest of the advocacy activities, ENGOs that had engaged in the activities fairly often and very often were coded 1, and 0 otherwise. Percentages indicate the percentage of ENGOs that had engaged in the activities fairly often and very often.

Figure 12.1. Autocratic Legitimation Strategies and CSOs Consultations Note: Linear prediction of v2cscnsult_osp by legitimation strategies, with a 95% confidence intervals.

Figure 12.2. Marginal Effects of Legitimation Strategies on CSOs Consultations Note: Marginal effects with a 95% confidence intervals.

Figure 13.1. Conditions Shaping Policy Advocacy under Authoritarianism