This thesis tries to fill in the research gap on whether the ideologies of armed groups can change. Tokdemir et al. (2021) discovered that an increase in the number of competitors leads to a polarisation of an armed group’s ideology, I theorise that a decrease in the number of competitors will lead to depolarisation of an armed group’s ideology, because then the armed group gains more access to the finite resource pool. I conducted a structured, focused within-case comparison over time using discourse analysis. Discourse analysis allows for a detailed and complex operationalisation of ideology. The case study focuses on the People’s War Group/Communist Party of India-Maoist, a communist group part of the Naxalite movement in India. For the independent variable, I use multiple secondary sources to establish the number of competitors in 1991 and 2008. For the dependent variable, I analyse multiple magazine publications from 1991 and 2008 written by the PWG/CPI-Maoist to establish their ideology. The findings of this research do not support the hypothesis. There is limited change over time in the ideology of the PWG/CPI-Maoist even though the number of competitors changed between 1991 and 2008. Suggestions for future research are given at the end.

To what extent can two 'Modes' be found amongst non-state armed groups (NSAGs)? The Modes theory hypotheses that NSAGs are organized into two distinct Modes, either the 'Imagistic' or the 'Doctrinal' Mode, with each Mode characterised by a particular organisational structure and type of resource extraction. Previous preliminary research on this had three limitations: 1) it used a sample of only five groups; 2) "Galton's problem", i.e. non-independence of groups was not taken into account, and 3) the theory of resource extraction was underdeveloped. The present research overcomes each of these limitations. First, can the patterns found in the 5-group case study be replicated with a larger sample? A dataset with 50 NSAGs was constructed, with variables that were derived from the theory and resembling those used for the 5-group case study. Analysis results showed that a few 'Doctrinal' variables were significantly correlated and 'Doctrinal' rituals had a frequency distribution as predicted. Yet, most correlations between all Modes variables were not significant, and analyses did not convincingly show that there were two distinct Modes. Second, to what extent does "Galton's problem" (autocorrelation) play a role amongst NSAGs regarding their Modes properties? Data on relations between armed groups and their geographical locations were added to the sample. Using Moran's I, it appeared that two to four Modes properties were significantly autocorrelated. With geographical region as covariate, analyses were redone. Results remained basically the same as without regional covariate. Third, what is the relationship between Modes and the type of Resource extraction? Resource types were operationalized and Resource type data were added to the sample of NSAGs. A theory on Risk, and a ROI (Return on Investment) and Risk index were developed. Regressions showed no change in ROI or Risk over time. Low ROI did not typically predict Doctrinal-related Modes properties, and high-ROI did not typically predict Imagistic-related Modes properties. The significance of relations between Modes properties and Resource types strongly depended on whether the analysis was a correlation or a regression. Overall, only a few relationships were significant and no consistent pattern emerged. Limitations of the present research, and suggestions for further research are discussed, as well as an overall assessment of the Modes theory.